Ysrcp

Table of Contents (Subject Area: Biochemistry)

Article

Authors

Pages in the Encyclopedia

Bioenergetics Enzyme Mechanisms Food Colors Glycoconjugates and Carbohydrates Ion Transport Across Biological Membranes Lipoprotein/Cholesterol Metabolism Membrane Structure Natural Antioxidants In Foods Nucleic Acid Synthesis Protein Folding Protein Structure Protein Synthesis Vitamins and Coenzymes

Richard E. McCarty and Eric A. Johnson Stephen J. Benkovic and Ann M. Valentine Pericles Markakis Eugene A. Davidson George P. Hess Alan D. Attie Anna Seelig and Joachim Seelig Eric A. Decker Sankar Mitra_ Tapas K. Hazra and Tadahide Maurice Eftink and Susan Pedigo Ivan Rayment Paul Schimmel and Rebecca W. Alexander David E. Metzler

Pages 99-115 Pages 627-639 Pages 105-120 Pages 833-849 Pages 99-108 Pages 643-660 Pages 355-367 Pages 335-342 Pages 853-876 Pages 179-190 Pages 191-218 Pages 219-240 Pages 509-528

Bioenergetics
Richard E. McCarty Eric A. Johnson
Johns Hopkins University

I. II. III. IV.

Catabolic Metabolism: The Synthesis of ATP Photosynthesis Origin of Mitochondria and Chloroplasts Illustrations of the Uses of ATP: Ion Transport, Biosynthesis, and Motility V. Concluding Statements

GLOSSARY
Adenosine 5 -triphosphate (ATP) The carrier of free energy in cells. Bioenergetics The study of energy relationships in living systems. Chloroplasts The sites of photosynthesis in green plants. Ion transport The movement of ions across biological membranes. Metabolism The total of all reactions that occur in cells. Catabolic metabolism is generally degradative and exergonic, whereas anabolic metabolism is synthetic and requires energy. Mitochondria Sites of oxidative (catabolic) metabolism in cells. Photosynthesis Light-driven synthesis of organic molecules from carbon dioxide and water. Plasma membrane The barrier between the inside of cells and the external medium.

BIOENERGETICS, an amalgamation of the term biological energetics, is the branch of biology and biochemistry that is concerned with how organisms extract energy from their environment and with how energy is used to fuel the myriad of life’s endergonic processes. Organisms may be usefully divided into two broad groups with respect to how they satisfy their need for energy. Autotrophic organisms convert energy from nonorganic sources such as light or from the oxidation of inorganic molecules to chemical energy. As heterotrophic organisms, animals must ingest and break down complex organic molecules to provide the energy for life. Interconversions of forms of energy are commonplace in the biological world. In photosynthesis, the electromagnetic energy of light is converted to chemical energy, largely in the form of carbohydrates, with high overall efficiency. The energy of light is used to drive oxidation– reduction reactions that could not take place in the dark. Light energy also powers the generation of a proton electrochemical potential across the green photosynthetic

99

100

Bioenergetics

FIGURE 1 Central role of adenosine 5 -triphosphate (ATP) in metabolism. Catabolic (degradative) metabolism is exergonic and provides the energy needed for the synthesis of ATP from adenosine 5 -diphosphate (ADP) and inorganic phosphate (Pi ). The exergonic hydrolysis of ATP in turn powers the endergonic processes of organisms.

membrane. Thus, electrical work is an integral part of photosynthesis. Chemical energy is used in all organisms to drive the synthesis of large and small molecules, motility at the microscopic and macroscopic levels, the generation of electrochemical potentials of ions across cellular membranes, and even light emission as in fireflies. Given the diversity in the forms of life, it might be expected that organisms have evolved many mechanisms to deal with their need for energy. To some extent this expectation is the case, especially for organisms that live in extreme environments. However, the similarities among organisms in their bioenergetic mechanisms are as, or even more, striking than the differences. For example, the sugar glucose is catabolized (broken down) by a pathway that is the same in the enteric bacterium Escherichia coli as it is in higher organisms. All organisms use adenosine 5 -triphosphate (ATP) as a central intermediate in energy metabolism. ATP acts in a way as a currency of free energy. The synthesis of ATP from adenosine 5 -diphosphate

(ADP) and inorganic phosphate (Pi ) is a strongly endergonic reaction that is coupled to exergonic reactions such as the breakdown of glucose. ATP hydrolysis in turn powers many of life’s processes. The central role of ATP in bioenergetics is illustrated in Fig. 1. Partial structures of several compounds that play important roles in metabolism are shown in Fig. 2. In this article, the elements of energy metabolism will be discussed with emphasis on how organisms satisfy their energetic requirements and on how ATP hydrolysis drives otherwise unfavorable reactions.

I. CATABOLIC METABOLISM: THE SYNTHESIS OF ATP
Metabolism may be defined as the total of all the chemical reactions that occur in organisms. Green plants can synthesize all the thousands of compounds they contain

Bioenergetics

101 of glucose by Pi is an unfavorable reaction, characterized by a G 0 of about 4 kcal/mol, at pH 7.0 and 25◦ C. (Note that the biochemist’s standard state differs from that as usually defined in that the activity of the hydrogen ion is taken as 10−7 M, or pH 7.0, rather than 1 M, or pH 0.0. pH 7.0 is much closer to the pH in most cells.) This problem is neatly solved in cells by using ATP, rather than Pi , as the phosphoryl donor: Glucose + ATP ←→ Glucose 6-phosphate + ADP. The G 0 for this reaction, which is catalyzed by the enzyme hexokinase, is approximately −4 kcal/mol. Thus the phosphorylation of glucose by ATP is an energetically favorable reaction and is one example of how the chemical energy of ATP may be used to drive otherwise unfavorable reactions. Glucose 6-phosphate is then isomerized to form fructose 6-phosphate, which in turn is phosphorylated by ATP at the 1-position to form fructose 1,6-bisphosphate. It seems odd that a metabolic pathway invests 2 mol of ATP in the initial steps of the pathway when ATP is an important product of the pathway. However, this investment pays off in later steps. Fructose 1,6-bisphosphate is cleaved to form two triose phosphates that are readily interconvertible. Note that the oxidation–reduction state of the triose phosphates is the same as that of glucose 6-phosphate and the fructose phosphates. All molecules are phosphorylated sugars. In the next step of glycolysis, glyceraldehyde 3-phosphate is oxidized and phosphorylated to form a sugar acid that contains a phosphoryl group at positions 1 and 3. The oxidizing agent, nicotinamide adenine dinucleotide (NAD+ ), is a weak oxidant (E 0 , at pH 7.0 of −340 mV). The oxidation of the aldehyde group of glyceraldehyde 3-phosphate to a carboxylate is a favorable reaction that drives both the oxidation and the phosphorylation. This is the only oxidation–reduction reaction in glycolysis. The hydrolysis of acyl phosphates, such as that of position 1 of 1,3-bisphosphoglycerate, is characterized by strongly negative G 0 values. That for 1,3-bisphosphoglycerate is approximately −10 kcal/mol, which is significantly more negative than the G 0 for the hydrolysis of ATP to ADP and Pi . Thus, the transfer of the acyl phosphate from 1,3-bisphosphoglycerate to ADP to form 3-phosphoglycerate and ATP is a spontaneous reaction. Since two sugar acid bisphosphates are formed per glucose metabolized, the two ATP invested in the beginning of the pathway have been recovered. In the next steps of glycolysis, the phosphate on the 3-position of the 3-phosphoglycerate is transferred to the hydroxyl residue at position 2. Removal of the elements of water from 2-phosphoglycerate results in the formation of an enolic phosphate compound, phospho(enol)pyruvate

FIGURE 2 Some important reactions in metabolism. Shown are the phosphorylation of ADP to ATP, NAD+ , NADH, FAD, FADH2 acetate, CoA, and acetyl CoA. For clarity, just the parts of the larger molecules that undergo reaction are shown. NAD+ , nicotinamide adenine dinucleotide; NADH, nicotinamide adenine dinucleotide (reduced form); FAD, flavin adenine dinucleotide; FADH2 , flavin adenine dinucleotide (reduced form); CoA, coenzyme A; AMP, adenosine monophosphate.

from carbon dioxide, water, and inorganic nutrients. The discussion of the complicated topic of metabolism is somewhat simplified by separation of the subject into two areas—catabolic and anabolic metabolism. Catabolic metabolism is degradative and is generally exergonic. ATP is a product of catabolic metabolism. In contrast, anabolic metabolism is synthetic and requires ATP. Fortunately, there are relatively few major pathways of energy metabolism. A. Glycolysis and Fermentation Carbohydrates are a major source of energy for organisms. The major pathway by which carbohydrates are degraded is called glycolysis. Starch, glycogen, and other carbohydrates are converted to the sugar glucose by pathways that will not be considered here. In glycolysis, glucose, a sixcarbon sugar, is oxidized and cleaved by enzymes in the cytoplasm of cells to form two molecules of pyruvate, a three-carbon compound (see Figs. 3 and 4). The overall reaction is exergonic and some of the energy released is conserved by coupling the synthesis of ATP to glycolysis. Before it may be metabolized, glucose must first be phosphorylated on the hydroxyl residue at position 6. Under intracellular conditions, the direct phosphorylation

102

Bioenergetics

FIGURE 3 Schematic outline of carbohydrate metabolism. Glucose is oxidized to two molecules of pyruvate by glycolysis in the cytoplasm. In mitochondria, pyruvate is oxidized by molecular oxygen to CO2 and water. The synthesis of ATP is coupled to pyruvate oxidation.

(PEP). The free energy of hydrolysis of PEP to form the enol form of pyruvate and Pi is on the order of −4 kcal/mol. In aqueous solution, however, the enol form of pyruvate is very unstable. Thus, the hydrolysis of PEP to form pyruvate is a very exergonic reaction. The G 0 for this reaction is −14.7 kcal/mol, which corresponds to an equilibrium constant of 6.4 × 1010 . PEP is thus an excellent phosphoryl donor and the formation of pyruvate is coupled to ATP synthesis. Since two molecules of pyruvate are formed per glucose catabolized, two ATP are formed. Thus the net yield of ATP is two per glucose oxidized to pyruvate. In some organisms, glycolysis is the only source of ATP. A familiar example is yeast growing under anaerobic (no oxygen) conditions. In this case, glucose is said to be fermented and ethyl alcohol and carbon dioxide (CO2 ) are the end products (Fig. 5). In contrast, all higher organisms can completely oxidize pyruvate to CO2 and water, using molecular oxygen as the terminal electron acceptor. The conversion of glucose to pyruvate releases only a small fraction of the energy available in the complete oxidation of glucose. In aerobic organisms, more than 90% of the ATP made during glucose catabolism results from the oxidation of pyruvate.

B. Oxidation of Pyruvate: The Citric Acid Cycle In higher organisms, the oxidation of pyruvate takes place in subcellular, membranous organelles known as mitochondria. Because mitochondria are responsible for the synthesis of most of the ATP in nonphotosynthetic tissue, they are often referred to as the powerhouses of cells. Mitochondrial ATP synthesis is called oxidative phosphorylation since it is linked indirectly to oxidative reactions. In the complete oxidation of pyruvate, there are five oxidation–reduction reactions. Three of these reactions are oxidative decarboxylations. The electron acceptor (oxidizing agent) for four of the reactions is NAD+ ; the oxidizing agent for the fifth is flavin adenine dinucleotide, or FAD. Knowing the oxidation–reduction potentials of the reactants in an oxidation–reduction reaction permits the ready calculation of the standard free energy change for the reaction. It may be shown that G 0 = −n F E 0 , (1)

where n is the number of electrons transferred in the reaction, F is Faraday’s constant (23,060 cal/V-equivalent), and E 0 is the difference between the E 0 value of the oxidizing agent and that of the reducing agent.

Bioenergetics

103 FADH2 is −38 kcal/mol. These two strongly exergonic reactions provide the energy for the endergonic synthesis of ATP. The details of carbon metabolism in the citric acid cycle are beyond the scope of this article. In brief, pyruvate is first oxidatively decarboxylated to yield CO2 , NADH, and an acetyl group attached in an ester linkage to a thiol on a large molecule, known as coenzyme A, or CoA. (See Fig. 2.) Acetyl CoA condenses with a four-carbon dicarboxylic acid to form the tricarboxylic acid citrate. Free CoA is also a product (Fig. 6). A total of four oxidation– reduction reactions, two of which are oxidative decarboxylations, take place, which results in the generation of the three remaining NADH molecules and one molecule of FADH2 . The citric acid cycle is a true cycle. For each two-carbon acetyl moiety oxidized in the cycle, two CO2 molecules are produced and the four-carbon dicarboxylic acid with which acetyl CoA condenses is regenerated. The mitochondrial inner membrane (Fig. 7) contains proteins that act in concert to catalyze NADH and FADH2 oxidation by molecular oxygen. [See reactions (2) and (3) above.] These reactions are carried out in many small steps by proteins that are integral to the membrane and that undergo oxidation–reduction. These proteins make up what is called the mitochondrial electron transport chain. Components of the chain include iron proteins (cytochromes and iron–sulfur proteins), flavoproteins (proteins that contain flavin), copper, and quinone binding proteins. The oxidation of NADH and FADH2 by molecular oxygen is coupled in mitochondria to the endergonic synthesis of ATP from ADP and Pi . For many years the nature of the common intermediate between electron transport and ATP synthesis was elusive. Peter Mitchell, who received a Nobel Prize in chemistry in 1978 for his extraordinary insights, suggested that this common intermediate was the proton electrochemical potential. He proposed in the early

FIGURE 4 A view of glycolysis. Glucose, a six-carbon sugar, is cleaved and oxidized to two molecules of pyruvate. There is the net synthesis of two ATP per glucose oxidized and two NADH are also formed.

The reduced form of NAD+ , NADH, is a strong reducing agent. The E 0 at pH 7.0 of the NAD+ –NADH couple is −340 mV, which is equivalent to that of molecular hydrogen. E 0 is the potential when the concentrations of the oxidized and reduced species of an oxidation–reduction pair are equal. Reduced FAD, FADH2 , is a weaker reductant than NADH, with an E 0 (pH 7.0) of about 0 V. In contrast, molecular oxygen is a potent oxidizing agent and fully reduced oxygen, water, is a very poor reducing agent. The E 0 (pH 7.0) for the oxygen–water couple is +815 mV. The oxidation of NADH and FADH2 results in the reduction of oxygen to water: H+ + NADH + 1 O2 → NAD+ + H2 O 2 and FADH2 + 1 O2 → FAD + H2 O. 2 (3) (2)

In both cases two electrons are transferred to oxygen, so that the n in Eq. (1) is equal to 2. Under standard conditions, the oxidation of 1 mol of NADH by oxygen liberates close to 53 kcal, whereas the G 0 for that of

FIGURE 5 Fates of pyruvate. In yeasts under anaerobic conditions, pyruvate is decarboxylated and reduced by the NADH formed by glycolysis to ethanol. In anaerobic muscle, the NADH generated by glycolysis reduces pyruvate to lactic acid. When O2 is present, pyruvate is completely oxidized to CO2 and water.

104

Bioenergetics

FIGURE 6 A view of the oxidation of pyruvate. The oxidation of pyruvate generates three CO2 , four NADH, and one FADH2 . The oxidation of NADH and FADH2 by the mitochondrial electron transport chain is exergonic and provides most of the energy for ATP synthesis.

1960s that electron transport through the mitochondrial chain is obligatorily linked to the movement of protons across the inner membrane of the mitochondrion. In this way, part of the energy liberated by oxidative electron transfer is conserved in the form of the proton electrochemical potential. This potential, µH+ , is the sum of contributions from the activity gradient and that of the electrical gradient: µH+ = RT ln [H+ ]a [H+ ]b + F ϕ, (4)

where R is the gas constant; T , the absolute temperature; a and b, the aqueous spaces bounded by the membrane; F, Faraday’s constant; and ϕ, the membrane potential. As Mitchell suggested, the mitochondrial inner membrane is poorly permeated by charged molecules, including protons. The membrane thus provides an insulating layer between the two aqueous phases it separates. Thus the transport of protons across the membrane generates an electrochemical potential. In the case of mitochondria, the membrane potential is the predominant component of the electrochemical of the proton. The total µH+ in actively respiring mitochondria is on the order of −200 mV, if one uses the convention that the inside space bounded by the membrane is negative. Electron transport from NADH and FADH2 to oxygen provides the energy for the generation of the electrochemical potential of the proton. The flow of protons down this

potential is exergonic and is the immediate source of energy for ATP synthesis. The proton-linked synthesis of ATP is catalyzed by a complex enzyme called ATP synthase. Remarkably similar enzymes are located in the coupling membranes of bacteria, mitochondria, and chloroplasts, the intracellular sites of photosynthesis in higher plants. Even though the reaction that they catalyze seems relatively straightforward (see Fig. 2), the ATP synthases contain a minimum of 8 different proteins and a total of about 20 polypeptide chains. ATP is formed in the aqueous space bounded by the mitochondrial inner membrane. This space is known as the matrix (see Fig. 7). Most of the ATP generated within mitochondria is exported to the cytoplasm where it is used to drive energy-dependent reactions. The ADP and Pi formed in the cytoplasm must then be taken up by the mitochondria. The inner membrane contains specific proteins that mediate the export of ATP and the import of ADP and Pi . One transporter catalyzes counterexchange transport of ATP out of the matrix with ADP in the cytoplasm into the matrix (Fig. 8). At physiological pH, ATP bears four negative charges, and ADP, three. Thus, the one-to-one exchange transport of ATP with ADP creates a membrane potential that is opposite in sign of that created by electrontransport-driven proton translocation. ATP/ADP transport costs energy and the direction of transport is poised by the proton membrane potential. In addition, phosphate

Bioenergetics

105

FIGURE 7 Diagrams of the structures of mitochondria and chloroplasts. The inner membrane of mitochondria and the thylakoid membrane of chloroplasts contain the electron transport chains and ATP synthases. Note that the orientation of the inner membrane is opposite that of the thylakoid membrane.

uptake into mitochondria is coupled to the electrochemical proton potential. The phosphate translocator (see Fig. 8) catalyzes the counterexchange transport of H2 PO2− and 4 hydroxide anion (OH− ). The outward movement of OH− causes acidification of the matrix, whereas the direction of proton transport driven by electron transport is out of the mitochondrial matrix and results in an increase in the pH of the matrix. In the total oxidation of glucose to CO2 and water, six CO2 are released and six O2 are reduced to water. For each pyruvate oxidized, four NADH and one FADH2 are generated. Since two molecules of pyruvate are derived by means of glycolysis from one molecule of glucose, a total of eight NADH and two FADH2 are formed by pyruvate oxidation. Four electrons are required for the reduction of O2 to two molecules of H2 O. Thus, pyruvate oxidation accounts for the reduction of five of the six molecules of

O2 in the complete oxidation of glucose. The sixth O2 is reduced to water by electrons from the NADH formed by the oxidation of triose phosphate in glycolysis. Fermentation, or anaerobic glycolysis, yields but 2 mol of ATP per 1 mol of glucose catabolized. In contrast, complete oxidation of glucose to CO2 and water yields about 15 times more ATP. Thus, it is understandable why yeasts and some bacteria consume more glucose under anaerobic conditions than when oxygen is present. In animals, glucose is normally completely oxidized. During strenuous exercise, however, the demand for oxygen by muscle tissues can outstrip its supply and the tissue may become anaerobic. Muscle contraction requires ATP, and rapid breakdown of glucose and its storage polymer, glycogen, takes place under anaerobiosis. Glycolysis would stop quickly if the NADH produced by the oxidation of triose phosphate were not converted back to NAD+ . In muscle cells under O2 -limited conditions, pyruvate is reduced by NADH to lactic acid (see Fig. 5), a source of muscle cramps during exercise. At rest, lactic acid is converted back to glucose in the liver and kidneys and returned to muscle tissues where it stored in the form of glycogen.

C. Oxidation of Fats and Oils, Major Metabolic Fuels Fats and oils are ubiquitous biological molecules that are major energy reserves in animals and developing plants. Fats and oils are esters of glycerol, a three-carbon compound with hydroxyl groups on all three carbons, and carboxylic acids with long hydrocarbon chains. The most common fats and oils contain fatty acids with straight chains with an even number of carbon atoms. Most often, the total number of carbons in a fatty acid in a triglyceride ranges from 14 to 18. The difference between a fat and an

FIGURE 8 ATP, ADP, and Pi transport in mitochondria. ATP is formed inside mitochondria. Most of the ATP is exported to the cytoplasm where it is cleaved to ADP and Pi . The mitochondrial inner membrane contains specific proteins that mediate not only ATP release coupled to ADP uptake, but also Pi uptake linked to hydroxide ion (OH− ) release.

106 oil is simply melting temperature. Oils are liquid at room temperature, whereas fats are solid. Familiar examples are olive oil and butter. The most significant reason for this difference in melting temperatures between fats and oils is the degree of unsaturation (double bonds) of the fatty acids they contain. The introduction of double bonds into a hydrocarbon chain causes perturbations in the structure of the chain that decrease its ability to pack the chains closely into a solid structure. Olive oil contains far more unsaturated fatty acids than butter does and is thus a liquid at room temperature and even in the cold. Regardless of the physical properties of triglycerides, they are the long-term energy reserves of higher organisms. Consider the fact that the complete oxidation of triglycerides to CO2 and water yields 9 kcal/g, whereas that of the carbohydrate storage polymers, starch and glycogen, yields just 4 kcal/g. When it is also remembered that fats and oils shun water, but glycogen and starch are more hydrophilic, triglycerides have an additional advantage over the glucose polymers as deposits of potential free energy. As hydrophobic moieties, fats and oils require less intracellular space than that required by the glucose polymers. The first step in the breakdown of triglycerides (Fig. 9) is their conversion by hydrolysis to their components, glycerol and fatty acids. Glycerol is a close relative of the threecarbon compounds involved in the catabolism of glucose and may be completely oxidized to CO2 and water by glycolysis and the tricarboxylic acid cycle. The fatty acids are first converted to CoA derivatives at the expense of the hydrolysis of ATP and then transported into mitochondria where they are broken down sequentially, two carbon units at a time, by a pathway known as β-oxidation (see Fig. 9). The fatty acyl CoA derivatives undergo oxidation at the carbon that is β to the carboxyl carbon from that of a saturated carbon–carbon bond to that of an oxo-saturated carbon bond. Enzymes that contain FAD or use NAD+ as the electron acceptors catalyze these reactions. As is the case in the oxidation of carbohydrates, the NADH and FADH2 generated by the β-oxidation of fatty acids are converted to their oxidized forms by the mitochondrial electron transport chain, which results in the formation of ATP by oxidative phosphorylation. Once β-oxidation is complete, the terminal two carbons of the fatty acid chain are then released as acetyl CoA. Oxidation and cleavage of the fatty acid continue until it is entirely converted to acetyl CoA. The conversion of a saturated fatty acid with 18 carbon atoms to 9 acetyl CoA produces 8 NADH and 8 FADH2 . The acetyl CoA is burned by the citric acid cycle to generate more ATP. The high caloric content of fats pays off to cells in the yield of ATP.

Bioenergetics

FIGURE 9 Oxidation of fatty acids. Fats and oils are hydrolyzed to form glycerol and fatty acids. CoA derivatives of the fatty acids are oxidized in mitochondria by NAD+ and FAD to β-oxo-derivatives. CoA cleaves these derivatives to yield acetyl CoA and a fatty acid CoA molecule that is two carbons shorter. The process continues until the fatty acid has been completely converted to acetyl CoA. The acetyl moiety is oxidized in the citric acid cycle to CO2 and water. The complete oxidation of a fatty acid of about the same molecular weight of glucose yields four times more ATP than that of glucose.

D. Catabolism of Proteins and Amino Acids In addition to containing carbohydrates and fats, diets may be rich in proteins. The catabolism of proteins results in the generation of their component parts, amino acids. When the dietary amino acid requirements of an individual are

Bioenergetics

107 Although some bacteria carry out photosynthesis without the evolution of oxygen, this article deals solely with oxygenic photosynthesis that takes place in higher plants and algae. In a purely formal sense, oxygenic photosynthesis may be represented as the reverse of the oxidative breakdown of a six-carbon carbohydrate, such as glucose. An equation that describes photosynthesis in part illustrates this relationship: 6CO2 + 12H2 O → C6 H12 O6 + 6O2 + 6H2 O, (5)

satisfied, the excess amino acids in the diet are catabolized to CO2 and water as a source of energy. Some amino acids are degraded to molecules that feed directly into glycolysis, and others result in the production of acetyl CoA. Excess nitrogen resulting from the catabolism of amino acids and other compounds that contain nitrogen is excreted in mammals in urine in the form of the simple organic compound urea. Some amino acids are precursors in the biosynthesis of other organic molecules. E. Summary In summary, organisms such as humans and other animals, many bacteria, fungi, and nongreen plants derive the energy they must have to power life from foodstuffs they obtain from their environment. The degradation of carbohydrates and the oxidation of fats are the major sources of energy for heterotrophic (other feeding) organisms. How these molecules that are essential to life are generated is the next subject considered.

II. PHOTOSYNTHESIS
From a purely thermodynamic standpoint, life is an improbable event. Consider, for example, the complex structures of organisms, not only at the macroscopic level, but also at the microscopic and atomic levels. These ordered structures can be formed and maintained only by the expenditure of energy. Within the ecosystem that we call the earth, the organic nutrients necessary to sustain the life of heterotrophs such as us are provided directly and indirectly by photosynthesis. In both quantitative and qualitative terms photosynthesis is the most significant biological process on Earth. Approximately 2 × 1011 tons of carbon dioxide are converted to organic compounds each year. It is to photosynthesis in prehistoric times that we owe the reserves of fossil fuels. The oxygen that we breathe is a direct result of photosynthesis, now and in prehistory. If the earth were an isolated system in a thermodynamic sense, life would be in jeopardy in that the energy reserves for life would be consumed. Without the input of energy from a source external to the earth, the planet must tend toward achieving equilibrium within its environment. Fortunately, the earth is not an isolated system. The hydrogen fusion reactor of the Sun bathes our planet in electromagnetic radiation, including visible light. A fraction of the solar energy that impinges on Earth is converted by photosynthesis to chemical energy in the form of organic molecules that heterotrophic organisms use to satisfy their continued need for energy. The process by which light energy is used to drive the otherwise unfavorable synthesis of these organic molecules is called photosynthesis.

where C6 H12 O6 refers to a six-carbon sugar. This equation in reverse describes the oxidative catabolism of a sixcarbon sugar such as glucose. Under standard conditions, the complete oxidation of glucose liberates 686 kcal/mol; the synthesis of a mole of glucose from carbon dioxide and water thus minimally requires the input of an equivalent amount of energy. In photosynthesis, visible light provides this energy. When it is considered that the only source of carbon for the tens of thousands of organic compounds synthesized in green plants is from the assimilation of carbon dioxide by means of photosynthesis, the inadequacy of Eq. (5) to describe photosynthesis, despite its usefulness, is readily apparent. Inspection of Eq. (5) reveals that photosynthesis is an oxidation–reduction process. Simply put, photosynthesis is the light-driven reduction of carbon dioxide to the oxidation–reduction level of a carbohydrate by using water as the electron and hydrogen donor. In the process, water is oxidized to molecular oxygen. As stated previously, water is a very poor reducing agent. However, water at an effective concentration of 55 M is readily available in the biosphere. Although organic compounds and inorganic molecules such as hydrogen sulfide are more powerful reducing agents than water is, their use in photosynthesis as the source of electrons for photosynthesis is restricted to certain species of bacteria. The thermodynamically very unfavorable reduction of carbon dioxide by water is driven by light. A. Light Reactions How the electromagnetic energy of light is converted to chemical energy in the form of reduced organic molecules is complex. Nonetheless, the first principles of energy conservation and conversions in photosynthesis may be simply depicted. All higher photosynthetic organisms contain two forms of the green pigment chlorophyll. More than 99% of the chlorophyll in chloroplasts, the organelles in which photosynthesis takes place, functions in a passive, purely physical manner. Organized in specific pigment– protein complexes within the photosynthetic membrane, these chlorophylls absorb visible light and transfer excitation energy to nearby chlorophylls with efficiencies very close to 100%. In a real sense, more than 99% of the

108 chlorophylls function only to gather light and as such they are often referred to as light-harvesting chlorophylls. Within picoseconds of the harvesting, the excitation energy is transferred to specialized chlorophyll molecules called reaction center chlorophylls. These reaction center chlorophylls are identical to the majority of the lightharvesting chlorophylls. Yet, rather than acting in a passive manner when they are excited, the reaction center chlorophylls perform photochemistry. The two reaction center chlorophylls are termed P700 and P680. The “P” stands for pigment and the numbers refer to their absorption maxima, in nanometers, in the red region of the spectrum. The reaction center chlorophylls were first detected by lightinduced bleaching at 680 and 700 nm. When the reaction center chlorophylls are excited, either directly or by resonance energy transfer from excited light-harvesting chlorophylls, an electron is transferred from the reaction center chlorophyll ensemble to an electron acceptor. These light-driven oxidation–reduction reactions occur within picoseconds and can operate with a quantum efficiency that is close to 100%. The reactions may be written as follows: P700∗ + FeS → P700+ + FeS− and P680∗ + Q → P680+ + Q− , (7) (6)

Bioenergetics

where the asterisks indicate the first excited singlet state of the reaction center chlorophyll, and FeS and Q are the redox active part of an iron–sulfur protein and a quinone, respectively, the first stable electron acceptors. P700+ and

P680+ are chlorophyll cation radicals and Q− is a half reduced quinone and FeS− is a reduced iron-sulfur protein. The reactions shown in Eqs. (6) and (7) cannot take place, in the direction shown, in the dark when the reaction center chlorophylls are in the unexcited, ground state. The G 0 for both these reactions is approximately +24 kcal/mol. The excited reaction center chlorophylls are, however, much stronger reducing agents than the ground state chlorophylls are. The E 0 of P700∗ is about 1.3 V more reducing than that of P700 in the ground state. These two electron transfer reactions are the only lightdriven reactions in photosynthesis and they set the entire process in motion. The electron transport chain of chloroplasts is illustrated in Fig. 10. Specific light-harvesting chlorophyll–protein complexes are associated with the reaction center chlorophyll– protein complexes in assemblies known as photosystems. Photosystem I (PS I) contains P700 and the FeS acceptor, and photosystem II (PS II), P680 and the quinone acceptor. Electron transfer in PS I generates a relatively weak oxidizing agent (P700+ , E 0 = +430 mV) and a strong reductant (FeS− , E 0 = −600 mV). The primary reductant generated in photosynthesis is nicotinamide adenine dinucleotide phosphate (NADP+ ), which, as the name suggests, differs from NAD+ by a single phosphate. While the physical properties of NADP+ and NAD+ are very similar, enzymes that use these pyridine nucleotides as substrates can discriminate between them by at least a factor of 1000. In general NAD+ is used in catabolic metabolism as we have seen for glycolysis and the tricarboxylic acid cycle. The reduced form of NADP+ , NADPH, is, in contrast,

FIGURE 10 Electron transport and ATP synthesis in chloroplasts. The jagged arrows represent light striking the two photosystems (PS I and PS II) in the thylakoid membrane. Other members of the electron transport chain shown are a quinone (Q), the cytochrome complex (b 6 f ), plastocyanin (PC), and an iron–sulfur protein (FeS). The chloroplast ATP synthase is shown making ATP at the expense of the electrochemical proton gradient generated by electron transport.

Bioenergetics

109 phosphate at the 1-position of the bisphosphorylated sugar acid is transferred to ADP to form ATP. The conversion of 3-phosphoglycerate to carbohydrates occurs by a pathway that is essentially the reverse of glycolysis. It must be emphasized, however, that glycolysis and photosynthetic carbon metabolism take place in separate intracellular compartments. Glycolysis occurs in the cytoplasm and uses NAD+ as the electron acceptor. The photosynthetic reduction of 3-phosphoglycerate occurs inside chloroplasts in the aqueous space known as the stroma. The enzymes in the two compartments are not the same even though they catalyze similar reactions. For example, the triose phosphate dehydrogenase in the cytoplasm is very specific for NAD+ , whereas that in the chloroplast stroma is equally specific for NADPH. Therefore, ATP is required for the reduction by NADPH of 3-phosphoglycerate to the oxidation level of a carbohydrate: ATP + 3-phosphoglycerate → ADP + 1,3-bisphosphoglycerate, (9)

used in biosynthesis, or anabolic metabolism. The E 0 of the NADP+ –NADPH redox pair is −340 mV. Thus, electron transfer from the reduced iron–sulfur protein of PS I to NADP+ is energetically a very favorable spontaneous reaction. It is NADPH that provides the electrons for CO2 reduction. The ultimate electron donor is water. Two water molecules are oxidized by PS II to yield four protons and molecular oxygen. Water is a very weak reducing agent. Thus, a strong oxidizing agent is needed for water oxidation. P680+ fits the bill. The midpoint potential of the P680+ –P680 redox pair is on the order of +1 V. Since the water–oxygen redox couple has an E 0 of +0.815 V, the oxidation of water by P680+ is an energetically spontaneous reaction. Water oxidation is catalyzed by a manganese-containing enzyme that is plugged into the energy-converting thylakoid membrane. So far, we have seen that the reduced FeS protein of PS I is converted to its oxidized form by passing electrons eventually to NADP+ . In PS II, P680+ is reduced to P680 with electrons extracted from water. For electron transport to continue, the electron acceptor of PS II, Q− , and the electron donor of PS I, P700+ , must be oxidized and reduced, respectively. The redox potential of the Q–Q− couple is about +0.05 V, whereas that of P700+ –P700 is near +0.450 V. Thus, electron transport from Q− to P700+ is energetically spontaneous with a free energy of 9.3 kcal/mol for each electron transferred. Electron transport from Q− to P700+ is mediated by a quinone, iron–sulfur, and a cytochrome protein complex in the thylakoid membrane. This protein, the cytochrome b6 f complex, is remarkably similar to the cytochrome bc1 complex of the mitochondrial electron transport chain. B. CO2 Reduction Linear electron transport in oxygenic photosynthesis is the reduction of NADP+ to NADPH by water, which results in the formation of molecular oxygen: 2H2 O + 2NADP+ → O2 + NADPH + 2H+ . (8)

and the bisphosphoglycerate is in turn reduced by NADPH: NADPH + H+ + 1,3-bisphosphoglycerate → NADP+ + Pi + glyceraldehyde 3-phosphate. (10)

NADPH is incapable of reducing CO2 by itself; ATP is also required. The CO2 acceptor in photosynthesis is the fivecarbon, phosphorylated sugar ribulose 1,5-bisphosphate. CO2 cleaves this sugar into 2 mol of the three-carbon sugar acid 3-phosphoglycerate, a compound that is also an intermediate in glycolysis. The enzyme that catalyzes this reaction, ribulose 1,5-bisphosphate carboxylase/oxygenase, or rubisco, is present in very high concentrations within chloroplasts, which makes it among the most abundant proteins in the biosphere. Recall that in glycolysis one of the two steps in which ATP is formed is the conversion of 1,3bisphosphoglycerate to 3-phosphoglycerate. The acyl

Since two 3-phosphoglycerates are generated for each CO2 assimilated, two NADPH and two ATP are required for reduction. This reaction is the only one in photosynthetic carbohydrate metabolism that is an oxidation– reduction reaction. Glyceraldehyde 3-phosphate is a sugar phosphate and may be readily converted within chloroplasts to many sugars and the glucose polymer starch. Some of the glyceraldehyde 3-phosphate is used in a complex series of reactions to regenerate the five-carbon acceptor of CO2 , ribulose 1,5-bisphosphate. In the process, one phosphate is cleaved from one of the sugar phosphate intermediates. Thus, ribulose 5-phosphate, the product of the cycle, must be phosphorylated by using ATP as the phosphoryl donor. As a consequence, three ATP and two NADPH are required for each CO2 taken up. Photosynthesis must satisfy the energy requirements of all living tissues in plants, including roots, stems, and developing fruit. Up to 75% of the triose phosphate formed is exported from the chloroplasts in leaf cells to the cytoplasm where it is converted to sucrose, a major product of photosynthesis. In most plants, sucrose is transported to the rest of the plant where it is either stored as starch or broken down by glycolysis and the citric acid cycle in exactly the same way as it is in animals to produce the ATP needed to sustain life.

110 C. ATP Synthesis ATP synthesis in chloroplasts is called photophosphorylation and is similar to oxidative phosphorylation in mitochondria. The light-driven transport of electrons from water to NADP+ is coupled to the translocation of protons from the stroma across the thylakoid membrane (the green, energy-converting membrane) into the lumen. Electron transport from Q− to P700+ is exergonic. Part of the energy released by electron transport is conserved by the formation of an electrochemical proton gradient. The cytochrome b6 f complex of chloroplasts functions not only in electron transport, but also in proton translocation. The active site of the oxygen-evolving enzyme is arranged so that the protons formed during water oxidation are released into the thylakoid lumen. These protons contribute to the electrochemical proton potential. The thylakoid membrane contains a protein that functions to transport Cl− across the membrane. Proton accumulation in the thylakoid lumen is electrically balanced in large part by Cl− uptake. As a result, thylakoids accumulate HCl and the membrane potential across the membrane is low. The pH inside the lumen during steady-state photosynthesis is about 5.0. One of the earliest experiments that supported the hypothesis that ATP synthesis and electron transport were linked by the electrochemical proton potential was carried out with isolated thylakoid membranes. Thylakoid membranes were placed in a buffer at pH 4.0 and after a few seconds the pH was rapidly increased to 8.0, which resulted in the formation of a proton activity gradient. This artificially formed gradient was shown to drive the synthesis of ATP from ADP and Pi . The experiments were carried out in the dark so that the possibility that electron transport contributed to the ATP synthesis was excluded. Thus, a proton activity gradient was proven capable of driving ATP synthesis. The thylakoid membrane enzyme that couples ATP synthesis to the flow of protons down their electrochemical gradient is called the chloroplast ATP synthase (see Fig. 10). This enzyme has remarkable similarities to ATP synthases in mitochondria and certain bacteria. For example, the β subunits of the chloroplast ATP synthase have 76% amino acid sequence identity with the β subunits of the ATP synthase of the bacterium E. coli. The reaction catalyzed by ATP synthases is nH+ + ADP + Pi + H+ → nH+ + ATP + H2 O, (11) a b where n is the number of protons translocated per ATP synthesized, probably three or four, and a and b refer to the opposite sides of the coupling membrane. Provided the electrochemical proton potential is high, the reaction is poised in the direction of ATP synthesis. In principle,

Bioenergetics

when the proton potential is low, ATP synthases should hydrolyze ATP and cause the pumping of protons across the membrane in the direction opposite that which occurs during ATP synthesis. ATP-dependent proton transport by the ATP synthase is of physiological significance in E. coli under anaerobic conditions in that it generates the electrochemical proton potential across the plasma membrane of the bacterium. This potential is used for the active uptake of some carbohydrates and amino acids. In contrast, ATP hydrolysis by the chloroplast ATP synthase in the dark has no physiological role and would be wasteful. In fact, the rate of ATP hydrolysis by the ATP synthase in thylakoids in the dark is less than 1% of the rate of ATP synthesis in the light. Remarkably, within 10–20 msec after the initiation of illumination, ATP synthesis reaches its steady-state rate. Thus, the activity of the chloroplast ATP synthase is switched on in the light and off in the dark. In addition to being the driving force for ATP synthesis, the electrochemical proton potential is involved in switching the enzyme on. Structural perturbations of the enzyme induced by the proton potential overcome inhibitory interactions with bound ADP as well as with a polypeptide subunit of the synthase. An additional regulatory mechanism that is unique to the chloroplast ATP synthase is reductive activation. Reduction of a disulfide bond in a subunit of the chloroplast ATP synthase to a dithiol enhances the rate of ATP synthesis, especially at physiological values of the proton potential. The electrons for this reduction are derived from the chloroplast electron transport chain.

III. ORIGIN OF MITOCHONDRIA AND CHLOROPLASTS
In animal, yeast, and fungal cells, DNA is present in two organelles, the nucleus and the mitochondria. In plant and algal cells, DNA is present in plastids (of which chloroplasts are one example) as well as in mitochondria and the nucleus. Unlike the DNA in the nucleus, which is packaged into chromosomes, plastid DNA and mitochondrial DNA are circular and thus resemble the DNA in prokaryotes (e.g., bacteria). Mitochondrial DNA is small and codes for relatively few mitochondrial proteins. Although mitochondria contain their own protein synthesis machinery, the majority of the hundreds of mitochondrial proteins are coded for by nuclear genes. These proteins are synthesized in the cytoplasm and imported into the mitochondria. Plastid DNA is somewhat larger than that of the mitochondrion and contains the genetic information for more chloroplast proteins. However, as is the case for mitochondria, most of the proteins in a chloroplast are coded by nuclear genes

Bioenergetics

111 carbon sugar) on the third position of the glycerol. Galactosyldiglycerides are absent in the membranes of animal, yeasts, and fungi but are present in the photosynthetic membranes of all organisms that carry out oxygenic photosynthesis. The lipid compositions of mitochondrial and chloroplast membranes are consistent with the engulfment hypothesis for the origin of these organelles.

and are synthesized in the cytoplasm. Proteins destined for mitochondria and chloroplasts have an extension on their N -terminal end that targets the proteins to the correct organelle and to the correct place within the organelle. These extensions, which, like the remainder of the proteins, are composed of amino acids, are usually cleaved off as the proteins find their proper place within the organelle. Remarkably, some proteins composed of more than one polypeptide may contain a polypeptide coded for by nuclear DNA and synthesized in the cytoplasm and another polypeptide that is coded for by mitochondrial or chloroplast DNA. Ribulose 1,5-bisphosphate carboxylase/oxygenase is a prominent example of such a protein in chloroplasts. The discovery that mitochondria and chloroplasts contain DNA, coupled with a wealth of sequence information about both DNA and proteins, added credence to the notion that these organelles arose from the engulfment of unicellular organisms by a primitive nucleated cell. Mitochondria may have been derived from a bacterium, and chloroplasts, from a unicellular alga. After the engulfment events, genes in the bacterium and alga coding for proteins that duplicated those in the nuclear genomes of the hosts were lost and other genes were transferred from the bacterial and algal genomes to the genomes of the hosts. The distribution of proteins and lipids within biological membranes is asymmetric. Thus, one side of a membrane is distinct from the other. The coupling membranes of mitochondria and chloroplasts are opposite to each other. Protons are ejected from mitochondria during respiratory electron transport but are taken up by thylakoids during light-driven electron transport. The catalytic portion of the ATP synthase is located on the outside of the thylakoid membranes, whereas that of the mitochondrial ATP synthase is present on the inside of the inner membrane. As seen in Fig. 7, the orientation of the coupling membranes of mitochondria and chloroplasts is consistent with the hypothesis that these organelles are of bacterial and algal origin. Each membrane in a cell has its distinct set of proteins and lipids. The most common membrane lipids are phospholipids. Phospholipids are diglycerides. Two of the three hydroxyls of glycerol are linked to long-chain fatty acids by ester bonds. The third position is occupied by phosphate. A number of different polar substituents are linked to the phosphate by anhydride bonds. The phospholipid composition of the mitochondrial inner membrane is virtually the same in plant mitochondria as in animal mitochondria and resembles that in the plasma membrane of some bacteria. The lipids in chloroplast membranes are very distinctive. The phospholipid content is unusually low and about 80% of the membrane lipids in thylakoids are diglycerides that have one or two galactose (a six-

IV. ILLUSTRATIONS OF THE USES OF ATP: ION TRANSPORT, BIOSYNTHESIS, AND MOTILITY
ATP powers most of the endergonic processes in cells. How the potential energy of the phosphoanhydride bond of ATP may be used to drive otherwise unfavorable reactions (Fig. 11) is discussed in this section. This discussion focuses on three major uses of ATP: the generation of ion gradients, biosynthesis, and movement. A. Ion Transport The plasma membrane is the barrier that separates the cytoplasm of cells from the exterior medium. All cells maintain a membrane potential that is negative. There is an excess of positive charge in the external medium in comparison with that in the cytoplasm. The membrane potential in plant cells can be as high as −200 mV. Energy is required to generate and maintain the membrane potential. All cells maintain gradients in ions across the plasma membrane. The intracellular K+ concentration is higher than that of the extracellular medium, and the concentration of Na+ , much lower. The free Ca2+ concentration in the cytoplasm is maintained at very low levels, 1000fold or more below the extracellular Ca2+ concentration. Often the intracellular proton concentration can be quite different from that in the medium. The pH in the cytoplasm of plant cells is close to 7.0, whereas that in the medium is about 5.0. Energy is needed to generate and maintain these ionic disequilibria. For example, the energy cost to generate a pH gradient of two pH units is equal to RT ln([H+ ]/[H+ ]), where the subscripts o and i o i stand for outside and inside the cell, respectively. At 25◦ C, the G for a 100-fold proton activity (pH 7.0 in versus pH 5.0 out) gradient is 2.7 kcal/mol. Plasma membranes of all higher organisms contain enzymes that are embedded in the membrane that act as ion pumps. That is, they catalyze the transport of ions against their electrochemical potential. In physiology, transport that is thermodynamically uphill is termed active transport to distinguish it from the spontaneous flow of ions down their electrochemical potential. The energy needed

112

Bioenergetics

FIGURE 11 Uses of ATP. The diagram shows some of the major processes in cells that are powered by ATP hydrolysis.

for the active transport of ions across the plasma membrane is provided by the hydrolysis of ATP to ADP and Pi . As much as 75% of cellular ATP may be consumed simply to generate and maintain ion gradients. The electrogenic ion pump in the plasma membrane of animal cells is the Na+ /K+ -ATPase. As shown in Fig. 12, three Na+ ions are transported out of the cell and two K+ ions are pumped in for each ATP that is hydrolyzed. Since three positively charged ions are exported, but only two imported, the Na+ /K+ -ATPase is electrogenic. The trans

plasma membrane potential is on the order of −50 mV. In addition, the pump keeps the intracellular Na+ concentration nearly 100-fold lower than that in the serum, and the intracellular concentration of K+ , about 30-fold higher than in serum. Indirectly, the Na+ /K+ -ATPase provides the energy for the active transport of amino acids and some carbohydrates into cells. The plasma membrane contains specific proteins that mediate the transport of these molecules in a manner that is obligatorily linked to the cotransport of

FIGURE 12 Some ion pumps in the plasma membrane. The Na+ /K+ -ATPase of animal cells uses the energy of ATP hydrolysis to move three Na+ ions out of the cells and two K+ ions in, which results in the generation of ion gradients and a membrane potential. Plant, yeast, and fungal cells do not have a Na+ /K+ -ATPase, but instead have a H+ -ATPase, as the electrogenic pump. The plasma membrane also contains a Ca2+ -ATPase that pumps Ca2+ out of cells to help keep the intracellular Ca2+ concentration low.

Bioenergetics

113 Inhibitors of the enzyme responsible for the acidification of the stomach are well known and equally well-advertised alleviators of “heartburn.” This enzyme is present in the parietal cells of the stomach and resembles the Na+ /K+ -ATPase. Instead of catalyzing the ATP-dependent exchange of Na+ and K+ , the stomach acid pump excretes H+ into the lumen of the stomach in exchange for K+ . B. Biosynthetic Use of ATP The input of energy in the form of the hydrolysis of ATP to either ADP and Pi or to adenosine monophosphate (AMP) and pyrophosphate powers the synthesis of biological molecules, including, as we have seen, carbohydrates in photosynthesis, proteins, DNA, RNA, and fatty acids. To delve into the role of ATP in biosynthesis in depth is not possible in this brief article. Aspects of fatty acid biosynthesis, however, reveal interesting principles of the energetics of biosynthetic pathways. Fatty acids are oxidized completely to CO2 and water by β-oxidation and the citric acid cycle. Acetyl CoA is the end product of β-oxidation of fatty acids and is the source of carbon for fatty acid biosynthesis. Yet, the pathways for fatty acid degradation and synthesis are so very different that they even occur within different compartments within cells. Fatty acid synthesis takes place in the cytoplasm of animal cells and in the plastids of plant cells, whereas βoxidation is located in mitochondria in both animal and plant cells. Often, the pathway for the synthesis of a compound differs significantly from that for its degradation. Among the reasons that the separation of synthetic and degradative pathways evolved are energetics and regulation. The oxidation of fatty acids to acetyl CoA is very exergonic. It is not feasible on energetic grounds to make fatty acids from acetyl CoA by reversing β-oxidation. Metabolism of carbohydrates and fats is regulated in mammals by a number of hormones, including insulin, glucagon, and epinephrine (adrenaline). Having separate pathways for the degradation and the biosynthesis makes it possible to turn off one pathway while up-regulating another. For example, glucagon and epinephrine selectively stimulate the breakdown of fats and fatty acids, whereas insulin has the opposite effect. The fine control of fatty acid metabolism that has evolved would clearly not be possible without the existence of separate pathways for biosynthesis and catabolism. CO2 is required for the synthesis of fatty acids. Yet, when fatty acid synthesis is carried out in the presence of radioactive CO2 , the fatty acid made is devoid of radioactivity. ATP is used to add CO2 to a precursor, and in a subsequent step in the pathway of fatty acid biosynthesis,

Na+ . Since the extracellular Na+ concentration is higher than that in the cytoplasm and the membrane potential is negative, the Na+ flows from outside to inside the cell. Assuming a membrane potential of −50 mV and a 100-fold Na+ concentration gradient, the flow of Na+ would liberate about 3.8 kcal/mol at 25◦ C. This exergonic flow of Na+ provides the energy needed for the active transport of the amino acid or carbohydrate. Although Na+ flux is the immediate source of energy for the active transport in Na+ -linked transporters, it is important to keep in mind that the ultimate energy source is ATP hydrolysis by the Na+ /K+ -ATPase. Plants, yeasts, and fungi do not contain a Na+ /K+ ATPase in their plasma membranes. Instead, they contain a H+ -ATPase that is the generator of the plasma membrane potential. The H+ -ATPase is structurally and mechanistically related to the Na+ /K+ -ATPase but translocates only H+ . The H+ -ATPase is capable of generating large electrochemical proton gradients. The imbalance in the Na+ and K+ concentrations between the inside and the outside of the plant cell is maintained by other mechanisms that include exchange transport of Na+ for H+ . The active transport of some organic molecules across the plasma membrane of plants, yeasts, and fungi is linked to the cotransport of H+ down its eletrochemical gradient into the cell. An important example of proton-linked transport is that of sucrose loading into the vascular element, the phloem, that transports sucrose from the leaves to the remainder of a plant. The concentration of sucrose in phloem cells near leaves that are actively carrying out photosynthesis can be 0.5 M or higher, whereas that in the intracellular space, just 0.001 M. The energy cost of generating this gradient is 3.7 kcal/mol at 25◦ C. The immediate source of energy is proton flow, and the ultimate source, ATP hydrolysis by the H+ -ATPase. The concentration of free Ca2+ (meaning that unbound to proteins and membrane lipids) in the cytoplasm of cells is normally maintained at a very low level. Under certain circumstances, however, transient increases in the cytoplasmic Ca2+ concentration are triggered. Ca2+ is a major player in the transmission of some hormonally induced signals in plants and animals. Muscle contraction is also induced by release of Ca2+ from internal membranes within muscle cells. The plasma membrane contains an enzyme that catalyzes the export of Ca2+ from the cytoplasm at the expense of ATP hydrolysis. The Ca2+ -ATPase has features that place it in the category of plasma membrane enzymes that also includes the Na+ /K+ -ATPase and the H+ -ATPase. The Ca2+ -ATPase functions to keep the cytosolic Ca2+ concentration low ( k3 . The amplitude of the burst can provide the concentration of active sites in an enzyme preparation. By varying the concentration of S, one can find values for k−1 /k1 , k2 , and k3 . There are many variations on transient kinetics, as will be illustrated in our case studies of individual enzymes.

III. ILLUSTRATIVE EXAMPLES
A. α-Chymotrypsin Alpha-chymotrypsin (Fig. 2) catalyzes the facile hydrolysis of peptide bonds, in particular those adjacent to the carboxyl group of aromatic amino acids (tryptophan, tyrosine, phenylalanine) as well as a variety of esters derived from similar N -acylated amino acids. The enzyme

630

Enzyme Mechanisms

FIGURE 2 The crystal structure of α-chymotrypsin showing the catalytic triad of amino acid side chains. [Adapted from Blevins, R. A., and Tulinsky, A. (1985). “The refinement and crystal structure of the dimer of α-chymotrypsin at ˚ 1.67 A resolution,” J. Biol. Chem. 260, 4264–4275.]

has been the subject of intensive mechanistic study, most of which occurred well before a crystal structure was available. A key insight was provided by studying the enzymecatalyzed hydrolysis of p-nitrophenyl acetate. Transient kinetic studies revealed burst kinetics (Fig. 3) with an initial rapid liberation of p-nitrophenolate followed by a slower steady-state rate. The biphasic time course is consistent with the existence of two intermediates (ES and acyl-E), with the second accumulating owing to its slower breakdown to product. The intermediate is a covalent enzyme species acylated at serine-195 (see Fig. 2), a fact initially revealed by chemically esterifying this enzyme residue specifically and irreversibly with diisopropylphosphorofluoridate. No burst kinetics is seen with amide substrates because the acylation step limits turnover. The same intermediate, however, is formed as shown by partitioning experiments in which an exogenous nucleophile such as hydroxylamine is added to compete with water in the deacylation step. The result revealed equivalent levels of

hydroxamate and acid products formed from either amide or ester substrates derived from a common amino acid, which implicated the presence of the intermediate in both enzyme-catalyzed processes.

FIGURE 3 Plot of the burst in hydrolysis of p-nitrophenyl acetate. The concentration of product is observed as a function of time. [From Fersht, A. (1999). Structure and Mechanism in Protein Science. W. H. Freeman and Company, New York. Used with permission.]

Enzyme Mechanisms

631

FIGURE 4 The pH dependence of kcat /K M and k cat for the α-chymotrypsin-catalyzed hydrolysis of esters and amides. [From Hammes, G. G. (1982). Enzyme Catalysis and Regulation. Academic Press, New York. Used with permission.]

The pH dependence of the steady-state kinetic parameters is shown in Fig. 4 and implicates the ionization of two groups in the free enzyme and one in the ES complex. These data combined again with chemical modification studies (now superseded by site-specific mutagenesis) implicated histidine-57 (pK a ∼ 7) and the N -terminal amino acid isoleucine (pK a ∼ 8.5). The latter forms a salt bridge with aspartate-194 that helps maintain the active structure of the enzyme; the former is involved in general acid–base chemistry at the active site. These data, along with further information derived from the reaction of specific substrates with the enzyme by using stopped-flow methods, led to the elucidation of a kinetic sequence that consistently implicated the acylation and deacylation of Ser195 assisted by His57 and Asp102. The crystal structure of chymotrypsin (Fig. 2) reveals that these three residues form a catalytic triad, a feature repeated for many hydrolytic enzymes. This triad operates within a well-defined binding site that is lined with nonpolar amino acids capable of van der Waals interactions with polypeptide substrates containing aromatic side chains. A plausible mechanism is outlined in Fig. 5 in terms of the chemistry occurring during the individual kinetic steps. The key features of this mechanism require the participation of the serine hydroxyl as a nucleophile whose attack on the carbonyl of the substrate is facilitated through proton abstraction by the imidazole nitrogen of His57 and its redonation to the amine-leaving group. Deacylation of the enzyme follows general base catalysis of water attack again by His57 and the return of the enzyme to its resting state. Catalysis of the chemical process through the participation of the side chains of an enzyme in proton, hydride, and electron transfer is a hallmark of enzyme catalysis and

can occur efficiently in the confines of the active site owing to the optimal alignment and juxtapositioning of the substrate for chemical reaction. B. Dihydrofolate Reductase Dihydrofolate reductase (DHFR) catalyzes the reduction of 7,8-dihydrofolate (H2 F) by nicotinamide adenine dinucleotide phosphate (reduced form) (NADPH) to form 5,6,7,8-tetrahydrofolate (H4 F), a key step in furnishing the parental cofactor needed for de novo pyrimidine and purine biosynthesis. The enzyme has been the target of antitumor and antimicrobial drugs. A complete kinetic scheme (Fig. 6) obtained primarily through transient kinetics has been described for the enzyme from Escherichia coli as well as other sources and provides a second case study as to how to define the catalytic process. Measurement of the rates of binding and dissociation of substrate and cofactors provided valuable insights into the identity of rate-limiting kinetic steps in the scheme shown in Fig. 6. Two procedures were used. In the first, direct observation of changes in the intrinsic enzyme or NADPH fluorescence upon ligand binding showed that the addition of ligand was biphasic in accord with the existence of two conformers, of which only one bound the ligand: DHFR1 k2 k−2

DHFR2 + L

k1 k−1

DHFR2 · L .

The rate of the initial fast phase and its amplitude are associated with the binding of L to DHFR2 (k1 , k−1 ) and the level of DHFR2 ; the rate of the second phase is the conversion of DHFR1 to DHFR2 (k2 ). The method was extended to the binding of a second ligand to binary

632

Enzyme Mechanisms

FIGURE 5 The mechanism of amide hydrolysis by α-chymotrypsin. [From Fersht, A. (1999). Structure and Mechanism in Protein Science. W. H. Freeman and Company, New York. Used with permission.]

DHFR2 · L complexes and revealed that the binding of various ligands was near the diffusion-controlled limit. In the second procedure a competitive trapping technique was employed in which the enzyme–ligand complex is mixed with an excess of a second ligand that competes for the binding site. With this method, k−1 is measured accurately when k T [T ] k1 [L 1 ], k−1 , and k−T . This DHFR · L 1 k−1 k1

DHFR + L 1

k T [T ] k−T

DHFR · T.

procedure identified a preferred pathway for dissociation of the product H4 F as the rate-limiting step in the steady-state cycle. The assistance of the cofactor NADPH

in promoting product dissociation is an unusual feature, though not limited to DHFR, and follows the rapid loss of NADP+ . Events around the chemical step of reduction/oxidation were monitored by directly observing the conversion of NADPH to NADP+ . The kinetics are again biphasic owing to the rapidity of the hydride transfer process; that the rapid phase is associated with the chemical step is verified by the observation of a kinetic deuterium isotope effect of 3 when the transferring hydrogen of the NADPH is replaced with deuterium. This step shows a pH dependence with a pK a of 6.5 that implicates the Asp125 (27 in E. coli) in the proton transfer events required to complete the reduction.

FIGURE 6 The kinetic scheme for conversion of H2 F to H4 F by DHFR, including the rate constants for each step at 25◦ C. In this scheme, NH represents NADPH and N+ represents NADP+ .

Enzyme Mechanisms

633 DHFR protein molecule. The protein fold through its complex vibrational modes apparently may couple some set of motions to a promotional vibration that fosters passage of the reactive ternary complex over the activation barrier. C. Phosphate Transfer Enzymes that catalyze the transfer of a phosphoryl moiety between two substrates have provided excellent examples of the use of isotopes in kinetic and stereochemical studies. The enzyme hexokinase, which promotes the conversion of glucose plus ATP to glucose-6-phosphate and ADP has been the subject of kinetic studies that suggested an ordered kinetic sequence with glucose being the first substrate to add and glucose-6-P the last product to be released. Specific information on the identity of rate-limiting steps and the steady-state levels of reaction intermediates was obtained by isotope trapping studies. In its simplest form, enzyme and isotopically labeled substrate (S ∗ ) are incubated (the pulse) and rapidly diluted into excess unlabeled substrate (the chase), and allowed to react for a chosen time. Then the reaction is stopped by a quenching reagent that jumps the pH or denatures the enzyme. From the amount of E · S ∗ converted to product versus that lost to dissociation (replacement by S gives nonlabeled product) the dissociation rate of S ∗ from E and other ES complexes can be calculated. This method has been used in the study of the partitioning of ES complexes in the steady state. In the case of hexokinase, the question was the partitioning of the functional E · glucose · ATP complex between product formation and substrate release. For glucose the relevant scheme is E + Glc∗
Glc koff ∗

Measurement of this step in the reverse direction (i.e., for DHFR · NADP+ · H4 F) coupled with determination of the overall equilibrium constant permitted construction of Fig. 6. The kinetic scheme served as the basis for the explanation of the contribution of various elements of the protein to its function. Site-specific mutagenesis is a technique in which one or more amino acids are replaced by other amino acids through alteration of the gene encoding the enzyme. For the mutant proteins, the same kinetic scheme was reconstructed to calculate the free energy differences arising from changes in the kinetic steps caused by the mutations. Replacing the hydrophobic residues such as Phe30 and Leu54 (Fig. 7) singly or pairwise with other amino acids revealed that the cumulative effect of two mutations was generally nonadditive in terms of the free energy associated with individual steps in Fig. 6, consistent with long-range interactions across the enzyme active site mediated by bound substrate and cofactor. The nonadditivity differed for each step in Fig. 6, which implicated differing conformations of the protein as arising throughout the catalytic cycle. Of particular interest was the discovery that changes in the amino acid sequence at loci outside the active site also strongly influence (by a factor of >102 ) the rate of the chemical step. In combination with dynamic NMR measurements and molecular mechanics calculations, this observation has been attributed to the importance for catalysis of long-range motions that occur across the entire

E · Glc∗

ATP koff

E · Glc∗ · ATP

kc k−c

E · Glc∗ -6-P · ADP

ADP koff

E · Glc∗ -6-P + ADP

E·

FIGURE 7 Crystal structure of DHFR from Lactobacillus casei with methotrexate (a strong inhibitor) and NADPH bound. Amino acid residues discussed in the text are labeled. [Adapted from Bolin, J. T. et al. (1982). “Crystal structures of Escherichia coli and ˚ Lactobacillus casei dihydrofolate reductase refined at 1.7 A resolution,” J. Biol. Chem. 257, 13650–13662.]

In this case the reaction is allowed to reach steady-state turnover, and the solution is either stopped by quench or chased by addition of excess unlabeled substrate followed by a delay sufficient for several turnovers then addition of quench. The presence of a difference in the level of the labeled product obtained by the two procedures represents the concentration of E · Glc · ATP∗ complex in the steady state, which is approximately 50% of E T , the total enzyme concentration. The observed steady-state and pretransient rates are consistent with steps kc and k−c being ADP at equilibrium relative to koff , which is typical for many phosphotransfer enzymes in which the chemical steps are generally not rate limiting. Additional information can be obtained by using the label in the second substrate (i.e., [γ -32 P]ATP) and following a similar protocol, which thereby allows calculation of the dissociation rate of ATP

634 from E · Glc · ATP. In this case E · Glc · ATP∗ is approxiATP mately 25% of E T , which requires that koff compete with ADP the dissociation of ADP (koff ) from E · Glc-6-P · ADP. In this manner the individual rate constants for hexokinase were largely determined and the order of substrate association was verified.

Enzyme Mechanisms

taining an 18 O label in the βγ bridging oxygen provides the necessary probe for finding this intermediate by means of the process below. In these experiments, isotopes are used as labels so that the fate of a particular atom may be followed throughout the course of the reaction.

O Ad O P O

O O P O O
18O

O P O O O P O O E Ad O P O
18O

O

Glu

O P O
18O

O O Glu E Ad O P O O
18O

O O P O
18O

E

Ad O P O

GluP

O P
18O

P O

O

Glu

E

O Ad O P O

18O

O
18O

O P
18O

P O

O

Glu

Isotopic labeling has also been cleverly used to demonstrate the existence of enzyme-bound intermediates that do not readily dissociate into solution. The enzyme glutamine synthetase catalyzes the formation of glutamine from ATP and ammonia possibly through a tightly bound glutamyl phosphate intermediate.
COOH CH2 CH2 H2N COOH H2N ATP NH3 CONH2 CH2 CH2 COOH ADP

If the formation of glutamyl phosphate were reversible and occurred in the absence of ammonia, then the presence of a symmetric torsion motion at the cleavage site might be used to
O C CH2 E H2N CH2 ATP E H2N OH O C CH2 CH2 ADP OPO3

The appearance of 18 O in the nonbridging oxygens of the β-phosphate can be measured by mass spectrometric and NMR methods. The extent of equilibration is partially inhibited by the presence of ammonia as required if glutamyl phosphate is a reaction intermediate. Isotopic labeling studies of phosphotranferase reactions culminated in the synthesis of ATP chiral at the γ -phosphorus. Chirality was achieved by the synthesis of [γ -16 O,17 O,18 O]ATP of one configuration, and the analysis of its chirality was achieved by stereochemically controlled transfer of the γ -phosphoryl moiety to (S)propane-1,2-diol where the absolute configuration was determined by a chemical/mass spectrometric sequence. The observation of inversion of configuration has been accepted as evidence of an “in-line” displacement mechanism at phosphorus by the two bound substrates; the observation of retention of configuration was used to implicate the existence of a phosphoryl enzyme intermediate in the phosphoryl transfer process. For hexokinase, our case study, the finding is one of inversion, consistent with a direct transfer mechanism. D. Triosephosphate Isomerase Triosephosphate isomerase (TIM) catalyzes the interconversion of D-glyceraldehyde-3-phosphate (G3P) and dihydroxyacetone phosphate (DHAP). The equilibrium

COOH

COOH

detect an isotopic exchange brought about by glutamyl phosphate formation. The synthesis of γ -18 O4 P-ATP con-

Enzyme Mechanisms

635
OH H 3H O
2 OPO3

lies far to the side of DHAP, hence the longer arrow pointing to that compound. The enzyme operates with a turnover number of ∼107 s−1 , which is nearly as fast as the diffusion-controlled limit. TIM is therefore called an almost perfectly evolved enzyme because no catalytic refinement could make the rate faster than it already is.
O H OH G3P OPO3
2

a small amount of O3 H 2 OPO3 H OH

oxidation

O3 H O OH

2 OPO3

HO O DHAP

OPO3

2

Many tools have been used to study the TIM mechanism, including X-ray crystallography and NMR, site-directed mutagenesis, and affinity labeling. Strong evidence for the mechanism, however, was supplied by studies using isotopic labeling of substrates. It was found that if the above reaction was carried out in tritiated water, one atom of tritium was stereospecifically incorporated into DHAP.
O H OH
2 OPO3
3

This result suggested that a single base was involved in the proton abstraction/proton addition step. If more than one base were involved, the chance that any label would not be washed out by the solvent and would be added to the deprotonated intermediate would be vanishingly small. In combination with other kinds of experiments, isotopic labeling was therefore invaluable in elucidating the mechanism of triosephosphate isomerase (shown below) in which B is a protein-derived base.
O H OH B B H OPO3
2

H O OH B OPO3
2

HO O

OPO3

2

H 2O

OH H 3H O OPO3
2

This result suggested that a base abstracts a proton from the substrate, and the proton then undergoes exchange with labeled protons from the solvent before being added back to form the product stereospecifically. The existence of a cis-enediol intermediate (shown below) would account for these observations, if the enzyme added the proton back to the same face of the enediol that it was abstracted from.

Isotopes can be used in another way to measure the energy barrier heights for various steps in the catalytic mechanism as noted above for the reaction catalyzed by dihydrofolate reductase. For example, if a proton transfer is involved in the rate-limiting step, then substitution of that proton with one of the heavier isotopes of hydrogen (deuterium or tritium) will cause the step to proceed more slowly. These so-called kinetic isotope effect experiments in combination with steady-state rate measurements in the case of TIM allowed the elucidation of the rate constants for partitioning of the cis-enediol intermediate and construction of a detailed kinetic scheme as shown above for dihydrofolate reductase. E. Aspartate Aminotransferase Many enzymes employ exogenous molecules known as cofactors to assist in executing their chemistry. Sometimes these cofactors are covalently bound to the enzyme and sometimes not. Many types of cofactors are known, and here we will focus on a well-studied example called pyridoxal phosphate (PLP), which often participates in the metabolism of amino acids. PLP, derived from vitamin B6 , is a covalently bound cofactor; it is attached to lysine residues by means of a Schiff base or imine linkage as shown at right.
(CH2)4 H O2 PO 3 N H PLP O OH CH3 O2 PO 3 N H H N H O CH3 Enzyme

OH H O CH2OPO3
2

One unresolved question was whether only a single protein base was involved (so that the transfer was from substrate to base and directly back to form product) or whether a different base was responsible for protonation as part of a more extensive proton relay. The nature of the protein base was explored by doing a similar experiment to the one described above but in the other direction; that is, by labeling the DHAP and observing its conversion to G3P. Although the equilibrium lies far to the side of the DHAP, trapping by irreversible oxidation by G3P dehydrogenase of any G3P formed was used to convert significant quantities of DHAP. If the DHAP was labeled at C1, a small but measurable amount of the label was transferred to C2.

enzyme-bound PLP

636 The substrates for most PLP-requiring processes are α-amino acids, and most of the processes take place at the α-carbon position, although some take place at the β- or γ -carbon. The enzymes which use PLP catalyze a wide range of reactions, including racemizations, decarboxylations, and amine transfers. In general, for all three of these classes of reactions at the α-carbon the substrate displaces the lysine and forms an aldimine intermediate with the PLP.
R H CO2 NH2
2

Enzyme Mechanisms

Lys-Enzyme N H O N H CH3
2

R N O3PO N H

H CO2 H O NH2

Lys-Enzyme

displaced lysine), with the pyridine ring of PLP acting as an electron sink. For the racemases, a proton is then delivered to the opposite face from the same or a different basic residue with the net result of inversion of configuration at the α-carbon. Attack of the active site lysine effects product release and regenerates the cofactor. The structure of one PLP-utilizing transaminase, aspartate aminotransferase, is shown in Fig. 8. This enzyme catalyzes the reversible transamination reaction shown below.
O CO2 H3N CO2 CO2 CO2 -ketoglutarate aspartate CO2 glutamate oxaloacetate H3N CO2 O CO2 CO2

O3PO

CH3

The now very acidic α-proton of the amino acid is abstracted by a basic amino acid residue (often the

FIGURE 8 Structure of an aspartate aminotransferase. The protein is a homodimer, with one covalently bound pyridoxal phosphate (shown in black) in each of the two subunits. The expanded view shows the cofactor in greater detail. [Adapted from Rhee, S. et al. (1997). “Refinement and comparisons of the crystal structures of pig cytosolic aspartate aminotransferase and its complex with 2-methylaspartate,” J. Biol. Chem. 272, 17293–17302.]

Enzyme Mechanisms

637 At the core of cytochrome P450 is an iron porphyrin, or heme, group, protoporphyrin IX, which is depicted below.
COO

In the transamination reaction, formation of the aldimine intermediate between aspartate and PLP and its deprotonation proceeds as described above for the racemases. However, reprotonation occurs not at the same carbon as in the racemization mechanism but at a position adjacent to the PLP heterocycle.
O2C H
2

H CO2 N H O N H
H H
2

O 2C H O3PO H N

H CO2 H O N H
H2 O

O 2C O
2

CO2 NH2 O N H

N N Fe N N

COO

O3PO

O3PO

Hydrolysis releases the product oxaloacetate and generates a new form of the cofactor called pyridoxamine. The reverse reaction is then carried out on the other substrate, α-ketoglutarate, forming glutamate and regenerating the PLP cofactor. Reactions at the β-position (for example, in threonine dehydatase) or the γ -position (in methionine-γ -lyase) also proceed by means of formation of an aldimine intermediate with the α-carbon of an α-amino acid. Such a survey of PLP-dependent enzymes illustrates the important point that one cofactor can be used for different kinds of transformations. The reactions described all go through a common aldimine intermediate, with the ultimate course of the reaction being controlled by the appropriate substrate specificity and positioning of amino acid side chains. This flexibility allows nature to expand its chemical repertoire with a relatively small set of cofactors. There are other organic cofactors such as thiamine pyrophosphate and biotin that participate in carbon– carbon bond formation and cleavage, cofactors that participate in reduction/oxidation, or redox, reactions such as nicotinamide and flavin moieties discussed in some of the earlier examples, and still others that are metal based such as vitamin B12 and porphyrin, which is our next topic. F. Cytochrome P450 A different kind of cofactor from PLP is responsible for the chemistry of cytochrome P450 (Fig. 9), an enzyme which oxidizes hydrocarbons. It is known as a mixed-function oxidase, or monooxygenase, because one oxygen atom from molecular oxygen is incorporated into the product while the other goes on to form water. Cytochrome P450 in the liver, for example, oxidizes and detoxifies many kinds of substances that would otherwise be poisonous. One such well-studied reaction, the hydroxylation of camphor, is depicted below.

O

O2, 2H , 2e

O OH

H 2O

In the structure in Fig. 9, the iron porphyrin is shown in black. Cytochrome P450 is a redox catalyst. The multiple available oxidation states allow the cofactor to accept and donate electrons during different stages of the catalytic cycle. Since the early 1970s, this enzyme and its relatives have been the subject of intense study by, among others, enzymologists, toxicologists, biophysical chemists, and inorganic chemists. The latter have tried to model the chemistry of cytochrome P450 with synthetic small molecules with the twin goals of mimicking its activity and understanding how the enzyme itself works. Working in parallel, biophysical chemists and enzymologists have performed many steady-state and pretransient kinetic studies such as the ones already discussed, which have contributed to a working model for the mechanism shown in Fig. 10. Although this mechanism is in some senses more complicated than those that we have discussed, the same concepts apply. Starting at the top of the cycle, in the resting state of the enzyme the iron is in the +3 oxidation state and is bound by water. Substrate docks to its specific binding site and displaces water to start the catalytic cycle, and an electron is then introduced to reduce the iron to the +2 oxidation state. The dashed line is meant to indicate association of the substrate with the active site, not an actual bond to the iron. The requirement that substrate bind before reduction occurs is a control feature which prevents formation of very active and potentially damaging species in the absence of substrate. Oxygen then binds and accepts an electron from the iron, and introduction of another electron and two protons allows one atom of dioxygen to be released as water, which leaves behind a very active high valent (formally iron 5+) species. What follows is known as a radical rebound step. A hydrogen atom is removed from the substrate and transferred to the terminal oxygen atom, which produces a substrate radical. The radical recombines with the new hydroxo moiety to form the hydroxylated product, which

638

Enzyme Mechanisms

FIGURE 9 The cytochrome P450cam structure. The bound heme is depicted in black, and the iron atom at the center of the heme appears as a sphere. [Adapted from Poulos, T. L., Finzel, B. C., and Howard, A. J. (1986). “Crystal structure of substrate-free Pseudomonas putida cytochrome P450,” Biochemistry 25, 5314–5322.]

is then displaced by water; this completes the catalytic cycle. One important line of investigation which has supported the radical rebound hypothesis is the use of radical clock substrate probes. These probes rearrange in a diagnostic way on a very rapid and calibrated time scale when a hydrocarbon radical is formed. In the case of P450, rearranged products have been isolated after oxidation and have been used as evidence of an intermediate substrate radical. In this way, even though the lifetime of the radical is too short for it to be observed directly, its character can be explored by the judicious choice of substrate analogues. Mechanistic proposals are under constant scrutiny and revision, and aspects of the foregoing mechanism have been challenged. In particular, the possibility has been suggested that a species other than a high valent iron-oxo (likely a hydroperoxo species) may be the active oxidant for some substrates. Debates such as these are a great strength of the study of enzyme mechanisms. Given all the tools which have been developed in this field, and the wealth of interesting problems to which these tools can be applied, the study of enzyme mechanisms should be considered a vital and evolving process. The answer to

the question of “how enzymes work” cannot be described fully in a single scheme.

IV. ORIGINS OF THE CATALYTIC EFFICIENCY OF ENZYMES
The source of the stereospecificity of enzyme-catalyzed reactions is clearly revealed by the fit of the substrate to the enzyme’s active site that spatially then directs the stereochemical course of the chemical events. The speed of these reactions has been attributed to the lowering of the activation energy for the process by the greater affinity of the enzyme for the transition state than that for the substrate. Although this proposal is an adequate rationale, it is often a necessary thermodynamic statement that does not offer insights into how the activation barrier is actually lowered. The preorganization of substrate and active site residues within a protein cavity converts an intermolecular process to intramolecular and may have both an enthalpic and an entropic advantage. The active site provides an environment in which the enzyme·substrate complex is

Enzyme Mechanisms

639 often accelerate, by large factors, the reaction over that in aqueous media. For DHFR in particular, molecular dynamics calculations, NMR measurements of solution structure, and kinetics measurements of mutant forms of the enzyme appear to support the importance of dynamic motions of the protein fold to trigger the reaction of an enzyme– substrate NAC. The mutations in question (for example Gly120 in Fig. 7) are well removed from the active site and underscore the role of the entire protein fold. The contribution of dynamic motions to the overall catalytic rate remains to be elucidated for the majority of enzymes. Their existence may explain why more rigid molecules such as imprinted polymers and catalytic antibodies do not generally exhibit the large rate accelerations noted with enzymes despite the fact that they too have converted an intermolecular process to an intramolecular process.

FIGURE 10 A currently accepted version of the catalytic cycle of cytochrome P450. The iron porphyrin is drawn as a parallelogram, the substrate is designated as S H and the product as S OH. See the text for a description. [Adapted from Mueller, E. J., Loida, P. J., and Sligar, S. G. (1995). Twenty-five years of P450cam research. In “Cytochrome P450: Structure, Mechanism, and Biochemistry” (P. R. Ortiz de Montellano, ed.), 2nd ed., pp. 83–124, Plenum, New York.]

SEE ALSO THE FOLLOWING ARTICLES
BIOCONJUGATE CHEMISTRY • BIOENERGETICS • BIOINORGANIC CHEMISTRY • BIOREACTORS • FIBEROPTIC CHEMICAL SENSORS • GENE EXPRESSION, REGULATION OF • LIPOPROTEIN/CHOLESTEROL METABOLISM • TRANSLATION OF RNA TO PROTEIN • VITAMINS AND COENZYMES

populated with cofactors that are poised for reaction. These structures, or NACs (near attack conformers), are similar in structure to the transition state so that only slight changes in bond distances and angles within the structures through the normal dynamic motions of the protein are sufficient to trigger the crossing of the reaction barrier. The enzyme’s active site is also preorganized in the sense that the locus of general acids/bases, nucleophiles, solvents, dipoles, hydrogen bonds, and so forth are fixed by the NAC to interact with the transition state. Molecular dynamic calculations sampling several enzyme classes suggest that the affinity of these enzymes for their transition states is little changed from that for the substrate. The enzymecatalyzed reaction also benefits in many cases due to the nonaqueous interaction of the active site cavity, which can

BIBLIOGRAPHY
Bugg, T. (1997). “An Introduction to Enzyme and Coenzyme Chemistry,” Blackwell Sci., London. Fersht, A. (1999). “Structure and Mechanism in Protein Science,” Freeman, New York. Hammes, G. G. (1982). “Enzyme Catalysis and Regulation,” Academic Press, New York. Ortiz de Montellano, P. R., ed. (1995). “Cytochrome P450: Structure, Mechanism, and Biochemistry,” 2nd ed., Plenum, New York. Page, M. I. (1984). “The Chemistry of Enzyme Action,” Elsevier, New York. Price, N. C., and Stephens, L. (1989). “Fundamentals of Enzymology,” Oxford Univ. Press, New York.

Food Colors
Pericles Markakis
Michigan State University I. II. III. IV. Introduction Natural Food Pigments Food Browning Color Additives in Foods

GLOSSARY
Anthocyanins Red, blue, and violet water-soluble plant pigments of a phenolic nature. Browning, food Darkening of foods as a result of enzymatic or nonenzymatic reactions. Caramel Brown coloring matter made by heating sugars dry or in solution. Carotenes Chiefly orange-yellow plant and animal pigments; some are provitamins A. Certification, color Submission of a sample of a listed color additive to the Food and Drug Administration and, after chemical analysis, issuance of a certificate permitting marketing of the batch from which the sample was taken; certain color additives are exempt from certification. Chlorophyll Green pigment of plants; chemically it is related to the red pigment of blood. Colorant Substance that colors or modifies the color of another substance. Excipient Inert substance used as a diluent or vehicle of a colorant. Heme Color-furnishing portion of the red pigment molecule of blood and meat. Lakes, color Water-insoluble pigments prepared by pre-

cipitating soluble dyes on an insoluble substratum, alumina in the case of food lakes. Listed color(ant)s Color additives that have been sufficiently evaluated to convince the Food and Drug Administration of their safety for the application intended.

FOOD COLORS are both the sensations evoked when light reflected from foods stimulates the retina of the eye and the particular food components involved in the process. These components, also known as food colorants, may be present in foods naturally, or formed during food processing, or intentionally added to foods, or all of these. This article deals with all groups of food colorants.

I. INTRODUCTION
Color is important for identifying foods, judging their quality, and eliciting aesthetic pleasure in our encounters with them. Because color is usually the first food attribute to strike the senses, its significance in food marketing is obvious (“eating” with the eyes). Thus, all food providers (growers, grocers, homemakers, chefs, and industrial food processors) do their best to present a food with an attractive

105

106 color. In certain instances, the original color of the food must be preserved, as is the case with most fruits and vegetables. In other instances, culinary art is required to create new, pleasing colors, as when turkey is roasted, bread is baked, or potato chips are fried. In still other instances, colors (colorants) are added to foods, as is done with many beverages and candies. The coloring matter of foods is discussed under three headings: natural food colors, food browning, and food color additives.

Food Colors

II. NATURAL FOOD PIGMENTS
Approximately 1500 colored compounds, also known as natural food pigments, have been isolated from foodstuffs. On the basis of their chemical structure, these food pigments can be grouped in the following six classes: heme pigments, chlorophylls, carotenoids, flavonoids, betalains, and miscellaneous pigments. A. Heme Pigments Heme (from the Greek for blood) is the basic chemical structure (Fig. 1) responsible for the red color of two important animal pigments: hemoglobin, the red pigment of blood, and myoglobin, the red pigment of muscles. Practically all the red color of red meat is due to myoglobin, since the hemoglobin is removed with the bleeding of the slaughtered animal. Other colored muscle compounds (cytochromes, vitamin B12 , flavoproteins) do not contribute significantly to the color of red meat. Myoglobin is a protein that facilitates the transfer of oxygen in muscles. It was the first protein to be fully elucidated with regard to the three-dimensional arrangement of its atoms. Hemoglobin, the oxygen-carrying pigment

of blood, is composed of four heme groups attached to four polypeptide chains. The myoglobin in meat is subject to chemical and color changes. Freshly cut meat looks purplish. On exposure to air, the surface of the meat acquires a more pleasing red hue (blooming of the cut). The color change is due to the oxygenation of myoglobin (an oxygen molecule is attached to the heme group in a fashion parallel to the oxygenation of hemoglobin). The oxygenated myoglobin is called oxymyoglobin. When meat is packed in plastic film, the oxygen permeability of the film should be sufficient to keep the myoglobin oxygenated. In both myoglobin and oxymyoglobin the heme iron is in the Fe2+ form. In the presence of oxygen, myoglobin is eventually oxidized to brown metmyoglobin, in which the heme iron is in the Fe3+ form. Both the oxygenation and oxidation processes are reversible. Severe oxidative deterioration may result in the formation of green pigments (sulfmyoglobin, cholemyoglobin). When meat is cooked, the protein moiety (globin) of myoglobin is denatured and the heme is converted chiefly to nicotinamide hemichrome, the entire pigment acquiring a brown hue. These changes are irreversible. Heated meat is also subject to the browning reactions discussed in Section III. A simplified scheme of the red-pigment changes in fresh and heated meat is shown in Fig. 2. In cured meats, in which nitrite is used, many reactions occur, some of which lead to color changes. Among the established reactions are the following: (1) the nitrite salt is converted to nitric oxide (NO), nitrate, and water; (2) the NO replaces the H2 O attached to the iron of heme and forms nitrosyl myoglobin, which is reddish; (3) on heating, the nitrosyl myoglobin is transformed to nitrosyl hemochrome, which has the familiar pink color of cured meats; and (4) any metmyoglobin present in the cured meat is similarly nitrosylated, reduced, and finally converted to nitrosyl hemochrome. B. Chlorophylls Several chlorophylls have been described. Two of them, chlorophyll a and chlorophyll b, are of particular interest in food coloration because they are common in green plant tissues, in which they are present in the approximate ratio 3 : 1, respectively. Their structures resemble that of heme since they are all derivatives of tetrapyrrole. An important difference is that the central metal atom is iron in heme and magnesium in the chlorophylls. Another difference is that the pyrrole unit IV in the chlorophylls is hydrogenated. In addition, the chlorophylls contain a 20-carbon hydrophobic “tail,” the phytyl group (Fig. 3). The chlorophylls are located in special cellular bodies, the chloroplasts, where they function as photosynthetic

FIGURE 1 Structure of heme.

Food Colors

107

FIGURE 2 Pigment changes in fresh and heated red meat.

agents. As food pigments, chlorophylls impart their green color to many leafy (spinach, lettuce, etc.) and nonleafy (green beans and peas, asparagus, etc.) vegetables and to unripe fruits. They are not very stable pigments, however. Ethylene, a gaseous plant hormone, destroys chlorophylls, and it is occasionally used to degreen fruits. The acids naturally present, formed, or added to plant tissues during food processing convert the bright green chlorophylls to dull olive brown pheophytins by replacing the magnesium of the molecule with hydrogen. Unfortunately, no fail-safe procedure has been proposed for preventing this discoloration in heated and stored green vegetables. Freezing storage is an effective method of preserving the green color of vegetables.

C. Carotenoids Many of the yellow, orange, and red colors of plants and animals are due to carotenoids, pigments similar to those of carrots. The basic structure of carotenoids is a chain of eight isoprenoid units. Certain isoprenoid derivatives with shorter chains (e.g., vitamin A) are also considered carotenoids. Most of the structural differences among carotenoids exist at the ends of the chain. Some carotenoids are hydrocarbons and are known as carotenes, while others contain oxygen and are called xanthophylls. The structures of several carotenoids, along with the foods or tissues in which they are present, are shown in Table I. Because of the numerous double bonds in the carotenoid molecule, a large number of cistrans isomers are theoretically possible. The carotenoids of foods, however, are usually in the all-trans form (Table I). Trans to cis transformation is possible and is accelerated by heat, light, and acidity. Carotenoids occur free or as esters of fatty acids or as complexes with proteins and carbohydrates; for example, in paprika, capsanthin is esterified with lauric acid. In live lobster, astaxanthin is complexed with protein; the astaxanthin–protein complex is blue-gray, the color of live lobster, but on heating, the complex is broken and the freed astaxanthin imparts its red color to the cooked lobster. Carotenoids are present in a large variety of foods, from yeast and mushrooms, to fruits and vegetables, to eggs, to fats and oils, to fish and shellfish. As fat-soluble substances, carotenoids tend to concentrate in tissues or products rich in lipids, such as egg yolk and skin fat, vegetable oils, and fish oils. Plants and microorganisms synthesize their own carotenoids, while animals appear to obtain theirs from primary producers. In the development of many fruits (e.g., citrus fruits, apricots, tomatoes) ripening is associated with the accumulation of carotenoids and the

FIGURE 3 Structure of chlorophylls a and b. [From Aronoff, S. (1966). In “The Chlorophylls” (L. P. Vernon and G. R. Seeley, eds.), Academic Press, New York.]

108
TABLE I Types of Carotenoids and Their Natural Sources Structure Name and source

Food Colors

β-Carotene (carrot, egg, orange, chicken fat)

Xanthophyll (vegetables, egg, chicken fat)

Zeaxanthin (yellow corn, egg, liver)

Cryptoxanthin (egg, yellow corn, orange)

Physalien (asparagus, berries)

Bixin (annatto seeds)

Lycopene (tomato, pink grapefuit, palm oil)

Capsanthin (paprika)

Astaxanthin (lobster, shrimp, salmon)

Torularhodin (Rhodotorula yeast)

Canthaxanthin (mushrooms)

β-Apo-8 -carotenal (spinach, orange)

Food Colors

109 purple, violet, magenta, and most of the red hues of flowers, fruits, leaves, stems, and roots are attributable to pigments chemically similar to the original “flower blues.” Two exceptions are notable: tomatoes owe their red color to lycopene and red beets owe theirs to betanin, pigments not belonging to the anthocyanin group. Anthocyanins are glycosides of anthocyanidins, the latter being polyhydroxyl and methoxyl derivatives of flavylium. The arrangement of the hydroxyl and methoxyl groups around the flavylium ion in six anthocyanidins common in foods is shown in Fig. 4. There are at least 10 more anthocyanidins in nature, practically always appearing as glycosides. The number of anthocyanins far exceeds that of anthocyanidins, since monosaccharides, disaccharides, and at times trisaccharides glycosylate the anthocyanidins at various positions (always at 3, occasionally at 5, and seldom at other positions). Eventual acylation with p-coumaric, caffeic, and ferulic acids increases the number of natural anthocyanins. An example of acylated anthocyanin is the dark purple eggplant pigment delphinidin, 3-[4-(p-coumaroyl)-Lrhamnosyl-(1 → 6)-D-glycosido] 5-D-glucoside. The color of anthocyanins is influenced not only by structural features (hydroxylation, methoxylation, glycosylation, acylation), but also by the pH of the solution in which they are present, copigmentation, metal complexation and self-association. The pH affects both the color and the structure of anthocyanins. In very acidic solution, anthocyanins are red, but as the pH rises the redness diminishes. In freshly prepared alkaline or neutral solution, anthocyanins are blue or violet, but (with the exception of certain multiacylated anthocyanins) they fade within hours or minutes. In acidic solution four molecular species of anthocyanins exist in equilibrium: a bluish quinoidal (or quinonoidal) base A, a red flavylium cation AH+ , a colorless carbinol pseudo-base B, and a colorless or yellowish chalcone C (Fig. 5). At very low pH (below 1), the red cation AH+ dominates, but as the pH rises to 4 or 5, the concentration of the colorless form B increases rapidly at the expense of AH+ , while forms A and C remain scarce. In neutral and alkaline solutions, the concentration of base A rises and its phenolic hydroxyls ionize, yielding unstable blue or violet quinoidal anions A− (Fig. 6). Although it is true that the reaction of most plant tissues pigmented with anthocyanins (fruits, flowers, leaves) is slightly acidic, pH alone cannot explain the vivid colors encountered in these tissues. One mechanism leading to the enhancement and stability of anthocyanin coloration is copigmentation, that is, the association of anthocyanins with other organic substances (copigments). This association results in complexes that absorb more

disappearance of chlorophyll. The intensity of the yellow color of certain animal products, such as egg yolk and milk fat or butter, depends on the carotenoid content of the feed the animals ingest. In view of this dependency, the seasonal variation in the color of these products is understandable. A nutritionally important interconversion of carotenoids is the formation of retinol (vitamin A) from

β-carotene and other carotenoids possessing a β-ionone ring and known as provitamins A. The stability of carotenoids in foods varies greatly, from severe loss to actual gain in carotenoid content during storage. Carotenoid losses amounting to 20 or 30% have been observed in dehydrated vegetables (e.g., carrots, sweet potatoes) stored in air. These losses are minimized when the dry product is stored in vacuum or inert gas (e.g., nitrogen), at low temperatures, and protected from light. The main degradative reaction of carotenoids is oxidation. Oxygen may act either directly on the double bonds or through the hydroperoxides formed during lipid autoxidation. Hydroperoxides formed during enzymatic lipid oxidation can also bleach carotenoids by a coupled lipid–carotenoid oxidation mechanism. On the other hand, certain vegetables, such as squash and sweet potatoes, in which carotenoid biosynthesis continues after harvesting, may manifest an increase in carotenoid content during storage. D. Flavonoid Pigments Hundreds of flavone-like pigments are widely distributed among plants. On the basis of their chemical structure, these pigments are grouped in several classes, the most important of which are listed in Table II. The basic structure of all these compounds comprises two benzene rings, A and B, connected by a heterocycle. The classification of flavonoids is based on the nature of the heterocycle (which is open in one class). Most of these pigments are yellow (Latin, flavus). One important exception is the anthocyanins, which display a great variety of red and blue hues. Because of the strong visual impact of anthocyanins on the marketing of fruits and vegetables, these pigments will be discussed in greater detail than other flavonoids. 1. Anthocyanins The name of these pigments was originally coined to designate the blue (kyanos) pigments of flowers (anthos). It is now known that not only the blue color, but also the

110

Food Colors

TABLE II Major Classes of Flavonoids Class Flavones Structure Example (source) Apigenin (chamomile)

Flavan-3-ols (catechins)

(−)-Epicatechin (cocoa)

Flavan-3, 4-diols

Leucocyanidin (peanut)

Flavanones

Naringenin, hesperidin (citrus fruits)

Flavonols

Quercetin (apples, grapes)

Flavanonols

Taxifolin (Prunus)

Isofiavones

Genistein (soybeans)

Anthocyanidins

Pelargonidin, cyanidin, delphinidin (berries, red apples, red grapes)

Chalcones

Butein (Butea)

Aurones

Aureusidin (Oxalis)

Food Colors

111

FIGURE 4 Six anthocyanidins common in foods. The electric charge shown at position 1 is delocalized over the entire structure by resonance.

visible light (they are brighter) and light of lower frequency (they look bluer—the bathochromic effect) than the free anthocyanins at tissue pH. Most of these copigments are flavonoids, although compounds belonging to other groups (e.g., alkaloids, amino acids, nucleotides) can function similarly. A stacked molecular complex between an acylated anthocyanin and a copigment (flavocommelin) is shown in Fig. 7. Self-association is the binding of anthocyanin molecules to one another. It has been observed that the complexes absorb more light than the sum of the single molecules. This explains why a 100-fold increase in the concentration of cyanidin 3,5-glucoside results in a 300fold rise in absorbance.

Certain anthocyanins form complexes with metals (e.g., iron, aluminum, magnesium), and the result is an augmentation of the anthocyanin color. At times the complexes involve an anthocyanin, a copigment, and a metal. A large number of the anthocyanins present in fruits and vegetables have been identified. It is not unusual for a plant tissue to contain several anthocyanins (17 in certain grape varieties), all genetically controlled. Table III shows the anthocyanidin moieties of anthocyanins in common fruits and vegetables. Generally, the attractive color of anthocyaninpigmented foods is not very stable. Canning of red cherries or berries results in products with considerable bleaching. Strawberry preserves lose one-half of their anthocyanin content after a few weeks on the shelf, although the browning reaction may mask the loss. And red grape juice is subject to extensive color deterioration during storage.

FIGURE 5 Four anthocyanin structures present in aqueous acidic solutions: R is usually H, OH, or OCH3 . Gl is glycosyl. [Adapted from Brouillard, R. (1982). In “Anthocyanins as Food Colors” (P. Markakis, ed.), Academic Press, New York.]

FIGURE 6 Absorption spectra recorded immediately after dissolving an anthocyanin (malvin chloride) in buffers of pH 2, 6, and 10. The absorption peaks at pH 6 and 10 disappeared within 1 to 3 hr. (Adapted from Brouillard, R. (1982). In “Anthocyanins as Food Colors” (P. Markakis, ed.), Academic Press, New York.]

112

Food Colors

tribution in foodstuffs, while flavonone glycosides are of special interest in citrus fruits. Catechins, or flavan-3-ols, are present mainly in woody tissues. Among common foods, tea leaves contain at least six catechins representing about 25% of the dry weight of tea leaves. Tea catechins are excellent substrates for the catechol oxidase that is present in tea leaves and participates in the conversion of green tea to black tea. The reddish brown color of tea brew is due to a mixture of
FIGURE 7 Stacked molecular complex of awobanin and flavocommelin; p-C. denotes p-coumaroyl. [From Osawa, Y. (1982). In “Anthocyanins as Food Colors” (P. Markakis, ed.), Academic Press, New York.]

TABLE III Anthocyanidins Present as Anthocyanins in Fruits and Vegetables Fruit or vegetable Apple (Malus pumila) Blackberry (Rubus fructicosus) Black currant (Ribes nigrum) Blueberry (lowbush,Vaccinium angustifolium; highbush, V. corymbosum) Cherry (sour, ‘Montmorency,’ Prunus cerasus; sweet, ‘Bing,’ P. avium) Cranberry (Vacinnium macrocarpon) Elderberry (Sambucus nigra) Fig (Ficus carica) Gooseberry (Ribes grossularia) Grape (red European. Vitis vinifera) Anthocyanidin Cyanidin Cyanidin Cyanidin. delphinidin Delphinidin, petunidin, malvidin, peonidin, cyanidin Cyanidin, peonidin Cyanidin, peonidin Cyanidin Cyanidin Cyanidin Malvidin, peonidin, delphinidin, cyanidin, petunidin, pelargonidin Cyanidin, delphinidin, peonidin, malvidin, petunidin Peonidin Cyanidin Cyanidin Cyanidin. delphinidin Delphinidin Cyanidin Cyanidin Cyanidin, peonidin Delphinidin, cyanidin Cyanidin Pelargonidin, little cyanidin Pelargonidin, cyanidin, delphinidin Cyanidin Cyanidin, pelargonidin Delphinidin Cyanidin, peonidin Pelargonidin, cyanidin, delphinidin, petunidin Pelargonidin, cyanidin

Exposure to high temperatures and contact with the oxygen of the air appear to be two factors affecting anthocyanin stability most adversely. Ascorbic acid accelerates the destruction of anthocyanins, and so does light. Certain oxidizing enzymes, such as phenol oxidase, and a hydrolyzing enzyme known as anthocyanase may contribute to the degradation of anthocyanin pigments. Oxidizing enzymes act on the anthocyanidin moiety, while anthocyanase splits off the sugar residue(s); the freed anthocyanidin is very unstable and loses its color spontaneously. Sulfur dioxide, which is used for the preservation of some fruit products (pulps, musts), bleaches anthocyanin pigments, but on heating of the fruit prduct in vacuum the SO2 is removed and the anthocyanin color reappears. Large concentrations of SO2 , combined with lime, decolorize anthocyanins irreversibly and are used in the preparation of maraschino cherries. Anthocyanins act as anodic and cathodic depolarizers and thereby accelerate the internal corrosion of tin cans. It is therefore necessary to pack anthocyanin-colored products in cans lined with special enamel. In aging red wines anthocyanins condense with other flavonoids and form polymeric (MW ≤ 3000) redbrown pigments (Fig. 8). On continued polymerization these pigments become insoluble and form sediments in bottled red wines. Anthocyanins possessing more than one acyl group show extraordinary color stability over a wide pH range. One of them, peonidin-3-(dicaffeyl sophoroside) 5-glucoside, isolated from ‘Heavenly Blue’ morning glory flowers (Ipomoea tricolor), has been shown to “produce a wide range of stable colors in foods and beverages which have a pH range of 2.0 to about 8.0.” United States patent 4,172,902 covers its use as a colorant in foods. 2. Other Flavonoids Among flavonoids other than anthocyanins, the catechins, flavonols, and leucoanthocyanidins have the widest dis-

Grape (‘Concord,’ Vitis labrusca)

Mango (Mangifera indica) Mulberry (Morus nigra) Olive (Olea europea) Orange (‘Ruby,’ Citrus sinesis) Passion fruit (Passiflora edulis) Peach (Prunus persica) Pear (Pyrus communis) Plum (Prunus domestica) Pomegranate (Punica granatum) Raspberry (Rubus ideaus) Strawberry (Fragaria chiloensis and F. virginiaca) Beans (red, black; Phaseolus vulgaris) Cabbage (red, Brassica oleracea) Corn (red, Zea mays) Eggplant (Solanum melongena) Onion (Alium cepa) Potato (Solanum tuberosum) Radish (Raphanus sativus)

Food Colors

113

FIGURE 10 Basic structures of leucoanthocyanidins (1), anthocyanidins (2), and dimeric leucoanthocyanidins (3).

FIGURE 8 Proposed structure for a polymeric red-brown pigment in aging red wine. [From Somers, T. C. (1971). Phytochemistry 10, 2184.]

pigments known as theaflavins and thearubigins. The structure of one of them is shown in Fig. 9. Flavonols, like anthocyanidins, exist almost exclusively as glycosides. Three common flavonols are kaempferol, quercetin, and myricetin, resembling pelargonidin, cyanidin, and delphinidin, respectively, in the hydroxylation pattern of the B ring. Flavonol glycosides impart weak yellow hues to apples, apricots, cherries, cranberries, grapes, onions, plums, potatoes, strawberries, tea, tomatoes, and other commodities. Leucoanthocyanidins are compounds of the general formula 1 shown in Fig. 10. They have no color of their own, but in acidic environments and at elevated temperatures they are converted to colored anthocyanidins (2). This reaction is in competition with the condensation to a dimeric leucoanthocyanidin (3). Low temperature favors the formation of the dimeric compound, which can polymerize to yield products with pronounced tanning properties.

The most common leucoanthocyanidins are leucopelargonidin, leucocyanidin, and leucodelphinidin, which are converted to the corresponding anthocyanidins. This conversion results in the undesirable “pinking” of certain products such as canned pears, canned banana puree, processed brussels sprouts, and beer. On the other hand, polymerization to tannins leads to astringency and the formation of haze in beer (insolubilization of beer proteins). E. Betalains Betalain is a relatively new term used to describe a class of water-soluble plant pigments exemplified by the red-violet betacyanins and yellow betaxanthins. (In a parallel fashion, flavonoids comprise the red-blue anthocyanins and the typical yellow flavonoids that some authors call anthoxanthins.) Betalains owe their name to the red beet (Beta vulgaris), from which they were originally extracted, and they are not as widely distributed as flavonoids. Other foods containing betalains include chard, pokeberries, and Indian cactus fruits. The major red pigment of red beets is betanin, and their major yellow pigment is vulgaxanthin (Fig. 11).

FIGURE 9 Structure of theaflavin.

FIGURE 11 Two major pigments of red beets.

114 Betalains are stable in the pH range 3.5–7.0, which is the pH range of most foods, but they are sensitive to heat, oxidation, and light. F. Miscellaneous Natural Food Colors There are several hundred additional natural pigments that are not as widely represented in foods as the previously discussed coloring substances. Among them are the quinones and xanthones, which are yellow pigments. An example of a quinone is juglone, which is present in walnuts and pecans. Mangiferin, a representative of xanthones, is found in mangoes. Tannins include two types of pale yellow to light brown compounds, characterized by their property to convert animal hides to leather. One type consists of condensed tannins, to which reference was made in relation to the leucoanthocyanidins, and the other type comprises hydrolyzable tannins, which are esters of a sugar, usually glucose, with gallic acid, ellagic acid, or both. Corilagin is an example of a gallotannin, in which glucose is esterified with three gallic acid molecules. A yellow pigment that has attracted much attention because of its toxicity to humans and nonruminant animals is gossypol. It is present in cottonseeds, which are used as animal feed and have been considered a potential source of protein for human use. Several biologically very important food constituents are colored, such as phytochrome (yellow), vitamin B2 (riboflavin, orange-yellow), and vitamin B12 (red), although their contribution to food coloration is negligible.

Food Colors

III. FOOD BROWNING
Foods may develop a variety of brown colors, from yellowbrown to red-brown to black-brown, during handling, processing, and storage. These colors are desirable in certain foods (e.g., coffee, beer, bread, maple syrup). In other foods, such as most dehydrated fruits and vegetables, dried eggs, and canned or dried milk, browning is detrimental. Even when desirable, browning should not be excessive, as in potato chips, french fries, and apple juice. Numerous reactions lead to browning in foods. Some of these may also generate flavors and/or alter the nutritional properties of foods. Conventionally, browning is discussed as enzymatic and nonenzymatic browning. A. Enzymatic Browning Several enzymes may initiate reactions that eventually produce brown colors in foods. For example, the action of ascorbate oxidase on ascorbic acid or of lipoxidase on lipids leads to carbonyl products that may either polymer-

ize or react with amino compounds and form brown products. Phenolase (or phenol oxidase), however, is the principal browning enzyme. This enzyme oxidizes o-diphenols to o-quinones, which, by nonenzymatic processes, are ultimately converted to brown polymers known as melanins. Melanins are formed in both animal and plant tissues. A typical substrate of phenolase in animals is tyrosine. This amino acid is converted to melanin by a series of reactions, some of which are shown in Fig. 12. In dark hair, skin, eyes, and other animal tissues, melanin is attached to proteins. Tyrosine is also a phenolase substrate in plant tissues (e.g., potatoes), but odiphenols and polyphenols are by far the most common substrates of enzymatic browning in foods of plant origin. The following phenolic compounds have been associated with enzymatic browning in some foods: chlorogenic acid, caffeic acid, and catechin in apples, apricots, peaches, and pears; 3,4-dihydroxyphenylethylamine in bananas; (−)epicatechin, (+)-catechin, (+)-gallocatechin, and (−)epigallocatechin galate in tea leaves and cocoa beans; catechins in grapes; and tyrosine and chlorogenic acid in potatoes. The structures of four of these phenolics are shown in Fig. 13. Many fresh fruits and vegetables brown slowly as they senesce. The enzymatic browning of these commodities is more rapid when they are subjected to processing, such as the pressing of apples in making cider or the peeling and cutting of potatoes in preparing potato products. Since enzyme, substrate, and oxygen must all be present for the development of this type of browning, elimination of any one of the three agents will prevent the browning. Heat inactivation of the enzyme, the exclusion of oxygen (by keeping the commodity under water or packaging it under vacuum or inert gas), and the selection of varieties poor in substrate content or enzyme activity are ways of preventing this discoloration. Also, storage at low temperature and the addition of sulfur dioxide, ascorbic acid, citric acid, sodium chloride, or combinations of these compounds will inhibit browning. B. Nonenzymatic Browning A number of chemical processes not involving enzymes may result in food browning. Briefly discussed here are the Maillard reaction, caramelization, ascorbic acid browning, and metalpolyphenol browning. 1. The Maillard Reaction (Maillard Browning) This reaction is actually a series of reactions occurring from the first encounter of a carbonyl compound with an amine compound to the formation of brown pigments. It is also known as the carbonyl–amine reaction, and its brown

Food Colors

115

FIGURE 12 Conversion of tyrosine to melanin, catalyzed in part by tyrosinase (T). DOPA, Dihydroxyphenylalanine. Only part of melanin is shown.

products are often called melanoidins, indicating their visual similarity to the melanins of enzymatic browning. The most common carbonyl compounds of foods involved in the Maillard reaction are reducing sugars, and the most common amine compounds are amino acids. The intermediate reactions and their relative velocities vary with the type of initial reactants and the conditions of

FIGURE 13 Four phenolic compounds involved in enzymatic browning.

the reactions. Among sugars, pentoses are more reactive than hexoses, and hexoses are more reactive than reducing disaccharides. When free amino acids react with sugars, lysine appears to be the most active among them. In peptides and proteins, the N-terminal amino acid is the most reactive, followed by a nonterminal lysine. Raising the temperature and/or the pH accelerates the Maillard reaction. Intermediate water activity appears to maximize this reaction. Several pathways have been proposed for the formation of melanoidins through the interaction of carbonyl and amine compounds. A simplified scheme is shown in Fig. 14. This scheme involves first the condensation of a carbonyl compound (an aldohexose in this scheme) with an amine to a Schiff base via an intermediate product (not shown). The Schiff base quickly cyclizes to an Nsubstituted glycosylamine. Up to this step the process is reversible because the glycosylamine can be hydrolyzed back to the initial reactants. The N-substituted glycosylamine is then rearranged to an N-substituted 1-amino1-deoxy-2-ketose, the Amadori compound, which is in equilibrium with its enol form. If the initial carbonyl compound is a 2-ketose (e.g., fructose), the corresponding Nsubstituted 2-amino-2-deoxy-1-aldose is formed by the Heyns rearrangement, which is analogous to the Amadori

116

Food Colors

FIGURE 14 Simplified scheme of melanoidin formation by the sugar–amine reaction (± means presence or absence).

rearrangement for aldoses. The Amadori and Heyns compounds are subsequently subjected to a variety of transformations, which may include deamination, dehydration, enolization, cyclization, and degradation. The products of those reactions finally condense or polymerize with or without the participation of an additional amino compound to form dark pigments of a colloidal nature and ill-defined composition—the melanoidins. Some of the

intermediate products are collectively called reductones, because they are strongly reducing compounds that account for the reducing properties of systems undergoing Maillard browning. It should be mentioned here that a side reaction of the Maillard browning results in the formation of flavorful compounds, such as those associated with the roasting of meat, coffee, or nuts and the baking of bread. This

Food Colors

117 oxidized to the ferric state for the blackish complex to appear. The stem end, which contains much less citric acid than the rest of the tuber, displays the deepest darkening. Canned or pickled cauliflower may turn dark due to the interaction of polyphenols in the tissue with iron from external sources. As already indicated, nonenzymatic browning is desirable in certain instances and undesirable in others. The availability of reactants and the type of conditions (temperature, pH, moisture) will determine the extent of browning. A chemical preservative often used to inhibit nonenzymatic (and enzymatic) browning is sulfur dioxide. An obvious way to prevent metal-polyphenol browning is to eliminate contact between susceptible tissues and reactive metals and use inoffensive equipment (stainless steel, glass-lined tanks, etc.)

side reaction, known as the Strecker degradation, occurs between α-amino acids and dicarbonyl compounds and leads to the formation of aldehydes possessing one less carbon atom than the corresponding initial amino acids. The newly formed aldehydes are responsible for the pleasing flavors. A significant consequence of the Maillard reaction is the loss of the nutritional value of the amino acid involved in the reaction. If the participating amino acid is essential, and especially if it is a limiting one, as is lysine in most cereal grains, the Maillard reaction can seriously lower the nutritive value of the food. Toasting, for example, may reduce to one-half the protein efficiency ratio of bread. 2. Caramelization This reaction leads to brown products when sugars are heated dry or in solution. Certain conditions of caramelization favor the formation of flavor compounds as well. The chemical transformations involved in caramelization are complex and poorly understood. They include dehydration, fragmentation, and polymerization. On the heating of pentoses, furfural is formed which polymerizes to brown products. Heating hexoses results in hydroxymethylfurfural, which polymerizes similarly. The large quantities of industrial caramel color that are added to beverages (cola drinks), baked goods, and confections are made by heating high-conversion corn syrups in the presence of catalysts (acids, alkalis, salts). 3. Ascorbic Acid Browning When ascorbic acid is heated in the presence of acids, furfural is formed. The latter, either by itself or after reacting with amino compounds, polymerizes to brown products. Citrus juices, especially their concentrates, develop browning, which has been attributed to ascorbic acid degradation. 4. Metal–Polyphenol Browning Polyphenolic compounds form complexes with certain metals. The polyphenols of fruits and vegetables most commonly chelate iron. The resulting iron complexes are bluish black pigments. Cutting apples with a nonstainless-steel knife results in darkening of both the blade and the cut surface of the apple. This darkening is independent of the enzymatic browning that might develop as a result of cutting. Wine makers avoid contact between the wine and iron implements because of the black iron–tannin precipitate that forms on such contact. Certain varieties of potatoes tend more than others to darken after cooking. This darkening is attributed to a complex between iron and chlorogenic acid. The iron of the tissue must first be

IV. COLOR ADDITIVES IN FOODS
Colored substances are added to foods to modify the appearnce of insufficiently colored or discolored foods and to create new foods. The following criteria must be met if a food color (colorant) is to be used; (1) it must be safe at the level and under the conditions of use; (2) it must be stable in the products in which is added; (3) it must not impart any offensive property (flavor, texture) to the product; (4) it must be easy to apply; (5) it must have a high tinctorial power; and (6) it must not be too costly. There are two classes of color additives, those that must be certified and those that are exempt from certification. Both are strictly controlled in the United States by regulatory statutes (Food Color Additives Amendments), but an official certificate is required for each commercial batch of color of the first group, while no such certificate is necessary for the second group. For certification the manufacturer must submit a sample of the batch to the Food & Drug Administration for chemical analysis. The results of the analysis are compared with the specifications for certified colors published in the Code of Federal Regulations. If the compliance is complete, a certificate is issued for that particular batch of color. When a new color is to be introduced, the petitioner is expected to provide data proving that the new additive is safe and effective. The safety tests are elaborate and expensive. They include chronic toxicity feeding tests with two animal species over several generations. The effectiveness tests include long-term color stability experiments in the foods to which the color is to be added. The food color additives subject to certification are listed in Table IV. The initials FD&C stand for the Food, Drug & Cosmetic Act under which these additives are regulated. The food color additives exempt from certification are listed in Table V. Generally, only synthetic organic

118
TABLE IV Permanently Listed Food Colors Subject to Certification Color FD&C Blue #1 Structure Uses a Used in amounts consistent with GMP

Food Colors

FD&C Green #3

Used in amounts consistent with GMP

FD&C Yellow #5

Used in amounts consistent with GMP

FD&C Red #3b

Used in amounts consistent with GMP

FD&C Red #40

Used in amounts consistent with GMP

Orange B

Coloring sausage surfaces or casings (150 ppm max. based on finished product)

FD&C Blue #2

Used in amounts consistent with GMP

FD&C Yellow #6

Used in amounts consistent with GMP

Citrus Red #2
N N OH

OCH3

In skins of oranges that are not intended for processing (at 2 ppm max., based on whole fruit)

OCH3

GMP. Good manufacturing practices. The FDA has recently banned the use of Red #3 in such products as cake frostings, certain processed foods, cough drops, and lipstick, in which the color is mixed with other additives reacting with it. This dye can still be applied directly to meat, nut products, fruit and fruit juices, candy, confections, and breakfast cereals. b a

Food Colors
TABLE V Permanently Listed Colors Exempt from Certification Colorant Caramel β-Carotene Annatto extract Paprika Paprika oleoresin Turmeric Turmeric oleoresin Saffron Fruit juice Vegetable juice Toasted, partially defatted, cooked cottonseed flour a 119

Uses a In foods, generally consistent with GMP In foods, generally consistent with GMP In foods, generally consistent with GMP In foods, generally consistent with GMP In foods, generally consistent with GMP In foods, generally consistent with GMP In foods, generally consistent with GMP In foods, generally consistent with GMP In foods, generally consistent with GMP In foods, generally consistent with GMP In foods, generally consistent with GMP

Colorant Cochineal extract; carmine Dehydrated beets Riboflavin Carrot oil β-Apo-8 -carotenal Titanium dioxide Grape skin extract Ferrous gluconate Canthaxanthin

Uses a In foods, generally consistent with GMP In foods, generally consistent with GMP In foods, generally consistent with GMP In foods, generally consistent with GMP In foods, generally not to exceed 25 mg/lb In foods, generally not to exceed 1% by weight In still and carbonated beverages and alcoholic beverages For coloring ripe olives, consistent with GMP In foods, generally not to exceed 30 mg/lb

GMP, Good manufacturing practices.

colorants are subject to certification, while natural organic and inorganic colors, such as paprika and titanium oxide are not. The colorant β-carotene is not subject to certification whether it is obtained from a natural source or it is synthetically produced. While synthetic food dyes are generally water-soluble, food lakes are water-insoluble. Food lakes are prepared by precipitating dyes on alumina. These lakes are useful for coloring water-repelling foods, such as fats and oils, certain gums, as well as packaging materials, e.g., plastic films, lacquers and inks, from which soluble dyes would leach out. Listing of a food dye does not necessarily imply listing the corresponding lake. Polymeric food dyes have been developed that cannot pass the gastrointestinal wall and are excreted virtually intact in the feces. Toxicity and efficacy tests must be completed before FDA approval is granted to these dyes. In recent years, plant tissue culture techniques have been applied to the production of food colors. Also the pigments of two fungi: Monascus anka and Monascus purpureus are being considered for use in foods. These fungal pigments have been used as food colors and medicines in the Far East for hundreds of years. The regulations regarding color additives can be found in the Code of Federal Regulations, Title 21, Parts 70–82. Changes in these regulations are published in the “Federal Register.” Additional information on color additives can be obtained from: Food & Drug Administration Division of Color & Cosmetics 200 C Street. S. W. Washington, DC 20204

SOME DOMESTIC SUPPLIERS OF COLOR ADDITIVES
Beatrice Foods Co., 156 W. Grand Ave., Beloit, WI 53511 BIOCON Inc., 518 Codell Dr., Lexington, KY 40509 COLORCON Inc., Moyer Blvd., West Point, PA 19486 Crompton & Knowles Co., 1595 MacArthur Blvd., Mahwah, NJ 07430 Hilton-Davis Co., 2235 Langdon Farm Rd., Cincinnati, OH 45237 H. K. COLOR Group, 155 Helen St., South Plainfield, NJ 07080 Hoffmann-LaRoche Inc., 304 Kingsland St., Nutley, NJ, 07110 Meer Corp., 9500 Railroad Ave., North Bergen, NJ 07047 Pylam Products Co., 1001 Stewart Ave., Garden City, NY 11530 Sethness Products Co., 2367 W. Logan Blvd., Chicago, IL 60647 Sun Chemical Corp., 441 Tompkins Ave., Staten Island, NY 10305 Warner-Jenkinson Co., 2526 Baldwin St., St. Louis, MO 63106 Whittaker, Clark & Daniels Inc., 1000 Coolidge St., South Plainfield, NJ 07080

SEE ALSO THE FOLLOWING ARTICLES
BIOPOLYMERS • NATURAL ANTIOXIDANTS IN FOODS • POLYMERS, SYNTHESIS

120

Food Colors
Hutchings, J. B. (1999). “Food Color and Appearance,” 2nd ed., Aspen Frederick, MD. Ilker, R. (1987). “In-vitro pigment production,” Food Technol. 41(4), 70. Markakis, P., (ed.). (1982). “Anthocyanins as Food Colors,” Academic Press, New York. Marmion, D. M. (1991). “Handbook of U.S. Colorants for Foods, Drugs and Cosmetics,” 3rd ed., Wiley-Interscience, New York. Walford, J., (ed.). (1984). “Developments in Food Colours,” Vols 1 and 2, Elsevier, London.

BIBLIOGRAPHY
Association of Official Analytical Chemists (1995). “Official Methods of Analysis” 16 ed., Washington, DC. Food & Drug Administration (1982). “Toxicological Principles for the Safety Assessment of Direct Food Additives and Color Additives in Food,” (The Redbook), Washington, DC. Francis, F. J. (1986). “Handbook of Food Colorant Patents,” Food & Nutrition Press (1999). Westport, CT.

Glycoconjugates and Carbohydrates
Eugene A. Davidson
Georgetown University School of Medicine I. II. III. IV. V. Monosaccharides Oligosaccharides Polysaccharides Glycoconjugates Analytical Methods

GLOSSARY
Aglycone Group attached to the hydroxyl of a furanose or pyranose sugar to form a glycoside. Can vary from a simple methyl group to other sugars to complex alkaloids. Aldose Polyhydroxy aldehyde with three or more carbon atoms. When of sufficient length will form fiveor six-membered ring (furanose or pyranose) hemiacetals. Designated D- or L- based on the configuration of the asymmetric center furthest from the carbonyl group relative to glyceraldehyde. Anomer Formation of the cyclic ring structure confers asymmetry on the original carbonyl carbon atom. The two anomers produced are alpha- (the newly formed hydroxyl group projects on the same side as the five- or six-membered ring) or beta- (the new hydroxyl projects on the opposite side of the newly formed ring). This is combined with D-, or L-; α-D-glucopyranose, β-Lfructofuranose, for example. Epimer Saccharide that differs from a reference sugar at

a single asymmetric center. Thus, D-galactose is the 4-epimer of D-glucose. Furanose Five-membered ring (derived from furan) with a single oxygen formed by reaction of the carbonyl carbon of an aldose or ketose with the appropriate hydroxyl group. Sugars commonly found as furanoses include ribose and 2-deoxyribose (in RNA and DNA, respectively), and fructose in combined form (sucrose, for example). Glycolipid Structure containing one or more saccharide units covalently attached to the primary hydroxyl group of N -acyl sphingosine (ceramide), a C-18 amino alcohol. Those containing sialic acid are termed gangliosides. Glycoprotein Protein with covalently attached saccharides. Linkage may be N - (amide nitrogen of asparagine residues), or O- (serine and threonine are most common). The saccharide units may be heterogeneous within the same protein, even at the identical substitution site. Glycosaminoglycan Polysaccharide chain found in

833

834 proteoglycans. Has a linear structure that features a repeating disaccharide unit comprised of glucosamine or galactosamine and a uronic acid or galactose. The amino sugars are generally N -acetylated and contain ester sulfate. The pattern of sulfation varies and may confer specific biological properties. Glycoside Acetal or ketal formed by substitution of the anomeric hydroxyl with an aglycone. This reaction fixes the configuration at the anomeric center. Linkages are generally acid labile. Ketose (ulose) Polyhydroxy ketone (at carbon 2) of four or more carbon atoms (see Aldose). Mutarotation The change in optical rotation associated with the equilibration, in solution, of the anomeric and ring forms of a saccharide beginning with a single crystalline form. Thus, dissolving α-D-glucopyranose in water allows formation of the β-anomer as well as the furanose forms. Since all have different optical rotations, the initial value changes until the equilibrium mixture is achieved. Oligosaccharide Sugars that contain 2–10 monosaccharide units. These may be linear or branched. The most common representatives are derived from aldoses and will contain a single reducing terminus (free, or potential, aldehydo group) and one or more nonreducing termini (linked to other sugars only at their anomeric hydroxyl group). Polysaccharide Structure containing more than 10 monosaccharide units. These may be the same (homopolysaccharide) or different (heteropolysaccharide), and the structure may be linear or branched. Natural diversity is very large. Proteoglycan Special category of glycoproteins in which the saccharide chain is covalently linked to protein via an initiating xylose residue, a specific core oligosaccharide, and is followed by a glycosaminoglycan chain. Pyranose Six-membered ring form of a saccharide formed by reaction of the appropriate hydroxyl group with the carbonyl carbon. Generally the most stable structure in solution. The conformer resembles cyclohexane and is generally a chair form with axial and equatorial substituents.
CHO H C OH

Glycoconjugates and Carbohydrates

CHO HO C H

CH2OH
D-Glyceraldehyde

CH2OH
L-Glyceraldehyde

Fischer projection formulas CHO H C OH HO CHO C H

CH2OH
D-Glyceraldehyde

CH2OH
L-Glyceraldehyde

Perspective formulas
FIGURE 1 Projection formulas of D- and L- glyceraldehyde.

drates extends beyond the food industry to textiles, pharmaceuticals, and a wide range of chemicals. The common definition of carbohydrates is polyhydroxy aldehydes or ketones (the carbonyl group is generally at C-2), their derivatives, oligomers and polymers. The term itself arises from the empirical formula of the compounds initially studied (glucose, for example) since that is represented by C(H2 O)n . An appreciation of the diversity available in such structures began about 1900 with the work of van’t Hoff on the tetrahedral carbon atom, extrapolated by Emil Fischer in his elucidation of the stereochemistry of glucose. During that series of experiments, Fischer was able to deduce the relative positions of all of the hydroxyls at the asymmetric centers and could relate glucose to glyceraldehyde (Fig. 1). However, he was unable to provide an absolute stereo structure for glyceraldehyde and thus had to decide between mirror image structures for natural glucose. That he chose the correct one may reflect little more than that he was right handed. He did synthesize all of the possible aldohexoses.

I. MONOSACCHARIDES
It is generally accepted that monosaccharides, or simple sugars, may have three to nine carbon atoms in a linear chain. Thus the nomenclature aldotriose, aldotetrose, aldopentose, etc., or ketotetrose, ketopentose, etc. The vast majority of naturally occurring sugars contain five or six carbon atoms. These saccharides are classified as D- or Lif they can be derived from D- or L-glyceraldehyde, respectively, or the four carbon ketoses (Fig. 2). The determining asymmetric center is that farthest from the carbonyl group—projection of the hydroxyl group to the right is D. As the chain length increases, new asymmetric centers

THE UTILIZATION of carbohydrates as food sources, sweetening agents, and clothing materials dates back several thousand years. Likewise, the production of beer and wine was known in ancient times although the relationship to saccharides and fermentation was not appreciated then. Manufacture of sugar as a sweetener including refinements such as decolorization with charcoal also has ancient origins. The commercial importance of carbohy-

Glycoconjugates and Carbohydrates

835
H H C C HO H H C C C OH O H OH OH

H C H HO H H C C C C

O OH H OH OH

CH2OH
D-Glucose, an aldohexose

CH2OH
D-Fructose, a ketohexose

FIGURE 2 Projection formulas of D-glucose and D-fructose.

are introduced. Accordingly, there are four aldotetroses, eight aldopentoses, four ketopentoses, eight ketohexoses, etc. (Fig. 3). Fischer recognized that sugars such as glucose existed in a ring form (i.e., a hemiacetal or hemiketal formed by addition of a hydroxyl group to the carbonyl carbon). Based on his knowledge of lactones, he assigned the participating hydroxyl to C-4 in the case of glucose, thus forming a five-membered (furanose) ring. The thermodynamics, however, show that the six-membered (pyranose) ring is the more stable form and that is the one predominating in solution. Note that the internal addition reaction at C-1 causes that carbon to become asymmetric and, therefore, two new isomeric structures are formed. To distinguish this asymmetry from that present at the other asymmetric centers, the term anomer is employed with the designation α or β reflecting projection of the newly formed hydroxyl group on the same or opposite side of the ring, respectively (Fig. 4). The ring structures (aldopyranose for the six-membered rings and aldofuranose for the five-membered rings) are, in aqueous solution, in ready equilibrium with the openchain, free aldehyde form. Thus, a solution of glucose in water contains five species: free aldehyde, two furanoses, and two pyranoses (Fig. 5). The aldehyde form represents about 0.025% of the total, with the bulk made up of the two pyranoses (the beta form predominates; see below). Although the six-membered ring is the energetically preferred form, the predominant ketohexose, D-fructose, is found in the furanose form in combined structures (sucrose, for example) although it too prefers the six-membered ring when free in aqueous solution. Other commonly occurring furanosides include ribose and deoxyribose as components of RNA and DNA (Fig. 6); galactose is also found as a furanoside in several plant and microbial polysaccharides. A key property of the free monosaccharides is their ability to be readily oxidized (initially via the available alde-

hydo function), especially in alkaline solution. This characteristic was used as an analytical tool for many years, and the term reducing sugar was employed to designate saccharides with a free (or potentially free, in aqueous solution) carbonyl function. Prior to the advent of specific enzymatic methods, this was the standard procedure for measurement of blood glucose levels (i.e., reducing sugar). Monosaccharides that differ from one another at a single asymmetric center other than the one derived from the carbonyl carbon are termed epimers (D-galactose is the 4-epimer of D-glucose, and D-mannose is the 2-epimer, for example; Fig. 7). Mirror-image structures are termed enantiomers. An important and characteristic feature of saccharides is their ability to interact with plane-polarized light. Thus, each sugar has a characteristic optical rotation that is a complex function of the asymmetric centers, their polarizability, solvent interactions, etc. The sign (+ or dextrorotatory, − or levorotatory) and magnitude cannot be determined a priori. Nor is it true that all sugars of the same configuration (D- or L-) will have the same sign of optical rotation. Thus, D-glucose is dextrorotatory, whereas D-fructose is levorotatory. Since monosaccharides tend to crystallize in a single anomeric form rather than as mixtures, dissolving them in water results in a change in the optical rotation from that of the pure anomer (either alpha or beta) to that of the thermodynamically defined equilibrium mixture. This phenomenon is termed mutarotation (Fig. 8). As the principles of conformational analysis became known, it was rapidly realized that the pyranose form of sugars adopted a chair structure analogous to that of cyclohexane (Fig. 9). In this spatial arrangement, equatorial hydroxyl groups are thermodynamically more stable (energy differences of about 1.5 kcal) than axial ones with the exocyclic hydroxymethyl group having a much larger effect. It is not surprising, therefore, that β-D-glucose, the allequatorial structure, is the most prevalent natural sugar and the most prevalent organic compound on earth (Fig. 10). As a philosophical aside, it may be inferred that the choice between D- and L-glucose was made very early and on a basis we do not understand since they are conformationally equivalent. It is also expected that other naturally occurring sugars such as D-mannose or D-galactose would have only a single axial hydroxyl. Of the 16 possible aldohexoses, D-glucose, D-mannose, and D-galactose are widely distributed in nature while the remainder are of laboratory interest only. Parenthetically, idose (three axial hydroxyls in the classic conformer) has never been obtained in crystalline form. Additional widely distributed sugars include D-xylose (all-equatorial aldopentose), D-ribose, and 2deoxy-D-ribose (backbone components of RNA and DNA,

836
CH2OH CO H H H C C C OH OH OH HO H H CH2OH CO C C C H OH OH H HO H CH2OH CO C C C OH H OH HO HO H CH2OH CO C C C H H OH

Glycoconjugates and Carbohydrates

CH2OH
D-Allulose (D-Paloose)

CH2OH
D-Fractose

CH2OH
D-Sorbose

CH2OH
D-Tagatose

CH2OH CO H H C C OH OH HO H

CH2OH CO C C H OH

CH2OH
D-Ribulose

CH2OH
D-Xytulose

CH2OH CO HO C H H

CH2OH CO C OH

CH2OH
L-Erythrulose (L-Threulose)

CH2OH
D-Erythrulose (D-Threulose)

CH2OH CO CH2OH Dihydroxyacetose CHO H H H H C C C C OH OH OH OH HO H H H CHO C C C C H OH OH OH H HO H H CHO C C C C OH H OH OH HO HO H H CHO C C C C H H OH OH H H HO H CHO C C C C OH OH H OH HO H HO H CHO C C C C H OH H OH H HO HO H CHO C C C C OH H H OH HO HO HO H CHO C C C C H H H OH

CH2OH
D-Allose

CH2OH
D-Altrose

CH2OH
D-Glucose

CH2OH
D-Mannose

CH2OH
D-Gulose

CH2OH
D-Idose

CH2OH
D-Galactose

CH2OH
D-Talose

(VIII)

(IX)

(X)

(XI)

(XII)

(XIII)

(XIV)

(XV)

CHO H H H C C C OH OH OH HO H H

CHO C C C H OH OH H HO H

CHO C C C OH H OH HO HO H

CHO C C C H H OH

CH2OH
D-Ribose

CH2OH
D-Arabinose

CH2OH
D-Xylose

CH2OH
D-Lyxose

(IV)

(V)

(VI)

(VII)

CHO H H C C OH OH HO H

CHO C C H OH

CH2OH
D-Erythrose

CH2OH
D-Threose

(II)

(III)

CHO H C OH

CH2OH
D-Glycaraldehyde

(I)

FIGURE 3 Structures of aldoses and ketoses up to six carbons in length. Monosaccharides with up to nine carbons are present in nature.

Glycoconjugates and Carbohydrates
5

837
HOCH2 H 4 O H
3

H H HO H H

C C C C C

OH OH H OH O

HO H HO H H

C C C C C

H OH H OH

O H
2

5

OH
1

HOCH2 H 4 H
3

O H
2

OH
1

H

H

OH H

OH O C H H H C C C OH OH OH H H H C

OH O

H

CH2OH -D-Glucose (Fischer projection) CH2OH O OH HO

CH2OH -D-Glucose (Fischer projection) CH2OH O OH

CH2 C C OH OH

CH2OH
D-Ribose

CH2OH 2-Deoxy-D-ribose
D-ribose

OH OH HO OH OH -D-Glucose -D-Glucose (Haworth projection) (Haworth projection)
FIGURE 4 Alpha- and beta-forms of D-glucose.

FIGURE 6 Structures of (right).

(left) and 2-deoxy-D-ribose

H H HO H H

C C C C C

OH OH H OH O
1

HO H O Pyranose rings HO H H C C H OH H OH OH

C C C C C

H OH H OH O

respectively). A large number of other sugars are present in nature. The vast majority of naturally occurring saccharides exist other than as free monosaccharides (only D-glucose and D-fructose are widely found free) and are present in disaccharides and complex oligomeric and polymeric structures in both plants and animals. About a dozen different sugars are present in mammals, all but glucose in combined form. Plants and microorganisms have very diverse saccharides, again all in combined forms. A. Derivatives—Natural and Laboratory The hydroxyl group formed as a result of ring closure represents a site for attachment of a broad variety of substituents. Compounds that are formed in such reactions are full acetals (ketals) and thus no longer undergo interconversion at the anomeric center. The configuration of such glycosides is therefore either alpha or beta depending on the relationship between the C-1 group and

CH2OH
-D-Glucose

CH2OH
-D-Glucose

38%

H

2

62%

HO 3 C H H
4 5 6

C C

CH2OH HO H HO H C C C C C OH H OH O H

1

CHO C C C C H H OH OH H HO H H

1 2 3 4 5 6

CHO C C C C OH H OH OH H HO HO H

1 2 3 4 5 6

CHO C C C C OH H H OH

H H HO H H

C C C C C

OH OH O H

~ 0.02% Furanose rings

HO HO H H

2 3 4 5 6

OH

H

CH2OH

CH2OH

CH2OH

CH2OH Less than 0.5%

CH2OH Less than 0.5%

D-Mannose

D-Glucose

D-Galactose

(epimer at C-2)

(epimer at C-4)
D-galactose

FIGURE 5 Equilibrium mixture of D-glucose in aqueous solution.

FIGURE 7 Relationship between D-glucose and (4-epimer), and D-mannose (2-epimer).

838
H
1C

Glycoconjugates and Carbohydrates
O

CH2OH H O H OH
D-Glucose.

H OMe

H HO H H

2 3 4 5 6

C C C C

OH H OH OH

H

HO

H OH Methyl α-D-glucopyranoside CH2OH H O H OH H H OMe

CH2OH

6 5

CH2OH C OH H C
2

HO
H O

H
4

H C OH

H OH Methyl β-D-glucopyranoside
FIGURE 11 Methyl glucopyranosides.

HO C 3 H
6 5

C OH

CH2OH O H C
2

6 5

CH2OH O H C
2

H C H 4 C OH HO C 3

H OH mutarotation H C H 4 C OH HO C 3 H

OH H

C

C OH

H OH -D-Glucopyranose

-D-Glucopyranose

FIGURE 8 Interconversion of glucose forms in aqueous solution. The optical rotation is initially reflective of the starting material (alpha or beta) and changes (mutarotates) until equlibrium is achieved.

Axis ax eq eq ax ax eq O ax eq eq ax ax Two possible chair forms eq eq

ax ax eq

Axis ax O eq ax eq

the projection of the ring; if on the same side, designate alpha, otherwise, beta (Fig. 11). A wide variety of natural and man-made derivatives (glycosides) are known with the substituents, aglycones, varying from simple methyl groups to complex organic molecules including other sugars (see below). In addition to substitution of the anomeric hydroxyl, many modifications of the hydroxyl loci are known in nature. Most prevalent are those in which the hydroxyl group at C-2 is replaced by an amino function, generally acetylated. The sugar 2-deoxy 2-acetamido-D-glucose (N acetylglucosamine) is distributed throughout nature and, in its polymeric form (chitin), forms the organic matrix of insect and arthropod exoskeletons. Hence, it is likely the second most prevalent organic molecule on earth. Other variations include oxidation (C-6 or C-1) to form carboxyl groups and loss of a hydroxyl to form deoxy sugars (Fig. 12). B. Chemical Transformations The chemistry of carbohydrates involves transformations at both the carbonyl function and the chain hydroxyl groups. The presence of multiple species of simple sugars in aqueous solution gives rise to complex reaction patterns wherein product distribution may well be determined by kinetic rather than thermodynamic considerations. 1. Reactions at the Carbonyl Group

FIGURE 9 Conformational structure of pyranose rings.

H HO HO

CH2OH H H O H OH H OH

FIGURE 10 All-equatorial structure of β-D-glucose.

The carbonyl function reacts directly with a variety of reagents either as the free aldehyde or keto group or as the corresponding hemiacetal or hemiketal. In early work, Fischer introduced phenylhydrazine, which condenses at

Glycoconjugates and Carbohydrates
Glucose family CH2OH H H OH HO H OH O H H HO H NH2 OH H H OH CH2OH O H H HO H NH C β -D-Glucose CH2 H H OH HO H OH O PO3 O OH H H HO H NH2
2

839
Amino sugars O H OH H H H H O NH2 OH CH2OH HO H OH O H H β -D-Galactosamine HO H H OH H H OH CH2OH O H 2N H β -D-Mannosamine OH

CH2OH OH H

β -D-Glucosamine CH2OH H H R O H H OH

CH3 N-Acetyl-β -D-Glucosamine Deoxy sugars CH2OH H H R HO H NH C O O H H OH R O CH3 C H HO OH H β -L-Fucose H H O CH3 H OH OH H OH OH α -L-Rhamnose H HO CH3 H H O H H OH

COO

β -D-Glucose 6-phosphate O C H H OH HO H OH O H H β -D-Glucuronate HO OH H O

Muramic acid

CH3 N-Acetylmuramic acid Acidic sugars CH3

O C O

O R H H OH C C OH OH

CH2OH OH H OH H H OH C O HO O H

CH2OH O H OH H
D-Glucono-δ

O O

C HN

H R H

H OH -lactone

H H

H OH

CH2OH

D-Gluconate

N-Acetylneuraminic acid (sialic acid)

FIGURE 12 Sugars related to D-glucose that occur in nature.

the carbonyl carbon and oxidizes the adjacent secondary hydroxyl (aldoses) or primary alcohol (ketoses) to allow a second condensation. The resulting derivatives, phenylosazones, are crystalline and were employed both for identification and determination of stereochemistry. Phenylhydrazine itself is a liver poison and it is generally believed that Fischer suffered from liver damage due to this reagent (no longer employed). A simple example of the utility of this conversion is the fact that D-glucose, D-mannose, and D-fructose all give the same phenylosazone showing that their relative stereochemistry at the remaining asymmetric centers (carbons 3,4,5) is identical (Fig. 13). Certain transformations may dictate that the carbonyl function remain available for later chemistry while reac-

tions are carried out elsewhere in the chain. This may be achieved by formation of dithioacetals or, in the ring form, by suitable substitution of the anomeric hydroxyl group (see below). 2. Reaction at the Anomeric Hydroxyl Group The hemiacetal nature of the anomeric hydroxyl group makes it the most reactive of that type. Direct oxidation can be carried out with several reagents, most classically, hypoiodite. This results in rapid oxidation to the aldonolactone and is only exhibited by aldoses. Alternatively, reaction with alcohols results in the formation of full acetals (glycosides). A very large variety of such structures have been made. Control of stereochemistry at the anomeric center is complex, involves the configuration at carbon two and kinetic factors. In general, both alpha- and beta-glycosides are formed. The configuration is very important in biological systems since enzymatic transformations are generally stereospecific with essentially all glycosidases having absolute specificity for either the alpha or beta form. Glycoside formation is generally carried out by reaction of the sugar with the appropriate alcohol in nonaqueous solution, typically with an acid catalyst.

CH H HO H H CHO OH C6H5NHNH2 H OH OH CH2OH C

NNHC6H5
C6H5NHNH2

NNHC6H5 HO H H OH H OH CH2OH Same phenylosazone

HO HO H H

CHO H H OH OH CH2OH

D-Glucose

D-Mannose

FIGURE 13 Formation of phenylosazone from D-glucose or D-mannose. The loss of asymmetry at C-2 yields identical products.

840 3. Reactions at Secondary Hydroxyls Substitution reactions at secondary hydroxyls are generally performed either for analysis of structure or to serve a protective function during other reactions. Etherification of the nonanomeric hydroxyls was an important structural tool in the analysis of oligosaccharide and polysaccharide structure. Methyl ethers have been employed for structural determination for more than 75 years. Thus, methyl ether formation in a polysaccharide results in substitution only at free hydroxyls. Subsequent analysis of the methylated derivatives reveals positions previously occupied in glycosidic linkage. Reagents used for this purpose have evolved from dimethylsulfate to the commonly employed method of Hakomori using sodium hydride and dimethylsulfoxide. Another frequently used ether substituent is the benzyl group, which is stable under a variety of reaction conditions but can be removed by catalytic hydrogenation. This is employed mainly during synthetic schemes where protection of specific hydroxyls is required. Currently, ether formation is used primarily as an adjunct to mass spectrometric analysis. The secondary hydroxyl groups can also be esterified. Acetyl substitution via acetic anhydride is often used, especially as a protective group. Tosyl substitution using toluenesulfonyl chloride is often employed since the properties of that ester (good leaving group) allow for SN2 displacement resulting in inversion of configuration. Acetate esters are readily removed under basic conditions (methoxide ion) when glycosides remain intact. 4. Reactions at the Hydroxymethyl Group The exocyclic nature of this group makes it the second most reactive of the hydroxyl functions. Tritylation is often employed to transiently block C-6 (ready removal under mild acid conditions). This is also the site of enzymatically catalyzed phosphorylation of glucose and other sugars, the first step in their metabolic utilization. Oxidation to a carboxylic acid is common. The resulting uronic acid is, for the common sugars, widely distributed in natural polysaccharides in both plants and animals. It is of interest that L-iduronic acid (the 5-epimer of D-glucuronic acid) and L-guturonic acid (the 5-epimer of D-mannuronic acid) are formed in nature after the respective precursor uronic acid has been incorporated into polymeric linkage. 5. Hydroxyl Pairs The proximity and defined stereochemistry of hydroxyl groups in the ring structures of saccharides allows for a number of selective reactions. These include formation of
H H OH HO H OH CH2OH O H OH H
Me2CO ZnCl2, H3PO4

Glycoconjugates and Carbohydrates
OCH2 Me2C OCH OH H H O O H O CMe2 H

FIGURE 14 1,2 5,6-diisopropylidene D-glucofuranose. Formed by reaction of D-glucose with acetone in the presence of a suitable catalyst. The furanose product dominates due to kinetic control of the reaction.

acetals or ketals and of a group of interactions restricted to hydroxyls on adjacent carbons. Reaction of glucose with acetone results in the formation of 1,2 5,6-diisopropylidene glucofuranose (Fig. 14). The furanose product arises because of the high reactivity of the exocyclic hydroxymethyl group thus favoring the formation of the five-membered ring. This is a commonly used protective scheme since the ketal function is readily cleaved by mild acid; in fact, selective cleavage of the 5,6 ketal can be accomplished. Acetals are likewise formed; reaction of glucose with benzaldehyde results in formation of the 4,6 benzylidene derivative, another useful synthetic intermediate. Adjacent hydroxyls that are cis can undergo several different reactions. Included in this group is complexation with borate to form a transient cyclic adduct (Fig. 15). This alters chemical reactivity and electrophoretic properties of the reactive sugar and has several applications. A key property of cis-hydroxyls is their susceptibility to carbon–carbon bond cleavage by metaperiodate. This oxidative reaction proceeds via a five-membered ring intermediate, generates a pair of aldehydes, and results in scission of the carbon chain (Fig. 16). It has been demonstrated that the cis orientation is necessary for this reaction since fixed ring structures with trans-hydroxyls do not react. The reaction has been utilized to identify saccharides in tissues via subsequent treatment of the generated aldehydes with a suitable amine, forming a colored Schiff base product (periodate-Schiff reaction). C. Biosynthesis The fixation of carbon dioxide via photosynthesis is the initiating reaction in saccharide synthesis in nature. Light

HO B HO

OH OH

HO R HO

HO B HO

O R O H2O

FIGURE 15 Borate ester formation. cis-Hydroxyls are preferred. The extent of reaction was originally monitored by following the change in conductivity of borate solutions on the addition of saccharide.

Glycoconjugates and Carbohydrates
CH2OH
C C OH HIO4 OH 2 C O HIO3 H2O

841

HO
4

O H OH H H

H O

H HO O O P O UDP-galactose O O P O O CH2 H H
NAD UDP-glucose 4-epimerase

HN O O H H OH OH N

C C

OH IO4 OH

( H2O)

C C

O I O

O O O

C C

O IO3 O

FIGURE 16 Action of periodate on vicinal diols. The proposed cyclic intermediate suggests that rigid structures with transhydroxyls will react poorly, a prediction confirmed experimentally.

energy is harvested via chloroplasts and used to provide chemical potential in the form of adenosine triphosphate, and reducing equivalents. The key intermediate, Dribulose 1,5-bisphosphate, fixes carbon dioxide (the dark reaction) yielding products that are ultimately converted to D-glucose via a series of reactions of phosphorylated sugar intermediates. All other naturally occurring sugars are derived from glucose in transformations that involve phosphorylated or nucleotide-linked sugars. Thus, glucose-6phosphate is converted to fructose-6-phosphate, which in turn is converted to mannose-6-phosphate; fructose6-phosphate is also aminated to form 2-deoxy 2-amino glucose-6-phosphate. Galactose is formed by epimerization at C-4 of uridine diphosphoglucose (Fig. 17), fucose by a series of reactions initiating with guanosine diphosphomannose, etc. Thus, the diversity in saccharides seen in the biosphere stems from a single precursor, D-glucose. This is, therefore, the only required dietary saccharide for man.

CH2OH H
4

O H OH H H

H O

HO HO O O P O UDP-glucose H OH OH O O P O O CH2 H O H HN O

N

H

FIGURE 17 Biosynthesis of D-galactose by epimerization of uridiine diphosphoglucose.

II. OLIGOSACCHARIDES
The ability of the anomeric hydroxyl group of a sugar to be substituted with another sugar (via one of its hydroxyl groups), and for this process to be iterated leads to formation of di- and higher saccharides. Because several hydroxyls are available for such linkages and the anomeric configuration can vary as well, the number of possible structures grows exponentially even for oligomers derived from the same sugar. This may explain, in part, why many biological recognition events involve saccharides as a ligand. Two disaccharides are widely distributed in nature: sucrose and lactose. Sucrose (α-D-glucopyranosyl-β-Dfructofuranoside) is the table sugar of commerce and a major industrial product (Fig. 18); primary sources are sugar cane and beets. This saccharide and its source material (cane molasses) serve as the basis for rum production; cruder precursors are important medium additives for the

production of antibiotics. The sweetness of sucrose, and other simple sugars, is a major aspect of their commercial importance. The fact that sucrose is a nonreducing sugar contributes to its stability and market value. Lactose (β-D-galactopyranosyl-1-4-D-glucopyranose) is the major sugar of milk (Fig. 19). It is of interest that the ability to utilize either sucrose or lactose as a source of calories is dependent on their enzymatic hydrolysis to the constituent monosaccharides in the intestine. About onethird of the oriental population lacks the requisite galactosidase and is thus lactose intolerant; this same problem occurs in some infants (developmentally related) and results in a “colicky” baby.
6 CH
2OH

H
4

5

O H
2 1

H OH
3

H

HOCH2
2

1

O H
3

H HO
4 6 5

HO

O

CH2OH

H

OH

OH

H

Sucrose -D-fructofuranosyl -D-glucopyranoside Fru( 2 1 )Glc
FIGURE 18 Structure of sucrose. Note that this sugar is nonreducing.

842
6 CH
2OH

Glycoconjugates and Carbohydrates
6 CH
2OH

OH

OH
6

HO
4

5

O H
2 1

H OH
3

H O 4 H

5

O H
2 1

HO

O O OH 1

H OH
3

OH H

O O OH ( 1

O HO

4

H

3

H OH Lactose ( form) -D-galactopyranosyl-(1 4)- -D-glucopyranose Gal( 1 4)Glc
FIGURE 19 Structure of lactose, the major sugar present in milk.

H

OH

4)-linked D-glucose units

FIGURE 21 The repeating unit of cellulose showing hydrogen bond interactions. The extended structure allows chains to stack via the relatively hydrophobic axial faces of the pyranose rings.

A. Homopolysaccharides Other naturally occurring disaccharides of interest are maltose (α-D-glucopyranosyl-1-4-D-glucopyranose), an intermediate in the digestion of starches, and trehalose (α-D-glucopyranosyl-α-D-glucopyranoside), the major sugar of insect blood (Fig. 20). Naturally occurring tri- and higher oligosaccharides (less than 10 sugar units) are rare. Raffinose (3), stachyose (4), and verbascose (5) are all assembled on a sucrose core with substituents originating on the glucose moiety. The major homopolysaccharides are cellulose, chitin, starches (amylose and amylopectin), glycogen, and xylans. Cellulose, a linear glucose polymer linked β-1-4 is the predominant natural product in the biosphere. The allequatorial structure allows for extensive hydrogen bonding, whereas the axial, somewhat hydrophobic faces favor a nonaqueous environment. The resulting aggregates have a highly ordered, quasi-crytalline arrangement (Fig. 21); hence, the unusual stability exhibited by the molecule, exemplified in trees and wooden artifacts. The broad distribution and low cost of cellulose has made it a major starting material for industrial development while the unmodified polymer remains the basis for wood, paper, and cotton. Chitin, identical in linkage to cellulose but composed of N -acetylglucosamine instead of glucose, is the major structural component of insect and crustacean exoskeletons as well as a cell wall component of molds and fungi. The structural comments regarding cellulose also apply generally to chitin, especially in terms of stability. Less industrial development has been done with this polymer, in part because shrimp shells may present a more difficult starting material than trees. Amylose and amylopectin are closely related, α-1-4linked glucose polymers. They are major constituents of starches and hence key nutrients worldwide. Amylose is a linear chain with up to a few hundred glucose units. Amylopectin has the same backbone chain but α-1-6 branches approximately every 20 glucose units. The branches have the same linkages as the main chain. This ramified structure allows for efficient packing in cells. Glycogen (Fig. 22), present in all higher animals, is closely related to amylopectin in that it has the same fundamental structure of a linear glucose chain with branches, and the same linkages. In this case, however, branches occur about every seventh residue, yielding a highly rebranched, tree-like envelope. This is essential for both packing in cells and for the rapid degradation of the polymer to provide critical metabolic intermediates. Glycogen serves as a primary energy reservoir in muscle and as a source of circulating glucose in the liver. It is of interest

III. POLYSACCHARIDES
Polymeric saccharides can have 10,000 or more sugar units, four or five different sugars, and a branched structure. The diversity that can be envisioned is huge and much of it is actually observed. Classification of polysaccharides differentiates homopolymers (a single sugar type) and heteropolymers (two or more sugar types) with subtypes reflecting linear or branched structures within each group. The physical and biological properties of each polymer depend on the architecture of the molecule as well as the enzymatic machinery available to interact with it. Thus, cellulose, a linear β-1-4 glucose polymer is a stable, highly organized polymer, indigestible by man, whereas starch, its α-linked counterpart is a major food source for all mammalian species.

6 CH

2OH

H O H
2 1

H
4

5

H OH
3

H O H
1

O OH
2

5

HOCH2 6 H
3

OH
4

HO

H

H

OH Trehalose -D-glucopyranosyl -D-glucopyranoside Glc( 1 1 )Glc

OH

H

FIGURE 20 Structure of trehalose, the predominant sugar present in the blood of insects.

Glycoconjugates and Carbohydrates

843 components necessitates, in many cases, that the saccharide chain have a defined three-dimensional conformation. Much work has been done to model the three-dimensional structure of proteins and many protein structures have been determined by X-ray crystallographic analysis. Much less has been done with complex saccharides in terms of either molecular dynamics or conformational analysis of larger structures. The following discussion is intended to be representative only. In addition to cellulose and xylan, plant cell walls contain a variety of complex heteropolysaccharides including other glucans. Detailed structures have only been determined for a limited number of these structures and their interactions, possibly covalent, with other plant components have rarely been determined in detail. Commercially important are pectins, polygalcturonides, which are key components of jellies and related uronides (gums) used as thickening agents in a broad variety of applications ranging from ice cream manufacture to the production of printing ink. Agars are galactans that are sulfated and also contain 3,6-anhydro-L-galactose, which have gelation properties. They have broad application in microbiology. Bacteria likewise produce a wide variety of complex, cell-surface polysaccharides. Strains of Streptococcus pneumoniae, the causative agent of bacterial pneumonia, are characterized by a capsular polysaccharide—more than 100 different types are known. Since the capsular material is immunogenic and protective, polyvalent vaccines have been developed that utilize the capsular polysaccharides from common strains. Similarly, organisms responsible for bacterial meningitis have a characteristic polysaccharide capsule that is also immunogenic and protective. Conversely, several bacteria have, as an exterior capsule, saccharides sufficiently similar to those produced in their animal host so as to serve as a mechanism for avoidance of host immune responses. Many gram-positive bacteria have, as essential cell wall components, a complex saccharide structure that contains muramic acid (a condensation product of N -acetylglucosamine and pyruvate) and other amino sugars and is cross-linked by a peptide (Fig. 23). A widely distributed heteropolysaccharide, hyaluronic acid, has both commerical and biological importance. This molecule, a repeating structure of D-glucuronic acid and N -acetylglucosamine with β-1-3 and β-1-4 linkages, is found in bacterial and animal sources, and it is one of the few complex saccharides not covalently linked to protein (Fig. 24). Molecular weights vary depending on source but often exceed two million. The polyanionic nature of the molecule leads to a relatively extended solution conformation. This coupled with the highly hydrophilic chemistry results in solutions with very high viscosity, an important physical property. This is utilized in treatment of

3 units

6–7 units

6 CH

2OH

H
4

O H OH H H OH
1

H ( 1 6) branch point O
6 CH
2

Branch

O

H
4

O H OH H H OH
1

H O

Main Chain

O

FIGURE 22 Schematic structure for glycogen and details of a typical branch point. The same linkages are present in amylopectin, which has less frequent branches.

that the biosynthesis of glycogen initiates on a core protein (glycogenin), which may be cleaved from the polysaccharide subsequent to polymerization. Xylans are a group of polymers based on a structure analogous to that of cellulose wherein xylose is the repeating unit. The simplest representative contains only D-xylose with β-1-4 linkages and is a common component of plant walls. Several heteropolysaccharides utilize the xylan backbone and have various other saccharides as branches. Xylans are often associated with cellulose in plant cell walls. B. Heteropolysaccharides The diversity of heteropolysaccharides is enormous. Distributed throughout the animal and plant kingdom, this class of macromolecules rivals proteins in diversity. Initially thought to function only as structural components, it is now realized that specific sequences may have information content and can serve as recognition or signaling elements. In addition, modifications of saccharides already incorporated into a polymeric chain (sulfation, for example) add an additional level of complexity. The ability of these materials to function as other than strictly physical

844

Glycoconjugates and Carbohydrates

MurNAc (Gly)5 GlcNAc
L-Lys L-Ala

GlcNAc
D-Glu

NH2 MurNAc
L-Ala

MurNAc (Gly)5 GlcNAc
L-Lys D-Ala L-Ala D-Glu

D-Ala

(Gly)5 NH2 GlcNAc MurNAc
L-Ala D-Glu L-Lys D-Ala

D-Glu L-Lys D-Ala D-Ala

NH2

(Gly)5 GlcNAc

NH2

(Gly)5

MurNAc
L-Ala

GlcNAc

D-Glu L-Lys D-Ala

Glu

MurNAc
4

GlcNAc GlcNAc MurNAc MurNAc
L-Ala L-Ala D-Glu

Gly

GlcNAc

D-Glu L-Lys

Glu Gly

Gly

L-Ala-D-Glu-L-Lys-D-Ala-L-Lys D-Ala

[L-Ala-D-Glu-L-Lys-D-Ala]2

D-Ala-L-Ala-D-Glu-L-Lys-D-Ala

FIGURE 23 Typical bacterial cell wall structures (peptidoglycans).

Glycoconjugates and Carbohydrates
CH2OH H COO H O H OH H GlcA H H OH ( 1 3) HO O H NH C CH3 GlcNAc H O H H O

845

HOCH2 H
( 1 4)

O O NH H C CH2

C CH NH

O

H OH H

H NH C CH2 O Asn

O
O

FIGURE 24 Repeating unit of hyaluronate. This polysaccharide is distributed throughout connective tissue and is the only mammalian polysaccharide not covalently attached to protein.

GlcNAc Examples:

osteoathritis of the knee (an injectable) and in eye surgery as a viscoelastic. In addition, cell surface receptors have been identified that recognize the saccharides in hyaluronate, and interaction with specific proteins is responsible for the aggregate properties of connective tissue proteoglycans (see below). This diversity illustrates that a single polysaccharide may have both informational and physical roles in nature.

Asn

Asn Asn

Glc GlcNAc Man Gal Neu5Ac Fuc GalNAc

IV. GLYCOCONJUGATES
Structural analysis of proteins has shown that up to half of naturally occurring proteins are subject to posttranslational modifications with the vast majority glycosylated. These covalent linkages involve several amino acids and have distinct structrual characteristics. In addition, a large number of lipids have covalently attached carbohydrate, necessary for their biological functions. The glycoproteins may be classified into two broad categories: N -linked and O-linked. A. N-Linked Glycoproteins N -linked glycoproteins have carbohydrate covalently attached to asparagine residues that occur in the sequence Asn-X-Ser/Thr, where X is any residue except proline. This is a necessary but not sufficient key for glycosylation since there are many examples of such sequences that are not glycosylated even when others on the same polypeptide are substituted with sugar. The linking sugar is invariably N -acetylglucosamine, which is the terminal saccharide of the attached unit (Fig. 25). The number of saccharides present in N -linked structures varies from about 7 to 20 or more; branching is universal with some structures having four separate branches (antennae). All of the saccharides have a common core structure: GlcNAcGlcNAc-Man3 . The first mannose is β-linked (unusual) and the other two mannoses are attached α-1-3 and α-1-6, thus forming the initial branch point. Unlike protein syn-

FIGURE 25 Schematic of a typical N-linked oligosaccharide. Note the core structure, which contains two N-acetylglucosamine and three mannosyl residues. This is present in all units of this type.

thesis wherein the amino acid sequence is controlled by the genetic one, the final structure of saccharides is rarely so conserved. Thus, a given protein with several glycosylation loci is quite likely to have differing saccharide structures, even at the same amino acid site—all of them will still contain the pentasaccharide core noted above. These various glycoforms give rise to a type of heterogeneity that is difficult to characterize completely and may have implications for function. The biosynthesis of these molecules is also unusual. The saccharide is preassembled, not as the final structure but as a common, 14-sugar, lipid-linked precursor that is transferred en bloc to the target asparagine in a cotranslational manner (Fig. 26). This saccharide unit (GlcNAc2 -Man9 Glc3 ) is trimmed to a GlcNAc2 -Man5 structure that is then modified by addition of either more mannosyl residues or by several sugars, including GlcNAc, Gal, NANA, and Lfucose. The latter category is generally termed complex as opposed to those which contain GlcNAc and Man only (high mannose). B. O-Linked Glycoconjugates O-linked glycoconjugates have substantial diversity in that the saccharide units may be covalently attached to serine, threonine, tyrosine, hydroxylysine, or hydroxyproline residues. In addition, the type of glycosyl substitution

846
5 GDP 5 GDP-

Glycoconjugates and Carbohydrates

α(1,2) UDP UDPUMP Tunicamycin UDPP P P P P P P

α(1,2) α(1,3) α(1,6) β(1,4) β(1,4) β(1,4) GDP Amphomycin Translocation GDPP P

CYTOSOL

UDP

UDP-

Dol ER Membrane
P P P P P P

(x 4) LUMEN OF ER β(1,4) β(1,4) = GlcNAc = Mannose = Glucose α(1,6) α(1,3) α(1,2) α(1,2)

(x 3) β(1,4) β(1,4) α(1,6) α(1,6) α(1,2) α(1,3) α(1,3) α(1,2)

α(1,2) α(1,2) α(1,3) α(1,3) α(1,2)

FIGURE 26 Biosynthesis of the 14-sugar, lipid-linked oligosaccharide, the universal precursor for N-linked glycosylation.

varies widely, from single sugars to extended polysaccharide chains. The following discussion highlights key features of these types but is not intended to provide full details. One major category of O-linked glycosylation is termed mucin type. This is characterized by linkage of the sugar (N -acetylgalactosamine in the alpha configuration) to serine or threonine hydroxyl groups (Fig. 27). There is no identifiable consensus amino acid sequence known which targets specific residues to be substituted. The saccharide units range from di- to intermediate size oligosaccharides (up to 10 sugars) and are very diverse. Additional sugars present include galactose, N -acetylglucosamine, L-fucose, and sialic acid; some of the saccharide units may be sulfated. Mannose is characteristically absent. These molecules are often found in epithelial secretions; the protein cores may be quite large with a single glycoprotein having an aggregate molecular weight of one mil-

lion with a hundred or more saccharide units covalently attached. Glycosylation of tyrosine residues is unusual but a key step in the biosynthesis of glycogen, the major storage glucan of liver and muscle. The core protein, glycogenin, is able to autoglucosylate and attaches a series of glucosyl residues to a single tyrosine in the protein. When the glucose chain has reached four (or more) units (all linked α-1-4), the resulting saccharide moiety is then recognized by glycogen synthase for continuation of glycogen formation. The final polysaccharide may have several thousand glucose residues. Currently about 20 proteins have been identified as collagens. Criteria for this classification include the presence of a triple helical domain (“collagen helix”) and the presence of hydroxyproline and hydroxylysine residues. The latter may also be glycosylated with either a single galactose residue or a disaccharide (glucopyranosyl

Glycoconjugates and Carbohydrates

847

O CH2 HO H H NH C CH2 GalNAc Examples: Ser/Thr Ser O O H OH H H O CH2 C CH NH
FIGURE 28 Core saccharide of proteoglycans.

( 1 (GlcA

3) ( 1 GalNAc)n

4) ( 1 GlcA

3) ( 1 Gal

3) ( 1 Gal Xyl

4) Ser

O

Chondroitin sulfate Core protein

FIGURE 27 Typical O-linked, mucin-type oligosaccharide.

1-4 galactosyl-hydroxylysine). The extent of glycosylation varies from as little as 1 per 1000 amino acids to 8 or more. It is suggested that addition of carbohydrate causes local disruption in the collagen helix, thus opening the protein structure into a more mesh-like conformation. Glycosylation on hydroxyproline residues is rare but has been reported in both plants and fungi. Interestingly, this does not occur in collagen where hydroxyproline is a prominent residue. An unusual form of O-glycosylation has recently been described wherein a single N -acetylglucosaminyl residue is attached to either a serine or threonine residue in target proteins. In contrast to other O-linked glycoproteins, the entities involved are cytosolic and not secreted. It has been suggested that this modification is reciprocal with phosphorylation, a common form of regulatory substitution. A major class of O-linked glycoconjugates is the proteoglycans, key components of the extracellular matrix of animals. These glycoconjugates are distinguished by the linking sugar (D-xylose, attached to serine), a linear core saccharide (Gal-Gal-Xyl; Fig. 28), and continuation of the saccharide chain as a linear polysaccharide, which contains alternating residues of an amino sugar (N -acetylglucosamine or N -acetylgalactosamine) and a uronic acid (D-glucuronic or L-iduronic acid) (Fig. 29). The saccharides are generally sulfated (may include N -sulfation instead of N -acetylation of the glucosaminyl residues) giving rise to chains of considerable structural

diversity. Examples include heparin, a natural anticoagulant, and the chondroitin sulfate proteoglycans. In the case of heparin, it has been established that the antithrombin activity resides in a specific pentasaccharide sequence within the structure with a defined pattern of sulfation and sugar components. This type of “information” is known to be present in other complex saccharides and broadens the function of these molecules beyond that of space occupancy and water and electrolyte management. It is interesting to note that the biosynthesis of the iduronosyl moiety in heparin and related heparan sulfate chains occurs at the polymer level by inversion of configuration at C-5 of already incorporated glucuronosyl residues. Extracellular proteoglycans such as those of the chondroitin and dermatan sulfate families are associated with organization of the fibrillar elements of connective tissues (primarily collagens), bone deposition and maintenance of tissue hydration. Glycolipids represent another diverse class of glycoconjugates. In this case, the saccharides are assembled on a nitrogenous lipid (ceramide) derived from a C-18 amino alcohol (sphingosine) by fatty acylation of the amino group (Fig. 30). The primary hydroxyl group at C-1 is the site of sugar attachment. The saccharides range up to 10 sugars

Ser/Thr

CH2OH O3SO COO H H OH H GlcA O H H OH ( 1 3) H O H NH C CH3 O O H H O H ( 1 4)

GalNAc4SO3

FIGURE 29 Structure of the repeating unit of chondroitin 4sulfate. Other glycosaminoglycan chains presents in proteoglycans include dermatan sulfate (L-iduronic acid replacing Dglucuronic acid), variants with sulfate in the 6-position, and the heparin–heparan sulfate family, which contains both uronic acids, glucosamine, and both N- and O-sulfate esters.

848
Sphingosine HO
3CH

Glycoconjugates and Carbohydrates

CH X O C

CH

(CH3)12

CH3

tain sialic acid are termed gangliosides (originally isolated from neural tissue) and are involved in development, especially in the nervous system.
Fatty acid n

CH

N H

V. ANALYTICAL METHODS
Classical methods for the determination of saccharide structures are rarely employed at this time. Thus, methylation and periodate analyses, once the mainstay of carbohydrate work, have been superseded by mass spectrometry and nuclear magnetic resonance (NMR). Consider the typical problem—that of defining the complete structure of an oligosaccharide (polysaccharide), free or combined. Unlike procedures for analysis of DNA or proteins where automated methods provide rapid and unambiguous sequences, sugar sequences are not readily determined. Needed are linkage positions, configurations, and branching as well as identification of the specific sugars. In some cases, the use of specific enzymes can provide information about the linkage configuration but these do not always work, rarely are quantitative, and often do not discriminate between linkages. The advent of NMR has permitted complete analysis of oligosaccharides as long as sufficient material is available (or spectrometer time). Thus, the linkage configuration for pyranosides is readily apparent from the coupling constant between the protons on carbons-1 and -2 of the sugar in question. The order of sugars and possible branching can be determined

CH2

O

X CH2OH X Glucose H H OH OH H OH O H H

Neutral glycolipids Glucosylcerebroside

Lactosylceramide (a globoside)

Di-, tri-, or tetrasaccharide

Glc

Gal

Neu5Ac

Ganglioside GM2

Complex oligosaccharide

Glc

Gal

GalNAc

FIGURE 30 Ceramide and glycosphingolipid structures.

and are extremely diverse. They are found on cell surfaces and function as receptors and immunologic determinants. The blood group ABO system is defined by specific sugars present on erythrocyte glycolipids; thus, type A is characterized by alpha-linked N -acetylgalactosamine, type B by alpha-linked galactose, etc. Glycosphingolipids that con-

TABLE I Representative Lectins and Their Ligandsa Lectin family and lectin Plant Concanavalin A Griffonia simplicifolia lectin 4 Wheat germ agglutinin Ricin Animal Galectin-1 Mannose-binding protein A Viral Influenza virus hemagglutinin Polyoma virus protein 1 Bacterial Enterotoxin Cholera toxin Abbreviation Ligand(s)

ConA GS4 WGA

Manα1—OCH3 Lewis b (Leb ) tetrasaccharide Neu5Ac(α2 → 3)Gal(β1 → 4)Glc GlcNAc(β1 → 4)GlcNAc Gal(β1 → 4)Glc Gal(β1 → 4)Glc High-mannose octasaccharide Neu5Ac(α2 → 6)Gal(β1 → 4)Glc Neu5Ac(α2 → 3)Gal(β1 → 4)Glc Gal GM1 pentasaccharide

MBP-A HA VP1 LT CT

Source: Weiss, W. I., and Drickamer, K. (1996). Structural basis of lectin-carbohydrate recognition. Annu. Rev. Biochem. 65, 441–473. a More than 100 lectins are known. X-ray diffraction analyses of crystalline lectins suggest a common protein architecture for the saccharide binding sites.

Glycoconjugates and Carbohydrates

849 atively low molecular weights and limited by solubility), weight average (sedimentation or light scattering), and viscosity average (a mixed function that is also a measure of solution conformation—a measurement of mean endto-end distance). Solution structures are often helical for repeating linear polysaccharides and heavily dependent on solvent (ionic strength) for polyelectrolytes.

directly using mass spectrometric methods. In this case, however, direct identification of a specific hexose may not be possible since all will have the same molecular weight. Therefore, such studies are generally combined with compositional analysis (acid-catalyzed hydrolysis followed by chromatographic separation) and prior knowledge of the type of saccharide involved. Some structural data may also be obtained by the use of specific lectins (Table 1). These are a class of carbohydrate-binding proteins, generally of plant origin, that show high specificity for one or another saccharide. This may include linkage configuration and recognition of oligosaccharide. Information desired for polymeric structures beyond that of components and linkages will generally include molecular weight and solution conformation. Since polysaccharides (including chains found in proteoglycans) are polydisperse, the normal definition of molecular weight (as that of a single chemical entity) does not apply. Ideally, one should determine the mole fraction of each species present as a function of degree of polymerization. Currently, there are no methods for achieving this goal. Rather, physical measurements are made that provide number average (osmotic pressure, suitable only for rel-

SEE ALSO THE FOLLOWING ARTICLES
BIOPOLYMERS • KINETICS (CHEMISTRY) • PHARMACEUTICALS • ORGANIC CHEMISTRY, SYNTHESIS • POLYMERS, STRUCTURE • PROTEIN STRUCTURE • PROTEIN SYNTHESIS • STEREOCHEMISTRY • THERMODYNAMICS

BIBLIOGRAPHY
Advances in Carbohydrate Chemistry and Biochemistry—yearly review volume published by Academic Press, New York. Pigman, W., and Horton, D. eds., (1970). “The Carbohydrates, Chemistry and Biochemistry,” 4 volumes, Academic Press, New York. Stick, Robert V. (2001). “Carbohydrates: the Sweet Molecules of Life,” Academic Press, New York.

Ion Transport Across Biological Membranes
George P. Hess
Cornell University I. Relationship Between Transmembrane Inorganic Ion Flux and Transmembrane Potential II. Mechanism of Transmembrane Inorganic Ion Flux III. Inorganic Ion Transport and the Perception of Light IV. Inorganic Ion Transport and Integration of Environmental Information V. Inorganic Ion Transport and the Rapid Transmission of Electrical Signals over Long Distances (up to ∼1 m) VI. Properties of the Protein (Potassium Channel) that Allows K+ but Not Na+ to Cross the Membrane VII. Inorganic Ion Transport against a Concentration Gradient at Expense of ATP Hydrolysis VIII. Conclusion and Outlook

GLOSSARY
Adenosine triphosphate (ATP) Upon hydrolysis within a cell, ATP gives rise to adenosine diphosphate and adenosine monophosphate with the liberation of energy ( G is about − 12 kcal/mol). Faraday constant F Has a value of 96,485 coulombs per mole. Guanosine 3 ,5 -cyclic monophosphate (cGMP) Is converted to 5 guanosine monophosphate by an enzyme, cGMP phosphodiesterase. Universal gas constant R Has a value of 8.314 joules K−1 mol−1 (K is degrees Kelvin).

THE MOVEMENT of inorganic ions across biological membranes of animals plays a central role in the percep-

tion and integration of, and reaction to, environmental signals by the organism. Examples include vision, the integration and processing of this information (brain function), and the reaction of the organism to this information, for instance, muscle contraction (Fig. 1). Cells have the ability to hydrolyze adenosine triphosphate (ATP). The energy thus released is used to transport sodium and potassium ions across the cell membrane against a concentration gradient. This process establishes the transmembrane voltage (Vm ) of the cell membrane. The transmembrane voltage is perturbed by the movement of inorganic ions (generally Na+ , K+ , Cl− , and Ca2+ ) along the concentration gradient and voltage difference across the cell membrane that occurs in signal transmission between cells. There are many different transmembrane channel-forming proteins, which are activated by (1) a concentration gradient of inorganic ions across the membrane, (2) the transmembrane voltage,

99

100

Ion Transport Across Biological Membranes

of specific inorganic ions across the membrane. Invariably, this transport of inorganic ions across the cell membrane is accompanied by changes in the transmembrane voltage. The equilibrium transmembrane potential for a specific inorganic ion, for instance K+ , is given by the Nernst equation: EK+ = RT [K+ ]o . ln ZF [K+ ]i (1)

FIGURE 1 An environmental stimulus, for instance, light, activates a protein-mediated reaction in the eye, leading to the transmembrane flux of inorganic ions, a change in the transmembrane voltage, and neurotransmitter release. The neurotransmitter diffuses across a junction between the nerve terminal of the axon and the cell body of the adjacent cell about 20–40 nm in length, called synapse. The neurotransmitter binds to receptors in the membrane of the postsynaptic cell. Excitatory neurotransmitters ( ) activate receptors that form cation-specific transmembrane channels. Inhibitory neurotransmitters ( ) activate receptors that form anion-specific transmembrane channels. Once the transmembrane voltage of the cell is changed by a critical amplitude and sign (by ∼ +20 mV), an all-or-none process occurs. Transmembrane Na+ and K+ channels in the axonal membrane open transiently, resulting in an electrical signal that travels down the axon and neurotransmitter is again secreted. This process repeats itself and is terminated when the neurotransmitter is released adjacent to receptors on the surface of muscle cells. In the case of muscle cells, the receptor is the muscle nicotinic acetylcholine receptor and the neurotransmitter acetylcholine. The voltage change in the muscle cell membrane initiates muscle contraction. (From Hess, G. P., and Grewer, C. (1998). “Methods in Enzymology” (G. Marriott, ed.), Vol. 291, pp. 443–474, Academic Press, New York.) The resulting flow of inorganic ions through the membrane of the muscle cell results in a change of its transmembrane voltage Vm and muscle contraction.

•

(3) the binding of specific ligands to a channel-forming protein. (4) Some proteins use the energy liberated by the hydrolysis of ATP to transport inorganic ions against a concentration gradient. Only one example of each of these various proteins will be mentioned. In each case, the protein chosen is the one about which we have the most information. The proteins that facilitate inorganic ion transport across biological membranes are discussed in an order that illustrates their function in the life of an organism.

R, T, and F are the molar gas constant, absolute temperature, and Faraday constant, respectively. Z represents the valence of the inorganic ion and the subscripts o and i indicate whether the ion is outside or inside the cell membrane. The transmembrane potential of neurons is around −60 mV. In the nervous system, changes in the transmembrane potential due to a change in the flux rate of inorganic ions can be propagated rapidly and over distances as long as several feet via the axon, a long projection of many nerve cells (Fig. 1). At the axonal terminal, the voltage change initiates a process leading to the flux of calcium ions into the nerve terminal. This results in the secretion of chemical signals, neurotransmitters, which bind to membrane-bound proteins, neurotransmitter receptors, on adjacent cells. Upon binding specific neurotransmitters, the receptors transiently open transmembrane channels. The channels are permeable to Na+ , K+ , or Cl− , depending on the receptor. The resulting changes in the transmembrane voltage may lead to propagation of a signal to an adjacent cell. Thus, this interplay between chemical reactions and transmembrane voltage changes plays a decisive role in the rapid communication between nerve (and nerve and muscle) cells and in nervous system function. In 1890, Max Planck derived the relationship between the rate of movement of inorganic cations and anions across a porous barrier and the resulting electric field. If one assumes a constant electric field and constant inorganic ion concentration, the Planck equation is easily integrated and can be used to estimate the transmembrane voltage change, Vm , that results from the flow of inorganic ions across cell membranes. The resulting Goldman equation is Vm = RT PK (K+ )o + PNa (Na+ )o + PCl [Cl− ]i . ln F PK (K+ )i + PNa (Na+ )i + PCl [Cl− ]o (2)

I. RELATIONSHIP BETWEEN TRANSMEMBRANE INORGANIC ION FLUX AND TRANSMEMBRANE POTENTIAL
Membranes surround the individual cells of animals and organelles within the cell. They are composed of lipids and proteins. Specific proteins are responsible for the transport

PK , PNa , and PCl− represent the permeability coefficient of the membrane for K+ , Na+ , and Cl− , respectively. [K+ ], [Na+ ], [Cl− ] represent the molar concentrations of the ions, and the subscript o or i indicates whether the ions are outside or inside the cell membrane. As usual, R, T, and F represent the molar gas constant, the absolute temperature, and the Faraday constant respectively. How is the rate of ion movement through a proteinformed channel across a cell membrane related to the

Ion Transport Across Biological Membranes

101 controlled movement of inorganic ions through the cell membrane and the resulting change in transmembrane voltage, Vm .

II. MECHANISM OF TRANSMEMBRANE INORGANIC ION FLUX
FIGURE 2 Minimum mechanism to account for the rates of a neurotransmitter (acetylcholine) receptor-mediated cation translocation and for receptor inactivation and reactivation as a function of acetylcholine concentration. The active (A) and inactive (I) forms of the receptor bind neurotransmitter (L) in rapidly achieved equilibria denoted by the microscopic equilibrium constants (K). Active receptor with two bound ligand molecules (AL2 ) converts rapidly (1 to 2 msec) to an open channel (AL2 ) with an equilibrium constant for channel opening (1/ · = kcl /ko p where ko p and kcl are the rate constants for channel opening and closing respectively ). AL2 permits the movement of inorganic Na+ and K+ ions through the membrane, where Jm , is the observed rate constant for the flux of inorganic ions (Na+ , K+ ) through the open receptor-formed transmembrane channel (see Eq. 3). In the continued presence of neurotransmitter, the receptors reversibly form inactive forms I in the 10- to 200-msec time region, depending on the receptor and the concentration of neurotransmitter. This process is called receptor desensitization. (Reproduced with permission from Cash, D. J., Aoshima, H., and Hess, G. P. (1981) Proc. Natl. Acad. Sci. USA 78, 3381–3322.)

transmembrane voltage? In the nervous system, signal transmission is regulated by the binding of chemical signals, neurotransmitters, to membrane-bound proteins, called receptors. Commonly, when two molecules of a neurotransmitter have bound to the receptor, the protein forms a transmembrane channel that remains open for a few milliseconds, allowing the receptor-specific passage of sodium, potassium, or chloride ions. The chemical reaction for many neurotransmitter receptors can be written as shown in Fig. 2; in this case the kinetic mechanism of the nicotinic acetylcholine receptor is used as an example. This receptor plays an important role in signal transmission between nerve cells in the brain and between nerve and muscle cells (Fig. 1). The specific reaction rate for the transmembrane flux of inorganic cations controlled by the nicotinic acetylcholine ¯ receptor, Jm , has a value of about 5 × 107 M−1 sec−1 at ◦ 14 C. The relationship between the permeability coefficient P for a specific inorganic ion M± (Eq. 2) and the ¯ specific reaction rate Jm is given by: ¯ PM± = JM± R0 (AL2 ). (3)

R0 represents the moles of specific receptors in the cell membrane and (AL2 ) the fraction of the receptors that are in the open-channel form. Equations 2 and 3, therefore, establish the important relationship between the receptor-

Here, the emphasis is on kinetic techniques used to obtain the information needed to understand the mechanism of protein-mediated reactions that allow the transport of inorganic ions across biological membranes. A combination of structural, thermodynamic, and kinetic information is required to achieve this understanding. The use of X-ray crystallography, NMR measurements, and electron microscopy, to obtain structural information about proteins is described in detail elsewhere in this Encyclopedia. In 1976, Neher and Sakmann developed the singlechannel current-recording technique. It is simple, convenient, and widely used for measuring the properties of a single, open receptor channel, such as its conductance, lifetime, and ion specificity. In brief, a glass pipet with an internal diameter of 1 to 2 µ is attached to the surface of a cell membrane (Fig. 3A, left). Gentle suction is applied to isolate a small membrane patch within the glass pipet from the rest of the membrane (Fig. 3A, right). A silver chloride wire running from near the tip of the glass pipet to electrical recording equipment allows one to record the current flowing through single receptor-formed channels within the membrane patch (Fig. 3B). In the illustration, the deviation of the current from the baseline represents the current flowing through a single nicotinic acetylcholine receptor-channel in the presence of 20 µM acetylcholine, which activates this channel. The Gaussian distribution of the current amplitude gave a peak centered at 3 pA. The exponential distribution of the time the channel remains open (the lifetime distribution) gave a value of 2.4 msec for the mean lifetime of the open channel, τop . This lifetime is a measure of the rate constant for channel closing 1/τop = kcl (Fig. 2). Additionally, the technique allows one to determine conveniently the ion specificity of the channel from measurements of the current amplitude. In the case of transmembrane channels that open upon binding a specific ligand, how does one determine the other parameters of the channel-opening reactions? These are the dissociation constant of the neurotransmitter from the site controlling channel opening, the rate constant for channel opening kop and, therefore, the equilibrium constant for channel opening, 1/ = kop /kcl . In principle, many of these constants can be determined from the lifetime of the closed state, the time interval between the openings of single receptor-channels (Fig. 3B). From the mechanism in Fig. 2, a plot of the number of closures observed

102

Ion Transport Across Biological Membranes

FIGURE 3 The single-channel current-recording technique. (A) The tip of a borosilicate glass pipet, with a tip opening of 1–2 µm is pressed against the membrane of a frog muscle cell (left). A slight negative pressure (20–30 cm H2 O) is applied to the pipet for several seconds to form the seal between the membrane and the pipet (right). (Reproduced from O. Hamill et al. (1995). In “Single-Channel Recording,” 2nd edition (B. Sakmann and E. Neher, eds.), p. 663, Plenum Press, New York.) (B) A typical current trace recorded using the single-channel technique, a rat myoball cell containing nicotinic acetylcholine receptors, and 20-µM acetylcholine (pH 7.2, 22◦ C, and Vm = −80 mV). (Reproduced from F. Sigworth (1983). In “Single-Channel Recording,” first edition (B. Sakmann and E. Neher, eds.), Plenum Press, New York.)

within a definite time interval versus the time the channel was closed is expected to give a three-exponential distribution. From this distribution, the three different lifetimes, reflecting the constants to be determined, can be calculated. This evaluation requires many measurements to be made, which take time, and it is restricted to measurements made at low concentrations of neurotransmitters. At higher concentrations of neurotransmitter, the receptor becomes inactive, desensitized (in the millisecond time region) (Fig. 2) and the signal to be measured disappears. Additionally, we now know that many receptors on the cell surface exist in two forms, which desensitize with different rates. The desired, and missing, information that supplements results obtained with the single-channel current–recording technique can now be obtained by using a transient kinetic method with a microsecond time resolution, the laserpulse photolysis (LaPP) technique. The usual rapid kinetic techniques that are suitable for investigating small molecules in solution had to be modified for use with membrane-bound proteins. The time resolution for equilibrating ligands in solution with membrane-bound proteins is less than might be expected. This is because a layer of water molecules (the diffusion layer) covers the membrane containing the proteins on the surface of relatively large objects like cells, or even membrane patches with diameters in the micrometer range. Ligands in the solution surrounding the membrane-bound receptors must diffuse through the diffusion layer and this process may become rate limiting. The steps needed to overcome this problem are illustrated in Fig. 4. Photolabile precursors of neurotransmitters (“caged” neurotransmitters) that are biologically inactive have been developed. A photolabile precursor of carbamoylcholine, a stable analog of acetylcholine that activates the nicotinic acetylcholine receptor, is shown in the inset to Fig. 4. Photolabile precursors of all the major neurotransmitters are now available. This photolabile precursor of carbamoylcholine ([N -(α-carboxy-2-nitrobenzyl)-

carbamoylcholine] is equilibrated with nicotine acetylcholine receptors on the surface of a cell (Fig. 4A). At zero time, the compound is photolyzed, using a laser, by a single pulse within about 100 µsec to give carbamoylcholine and a biologically inert side-product, a 2-nitroso-α-keto carboxylic acid (Fig. 4 reaction). An optical fiber carries the light beam to the cell, which is attached to a currentrecording electrode (Fig. 4A). The technique for recording the current from all the receptors on the cell surface with high precision uses the same equipment as is used in the single-channel current-recording technique (Fig. 3). The increase in current that results when carbamoylcholine is liberated on the cell surface, due to the photolysis of caged carbamoylcholine is shown in Fig. 4B. The current is due the opening of receptor-channels on the cell surface and the flow of inorganic ions through them. In a different and slower time zone, the current then decreases due to receptor desensitization. In experiments with different neurotransmitter [glutamate, serotonin, γ -aminobutyric acid (GABA), and glycine] receptors, conditions could be obtained in which the rise of the current follows a single exponential rate law. The observed rate constant for the rise time, kobs , is related to the rate constants for channel opening (kop ) and closing (kcl ), the concentration of the ligand L that activates the transmembrane channel, and the dissociation constant of L, that is, K1 (Fig. 2): kobs = kcl + kop [L/(L + K1 )]2 . (4)

The relationship between kobs and the concentration of neurotransmitter is given in Fig. 4C. The slope of the line gives the value of the rate constant for channel opening (kop ) and the intercept on the ordinate gives the rate constant for channel closing (kcl ). This section outlined some approaches used to study the mechanism of proteins that transport inorganic ions across biological membranes. In the next section the properties of some individual proteins will be discussed.

Ion Transport Across Biological Membranes

103

FIGURE 4 A BC3 H1 cell (∼20-µm diameter) containing nicotinic acetylcholine receptors, attached to an electrode for whole-cell current recording, is equilibrated with caged carbamoylcholine. A cell-flow device was used to equilibrate the cell surface with ligands in a solution flowing over the cell from the left. (A) A Candela SLL500 dye laser is used. Rhodamine 640 or sulforhodamine 640 laser dye, together with a second harmonic generator, produces wavelengths of 328 and 318 nm, respectively. The laser beam is introduced from an optical fiber of 200-µm diameter. The fiber is adjusted to be ∼400 µm away from the cells so that the area illuminated around the cell has a diameter of 300 to 400 µm. The energy of the laser pulse emerging from the fiber is ∼ 500 µJ and the pulse length is 600 nsec. By projecting visible light through the optical fiber, the cell is illuminated and the fiber properly positioned. Inset: Photolysis at 328 nm (using the Candela dye laser) of caged carbamoylcholine liberates a 2-nitroso-α-ketocarboxylic acid and carbamoylcholine, a stable and well-characterized analog of acetylcholine. The wavelength of 328 nm was chosen to avoid cell damage at lower wavelengths and too low a product yield at higher wavelengths. Current is recorded in the whole-cell configuration. (B) Whole-cell current induced by 200 µM released carbamoylcholine at pH 7.4, 22–23◦ C, and −60 mV. The points represent the digitized current data; the parameters of the heavy solid line were calculated from a first-order plot of the rise time and an observed first-order rate constant of 2140 sec−1 . (C) Determination of kop , kcl , and K1 for the opening of nicotinic acetylcholine receptor channels in BC3 H1 cells at pH 7.4, 23◦ C and −60 mV. The values of kobs determined from experiments shown in B are plotted according to kobs = kcl + kop L2 /(L + K1 )2 . The values of kop , kcl , and K1 are 12,000 sec−1 , 500 sec−1 , and 210 µM, respectively. (Reproduced with permission from Hess, G. P., and Grewer, C. (1998). In “Methods in Enzymology,” (G. Marriott, Ed.), Academic Press, New York.

III. INORGANIC ION TRANSPORT AND THE PERCEPTION OF LIGHT
We shall now follow a cascade that consists of the perception of an environmental signal, the movement of ions across membranes and the resulting changes in transmembrane voltage, and the reaction of the organism to the signal (Fig. 1). Because we know most about the sensory cells of the eye, this system is used as an example, and particularly the role of rhodopsin. Opsin, which loops back and forth across the cell membrane, and rhodopsin are transmembrane proteins. Unwin and Henderson used electron microscopy to determine the ˚ structure of bacterial rhodopsin at 7 A resolution. More ˚ recently, the crystal structure of bovine rhodopsin at 2.8 A resolution was solved by Palczewski and colleagues.

Perception of light is initiated by a photochemical reaction in the retinal cells of the eye, the cis–trans isomerization of 11-cis-retinal (Fig. 5A). 11-Cis-retinal is attached by a Schiff-base linkage between its aldehyde group and the ∈-amino group of a lysine residue in the protein opsin, thus forming rhodopsin. 11-Cis-retinal absorbs light in the visible wavelength region, with an absorption maximum at 500 nm and a molar extinction coefficient of 40,000 cm−1 M−1 . The absorption of a single photon of light by 11-cis-retinal, and the subsequent conversion to all-trans-retinal, initiates a series of reactions. These include a conformational change of rhodopsin, and lead to a change in the rate at which sodium ions cross the membrane (see the next paragraph). This results in a change of the voltage across the cell membrane. The same conformational change of rhodopsin, initiated by the absorption of

104

Ion Transport Across Biological Membranes

Vm at about −40 mV. (7) Lowering the concentration of cGMP in the cell lowers the concentration of open Na+ conducting channels, which to remain open must have cGMP bound to them. (8) The closing of cGMP-activated transmembrane channels decreases the flux rate of Na+ into the cell and, therefore, changes the voltage across the membrane. (9) The changes in Vm result in the release by the retinal cell of chemical signals (neurotransmitters) adjacent to another cell. (10) The neurotransmitters bind to receptors on an adjacent cell that transiently form transmembrane channels, allowing cations or anions, depending on the receptor, to move through the cell membrane.

IV. INORGANIC ION TRANSPORT AND INTEGRATION OF ENVIRONMENTAL INFORMATION
In the previous section, the transport of ions across the membrane was initiated by an environmental signal, namely, light. This resulted in a change in the transmembrane voltage and a subsequent influx of calcium ions into the nerve terminal of a sensory cell (Fig. 1), resulting in the release of a chemical signal. In general, these chemical signals (neurotransmitters) are released from the nerve terminals of one cell and diffuse across a gap between cells called a synapse (Fig. 1). In the membrane of the adjacent cell, about 20–40 nm removed from the nerve terminal, the neurotransmitters bind to transmembrane neurotransmitter receptors. On binding the neurotransmitter, these transmembrane proteins, transiently (a few milliseconds) open a transmembrane channel (Fig. 1). These channels are specific for inorganic cations, sodium, or potassium, and some are also specific for calcium or chloride ions, depending on the receptor. If the resulting change in the transmembrane voltage, Vm , is of appropriate sign and magnitude (∼ + 20 mV) the voltage change will be propagated along the axon of the cell (Fig. 1), by a process in which sodium and potassium ions move across the membrane. This is described in detail at the end of this section. The major neurotransmitters are listed in Table I. Typical excitatory neurotransmitters are acetylcholine, glutamate, and serotonin. They bind to receptors that are named in the same way, for instance, the acetylcholine receptor, glutamate receptor, etc. These receptors on binding their specific neurotransmitter form transmembrane channels specific for the inorganic sodium and potassium cations. They are called excitatory receptors because they shift the transmembrane potential of the cell membrane to more positive values. A change in Vm of about +20 mV is required for the electrical signal to be transmitted to the nerve terminal where the release of a neurotransmitter is elicited. Typical inhibitory neurotransmitters are glycine

FIGURE 5 (A) Light-induced transformation of 11-cis-retinal to all-trans-retinal in visual pigments. 11-Cis-retinal is attached through a Schiff-base linkage of lysine 256 in rhodopsin. (B) Rhodopsin has a molecular weight of 40,000. Seven transmembrane helices are embedded in the cell membranes. 11-Cis-retinal lies near the center of the lipid membrane. The structure is based on the three-dimensional reconstruction of electron microscope images by Henderson and Unwin. (Reproduced with permission from Nelson, D. L., and Cox, M. M. p. 460. Figs. 13–22. “Lehninger Principles of Biochemistry” (2000). 3rd edition, Worth Publishers p. 460.)

light, leads to the release of all-trans-retinal. The all-transretinal is transformed to all-trans-retinol (Vitamin A) by pigment epithelia cells in the eye. It is a precursor in the synthesis of 11-cis-retinal, which is then transported back to cells containing opsin and reattached to the ∈-amino group of lysine 296 of opsin. The steps involved in the absorption of light by rhodopsin that lead to changes in the flux of sodium ions across the cell membrane, and to signal transmission, can be summarized as follows. (1) Light is absorbed. (2) Subsequently, the isomerization of retinal results in (3) a conformational change of rhodopsin; leading to (4) the activation of the enzyme cyclic-guanosine monophosphate (cGMP) phosphodiesterase. This results in (5) the hydrolysis of cyclic guanosine monophosphate (cGMP) to 5 guanosine monophosphate (5 -GMP). (6) cGMP activates a protein that forms Na+ -specific channels in the retinal cell membrane and maintains the transmembrane potential

Ion Transport Across Biological Membranes
TABLE I Major Neurotransmitters Excitatory Inhibitory

105 of the receptor spanning the membrane. The receptor consists of 5 subunits, each with a molecular weight of about 50 kiloDaltons. Together the subunits form a funnelshaped structure, projecting about 60 A above the plane of the membrane. The opening of the channel is about 50 A in diameter and narrows to a channel through the membrane about 30 A in length and 8 A wide. What is the relationship between the neurotransmitter concentration, the fraction of receptor-channels that open, the conductance of the channel, and the length of time the channels remain open (Fig. 2)? If we can answer these questions, and can also determine the concentration of receptor sites exposed to the neurotransmitter, we can calculate the change in transmembrane voltage (Eqs. 1–3) and predict whether or not a signal will be transmitted. The techniques for elucidating the mechanism of the reactions that allow one to determine the rate of transmembrane flux of inorganic ions and, therefore, the change in Vm have been discussed in Section I. It is important to mention that each cell is contacted by as many as 1000 different projections from other cells and each of these projections may release a different neurotransmitter from the nerve terminal. The mammalian brain contains an estimated 1012 cells; its information content is believed to exceed, by far, that of the largest supercomputers. It appears that the techniques are now at hand to (a) investigate the kinetic mechanism of the reactions by which neurotransmitter receptors determine the rate of flow of inorganic ions across the membrane, and (b) determine whether or not a signal will be transmitted to another cell. The chemical mechanism of receptor-mediated transport of inorganic ions across the membrane of neurons (nerve cells) by a few neurotransmitter receptors has been investigated. The mechanisms by which the many receptor isoforms facilitate the flow of inorganic ions across the cell membrane are still unknown. Also virtually unknown are the effects of diseases, causing mutations of receptors, and the influence of hundreds of drugs including anticonvulsants, antidepressants, anesthetics, and abused drugs, on the mechanisms.

O H3C O O glutamate H2N HO NH serotonin CH2 CO2 O NH3 CH2 O N(CH3)3 O

O O NH3

acetylcholine

GABA O NH3

glycine

and γ -aminobutyric acid (GABA). They bind to receptors that form transmembrane channels that are specific for chloride ions. They are called inhibitory receptors because they shift the transmembrane potential to more negative values and counteract the action of excitatory receptors. The neurotransmitter receptors are believed to belong to one large family of transmembrane, channel-forming proteins. A three-dimensional model of the nicotinic acetylcholine receptor from the electric organ of a fish, Torpedo species, has been proposed by Nigel Unwin based on electron diffraction images. Figure 6 gives a side view

FIGURE 6 Three-dimensional model of the nicotinic acetylcholine-gated ion channel. The structure is based on a threedimensional reconstruction of electron microscope images by Unwin and his colleagues. The receptor-channel complex consists of five subunits, all of which contribute to forming the pore. When two molecules of acetylcholine bind to portions of the α-subunits exposed to the membrane surface, the receptor-channel changes conformation. This opens a pore in the portion of the channel embedded in the lipid bilayer, and both K+ and Na+ flow through the open channel down their electrochemical gradients. (Reproduced with permission from Kandel, E. R., Schwartz, J. H., and Jessel, T. M. (2000). “Principles of Neuronal Science,” 4th edition. McGraw Hill, New York.)

V. INORGANIC ION TRANSPORT AND THE RAPID TRANSMISSION OF ELECTRICAL SIGNALS OVER LONG DISTANCES (UP TO ∼1 m)
The movement of inorganic ions across a cell membrane that initiates an electrical signal is only one step in signal transmission in organisms. The rapid transmission of this signal, over distances of up to 1 m, by the axon of cells is discussed next.

106 How is the signal that was initiated by neurotransmittermediated reactions propagated? In general, cell membranes are more permeable to potassium ions, which are at a higher concentration inside the cell membrane than outside and bound to immobile cations, mainly amino acids and proteins. The outflow of potassium ions along their concentration gradient is counteracted by the transmembrane voltage that is created by this outflow. The equilibrium potential for potassium ions, for instance, EK+ , is given by the Nernst equation (Eq. 1). The resting transmembrane potential of neurons is around −60 mV. When the flux of ions through neurotransmitter receptor channels results in a change (of ∼20 mV) in the transmembrane voltage to a more positive Vm value, the electrical signal is propagated along the axon of the cell within 1 msec. This occurs in the following way. At a critical value of Vm , specific voltage-dependent Na+ transmembrane channels in the axonal membrane open, allowing sodium ions to flow inside the axon (Fig. 1). As the inward flow of sodium ions changes Vm to even more positive values, the Na+ -specific channels are inactivated. Transmembrane K+ channels open, changing Vm again to more negative values. This in turn leads again to the opening of voltage-dependent Na+ -specific channels. These changes in transmembrane voltage, called action potentials, are propagated along the axon to the nerve terminal adjacent to another cell. Axons range from 0.1 mm to over 1 m in length and can convey electrical signals within 1 msec. When the signal arrives at the nerve terminal, Ca2+ -specific transmembrane channels open and the influx of calcium ions leads to the secretion of neurotransmitter. Typically, the neurotransmitter diffuses across the synaptic cleft, over a distance of 20–40 nm, and binds to the neurotransmitter receptors on the adjacent cell. Thus the signal is transmitted between the ∼1012 cells of the mammalian nervous system.

Ion Transport Across Biological Membranes

FIGURE 7 The selectivity filter of the potassium channel based on the X-ray crystallographic structure determination by MacKinnon and colleagues. The potassium channel is tetrameric with a hole in the middle that forms the ion pore. Each subunit forms two transmembrane helices, the inner and the outer helix. The pore helix and loop regions build up the ion pore in combination with the inner helix. The black spheres in the middle of the channel represent potassium ions. (Reproduced with permission from Branden, C., and Tooze, J. (1999). Fig. 12.11, p. 233. In “Introduction to Protein Structure,” 2nd edition, Garland Publishing, New York.)

VI. PROPERTIES OF THE PROTEIN (POTASSIUM CHANNEL) THAT ALLOWS K+ BUT NOT Na+ TO CROSS THE MEMBRANE
Recently, we have learned some of the properties of one channel that plays a central role in rapid signal transmission. The K+ channel from bacteria was crystallized, after a cytoplasmic tail of 33 residues was removed, and MacKinnon and colleagues have determined its structure at a resolution of 3.2 A (Fig. 7). This work represents the first high-resolution, X-ray diffraction study of an ionselective channel. Although the structure was obtained with the bacterial channel, K+ channels with similar se-

quences and properties are also present in other organisms. The bacterial K+ channel contains 158 amino acid residues. Four subunits are arranged around a central axis to form the channel. The K+ channel has two transmembrane helices. The structure gives an important indication as to how the ˚ channel allows the larger potassium ions (radius 1.35 A) 8 −1 to move through the channel at 10 ions sec , which approaches the diffusion-limited rate, but essentially prevents smaller sodium ions (radius 0.95 A) to pass. It has been estimated that K+ ions move through this channel at least 10,000 times faster than sodium ions. The ˚ ion pore is about 45 A long. Three cation-binding sites have been identified in it, two within a selectivity filter ˚ and separated by about 7.5 A and one in the cavity of the pore (Fig. 7). Both the concentration gradient and the electromotive force provide the driving force moving the ions through the channel. To pass through the channel, the inorganic ions have to pass through a selectivity filter. The theory is that ions that enter the selectivity filter must be dehydrated and that interaction of the dehydrated K+ with carbonyl oxygens from amino-acid residues inside the filter compensates for the dehydration. The distances between the carboxyl oxygens of the protein and two potassium ions in the filter are optimal to compensate for the cost of dehydration of potassium ions, but they would not be optimal for the dehydration of sodium ions.

Ion Transport Across Biological Membranes

107 transmembrane membrane ion transport that requires the hydrolysis of ATP. (2) Fast activated channels. For activation, these just require the binding of a ligand to the channel protein or a change in transmembrane voltage. The examples given are the neurotransmitter-activated channels and the voltage-activated K+ channel. The development of techniques using crystallography, NMR, electron diffraction, and molecular biology to produce specific proteins in large amounts to determine the structure of transmembrane channels formed by proteins is a very active field. It is expected that an increasing number of high-resolution transmembrane structures will be forthcoming in the next few years. These structures, together with kinetic measurements, are expected to give detailed information about the mechanism by which inorganic ions are transported across the cell membrane. The single-channel current-recording technique is ideally suited for studying channels that open because of the concentration gradient of inorganic ions and the resulting voltage changes. The change in transmembrane voltage determines whether or not a signal is propagated. Rapid chemical kinetic techniques with a 100-µsec time resolution, and suitable for investigations of ligand-gated ion channels on cell surfaces, are also now available. They are expected to provide additional information about ligand-gated ion channels and their mechanism of action. The ability to determine the effect of neurotransmitter concentration on the rate of transmembrane ion flux and, therefore, the change in transmembrane voltage is expected to provide important insight into how cells perceive, store, and transmit information. It is also expected to indicate how the receptor mechanism is changed by diseases of the nervous system and by the hundreds of drugs that affect the mechanism of these proteins. This information is expected to be essential in devising strategies for curing mental diseases and overcoming drug addiction.

VII. INORGANIC ION TRANSPORT AGAINST A CONCENTRATION GRADIENT AT EXPENSE OF ATP HYDROLYSIS
Most cells have a high concentration of K+ and a low concentration of Na+ inside the cell membrane relative to the concentration of these ions bathing the cell. As discussed, the rapid flow of these ions along their concentration gradient across the cell membrane is used in signal transduction between the cells of an organism, and plays an important role in life. To maintain the concentration gradient of sodium and potassium ions and thus the resting membrane potential of the cells, energy is needed. In 1979, Skou and Norby discovered an enzyme that reestablishes the original concentration gradient. It is called the Na+ –K+ ATPase where ATP stands for adenosine triphosphate, an energy source. Sodium ions are moved from inside of the cell to the outside and potassium ions are moved in the opposite direction. With each cycle of the enzyme, one molecule of ATP is hydrolyzed and two potassium ions are moved into the cell and three sodium ions moved out. Na+ –K+ ATPase is a membrane protein with two subunits spanning the membrane. The current hypothesis is that the enzyme reacts with ATP to give a phosphorylated enzyme and ADP (adenosine diphosphate). The conversion of ATP to ADP (adenosine diphosphate) and Pi (inorganic phosphate has been formulated as follows: ATP + Enzyme → ADP + P-Enzyme (phosphorylated enzyme) P-Enzyme + H2 O → Enzyme + Pi The resulting enzyme has a high affinity for K+ and a low affinity for Na+ . Hydrolysis of the phosphorylated enzyme results in the liberation of inorganic phosphate and the regeneration of the enzyme form with high affinity for Na+ and low affinity for K+ . The net result is the movement of two K+ ions into the cell and three Na+ ions out of the cell for each ATP hydrolyzed. About 25% of the energy consumption of a human at rest is used to maintain the resting concentration of sodium and potassium ions in cells.

SEE ALSO THE FOLLOWING ARTICLES
BIOENERGETICS • BIOINORGANIC CHEMISTRY • LIPOPROTEIN/CHOLESTEROL METABOLISM • MEMBRANE STRUCTURE • NUCLEIC ACID SYNTHESIS • PHOTOCHEMISTRY, MOLECULAR • PROTEIN STRUCTURE • PROTEIN SYNTHESIS

VIII. CONCLUSION AND OUTLOOK
It should be mentioned that examples of two types of ion channels have been given. (1) So-called slow channels involve second messengers. The examples given here are the light-activated channels that are opened by cGMP and the

BIBLIOGRAPHY
Doyle, D. A., Morais, C. J., Pfuetzner, R. A., Kuo, A., Gulbis, J. M., Cohen, S. L., Chait, B. T., and MacKinnon, R. (1988). “ The structure of the potassium channel: Molecular basis of K+ conduction and selectivity,” Science 280, 669–677.

108
Hammes, G. G. (2000). “Thermodynamics and Kinetics for the Biological Sciences,” Wiley Interscience, New York. Hess, G. P., and Grewer, C. (1998). Development and application of caged ligands for neurotransmitter receptors in transient kinetic and neuronal circuit mapping studies. In “Methods in Enzymology” (G. Marriott, ed.), Vol. 291, pp. 443–474, Academic Press, New York. Koester, J., and Siegelbaum, S. A. (2000). Ion channels. In “Principles of Neural Science” (E. R. Kandel, J. H. Schwartz, and T. M. Jessel, eds.) 4th edition, pp. 105–125, McGraw-Hill, New York. Nelson, D., and Cox, M. M. (2000). Biological membranes and transport. In “Lehninger Principles of Biochemistry,” 3rd edition, pp. 389–436, Worth Publishers, New York.

Ion Transport Across Biological Membranes
Palczewski, K., Kumasada, T., Hori, T., Behnke, C. A., Motoshima, J., Fox, B. A., Le Trong, I., Teller, D. C., Okada, T., Stenkamp, R. E., Yamamoto, M., and Miyano, M. (2000). “Crystal structure of rhodopsin: A G protein-coupled receptor,” Science 289, 739–745. Penner, R. (1995). A practical guide to patch clamping. In “Singlechannel Recording” (B. Sakmann, and E. Neher, eds.), 2nd edition, pp. 3–31, Plenum Press, New York. Subramaniam, S., and Henderson, R. (2000). “Crystallographic analysis of protein conformational changes in the bacteriorhopsin photocycle,” Biochim. Biophys. Acta 11460, 157–165. Unwin, N. (2000). “Nicotinic acetylcholine receptor and the structural basis of fast synaptic transmission,” Philos. Trans. R. Soc. Lond. B. Biol. Sci. 355, 1813–1829.

Lipoprotein/Cholesterol Metabolism
Alan D. Attie
University of Wisconsin I. II. III. IV. V. VI. VII. VIII. IX. Lipid Absorption Plasma Lipoprotein Structure Chylomicron Metabolism VLDL Metabolism (Endogenous Triglyceride Metabolism) IDL and LDL Metabolism Apolipoproteins Mediate Lipoprotein Metabolism The LDL Receptor Familial Hypercholesterolemia How Do Sterols Regulate Gene Expression? X. Other Lipoprotein Receptors XI. HDL and “Reverse Cholesterol Transport” XII. Tangier Disease and Familial Hypoalphalipoproteinemia XIII. Lp(a) and apo(a) XIV. CommonIsoforms of Apolipoprote in E(apoE) XV. Apo-E and Alzheimer’s Disease XVI. Familial Combined Hyperlipidemia and Hypertriglyceridemia XVII. Treatment of Lipoprotein Disorders XVIII. Final Perspective

GLOSSARY
ABCA1 A membrane lipid transport protein involved in cholesterol and/or phospholipid efflux from cells. Its action is necessary for the extracellular assembly of lipoprotein particles. Apo-A1 An apolipoprotein principally associated with HDL, an activator of lecithin cholesterol:acyltransferase. It interacts with cells to mediate delivery of cholesterol ester from HDL particles. Apo-B100 An apolipoprotein associated with VLDL and LDL particles, synthesized in the liver; it is a ligand for the LDL receptor. Apo-B48 An apolipoprotein associated with chylomicrons, synthesized in the intestine; it is a truncated

form of apo-B100 and does not bind to the LDL receptor. Apo-E An apolipoprotein principally associated with VLDL and chylomicrons; responsible for the receptormediated clearance of IDL and chylomicron remnants. It is a ligand for most members of the LDL receptor superfamily. The apo-E4 isoform is associated with increased risk of Azheimer’s disease. Apolipoproteins The protein components of plasma lipoproteins. Bile acids Detergent-like molecules formed from cholesterol. They are secreted by the liver and, together withcholesterol and phospholipids, form bile. Bile forms micelles which emulsify lipids in the intestinal lumen aiding in their absorption.

643

644 Cholesterol ester transfer protein (CETP) A bloodborne protein that catalyzes the exchange of triglycerides in VLDL for cholesterol esters in HDL. Chylomicron remnants Chylomicron particles that have been depleted of triglyceride after the lipoprotein lipase-mediated hydrolysis of their triglycerides. Chylomicrons Lipoprotein particles produced in the intestine to package and secrete dietary lipids. Chylomicrons are secreted into the mesenteric lymph. Familial hypercholesterolemia An elevation in LDL cholesterol due to mutations at the LDL receptor locus. Familial hypoalphalipoproteinemia A deficiency in HDL due to mutations at the ABCA1 locus. Fibric acid A class of drugs used to decrease the rate of VLDL triglyceride secretion. Gallstones Large crystals, usually made of cholesterol, that form in the biliary tract and/or gall bladder when cholesterol levels in bile are too high relative to phospholipid and bile acids. HDL High density lipoprotein; carries about 20% of plasma cholesterol. Its levels negatively correlate with risk of coronary heart disease. Is thought to mediate “reverse cholesterol transport.” HMG-CoA reductase A major rate-limiting step in the cholesterol biosynthetic pathway; catalyzes the conversion of HMG-CoA to mevalonate. IDL Intermediate density lipoprotein, also called “VLDL remnant”; the VLDL particle that has been depleted of triglyceride through the action of lipoprotein lipase. LDL Low density lipoprotein; carries approximately two-thirds of plasma cholesterol. Its levels are positively correlated with risk of coronary heart disease. It is formed in the circulation from the catabolism of VLDL. LDL receptor Receptor expressed in most tissues and mainly necessary for normal clearance of LDL from the bloodstream. Lipoprotein lipase Enzyme that catalyzes the hydrolysis of VLDL and LDL triglycerides to free fatty acids and glycerol; present on the luminal surface of the capillary endothelium. Lipoproteins Particles that transport lipids in the bloodstream. Lp(a) An LDL-like lipoprotein particle in which apoB100 is connected through a disulfide linkage to apo(a), a plasminogen-related protein. Microsomal triglyceride transfer protein (MTP) A protein in the endoplasmic reticulum necessary for secretion of VLDL and chylomicrons. Nicotinic acid A coenzyme precursor that at pharmacological doses helps to lower triglyceride levels and raise HDL levels.

Lipoprotein/Cholesterol Metabolism

Reverse cholesterol transport A hypothetical pathway by which cholesterol is transported from extrahepatictissues to the liver via HDL. Scavenger receptors Receptors that bind to a wide range of molecules, including modfied forms of LDL. They are involved in the accumulation of cholesterol in macrophages and smooth muscle cells in the arterial wall. SREBP cleavage activating protein (SCAP) A cholesterol-sensing protein that, in the absence of cholesterol, escorts SREBP from the endoplasmic reticulum to the Golgi, where it is proteolytically activated. Statins Cholesterol-lowering drugs that act by inhibiting cholesterol synthesis and upregulating the LDL receptor. Sterol responsive element binding protein (SREBP) A transcription factor that postively regulates sterolresponsive genes, is attached to the endoplasmic reticulum membrane, and is released through a cholesterolsensitive proteolysis step. Tangier disease A severe HDL deficiency syndrome caused by homozygous mutations at the ABCA1 locus. VLDL Very low density lipoprotein; a triglyceride-rich lipoprotein assembled in the endoplasmic reticulum and Golgi of hepatocytes and then secreted into the bloodstream. While in the circulation, it gives rise to IDL and then LDL after it loses triglyceride.

ANIMALS TRANSPORT LIPIDS in an aqueous environment at concentrations up to one million times their solubility in water. They accomplish this task by surrounding water-insoluble lipids with amphipathic lipids and proteins to form plasma lipoproteins. A major part of lipid transport is to supply energy for muscle contraction and to deliver lipids to adipose tissue for storage. Disorders in lipoprotein metabolism are a major risk factor for premature coronary heart disease throughout the world. These disorders arise from dysfunction in apolipoproteins, particular enzymes, and lipid transfer proteins, and also secondary to disorders in carbohydrate metabolism.

I. LIPID ABSORPTION
Animals absorb and transport large quantities of lipids. A major challenge is posed by the solubility characteristics of lipids. Because they are essentially insoluble in water, they must be packaged in order to move through an aqueous environment. Since many lipids are amphipathic (have water-soluble and water insoluble moieties), they have detergent-like properties that can be harmful to membranes.

Lipoprotein/Cholesterol Metabolism

645 Although the intestine has a large capacity for lipid absorption and esterification, it is not a lipid storage organ. It must therefore package and export absorbed lipids in order that they not accumulate. To accomplish this function, the intestine assembles the lipids into specialized particles called plasma lipoproteins.

II. PLASMA LIPOPROTEIN STRUCTURE
Plasma lipoproteins are uniquely endowed with the ability to transport large quantities of water-insoluble lipids through an aqueous environment. This because the nonpolar lipids (triglyceride and cholesterol ester; Fig. 2) are “buried” in the core of the lipoprotein, surrounded by a monolayer of amphipathic lipids, phospholipid, and unesterified cholesterol (Fig. 3). In addition to core and surface lipids, lipoproteins carry proteins termed apolipoproteins (Tables III and IV). These proteins stabilize the lipoprotein particles and carry out particular functions such as receptor recognition and activation of particular enzymes. Lipids are less dense than water. Consequently, because they are complexed with lipid, plasma lipoproteins tend to float when plasma is subjected to ultracentrifugation (Table I). In contrast, other blood proteins sediment in the centrifuge. Lipoproteins float at distinct buoyant densities and are named according to their flotation behavior. The lipoprotein classes are (Table II) chylomicrons, verylow density lipoprotein (VLDL), low density lipoprotein (LDL), and high density lipoprotein (HDL). Chylomicrons and VLDL are primarily triglyceride carriers, while LDL and HDL are primarily cholesterol (mostly cholesterol ester) carriers. Key Points about the Distribution of Apolipoproteins r Only HDL has no apo-B. r For practical purposes LDL can be considered to have r The other apolipoproteins overlap and, as described

FIGURE 1 The major lipid components of bile. Bile acids are effective detergents and, together with phophatidylcholine and cholesterol, form micelles in the intestinal lumen. These micelles solubilize lipids and aid in their absorption by the intestinal mucosal cells.

Dietary cholesterol enters the intestinal lumen and is solubilized in a bile acid micelle. Cholesterol is otherwise quite insoluble in water (solubility limit ≈ 1 µg/liter). The transport of lipids into the intestinal epithelial cells requires their solubilization in bile acid micelles. First, the solubilization facilitates the hydrolysis of fatty acid ester bonds by the intestinal lipases. Second, the transport into the enterocytes requires the formation of a properly structured bile acid micelle. Bile is comprised of three lipid components: (1) bile acids, (2), cholesterol, and (3) phosphatidylcholine (PC; lecithin; Fig. 1). An excess of cholesterol relative to the two biliary amphipathic lipids (PC and bile acids) can lead to the formation of cholesterol precipitates, more commonly known as gallstones. It is important to remember that ordinarily the major component of dietary fat is always triglyceride, not cholesterol. For example, milk and butter have very little cholesterol but are very high in triglyceride. Triglyceride (and other glycerolipids, such as phospholipids) are hydrolyzed in the intestinal lumen to yield monoglycerides and free fatty acids. These lipolysis products are then absorbed by the intestinal epithelial cells and resynthesized as triglycerides, phospholipids, and cholesterol esters. Thus, the intestine mediates both lipolysis (in the lumen) and re-esterification (within the epithelial cells) of dietary lipids. The ability of the intestine to re-esterify monoglycerides and cholesterol is essential for net lipid absorption. This maintains a gradient that drives net lipid absorption from the intestinal lumen. Indeed, pharmaceutical companies have sought to develop inhibitors of cholesterol absorption that function by inhibiting the intestinal enzyme responsible for cholesterol esterification, acyl-CoA:cholesterol acyltransferase (ACAT).

only apo-B.

below, readily transfer between the various lipoprotein particles while circulating in the bloodstream. It is useful to make a distinction between exogenous and endogenous lipid transport. Exogenous lipid transport refers to dietary fat. Endogenous lipids are those synthesized by the liver and adipose tissue from substrates that have already been absorbed and metabolized in these tissues. The exogenous pathway refers to the absorption of lipids in the intestine and the metabolism of chylomicrons. Lipids from chylomicrons can eventually mix with the endogenous lipid pools in the liver and adipose tissue.

646

Lipoprotein/Cholesterol Metabolism

FIGURE 2 Lipids of plasma lipoproteins. Virtually all of the triglyceride and cholesterol ester of a lipoprotein is in the interior. Some unesterified cholesterol can also exist in the interior. However, all of the phospholipid is at the surface of the particle and surrounds the hydrophobic core.

III. CHYLOMICRON METABOLISM
Since the intestine is primarily an absorptive organ, it must have the means of exporting newly absorbed lipids. The enterocyte re-esterifies fatty acids and monoglycerides to form triglycerides and phospholipids. Absorbed cholesterol is esterified to form cholesterol esters. Under ordinary circumstances, the vast majority of the core lipids in the chylomicron are triglycerides; however, after a cholesterol-rich meal, the intestine manufactures choles-

terol ester-rich chylomicron particles. Triglycerides and cholesterol esters are then packaged into the core of chylomicrons, which are secreted into the lymphatics. By secreting chylomicrons into the lymphatics, they gain entrance into the general circulation via the thoracic duct. This guarantees that extrahepatic tissues, principally adipose tissue and muscle, are the first to be exposed to the newly secreted chylomicrons—if chylomicrons were secreted directly into the bloodstream, they would first be delivered to the liver via the portal vein. Chylomicrons are very large (up to 1 µm in diameter). Thus a plasma sample containing chylomicrons is milky in appearance. A sample of fasting plasma is typically clear in appearance, even if there is an elevation in LDL particles—LDL particles are not large enough to scatter light. In terms of net transport, the bulk of lipid flux is triglyceride → adipose tissue and muscle. How is the triglyceride
TABLE I Buoyant Density and Size of Plasma Lipoproteins Class Chylomicrons VLDL IDL LDL HDL2 HDL3 Density (g/ml) 0.93 0.93–1.006 1.006–1.019 1.019–1.063 1.063–1.125 1.125–1.210 Diameter (nm) 75–1200 30–80 25–35 18–25 9–12 5–9

FIGURE 3 The domain structure of a plasma lipoprotein. The nonpolar lipids triglyceride and cholesterol ester are surrounded by the amphipathic lipids phospholipid and cholesterol. The latter are stabilized by apolipoproteins. These proteins have amphipathic α-helix and amphipathic β-sheet secondary structures.

Lipoprotein/Cholesterol Metabolism
TABLE II Chemical Composition of Plasma Lipoproteinsa Class Chylomicrons VLDL IDL LDL HDL2 HDL3 Chol 2 7 9 8 5 4 PL 7 18 19 22 33 35 Protein 2 8 19 22 40 55 TG 86 55 23 6 5 3 CE 3 12 29 42 17 13

647 r They are the package of dietary fat. Thus, few r They are usually >80% triglyceride. r They are secreted into the lymphatics, rather than the r They are acted upon by lipoprotein lipase on the r After depletion of their triglyceride via the lipoprotein

chylomicrons are made in fasted people.

bloodstream.

surface of adipose and muscle capillaries.

a Values represent percent of dry mass. Chol, cholesterol; PL, phospholipids; TG, triglycerides; CE, cholesterol esters.

lipase reaction, the resulting remnants are cleared from the circulation. r Chylomicrons are not converted into LDL. Lipoprotein lipase catalyzes the complete hydrolysis of triglycerides to free fatty acid and glycerol. This reaction occurs in the bloodstream, while the chylomicron particle is in proximity to the endothelial surface of the capillary wall. Cholesterol ester is not a substrate for lipoprotein lipase. Therefore, with the selective loss of triglycerides, the particle becomes more enriched in cholesterol ester. The free fatty acids are then taken up and oxidized in muscle or re-esterified and stored as triglyceride droplets in adipose tissue. The abundance of the adipose form of lipoprotein lipase is increased by insulin, thus promoting uptake of free fatty acids and storage as cytoplasmic triglyceride droplets (see Fig. 4). A chylomicron that has been depleted of its triglyceride core is termed a chylomicron remnant. Chylomicron

transferred from the chylomicron particle to these tissues? As was the case with the transfer of lipids from the intestinal lumen to the intestinal epithelial cells, this transfer begins with a lipolytic reaction to convert an oily lipid, triglyceride, into an amphipathic lipid, free fatty acid. In this case, the reaction is catalyzed by lipoprotein lipase, an enzyme that resides at the luminal surface of the capillary endothelium in adipose tissue and muscle. Main Features of Chylomicron Metabolism r Chylomicrons are exclusively synthesized in the

intestine.

FIGURE 4 Chylomicron pathway. The intestine secretes chylomicron particles into the lymphatics. They gain entrance into the general circulation through the thoracic duct. Lipoprotein lipase, on the luminal surface of adipose and muscle capillary endothelial cells, hydrolyzes the triglyceride core to free fatty acids and glycerol. The free fatty acids are re-esterified and stored as triglycerides in adipose tissue or undergo β-oxidation in muscle. The lipid-depleted chylomicrons, chylomicron remnants, are cleared by the liver through a pathway that depends on apolipoprotein-E as a ligand for cellular receptors.

648 remnants are cleared from the circulation, primarily by the liver.

Lipoprotein/Cholesterol Metabolism

IV. VLDL METABOLISM (ENDOGENOUS TRIGLYCERIDE METABOLISM)
VLDL assembly and secretion is similar to the corresponding pathway for chylomicrons. Triglycerides and cholesterol esters are packaged into the core of the lipoprotein particle. However, in contrast to intestinal chylomicron secretion, hepatocytes secrete VLDL directly into the bloodstream. In the bloodstream, VLDL is acted upon by lipoprotein lipase, delivering its triglyceride cargo to muscle and adipose tissue. The resulting VLDL remnant particle, also termed IDL (intermediate density lipoprotein), is further metabolized as discussed below. In both the liver and the intestine, triglyceride incorporation into lipoprotein particles requires the action of microsomal triglyceride transfer protein (MTP). MTP resides in the lumen of the endoplasmic reticulum (ER) and facilitates the transfer of triglycerides and cholesterol ester from the cytoplasmic side of the ER to the interior of the lipoprotein particle. Mutations in MTP cause abetalipoproteinemia, an inability to secrete chylomicrons and VLDL. In humans, the liver is a major lipogenic tissue. The liver is able to transform excess carbohydrate or protein into fat (remember, when we eat too much of anything we get fat!). Fatty acid substrates for hepatic triglyceride production are derived from three sources (Fig. 5): (1) a continuous supply of albumin-bound fatty acid to the liver, primarily from adipose tissue triglyceride stores (after a meal this source drops, due to the antilipolytic action of insulin), (2) dietary fat already transported in chylomicrons delivered to the liver, and (3) carbohydrate in excess of the liver’s capacity for storage as glycogen.

FIGURE 5 Sources of fatty acid for hepatic triglyceride synthesis. (1) Adipose tissue lipolysis, catalyzed by hormone-sensitive lipase, provides fatty acids that travel through the bloodstream to the liver. (2) Chylomicron remnants still carry some triglyceride and are cleared from the circulation by the liver. (3) Carbohydrate is converted to fatty acids when glycogen stores are maintained.

LDL is cholesterol rich. The production of a cholesterolrich lipoprotein from a triglyceride-rich lipoprotein occurs by selective removal of triglyceride from VLDL. In summary, dietary fat is packaged into chylomicrons in the intestine. Dietary cholesterol is also packaged into chylomicrons, but a substantial fraction is delivered to the liver via chylomicron remnants. This cholesterol can then compete for the core of VLDL and appear in the bloodstream as cholesterol ester-enriched VLDL particles. Excess substrate in any form is converted into TG for export by the liver. For example, in some people, diets high in simple carbohydrates (e.g., fructose and sucrose) can lead to hypertriglyceridemia.

V. IDL AND LDL METABOLISM
Unlike chylomicron remnants, IDL has two competing metabolic fates: (1) uptake by the liver and (2) further processing to become LDL (Fig. 6). Since IDL can be cleared by the liver or can be processed to become LDL, this branch point represents an important stage where LDL concentrations can be regulated. Inefficient clearance of IDL tends to lead to increased LDL production. LDL is formed in the bloodstream through the catabolism of VLDL at the surface of blood vessels. In addition, VLDL is a triglyceride-rich lipoprotein, while

VI. APOLIPOPROTEINS MEDIATE LIPOPROTEIN METABOLISM
Chylomicrons and VLDL carry many apolipoproteins, among them apo-B and apo-E. Apo-B is a large (MW = 516 kDa) protein stably associated with the lipoprotein particle. The intestine produces a truncated form of the apoB, termed apo-B48, which is missing the carboxy-terminal half of apoB. The mechanism for the truncation involves a unique RNA editing event in the intestine that changes a Gln codon (CAA) to a STOP codon (UAA). Since only VLDL is converted to LDL, this means all apo-B in LDL is apo-B100 (Fig. 7).

Lipoprotein/Cholesterol Metabolism

649

FIGURE 6 The VLDL → IDL → LDL pathway. VLDL is secreted by the liver directly into the bloodstream. Like chylomicrons, VLDL triglyceride is a substrate for lipoprotein lipase and is hydrolyzed while at the luminal surface of adipose tissue and muscle capillaries. The VLDL remnant (IDL), unlike the chylomicron remnant, can give rise to LDL. Some IDL is also directly cleared by the liver. LDL is cleared by the liver and by extrahepatic tissues.

r VLDL is synthesized in the liver and secreted directly r As occurs with chylomicrons, VLDL is acted upon by

into the bloodstream.

lipoprotein lipase, on the surface of adipose and muscle capillaries. Unlike chylomicrons, VLDL remnants (IDL) are both cleared from the circulation and converted to LDL. This branch point is a factor that determines the rate of LDL production.
FIGURE 7 Chylomicrons contain a different form of apoB than VLDL or LDL. In the intestine, an RNA editing event introduces a stop codon in apo-B, resulting in a truncated protein product, apo-B48. VLDL is secreted with full-length apoB, apo-B100, and thus gives rise to LDL particles with apo-B100. The receptorbinding domain of apoB is at the C-terminal half of the protein, thus apo-B48 cannot bind to the LDL receptor; chylomicrons depend upon apo-E for receptor binding. Note: the particles are not drawn to scale; chylomicrons are about five times larger than LDL particles.

Main Features of VLDL and IDL Metabolism r The liver is an important lipogenic tissue and exports

newly synthesized triglyceride in VLDL particles.

Whether or not a particle carries apo-B48 or apo-B100 has physiological importance. The carboxy-terminal half of apo-B is required for receptor recognition. Thus, apoB100 can bind to cellular receptors (described below); apo-B48 cannot. For this reason, chylomicron remnants cannot be cleared from the circulation via apo-B48. Instead, chylomicron remnant clearance is mediated by another receptor ligand, apo-E. Indeed, genetic deficiency of apo-E leads to massive accumulation of cholesterol esterrich chylomicron remnants and IDL in the bloodstream. If chylomicrons and VLDL contain apo-E, why are they not cleared before they are acted upon by lipoprotein lipase? One explanation is that the apo-E on chylomicrons and VLDL is masked by another protein, apoC-III. During

650 lipolysis, the apoC-III protein falls off the particle, thereby exposing apo-E and permitting it to mediate particle clearance. Some Key Points r In terms of mass movement, export of triglyceride r r r r r r

Lipoprotein/Cholesterol Metabolism

(TG) from the liver and intestine is where the action is. This means net movement to extrahepatic cells. Most of the plasma TG (in the fasting state, when no chylomicrons are present) is carried in VLDL. Thus, if fasting plasma TG is high, VLDL is high. Dietary fat is packaged into chylomicrons in the intestine. Excess endogenous substrate is converted into TG for export by the liver. Chylomicrons and VLDL are depleted of TG in the circulation, thus accomplishing their mission of delivering TG to peripheral cells. Chylomicron remnants are not converted to LDL; they are cleared by the liver. Only VLDL can be converted to LDL. In humans, most of the plasma cholesterol is carried in LDL. If plasma cholesterol is elevated without elevation of plasma TG, then LDL is high.

VII. THE LDL RECEPTOR
The discovery of the LDL receptor pathway by Michael S. Brown and Joseph L. Goldstein represents the most significant triumph in the field of atherosclerosis research. In an extraordinary collaboration begun in 1972, they discovered that cells possess a high-affinity receptor that binds to the apo-B100 moiety of LDL. (They were awarded the Nobel Prize in 1985.) Binding of LDL to its receptor results in rapid endocyhtosis and the formation of an endocytic vesicle (Fig. 8). The LDL and the receptor separate while in this vesicle and part ways; the receptor recycles and returns to the cell surface, while the LDL particle is delivered to the lysosome, where the protein and lipid moieties are degraded. Hydrolysis of LDL cholesterol esters in the lysosome results in the release of free cholesterol, which exits the lysosome and exerts three important regulatory functions: 1. It suppresses cellular cholesterol synthesis by reducing the levels of the rate-limiting enzymes in the cholesterol biosynthetic pathway, principally 3-hydroxy-3-methyl-glutaryl-coenzyme A reductase (HMG-CoA reductase). 2. It enhances the re-esterification of cholesterol for storage in a cytoplasmic lipid droplet.

FIGURE 8 The LDL receptor pathway. LDL is internalized via receptor-mediated endocytosis. The endosomes are a sorting compartment; the receptor recycles to the plasma membrane, while the LDL is delivered to the lysosomes, where the cholesterol esters are hydrolyzed by lysosomal lipases. The free cholesterol then exits the lysosome and is able to inhibit de novo cholesterol synthesis by reducing the abundance of several cholesterol biosynthetic enzymes (e.g., HMG-CoA reductase) and the LDL receptor. Cells protect themselves from cholesterol toxicity by re-esterifying cholesterol to form a cytoplasmic cholesterol ester droplet. [From Brown, M. S., and Goldstein, J. L. (1986). Science 232, 34–47.]

3. It inhibits production of new LDL receptor, thus diminishing the further supply of cholesterol to the cell. The LDL receptor pathway assures a constant steadystate level of cellular cholesterol. This is accomplished both by adjusting cellular cholesterol synthesis according to ambient LDL levels and by altering LDL receptor number to limit the amount of LDL getting into cells. Like free fatty acids, unesterified cholesterol can be toxic to cells. The formation of cholesterol esters protects cells from cholesterol toxicity. About two-thirds of LDL is catabolized by the liver. The rest is cleared by just about all other tissues. Steroidproducing tissues are especially active in LDL uptake. Adrenal cells (and presumably ovarian and testicular cells) do not synthesize cholesterol at rates sufficient to support high rates of steroidogenesis. They supplement their cholesterol supply by consuming cholesterol carried on LDL and HDL. The level of LDL receptor activity is affected by the steady-state level of cholesterol in a cell. Thus, any factors

Lipoprotein/Cholesterol Metabolism

651

that increase or decrease the cholesterol level of a cell will affect the rate of LDL clearance from the circulation. This means that nutritional factors (proportion and type of dietary fat), hormonal status, pharmacological factors (drugs that inhibit cholesterol synthesis), and agents that affect bile acid metabolism all affect plasma cholesterol by influencing the level of expression of the LDL receptor.

VIII. FAMILIAL HYPERCHOLESTEROLEMIA
Brown and Goldstein studied LDL metabolism in cells from patients with a common metabolic inherited disorder called familial hypercholesterolemia (FH). People homozygous for this mutation have a 6- to l0-fold elevation of LDL levels, are born with detectable atherosclerosis, and usually do not survive childhood without a myocardial infarction. Heterozygotes have two- to fourfold elevations in LDL and suffer from coronary heart disease (CHD) during middle age (85% of FH heterozygotes have a heart attack before the age of 60.) FH is the most common inherited metabolic disorder in humans, with a gene frequency of 1 in 500, i.e., 1 in 500 people is a heterozygote for FH. Brown and Goldstein discovered that FH cells from homozygote donors showed little or no LDL-binding activity. FH cells from heterozygotes possessed about 50% of normal activity. They concluded that mutations in the gene encoding the LDL receptor are the molecular basis for the FH disease. The inability to clear LDL through the normal LDL receptor pathway causes hypercholesterolemia atherosclerosis. The loss of LDL receptor activity readily explains the inefficient clearance of LDL and the hypercholesterolemia of FH patients (Fig. 9). In addition to defective catabolism of LDL, there is also LDL overproduction for the following reason. IDL is also cleared through the LDL receptor. Diminished LDL receptor activity leads to prolonged circulation of IDL, giving it a greater opportunity to be converted to LDL. IDL is at an important branch point in lipoprotein metabolism; it can be directly cleared from the circulation or it can be further processed to become LDL. The LDL receptor can also regulate the secretion of VLDL. It promotes reuptake of newly secreted VLDL. Also, within the secretory pathway, the LDL receptor promotes the degradation of newly synthesized apo-B100. The principal regulatory site in the cholesterol biosynthetic pathway is the step catalyzed by the enzyme 3-hydroxy-3-methylglutaryl-CoA reductase (HMG-CoA reductase; Fig. 10). Regulation of HMG-CoA reductase by LDL cholesterol occurs principally through the regulation of the level of its mRNA. In analogy to genes reg-

FIGURE 9 Kinetic mechanism for LDL overproduction in familial hypercholesterolemia. VLDL catabolism gives rise to IDL. IDL has two competing fates. It can be cleared by the liver or continue to be processed to become LDL. In the absence of the LDL receptor, IDL clearance is sluggish, thus a large proportion of IDL is converted to LDL.

ulated by steroid hormones, nuclear transcription factors bind to DNA sequences upstream from the reductase gene and regulate transcription. In addition to transcriptional regulation, cholesterol and certain other sterols diminish HMG-CoA reductase by a second mechanism; they hasten the degradation of the enzyme. The protein is located at the

FIGURE 10 Regulated steps in cholesterol synthesis pathway. Step 1 is catalyzed by cytosolic acetoacetyl-CoA synthase. Steps 2 and 3 are catalyzed by HMG-CoA synthase and HMGCoA reductase, respectively. The later two enzymes are transcriptionally regulated by SREBP. Cholesterol feeds back on its own synthesis by decreasing the abundance of enzymes 2 and 3. HMG-CoA reductase is the target of widely used cholesterollowering drugs known as “statins.” Between mevalonate and cholesterol are more than 30 steps and branch points to nonsteroidal isoprenoid molecules.

652 ER membrane and has several transmembrane domains that are necessary in order for cholesterol to stimulate the degradation of the protein.

Lipoprotein/Cholesterol Metabolism

IX. HOW DO STEROLS REGULATE GENE EXPRESSION?
Cholesterol regulates its own formation by inhibiting the transcription of several genes in the cholesterol pathway, most notably HMG-CoA synthase and HMG-CoA reductase. For many years it was also known that polyunsaturated fats decrease the level of cholesterol synthesis. Now we know how these regulatory events occur. The transcription of the cholesterol-regulated genes is regulated by a regulatory region that is upstream (before the transcriptional start site) of these genes. A special DNA sequence termed sterol responsive element (SRE) determines the responsiveness of these genes to regulation by cholesterol. How does cholesterol inhibit the transcription of genes with SREs? A transcription factor that binds to SREs is termed sterol regulatory element binding protein (SREBP). This protein turns on the transcription of genes with SREs in front of them, thus is a positive transcription factor. SREBP is found in the nuclear and endoplasmic reticulum membrane in an inactive form. To be activated, it must be cleaved from the membrane and released so that it can enter the nucleus and turn on transcription. The key to cholesterol regulation is that cholesterol (or more likely, a metabolite of cholesterol) inhibits this cleavage event. The “sensing” of cholesterol is carried out by another protein, sterol cleavage activated protein (SCAP), a protein that binds to SREBP. In sterol-depleted cells, SCAP escorts SREBP from the ER to the Golgi, where it is activated by proteolytic cleavage. This transport step is blocked by sterols. Interestingly, unsaturated fatty acids also inhibit SREBP activation, thus explaining how they inhibit cholesterol synthesis (Fig. 11). The LDL receptor is also regulated by an SRE. This explains why cholesterol downregulates the activity of the LDL receptor. Many genes in fatty acid and triglyceride synthesis are regulated by SREs. The list is growing; thus the importance of SREBP in physiology will be enlarged in the future. The expression of SREBP is enhanced by insulin. This helps to understand how insulin promotes lipogenesis through through global activation of expression of numerous lipogenic enzymes. In some individuals on high carbohydrate diets, plasma VLDL levels rise, a consequence of an abnormally high rate of de novo lipogenesis. The hyperinsulinemia that accompanies insulin resistance (see

FIGURE 11 Transcriptional regulation of sterol-responsive genes. A transcription factor termed sterol regulatory element binding protein (SREBP) binds to the SREs and enhances transcription. However, the SREBP is held captive bound to the endoplasmic reticulum membrane. Only when it is released by proteolytic cleavage does it travel to the nucleus, where it regulates sterol-responsive genes. The protein traverses the membrane twice and is cleaved by the successive actions of two proteases. The proteolysis step occurs in the Golgi. Transport of SREBP to the Golgi requires a second protein, SCAP. The transport step is inhibited by cholesterol through a sterol-sensing function of SCAP. Thus, cholesterol regulates gene expression by controlling the activation of a membrane-bound transcription factor, SREBP.

Section XVI) is also associated with increased levels of VLDL. This might be a consequence of insulin-mediated stimulation of SREBP expression. Patients lacking the LDL receptor do not accumulate chylomicron remnants in the bloodstream. Since chylomicron remnant clearance is mediated by apo-E, it has been postulated that a separate receptor is responsible for chylomicron remnant clearance, a receptor that, in contrast to the LDL receptor, binds to apo-E, but not to apo-B100. Several additional members of the LDL receptor family have been identified (Table V). The first of these, the LRP, participates in chylomicron remnant clearance and plays a major role in that process when the LDL receptor is absent or dysfunctional.

Lipoprotein/Cholesterol Metabolism
TABLE III Properties of the Apolipoproteins Apolipoprotein Apo-A1 Apo-A2 Apo-A4 Apo-B100 Apo-B48 Apo-C1 Apo-C2 Apo-C3 Apo-D Apo-E CETP Lp(a) Molecular weight (Da) 28,300 17,000 46,000 516,000 264,000 6,500 8,800 8,750 33,000 35,000 74,000 400–700 Concentration (mg/dl) 90–130 30–50 10–30 80–100 — 4–7 3–8 8–15 10 3–6 — 0.4–80 Tissue origin Intestine, liver Intestine, liver Intestine, liver Liver Intestine Liver Liver Liver Liver, intestine, pancreas, kidney, adrenals, brain Liver, macrophages, brain, Adrenal Spleen, liver, small intestine, adrenal Liver, testes, brain Function LCAT activator Unknown Unknown Cholesterol, triglyceride transport, receptor recognition Triglyceride, cholesterol transport LCAT activator LPL activator LPL inhibitor Unknown Receptor recognition Cholesterol ester transfer Unknown

653

X. OTHER LIPOPROTEIN RECEPTORS
A. LRP/α2 -Macroglobulin Receptor The lipoprotein receptor-related protein (LRP)/α2 macroglobulin receptor (LRP) has a mass about five times larger than the LDL receptor. It contains many of the same structural domains as the LDL receptor, including a cysteine-rich repeat domain that binds to lipoproteins. In vitro studies have shown that the LRP does not bind to apoB100, but binds to VLDL particles that have been enriched in apo-E. The binding of LRP to its ligands is dependent upon cell surface proteoglycans. These are proteins containing sulfated carbohydrate residues. In addition to binding lipoproteins, the LRP also binds to α2 -macroglobulin, a plasma protein that binds to proteases and, upon binding, is cleared by the liver. The LRP also binds to plasminogen activators and their inhibitors, such as tissue-type plasminogen activator, urokinasetype plasminogen activator, and their corresponding inhibitors.

B. Cubulin Cubulin is a 460-kDa protein localized to epithelial cells, such as those of the kidney. It plays a role in albumin reabsorption; its absence leads to loss of albumin in the urine. Cubulin also mediates the uptake of the major HDL protein apo-A1 by the kidney. C. Megalin (gp330) This receptor is a member of the LDL receptor superfamily. It is the antigen that elicits an autoimmune response leading to a disease called Heymann’s glomerulonephritis. It interacts with cubulin and is thought to function together with cubulin in the uptake of apo-A1 by the kidney. D. A VLDL Receptor (?) A cDNA encoding another member of the LDL receptor family is expressed in muscle and adipose tissue, but not in liver. It binds to apo-E-containing lipoproteins in vitro, but not to LDL. Based on its tissue localization and binding properties, this protein has been postulated to be a receptor for triglyceride-rich lipoproteins. E. CD-36 CD-36 is expressed in a variety of tissues, including adipocytes and macrophages. It functions in fatty acid transport and, like scavenger receptors, binds a variety

TABLE IV Apolipoproteins of the Plasma Lipoproteins Chylomicrons VLDL LDL HDL A-1, A-2, A-4, B-48, C-1, C-2, C-3, E B-100, C-1, C-2, C-3, E, trace amounts of A-1, A-2, and A-4 90–98% apo B-100 A-1, A-2, sometimes, E

654
TABLE V Lipoprotein Receptors Name LDL receptor (LDLR) LDL receptor-related protein/ α2 -macroglobulin receptor (LRP) Ligands LDL, apo-E, hepatitis C virus, Rous sarcoma virus subgroup A, human rhinovirus Apo-E, lipoprotein lipase, hepatic lipase, thrombospondin, Pseudomonas exotoxin A, α2 -macroglobulin, receptor-associated protein (RAP), lactoferrin, t-PA, u-PA, t-PA:PAI-1, u-PA:PAI-1, elastase-a1-antitrypsin apo-E, Reelin LDL, apo-J/clusterin, apolipoproteinH/β2-glycoprotein-I, PAI-1, prourokinase, lipopoprotein lipase, aprotinin, transcobalamin-vitamin B12, vitamin D-binding protein, retinol-binding protein, PTH, insulin, β2-microglobulin, epidermal growth factor, prolactin, lysozyme, cytochrome c, thyroglobulin, plasminogen, albumin, polybasic drugs, RAP, Ca2+ Apo-E, apo-B100, Reelin, RAP

Lipoprotein/Cholesterol Metabolism

Tissue locations Ubiquitously expressed Liver, brain, lung, adrenal, intestine, kidney, placenta, ovary, testis

VLDL receptor gp330/megalin (VLDLR)

Smooth muscle, brain, adipose, adipose, adrenal, testis, ovary, placenta, lung Kidney, lung, thyroid, parathyroid, epididymis, ileum, placenta, thymus

Apo-E receptor R2 (apoER2; LR7/8B)

Brain, testes, ovary, placenta

of ligands, including senescent neutrophils, collagen, and malaria-infected erythrocytes. F. An HDL Receptor, SR-B1 Unlike LDL, HDL can deliver cholesterol to certain tissues (liver and steroidogenic tissues like adrenal, ovary, and leydig cells) without the HDL particle being internalized. The SR-B1 protein mediates the docking of HDL with cells and the delivery of cholesterol ester from the core of the HDL particle to the cells. This receptor is in the scavenger receptor family and is able to bind other lipoproteins in addition to HDL. The discovery that LDL receptor deficiency in familial hypercholesterolemia is associated with premature atherosclerosis was initially paradoxical. A hallmark of atherosclerotic lesions is the formation of “foam cells”— macrophages and smooth muscle cells that accumulate cytoplasmic cholesterol ester droplets. If these cells lack the LDL receptor, how do they import excessive amounts of cholesterol? Chemical modification of LDL (e.g., acetylation) abolishes its ability to bind to the LDL receptor. However, the modified LDL binds to other cell surface receptors, termed scavenger receptors. Unlike the LDL receptor, scavenger receptors are not downregulated by cholesterol. Like the LDL receptor, uptake through scavenger receptors leads to accumulation of LDL-derived cholesterol as cytoplasmic cholesterol ester droplets. Are there any physiological counterparts to in vitro chemical modification of LDL? The polyunsaturated fatty acyl chains of LDL lipids can be oxidized, leading to cleavage and formation of chemically reactive short-chain fatty aldehydes. These aldehydes react with the protein moiety of LDL, apo B-100, converting it to a ligand for scavenger

receptors. The discovery of LDL oxidation and its implications for atherosclerosis has ignited a strong interest in the potential role of dietary antioxidants (e.g., vitamins C and E) in the prevention of atherosclersosis. Three scavenger receptors for oxidized LDL have been identified; SR-A1, SR-A2, and CD36 (Table VI). These receptors have also been implicated in the phagocytosis of damaged or apoptotic blood cells (e.g., lymphocytes and erythrocytes). CD36 has been implicated in fatty acid uptake into cells. The scavenger receptors can bind a wide array of molecules other than lipoproteins (e.g., polyanions, oligonucleotides, bacterial endotoxin, and crocidolite asbestos).

XI. HDL AND “REVERSE CHOLESTEROL TRANSPORT”
Epidemiological studies show a strong inverse relationship between HDL levels (HDL cholesterol or apo-A1) and risk of CHD. Even a small increase in HDL is significantly correlated with a reduction in the risk of premature heart disease. Unlike VLDL and chylomicrons, HDL is formed from its protein and lipid components in the bloodstream and interstitial fluids. The major apolipoproteins of HDL are apo-A1 and apo-A2. These proteins are secreted from hepatocytes and intestinal epithelial cells independently and also as minor components of VLDL and chylomicrons. Apo-A1 and apo-A2 bind to phospholipids. Phospholipids are available from the surface of VLDL after lipolysis. In addition, cells are able to efflux phospholipids through the action of ABCA1. ABCA1 is a membrane transport protein belonging to a large family of ATP-binding cassette proteins. These

Lipoprotein/Cholesterol Metabolism
TABLE VI Scavenger Receptor Family Name Class A SR-AI, SR-AII Ligands Tissue locations

655

MARCO (SR-AIII) Class B SR-BI

Acetylated LDL, oxidized LDL, polyanions, crocidolite asbestos, bacterial endotoxin, lipoteichoic acid Bacteria HDL, LDL, modified lipoproteins, anionic phospholipids, acetyl LDL

Macrophages, (Kupffer cells, histiocytes, microglial cells), some endothelial cells (low level) Macrophages Highest expression in steroidogenic cells (adrenal, ovary, testis) and hepatocytes, lower level expression seen in absorptive epithelial cells of the proximal small intestine, lactating mammary gland, low levels observed in all cultured cells Adipose, macrophages, epithelial cells, monocytes, certain endothelial cells, platelets Drosophila macrophages Drosophila macrophages

CD36

Croquemort Class C SR-CI Class D Microsialin (CD68) Class E LOX-1 (SR-E1) Class F SREC FcgR2-B2

HDL, LDL, modified lipoproteins, anionic phospholipids, fatty acids, collagen, malariainfected erythrocytes Apoptotic cells Multiple polyanions, including modified LDLs, poly(I) Modified lipoproteins Oxidized LDL Oxidized LDL Oxidized LDL

Macrophages, Kupffer cells, endothelial cells Endothelial cells Endothelial cells Macrophages

proteins include the cystic fibrosis transmembrane receptor, the sulfonylurea receptor (a pancreatic β-cell protein involved in insulin secretion), and the multidrug resistance transporter proteins. It is thought that ABCA1 mediates the efflux of phospholipid and/or cholesterol from cells, thus making them available for association with apolipoproteins to form HDL. Its crucial role in this process was established by the discovery that two types of severe inherited HDL deficiency syndromes are caused by mutations in ABCA1 (see Fig. 12).

XII. TANGIER DISEASE AND FAMILIAL HYPOALPHALIPOPROTEINEMIA
Tangier disease is a rare recessive disorder in which patients have almost no HDL. Cholesterol ester accumulates in macrophages and macrophage-rich tissues like spleen and liver. Familial hypoalphalipoproteinemia (FHA) is a very common dominant disorder in which people have low HDL (typically 41◦ C (liquid crystalline phase). The segmental order parameters are approximately constant for the first nine chain segments, but decrease towards the central part of the bilayer. The chain ordering can be explained on the basis of the rotational isomeric model for hydrocarbon chains. In the region of constant order parameters, trans-gauche isomerizations occur only in complementary pairs, leaving the hydrocarbon chains essentially parallel to each other. This leads to well-ordered bilayers with disordered hydrocarbon chains. The decrease of the order parameter in the central region is due to increasing contributions of gauche states. The total number of gauche isomeric states was estimated to be between 3 and 6 per chain. The quantitative evaluation of the deuterium data further yields the thickness of the hydrocarbon region of the bilayer. For DPPC bilayers in the liquid crystalline

˚ phase, an average thickness of about 30 A and a thermal expansion coefficient of −2.5 10−3 /K was derived, in good agreement with X-ray diffraction experiments. It should be noted that this approach is also valid for determining the bilayer thickness of highly unsaturated membranes. The DPPC order parameter profile (Fig. 5) serves as a “gold standard” for molecular dynamics simulations of bilayers. Similar order profiles have now been established for a large variety of pure lipid membranes as well as intact biological membranes. B. Incorporation of cis-Double Bonds As mentioned before, almost all biological lipids contain unsaturated fatty acyl chains at the sn-2 position. The incorporation of the cis-double bond introduces a kink in the otherwise straight chain and reduces the gelto-liquid crystal phase transition temperature by some

362

Membrane Structure

FIGURE 4 Variation of the quadrupole splittings of the α- and the β-segment in the headgroup of phosphatidylcholine bilayers upon binding of cationic (open symbols) and anionic (solid symbols) substances. The difference between the quadrupole splittings of the α-segment with, ( ν(Xb )), and without ( νi◦ ) guest molecule is plotted versus that of the βsegment under identical conditions, ( ν(Xb ) − νi◦ )/ Xb = mi , where Xb , is the mole fraction of bound guest molecule and i stands for α or β. Metal ions ( ) [Altenbach and Seelig (1984). Biochemistry 23, 3913; Macdonald and Seelig (1987). Biochemistry 26, 1231, 6292]; drugs ( ) [Boulanger et al. (1981). Biochemistry 20, 6824; Seelig et al. (1988). Biochim. Biophys. Acta 939, 267; Bauerle and Seelig (1991). Biochemistry 30, 7203]; amphiphilies (∇) [Scherer and Seelig (1989). Biochemistry 28, 7720]; peptides ( ? ) [Beschiaschvili and Seelig (1991). Biochim. Biophys. Acta 1061, 78; Kuchinka and Seelig (1989). Biochemistry 28, 4216; Roux et al. (1989). Biochemistry 28, 2313; Spuhler et al. (1994). J. Biol. Chem. 269, 23904; Wieprecht et al. (2000). Biochemistry 39, 442; Schote et al. (2000). Pharm. Res.]; electrically neutral detergent ( ) [Wenk and Seelig (1997). Biophys. J. 73, 2565]; inorganic anion ( ) [Macdonald and Seelig (1988). Biochemistry 27, 6769]; peptides ( ) [Schote et al. (2000). Pharm. Res.]; amphiphiles ( ) [Scherer and Seelig (1989). Biochemistry 28, 7720]; phospholipids ( ✉) [Marassi and Macdonald (1992). Biochemistry 31, 10031; Scherer and Seelig (1989). Biochemistry 28, 7720; Pinheiro et al. (1994). Biochemistry 33, 4896]. The slope is characteristic for the sign of the electric charge and is m = −0.52 ± 0.01 for cations and m = −1.01 ± 0.05 for anions [cf. Scherer and Seelig (1989). Biochemistry 28, 7720].

*

40◦ C compared to the saturated lipid. The influence of the cis-double bond was investigated in bilayers composed of 1-palmitoyl-2-oleoyl-3-sn-phosphocholine (POPC) by either labeling the saturated palmitoyl chain or synthe-

sizing deuterated cis-unsaturated oleic acid. As seen in Fig. 5 and Fig. 6, the shape of the order profile of the palmitic acyl chain is similar to that observed for the fully saturated DPPC (Fig. 5), but the magnitude of the order

FIGURE 5 Order parameters |SCD | as a function of the labeled carbon atom for 1,2-dipalmitoyl-3-sn-phosphocholine ( ✉) and for 1-palmitoyl-2-oleoyl-3-sn-phosphocholine ( ) at 42◦ C. The sn-1 chains are labeled. [From Seelig and Seelig (1977). Biochemistry 16, 45.]

FIGURE 6 The effect of a cis- and a trans-double bond on the order parameter profile. Bilayers of 1-palmitoyl-2-oleoyl-sn-glycero3-phosphocholine labeled at different positions in the sn-1 ( ) and sn-2 chain ( ) measured at 27◦ C; 1-palmitoyl-2-elaidoyl-snglycero-3- phosphocholine labeled in the sn-2 chain (♦) measured at 40◦ C. [From Seelig and Waespe-Sarcevic (1978). Biochemisty 17, 3310.]

Membrane Structure

363 in the sn-2 chain than in lipids with only one double bond. C. Effect of Cholesterol on the Order and Motion of the Lipid Hydrocarbon Chains The influence of cholesterol on the order and mobility of lipid bilayers was investigated with both selectively deuterated lipids and deuterated cholesterol. Addition of 50% cholesterol to DPPC and DMPC bilayers was shown to lead to an almost twofold increase of the quadrupole splitting of the labeled fatty acyl chain segment compared to that of a cholesterol-free bilayer. When [3-2 H] cholesterol was added to a nondeuterated DPPC bilayer, again a large quadrupole splitting of the cholesterol probe was observed. Both probes lead to the conclusion that a high concentration of cholesterol induces an essentially all transconformation in those hydrocarbon chain segments which are in contact with the rigid steroid frame. This condensing effect of cholesterol was also observed in monolayerand neutron-diffraction experiments. The effect of cholesterol on the order parameter profile of individual fatty acyl chains in DPPC bilayers was simulated by means of molecular dynamics calculations as displayed in Fig. 7. A distinct plateau with an order parameter of the C D bond vector of SCD = −0.4 was detected. This means that the hydrocarbon chains are almost fully extended and that the order parameter of the long molecular axis, Smol = −2 SCD , approaches its maximum value of Smol = 1. In highly unsaturated lipid mixtures, typical for nerve and retinal membranes, cholesterol induces an increase in the order of both saturated and polyunsaturated hydrocarbon chains. However, the increase in order is about a factor of 2 smaller for polyunsaturated than for monounsaturated lipids.

parameters is distinctly smaller in the unsaturated system. This demonstrates that the presence of a cis-double bond causes a more disordered conformation of the hydrocarbon chains. Considering the relative flexibility within the palmitic acyl chain, the deuterium resonance data indicate a local stiffening of those segments which are located in the vicinity of the rigid cis-double bond. An increase in temperature leads to a further decrease of the order parameters. The 2 H-NMR spectrum of POPC membranes deuterated at the C-9, and C-10 positions of the oleic acyl chain shows two quadrupolar splittings, the larger corresponding to the C-9 deuteron and the smaller to the C-10 deuteron (Fig. 6). The presence of two quadroupole splittings at the same rigid segments is caused by a tilting of the cis-double bond with respect to the bilayer normal which produces different orientations for the C–2 H vectors of the 9- and 10-carbon atoms. The angle between the bilayer normal and the C C bond vector was found to be 7 to 8◦ . The order parameter of 1-palmitoyl-2-elaidoyl-snglycero-3-choline deuterated at the C-9 and C-10 transdouble bond of the elaidic acid chain is also included in Fig. 6. Due to the symmetry of the trans-double bond the two C D vectors at the C-9 and C-10 position make the same angle with the C C vector axis. They give rise to the same quadrupole splitting, and the evaluation of the order parameter of the C C axis is straightforward. Taking into account the different geometries, the molecular ordering and the angular fluctuations of the cis- and transdouble bonds are identical. In addition, there are no quantitative differences between sn-1 and sn-2 chain segments at this position in the bilayer. The segmental fluctuations around the bilayer normal thus only depend on the distance from the lipid–water interface but not on the specific segment geometry. The effect of cis-unsaturation was also investigated for the glycosphingolipid N -(oleoyl-d33)galactosylceramide incorporated at low concentration into liquid crystalline liposomes composed of 1,2-dimyristocyl-3-snphosphatidylcholine (DMPC) and POPC using the perdeuterated oleoyl chain as the reporter element. The primary effect of cis-9,-10 unsaturation in glycosphingolipids proved to be similar to that of cis-unsaturation in glycerolipids. It was further shown that the overall dynamics of N -(oleoyl)galactosylceramide in fluid phospholipid membranes was very similar to that of glycerolipids with comparable acyl chains. Increasing sn-2 unsaturation from one to six double bonds in PC leads to an inhomogeneous disordering along the neighboring perdeuterated sn-1 chain. As a consequence, the effect of a temperature increase leading to a decrease in the average chain length is somewhat less pronounced in lipids with three or more double bonds

FIGURE 7 Effect of cholesterol. Order parameter of the sn-2 chain in DPPC bilayers without ( ) and with ( ) 50 mol% cholesterol as function of carbon atom. [From Smondyrev and Berkowitz (1999). Biophys. J. 77, 2075.]

364 As far as lipid headgroups are compared, addition of cholesterol increases the chain order in the sequence 18:0– 18:1 PS < 18:0–18:1 PC < 18:0–18:1 PE for the monounsaturated lipid mixture and in the sequence 18:0–22:6 PS < 18:0–22:6 PE 18:0–22:6 PC for polyunsaturated mixtures. The cholesterol-induced variation of order parameters as a function of the chemical nature of the lipid species suggests a cholesterol-induced formation of lipid microdomains with a headgroup- and fatty-acyl-chaindependent lipid composition. In particular, under physiological conditions, the formation of PC-enriched microdomains has been proposed in which the saturated sn-1 chain is preferentially oriented toward the cholesterol molecule. The lifetime of a lipid molecule in a given cluster, however, is less than 10−4 s, and the cluster radius is probably smaller than 25 nm. In a natural membrane the effect of cholesterol is very similar as in model membranes. This was shown, for example, for Acholeplasma laidlawii membranes by the incorporation of perdeuterated and selectively deuterated fatty acids.

Membrane Structure

V. PHASE BEHAVIOR OF LIPIDS AND MEMBRANE DOMAIN FORMATION
A large number of phase diagrams for binary mixtures combining cholesterol with different saturated and unsaturated phosphatidylcholines have been established. Cholesterol at different bilayer concentrations can promote or suppress lateral segregation of phospholipids of differing acyl chain length. Addition of 50 mol% cholesterol to selectively deuterated DPPC bilayers leads to an elimination of the gel-toliquid crystal phase transition at 41◦ C. In contrast, cholesterol is also found to enhance the tendency of the PC components to exhibit lateral segregation. These seemingly contradictory effects of cholesterol can be readily explained in light of the cholesterol–phospholipid phase diagrams. The effect of cholesterol on the thermotropic phase behavior of aqueous dispersions of different lipids has been extensively investigated by means of differential scanning calorimetry. The results show an inverse correlation between the strength of intermolecular phospholipid–phospholipid interactions, as manifested by the gel-to-liquid crystalline phase transition temperatures of the pure phospholipids, and the miscibility of cholesterol with the respective bilayer (particularly gel-state bilayers). The miscibility of cholesterol with lipids carrying identical fatty acyl chains decreases in the order: PC ∼ PG ∼ SM > PS > PE > diglucosyl- and monoglucosyl-diacylglycerol > GalCer. However, if the

higher melting components are dispersed as minor components of total lipid in a host matrix consisting of, for example, 1-stearoyl-2-oleoyl-phosphatidylcholine and cholesterol, neither short-chain nor long-chain cerebrosides or sphingomyelins show phase separation in the physiological temperature range despite their high phase transition temperatures. Mixtures of cholesterol and sphingolipids have recently attracted attention since spingolipid–cholesterol domain formation has been observed in mammalian cell membranes upon cooling to 4◦ C and extraction with Triton X-100. This phenomenon has also been termed “lipid raft” formation. Lipid rafts exhibit a high lateral packing density and are suggested to entail a sorting of GPI-anchored proteins. The bulky intrinsic proteins remain in the fluid phase. At room temperature, lipid rafts are no longer detectable. Nevertheless, they are assumed to prevail as microdomains at growth temperature and to be relevant for membrane trafficking and protein sorting in mammalian cells. Although domain formation is now a common theme among biologists, an unambiguous physical–chemical characterization of domains under physiological conditions is still missing. On physical grounds, domain formation is most likely to occur in mixtures of lipids with widely different gel-to-liquid crystal phase transition temperatures. Phase separation will occur if the measuring temperature is below the phase transition temperature of one of the components of the mixture and if this component constitutes a major lipid fraction. Lipids exhibiting high phase transition temperatures generally have long saturated acyl chains and small headgroups, or headgroups that may interact via hydrogen bonding. Typical examples are sphingolipids, glycosphingolipids, or long-chain phosphatidylethanolamines. As a further mechanism, electrostatic interactions of anionic lipids with cationic compounds may also induce domain formation. Due to the biochemical complexity of biological membranes, the molecular mechanisms responsible for phase separation are not easily distinguished experimentally. The difficulty in understanding the diverging results arises, on the one hand, from the use of techniques differing in spatial (nanometers to micrometers) and temporal (nanoseconds to tens of seconds) resolution and, on the other hand, from the application of different experimental conditions. For technical reasons, domain formation was generally investigated at unsphyiological temperatures using lipids with bulky reporter groups. Both factors may affect the phase behavior of lipids. Further experiments are therefore required to test whether oganizational processes are induced by lipid domain formation under physiological conditions.

Membrane Structure

365 cross-sectional area of the fatty acyl chains. An area expansion upon membrane penetration of amphiphilic compounds was also shown with molecular dynamics simulation for local anesthetics and peptides. The observation of an area increase upon insertion of local anesthetics is consistent with the phenomenon of pressure reversal of local anesthesia, which may be due to the anisotropic compression of lipid membranes under hydrostatic pressure and the consequent release of anesthetic molecules. B. Order and Fluidity in the Presence of Transmembrane Proteins Hydrophobic transmembrane peptides aggregate in aqueous solution and therefore do not enter a lipid membrane spontaneously. In model membranes, peptide insertion is achieved by cosolubilization of peptide and lipid in an organic solvent (detergent solution) and subsequent evaporation of the solvent (equilibrium dialysis against detergent-free buffer). Reconstitution studies show that transmembrane peptides and proteins barely perturb the lipid bilayer order, suggesting a fluid-like match between the lipid acyl chains and the outer protein surface. The investigation of hydrophobic transmembrane peptides of different lengths has led to the conclusion that the average thickness of the lipid bilayer is significantly perturbed only in cases of a large mismatch between peptide length and membrane thickness. When the hydrophobic part of the peptide was larger (smaller) than that of the pure bilayer, the membrane thickness was increased (decreased). Larger intrinsic membrane proteins may span the membrane with several helices and perform functional tasks that can be quantified by biochemical assays. Two different approaches have been employed to study the lipid–protein interaction. One is to purify and delipidate transmembrane proteins and to reconstitute them with selectively deuterated lipids; the other is to incorporate deuterated fatty acids or other deuterated substrates into biological membranes by means of the biosynthetic pathway. In the latter case, the intact biological membrane is compared with aqueous bilayer dispersions formed from the extracted lipids. In the following we will discuss examples for the two types of assays. Cytochrome C oxidase catalyzes the transfer of electrons from cytochrome C to molecular oxygen and is one of the best investigated intrinsic membrane proteins. The beef-heart enzyme can be purified in an almost lipid-free form and can be functionally reconstituted by incorporation into different lipid systems since the natural lipid composition is usually not required for reconstitution of an active enzyme (see Fig. 9). The interaction of cytochrome C oxidase with lipid membranes has been investigated by means of spin-label

VI. INTERACTION OF MEMBRANE LIPIDS WITH AMPHIPHILIC MOLECULES AND TRANSMEMBRANE PROTEINS
A. Lipid Order Parameter in the Presence of Amphiphilic Molecules The outer lipid membrane surface of eukaryotic cells is generally uncharged. Amphiphilic, water-soluble molecules such as local anesthetics, viral or antibiotic peptides, or peptide toxins therefore partition into the bilayer interface because of their hydrophobicity. All these compounds are found to decrease the order of lipid membranes. This is illustrated in Fig. 8 which shows the effect of incorporating the cationic peptide fragment 828–848 from the carboxy-terminus of the envelope glycoprotein gp41 of HIV-1 (P828) into bilayers composed of 1-stearoyl(d35)2-oleoyl-sn-glycero-3-phosphoserine. A modest reduction of the lipid chain order near the glycerol backbone and a significant reduction towards the bilayer center are observed, indicating a decrease in the lateral packing density of the membrane and a corresponding increase of the

FIGURE 8 Influence of peptide P828S on the hydrocarbon chain order of 1-stearoyld35 -2-oleoyl-sn-glycero-3-phosphoserine at 32◦ C. The smoothed order parameter profile derived from dePaked nuclear magnetic resonance powder patterns has lost the information characteristic for the beginning of the fatty acyl chains seen in Fig. 5. (A) 2 H NMR order parameter profiles of SOPS-d35 in the absence of P828s ( ) and at lipid/peptide molar ratios of 20:1 ( ✉) and 10:1 ( ), respectively. (B) The peptideinduced difference in order parameters along the chain at molar lipid/peptide ratios of 20:1 ( ❤ and 10:1 ( ). Peptide-induced or) der changes are largest in the bilayer center, suggesting that the peptide acts as a spacer that is located in the membrane’s interface region. [From Smondyrev and Berkowitz (2000). Biophys. J. 78, 1672.]

366

Membrane Structure

FIGURE 9 Variation of the phosphorus T1 relaxation time with temperature. Pure POPC dispersed in 50 mM Tris buffer, pH 7.4, ). containing 1 mM EDTA ( ❤ Cytochrome C oxidase functionally reconstituted with [α-2 H2 ]POPC in 50 mM Tris, pH 7.4, and 10 mM EDTA. The higher temperatures (>30◦ C) were measured last (•). Same sample as in (A) but resuspended and washed in additional 10 mM EDTA after the measurement at 45◦ C ( ) [From Tamm and Seelig (1983). Biochemistry 22, 1474.]

electron paramagnetic resonance (epr) and by 2 H-, 14 N-, and 31 P-NMR experiments. The spin label method showed two motionally distinct lipid populations, with the slower component being attributed to the lipids interacting directly with the protein (“boundary lipids”). In contrast, NMR measurements of cytochrome C oxidase functionally reconstituted with headgroup and chain deuterated lipids revealed only one homogeneous population of lipids. The anisotropy of the segmental movements characterized by means of the residual 2 H and 14 N quadrupole splittings and the 31 P chemical shielding anisotropy as well as the segmental fluctuations, determined by measuring the 2 H- and 31 P spin-lattice (T1 ) relaxation times (Fig. 10), closely resemble those of pure lipid bilayers. Taken together, the anisotropy parameters as well as the T1 relaxation times provide no evidence for any strong polar or hydrophobic interaction between the lipid and the protein, neither in terms of a conformational change of the headgroup nor in terms of a significant immobilization of individual segments. The only noticeable difference between the NMR spectra of reconstituted membranes and pure lipid bilayers was a line broaden-

ing in the presence of protein, which probably arises from slower motions. Similar results were obtained in reconstitution experiments with lipophilin and proteolipid apoprotein-lecithin systems, sarcoplasmic reticulum Ca2+ , Mg2+ -ATPase, rhodopsin, and glycophorin. In all these cases deuterium NMR revealed only one lipid population while the epr spectra (as far as available) showed two components. The results further show that proteins either disorder or have little effect on hydrocarbon chain order in membranes above the gel-to-liquid crystal phase transition temperature, Tc , of the pure lipids. The question as to how an intrinsic protein affects the lipid environment was also investigated in systems containing a relatively low amount of lipid such as in partially delipidated cytochrome C oxidase surrounded by only 130 lipid molecules or in the crystalline lipovitellin/phosvitin complex containing about 100 phospholipid molecules in an interior cavity. In both systems the lipids remain in a fluid phase. Only when the lipid pool of cytochrome C oxidase was reduced to 6 to 18 molecules was a distinct broadening of the 2 H-NMR linewidth observed, indicating a lipid motion which was no longer axially symmetric. But even under these conditions, the total width of the spectrum was still considerably narrower than that observed for immobilized phospholipids in solid crystals. A second, much-debated question is whether or not cardiolipins form a long-lived complex with cytochrome C oxidase. To answer this question, the remaining lipids in partially (130 lipids per protein) and highly delipidated (“lipid-depleted”; 6 to 18 lipids per protein) cytochrome C oxidase were analyzed. In the partially delipidated preparation, approximately 11 cardiolipins, 54 phosphatidylethanolamines, and 64 phosphatidylcholines were found; in the “lipid-depleted” state, the corresponding numbers are 1 or 2 cardiolipins, 3 to 8 phosphatidylethanolamines, and 2 to 8 phosphatidylcholines. This result supports a fast exchange (>104 s−1 ) and is in contrast to earlier contentions that cardiolipin is the only remaining lipid in “lipid-depleted” cytochrome C oxidase. However, recent X-ray results show that the residual lipids in cytochrome C oxidase crystals are also heterogeneous and may not even contain cardiolipin. The random distribution of the remaining lipids is in accordance with a fast exchange between lipids on and off the protein surface and suggests that cardiolipin (which may have a potential role in electron transfer reactions) is at best interacting transiently rather than permanently with cytochrome C oxidase. The results obtained in reconstitution studies were confirmed with natural membranes. The natural systems investigated are, for example, Acholesplasma laidlawii

Membrane Structure

367

(grown on a medium supplemented with specifically deuterated or perdeuterated fatty acids), cardiolipin- or glycerol-auxotroph Escherichia coli (grown in tissueculture medium containing selectively deuterated fatty acids or phosphatidyl glycerol), and mouse fibroblast L-M cells (grown in tissue-culture medium containing selectively deuterated choline or ethanolamine). The membranes of these systems showed very similar fatty acid and headgroup motion, ordering, and orientation as the membranes formed from the extracted lipids without protein. No long-lived lipid–protein complexes were observed for neutral or negatively charged lipids.

SEE ALSO THE FOLLOWING ARTICLES
BIOENERGETICS • ELECTRON TRANSFER REACTIONS • ENERGY TRANSFER, INTRAMOLECULAR • ION KINETICS AND ENERGETICS • ION TRANSPORT ACROSS BIOLOGICAL MEMBRANES • LIPOPROTEIN/CHOLESTEROL METABOLISM • PROTEIN SYNTHESIS

BIBLIOGRAPHY
Davis, J. H. (1983). “The description of membrane lipid conformation, order and dynamics by 2H-NMR,” Biochim. Biophys. Acta 737, 117– 71. Devaux, P. F., and Seigneuret, M. (1985). “Specificity of lipid-protein interactions as determined by spectroscopic techniques,” Biochim. Biophys. Acta 822, 63–125. Divecha, N., Clarke, J. H., Roefs, M., Halstead, J. R., and D’Santos, C. (2000). “Nuclear inositides: inconsistent consistencies,” Cell. Molec. Life Sci. 57, 379–393. Koynova, R., and Caffrey, M. (1998). “Phases and phase transitions of the phosphatidylcholines,” Biochim. Biophys. Acta 1376, 91–145. Mitchell, D. C., Gawrisch, K., Litman, B. J., and Salem, N., Jr. (1998). “Why is docosahexaenoic acid essential for nervous system function?” Biochem. Soc. Trans. 26, 365–370. Muniz, M., and Riezman, H. (2000). “Intracellular transport of GPIanchored proteins,” Embo J. 19, 10–15. Op den Kamp, J. A. F., Roelofsen, B., and van Deenen, L. L. M. (1985). “Structural dynamic aspects of phsophatidylcholine in the human erythrocyte membrane,” Trends Biochem. Sci. 10, 320–323. Seelig, J. (1977). “Deuterium magnetic resonance: theory and application to lipid membranes,” Q. Rev. Biophys. 10, 353–418. Seelig, J. (1978). “[2 H]Hydrogen and [31 P]phosphorus nuclearmagnetic-resonance and neutron-diffraction studies of membranes,” Biochem. Soc. Trans. 6, 40–42. Seelig, J. (1978). “31 P nuclear magnetic resonance and the headgroup structure of phospholipids in membranes,” Biochim. Biophys. Acta 515, 105–140. Seelig, J. (1993). “Phospholipid headgroups as sensors of electric charge,” In “New Developments in Lipid–Protein Interactions and Receptor and Function” (K. W. A. Wirtz, ed.), Plenum Press, New York. Seelig, J. (1995). “Metal ion interactions with lipids,” In “Handbook of Metal–Ligand Interactions in Biological Fluids: Bioinorganic Chemistry,” Marcel Dekker, New York. Seelig, J., and Browning, J. L. (1978). “General features of phospholipid conformation in membranes,” FEBS Lett. 92, 41–44. Seelig, J., and Macdonald, P. M. (1987). “Phospholipids and proteins in biological membranes: 2 H-NMR as a method to study structure, dynamics, and interations,” Acc. Chem. Res. 20, 221–228. Seelig, J., and Seelig, A. (1980). “Lipid conformation in model membranes and biological membranes,” Q. Rev. Biophys. 13, 19–61. Seelig, J., Seelig, A., and Tamm, L. (1982). “Nuclear magnetic resonance and lipid–protein interactions,” In “Lipid-Protein Interactions” (P. Jost and O. H. Griffith, eds.), pp. 127–148, John Wiley & Sons, New York.

VII. CONCLUDING REMARKS
Solid-state NMR measurements have shown that functional biological membranes are in the liquid crystalline state and that structural features of lipids in the crystalline phase are essentially carried over into the liquid crystalline state. An order parameter profile comparable for the most diverse membranes has been established. The absolute values of order parameters may, however, vary as much as a factor of two as a consequence of the large variation in lipid composition encountered in biological membranes. Membrane ordering decreases upon increasing the temperature, introducing one or several cis-double bonds into a saturated fatty acyl chain, or upon adding an amphiphilic guest molecule. In contrast, it increases up to twofold upon addition of 50% cholesterol. Transmembrane proteins barely influence the lipid order, as they perfectly match the lipid bilayer properties. Due to the action of enzymes (e.g., phospholipases) the lipid packing density and hence the membrane order may vary with time and, in turn, may modulate the function of membrane proteins. A conformational change in a membrane protein may further be induced by an outof-plane rotation of the phospholipid headgroup dipole resulting in the development of a storage electric field across the membrane, which changes the protein structure. NMR measurements have further demonstrated that a fast exchange of lipid molecules is observed between the boundary of transmembrane proteins and the bulk lipid phase. At present, no physical–chemical evidence for the formation of domains or microdomains with lifetimes >10−4 s has been obtained under physiological conditions.

Natural Antioxidants In Foods
Eric A. Decker
University of Massachusetts I. Free Radical Scavengers II. Metal Chelators III. Antioxidant Enzymes

GLOSSARY
Antioxidant A compounds that can inhibit oxidative processes. Free radical A compound with an unpaired electron that can promote oxidative reaction. Free radical scavenger A compound that can absorb a free radical to decrease the radical energy thus making it less likely to cause oxidation. Metal chelators Compounds that can bind metals and decrease their reactivity. Phenolic A group of chemical compounds primarily found in plants that act as antioxidant and are beneficial to health.

ATMOSPHERIC (TRIPLET) oxygen is a low energy biradical (i.e., contains two unpaired electrons). However, during metabolism of oxygen as well nitrogen, alterations can occur to produce highly reactive oxygen and nitrogen species that will react with and cause damage to biomolecules. In foods, this can cause oxidation of lipids, pigments, vitamins, and proteins, leading to offflavor formation, discoloration, and loss of important nutrients. Foods, which are derived from a variety of different biological tissues, contain a host of different antioxidant defense systems to prevent the damaging effect of reactive

oxygen and nitrogen species. However, during the processing of biological tissues into foods, the formation of oxidizing species can increase and antioxidant systems can be overwhelmed leading to uncontrolled oxidative reactions resulting in loss of quality, decrease in shelflife, and formation of potentially toxic oxidation products. To protect food quality and safety, antioxidants are often added to processed foods. These antioxidants can be synthetically derived compounds, such as butylated hydroxytoluene and ethylene diaminetetraacetic acid. Concern over the use of synthetic food additives has driven the food industry to find effective natural antioxidants additives that are derived from biological sources. In addition, efforts to decrease oxidative deterioration have focused on the development of food processing techniques that preserve endogenous antioxidants and nutritional schemes that increase natural antioxidants in animal-derived foods. In addition to the association of natural antioxidants with food quality, these compounds have also been associated with health benefits. The association of the protective effects of fruits and vegetables in the diet against diseases, such as cancer and cardiovascular disease, has been established for years. Comprehensive reviews on the consumption of fruits and vegetables with cancer rates have shown that 60–85% of the studies have a statistically significant association with the decrease of cancer

335

336 incidence. Individuals who consume the highest amount of fruits and vegetables have half the cancer rate as those who consume the least amount. A similar association has been seen with cardiovascular disease, with 60% of the studies reviewed showing statistically significant protective effects. The consumption of an ample supply of fruits and vegetables provides a wide variety of phytochemicals that have been shown to have health benefits and antioxidant activity. The natural antioxidants with health benefits include ascorbic acid, α-tocopherol, β-carotene, and plant phenolics.

Natural Antioxidants In Foods

I. FREE RADICAL SCAVENGERS
A. Phenolic Antioxidants Phenolics are compounds that have a hydroxyl group associated with an aromatic ring structure. There are numerous variations of both natural and synthetic phenolics (see

Fig. 1 for examples). Natural phenolics are found predominately in the plant kingdom. Vitamin E or α-tocopherol is a plant phenolic required in the diet of humans and other animals. Phenolic compounds primarily inhibit lipid oxidation through their ability to scavenge free radicals and convert the resulting phenolic radicals into a low-energy form that does not further promote oxidation. Chemical properties, including ability of the antioxidant to donate hydrogen to the oxidizing free radical, decrease the energy of the antioxidant radical, and prevent autoxidation of the antioxidant radical into additional free radicals, will influence the antioxidant effectiveness of a free radical scavenger (FRS). In addition, physical partitioning of phenolics will also influence their reactivity. Initially, antioxidant efficiency is dependent on the ability of the FRS to donate a hydrogen to a high energy free radical. As the oxygen–hydrogen bond energy of the FRS decreases, the transfer of the hydrogen to the free radical is more energetically favorable and thus more rapid. The ability

FIGURE 1 Chemical structures of some examples of phenolic antioxidants.

Natural Antioxidants In Foods

337 associated with the prevention of diseases such as cancer and atherosclerosis. Plant foods high in phenolics include cereals, legumes, and other seeds (e.g., sesame, oats, soybeans, and coffee); red-, purple-, and blue-colored fruits (e.g., grapes, strawberries, and plums); and the leaves of herbs and bushes (e.g., tea, rosemary, and thyme). Many natural phenolics are capable of inhibiting oxidative reactions. However, because phenolics have such a wide array of chemical structures, it is not surprising that antioxidant activities and health benefits vary greatly. Knowledge of antioxidant activity, antioxidant mechanisms, and health benefits of plant phenolics is just beginning to be understood. This section focuses on the best studied of the plant phenolics. Tocopherols and tocotrienols are a group of phenolic FRS isomers (α, β, δ and γ ; see Fig. 1 for the structure of α-tocopherol) originating in plants and eventually ending up in animal foods via the diet. Interactions between tocopherols and fatty acid peroxyl radicals lead to the formation of fatty acid hydroperoxides and several resonance structures of tocopheroxyl radicals. Tocopheroxyl radicals can interact with other compounds or with each other to form a variety of products. The types and amounts of these products are dependent on oxidation rates, radical species, lipid state (e.g., bulk vs. membrane lipids), and tocopherol concentrations. Under condition of low oxidation rates in lipid membrane systems, tocopheroxyl radicals primarily convert to tocopherylquinone. Tocopherylquinone can form from the interaction of two tocopheroxyl radicals leading to the formation of tocopherylquinone and the regeneration of tocopherol. Tocopherylquinone can also be regenerated back to tocopherol in the presence of reducing agents (e.g., ascorbic acid). An additional reaction that can occur is the interaction of two tocopheroxyl radicals to form tocopherol dimers. Tocopherol is found in plant foods especially those high in oil. Soybean, corn, safflower, and cottonseed oil are good sources of α-tocopherol as are whole grains (in particular wheat germ) and tree nuts. All tocopherol isomers are absorbed by humans, but α-tocopherol is preferentially transfered from the liver to lipoproteins, which in turn transports α-tocopherol to tissues. For this reason, α-tocopherol is the isomer most highly correlated with vitamin E activity. Tea is an important source of dietary antioxidants for humans because it is one of the most common beverages in the world with annual consumption of over 40 liters/ person/year. Phenolics in tea are mainly catechin derivatives, including catechin (Fig. 1), epicatechin, epicatechin gallate, gallocatechin, epigallocatechin gallate, and epigallocatechin. Tea originates from leaves harvested from the bush, Camellia sinensis. Processing of tea leaves

of a FRS to donate a hydrogen to a free radical can sometimes be predicted from standard one electron reduction potentials (E◦ ). If a compound has a reduction potential lower than the reduction potential of a free radical found in a food or biological tissue (e.g., fatty acid based peroxyl radical), it can donate hydrogen to that free radical unless the reaction is kinetically unfeasible. For example, FRS including α-tocopherol (E◦ = 500 mV), urate (E◦ = 590 mV), catechol (E◦ = 530 mV), and ascorbate (E◦ = 282 mV) all have reduction potentials below peroxyl radicals (E◦ = 1000 mV, a common free radical in lipid oxidation reactions) and therefore can convert the peroxyl radical to a hydroperoxide through hydrogen donation. The efficiency of an antioxidant FRS is also dependent on the energy of the resulting antioxidant radical. If a FRS produces a low energy radical then the likelihood of the FRS radical to promote the oxidation of other molecules is lower and the oxidation reaction rate decreases. Phenolics are effective FRS because phenolic free radicals have low energy due to delocalization of the free radical thoughout the phenolic ring structure. Standard reduction potentials can again be used to help illustrate this point. Radicals on α-tocopherol (E◦ = 500 mV) and catechol (E◦ = 530 mV) have lower reduction potentials than polyunsaturated fatty acids (E◦ = 600 mV), meaning that their radicals do not posses high enough energy to effectively promote the oxidation of unsaturated fatty acids. Effective phenolic anioxidants FRS also produce radicals that do not react rapidly with oxygen to form hydroperoxides that could autoxidize, thus depleting the system of antioxidants. Antioxidant hydroperoxides are also a problem because they can decompose into radicals that could promote oxidation. Thus, if antioxidant hydroperoxides did form, this could result in consumption of the antioxidant with no net decrease in free radicals numbers. Antioxidant radicals may undergo additional reactions that remove radicals from the system, such as reactions with other antioxidant radicals or lipids radicals to form nonradical species. This means that each FRS is capable of inactivating at least of two free radicals, the first being inactivated when the FRS interacts with the initial oxidizing radical, and the second, when the FRS radical interacts with another radical via a termination reaction to form a nonradical product. Phenolic compounds that act as antioxidants are widespread in the plant kingdom. Plant phenolics can be classified as simple phenolics, phenolic acids, hydroxycinnamic acid derivatives, and flavonoids. In addition to the basic hydroxylated aromatic ring structure of these compounds, plant phenolics are often associated with sugars and organic acids. The consumption of natural plant phenolics have been estimated to be up to 1 g per day. Overall, the presence of phenolics in the diet has been positively

338 involves either blanching to produce green tea or fermenting to produce oolong or black tea. The fermentation process allows polyphenol oxidase enzymes to react with the catechins to form the condensed polyphenols that are responsible for the typical color and flavor of black teas. Green tea leaf extracts contain 38.8% phenolics on a dry weight basis with catechins contributing over 85% of the total phenolics. Condensation of catechins can decrease their solubility; therefore black tea extracts contain less phenolics (24.4%) of which 17% are catechins and 70% are condensed polyphenols (thearubigens). Extraction of phenolics with water from the leaves of rooibos (Aspalathus linearis) resulted in increased antioxidant activity with increasing extraction temperature and time, suggesting that brewing techniques could influence the antioxidant phenolic content of teas. Ingestion of dietary phenolics from tea has been associated with cancer prevention, and absorption of dietary tea phenolics has been reported. Grapes and wines are also significant sources of dietary phenolic antioxidants. Grapes contain a wide variety of phenolics including anthocyanins, flavan-3-ols (catechin), flavonols (quercetin and rutin), and cinnamates (Sglutathionylcaftaric acid). As with many fruits, the majority of grape phenolics are found in the skin, seeds, and stems (collectively termed pomace). During extraction of juice, the pomace is left in contact with the juice for varying times in order to produce products of varying color, with increasing contact time resulting in increased phenolic extraction and, thus, darker color. Therefore, white grape juices and wines have lower phenolics contents (119 mg of gallic acid equivalents/L) than red wines (2057 mg of gallic acid equivalents/L). As would be expected, red grape juice and wines have greater antioxidant capacity due to their higher phenolic content. Both grape juice and wines have been suggested to have positive heath benefits, however, their phenolic compositions are not the same due to differences in juice preparation and changes in phenolic composition that occurs during both fermentation and storage. The primary phenolics in soybeans are classified as isoflavones. Included among the soybean isoflavones are daidzein (Fig. 1), genistein, and glycitein, and the glycosolated counterparts daidzin, genistin, and glycitin. Unlike the phenolics in grapes and tea, soybean isoflavones are associated with proteins and, therefore, are found in soy flour and not in soybean oil. The concentration of isoflavones in soybeans varies with the environmental conditions under which the beans were grown. In addition, isoflavone concentrations in soy-based foods are altered during food processing operations such as heating and fermentation. Beside whole soybeans, isoflavones are found in soy milk, tempeh, miso, and tofu at concentrations ranging from

Natural Antioxidants In Foods

294–1625 µg/g product. Genistein and daidzein are absorbed into human plasma from products such as tofu and soy-based beverages. Bioavailability is low, with only 9–21% of the isoflavones being absorbed. Over 90% of the absorbed isoflavones are removed from the plasma within 24 hours. Herbs and spices often contain high amount of phenolic compounds. For example, rosemary contains carnosic acid, carnosol, and rosmarinic acid. Crude rosemary extracts are a commercially important source of natural phenolic antioxidant additives in foods meats, bulk oils, lipid emulsions, and beverages. B. Ascorbate Ascorbic acid (vitamin C; Fig. 2) acts as a water-soluble free radical scavenger in both plant and animal tissues. Like phenolics, ascorbate (E◦ = 282 mV) has a reduction potential below peroxyl radicals (E◦ = 1000 mV) and thus can inactivate peroxyl radicals. In addition, ascorbate’s reduction potential is lower than the α-tocopherol radical (E◦ = 500 mV), meaning that ascorbate may have an additional role in the regeneration of oxidized α-tocopherol. Interactions between ascorbate and free radicals result in the formation of numerous oxidation products. Although ascorbate seems to primarily play an antioxidant role in living tissues, this is not always true in food systems. Ascorbate is a strong reducing agent especially at low pH. When transition metals are reduced, they become very active prooxidants that can decompose hydrogen and lipid peroxides into free radicals. Ascorbate also causes the release of protein-bound iron (e.g., ferritin), thus promoting oxidation. Therefore, ascorbate can potentially exhibit prooxidative activity in the presence of free transition metals or iron-binding proteins. This does not typically occur in living tissues due to the tight control of free metals by systems that prevent metal reduction and reactivity. However, in foods the typical control of metals can be lost by processing operations that cause protein denaturation. Thus in some foods, ascorbate my act as a prooxidant and accelerate oxidative reactions. Ascorbate is found in numerous plant foods including green vegetables, citrus fruits, tomatoes, berries, and potatoes. Ascorbate can be lost in foods due to heat processing and prolonged storage. Transition metals and exposure to air will also cause the degradation of ascorbic acid. C. Thiols 1. Glutathione Glutathione (Fig. 2) is a tripeptide (γ -glutamyl-cysteinylglycine) where cysteine can be in either the reduced or

Natural Antioxidants In Foods

339

FIGURE 2 Chemical structures of miscellaneous natural antioxidants.

oxidized glutathione state. Reduced glutathione inhibits lipid oxidation directly by interacting with free radicals to form a relatively unstable sulfhydryl radical or by providing a source of electrons, which allows glutathione peroxidase to enzymically decompose hydrogen and lipid peroxides. Total glutathione concentrations in muscle foods range from 0.7–0.9 ug/kg. Oral administration of 3.0 of glutathione to seven healthy adults did not result in any increases in plasma glutathione or cysteine concentrations after 270 minutes. The bioavailability of glutathione in rats has also been reported to be low. Lack of, or low absorption of, glutathione may be due to the hydrolysis of the tripeptide by gastrointestinal protease.

2. Lipoic Acid Lipoic (thioctic) acid (Fig. 2) is a thiol cofactor for many plant and animal enzymes. In biological systems, the two thiol groups of lipoic acid are found in both reduced (dihydrolipoic acid) and oxidized forms (lipoic acid). Both the oxidized and reduced forms of the molecule are capable of acting as antioxidants through their ability to quench singlet oxygen, scavenge free radicals, chelate iron, and, possibly, regenerate other antioxidants such as ascorbate and tocopherols. Lipoic and dehydrolipoic acids can protect LDL, erythrocytes, and cardiac muscle from oxidative damage.

340 Although lipoic acid has been found in numerous biological tissues, reports on its concentrations in foods are scarce. Lipoic acid is detectable in wheat germ (0.1 ppm) but not in wheat flour. It has been detected in bovine liver kidney and skeletal muscle. Oral administration of lipoic acid (1.65 g/kg fed) to rats for five weeks resulted in elevated levels of the thiol in liver, kidney, heart, and skin. When lipoic acid was added to diets lacking in vitamin E, symptoms typical of tocopherol deficiency were not observed suggesting that lipoic acid acts as an antioxidant in vivo. However, lipoic acid was not capable of recycling vitamin E in vivo, as determine by the fact that α-tocopherol concentrations are not elevated by dietary lipoic acid in vitamin E deficient rats. D. Carotenoids Carotenoids are a chemically diverse group (>600 different compounds) of yellow to red colored polyenes consisting of 3–13 conjugates double bonds and in some cases, six carbon hydroxylated ring structures at one or both ends of the molecule. ß-Carotene is the most extensively studied carotenoid antioxidant (Fig. 2). ß-Carotene will react with lipid peroxyl radicals to form a carotenoid radical. Whether this reaction is truly antioxidative seems to depend on oxygen concentrations, with high oxygen concentrations resulting in a reduction of antioxidant activity. The proposed reason for loss of antioxidant activity with increasing oxygen concentrations involves the formation of carotenoid peroxyl radicals that autoxidizes into additional free radicals. Under conditions of low oxygen tension, the carotenoid radical would be less likely to autooxidize and thus could react with other free radicals thereby forming nonradical species with in a net reduction of radical numbers. The major antioxidant function of carotenoids in foods is not due to free radical scavenging but instead is through its ability to inactivate singlet oxygen. Singlet oxygen is an excited state of oxygen in which two electrons in the outer orbitals have opposite spin directions. Initiation of lipid oxidation by singlet oxygen is due to its electrophillic nature, which will allow it to add to the double bonds of unsaturated fatty acids leading to the formation of lipid hydroperoxides. Carotenoids can inactivate singlet oxygen by both chemical and physical quenching. Chemical quenching results in the direct addition of singlet oxygen to the carotenoid, leading to the formation of carotenoid breakdown products and loss of antioxidant activity. A more effective antioxidative mechanism of carotenoids is physical quenching. The most common energy states of singlet oxygen are 22.4 and 37.5 kcal above ground state. Carotenoids physically quench singlet oxy-

Natural Antioxidants In Foods

gen by a transfer of energy from singlet oxygen to the carotenoid, resulting in an excited state of the carotenoid and ground state, triplet oxygen. Harmless transfer of energy from the excited state of the carotenoid to the surrounding medium by vibrational and rotational mechanisms then takes place. Nine or more conjugated double bonds are necessary for physical quenching, with the presence of six carbon oxygenated ring structures at the end the molecule increasing the effectiveness of singlet oxygen quenching. In foods, light will activate chlorophyll, riboflavin, and heme-containing proteins to high energy excited states. These photoactivated molecules can promote oxidation by direct interactions with an oxidizable compounds to produce free radicals, by transferring energy to triplet oxygen to form singlet oxygen or by transfer of an electron to triplet oxygen to form the superoxide anion. Carotenoids inactivate photoactivated sensitizers by physically absorbing their energy to form the excited state of the carotenoid that then returns to the ground state by transfer of energy into the surrounding solvent.

II. METAL CHELATORS
A. Ethylene Diamine Tetraacetic Acid Transition metals will promote oxidative reactions by hydrogen abstraction and by hydroperoxide decomposition reactions that lead to the formation of free radicals. Prooxidative metal reactivity is inhibited by chelators. Chelators that exhibit antioxidative properties inhibit metal-catalyzed reactions by one or more of the following mechanims: prevention of metal redox cycling; occupation of all metal coordination sites thus inhibiting transfer of electrons; formation of insoluble metal complexes; stearic hinderance of interactions between metals and oxidizable substrates (e.g., peroxides). The prooxidative/antioxidative properties of a chelator can often be dependent on both metal and chelator concentrations. For instance, ethylene diamine tetraacetic acid (EDTA) can be prooxidative when EDTA:iron ratios are ≤1 and antioxidative when EDTA:iron is ≥1. The prooxidant activity of some metal-chelator complexes is due to the ability of the chelator to increase metal solubility and/or increase the ease by which the metal can redox cycle. The most common metals chelators used in foods contain multiple carboxylic acid (e.g., EDTA and citric acid) or phosphate groups (e.g., polyphosphates and phytate). Chelators are typically water soluble but many also exhibit some solubility in lipids (e.g., citric acid), thus allowing

Natural Antioxidants In Foods

341

it to inactivate metals in the lipid phase. Chelator activity is pH dependent with a pH below the pKa of the ionizable groups resulting in protonation and loss of metal binding activity. Chelator activity is also decreased in the presence of high concentrations of other chelatable nonprooxidative metals (e.g., calcium), which will compete with the prooxidative metals for binding sites.

III. ANTIOXIDANT ENZYMES
A. Superoxide Anion Superoxide anion is produced by the addition of an electron to molecular oxygen. Superoxide anion can promote oxidative reactions by (1) reduction of transition metals to their more prooxidative state, (2) promotion of metal release from proteins, (3) through the pH dependent formation of its conjugated acid which can directly catalyze lipid oxidation, and (4) through its spontaneous dismutation into hydrogen peroxide. Due to the ability of superoxide anion to participate in oxidative reactions, the biological tissues from which foods originate will contain superoxide dismutase (SOD). Two forms of SOD are found in eukaryotic cells, one in the cytosol and the other in the mitochondria. Cytosolic SOD contains copper and zinc in the active site. Mitochondrial SOD contains manganese. Both forms of SOD catalyze the conversion of superoxide anion (O2− ) to hydrogen peroxide by the following reaction. 2O2− + 2H+ → O2 + H2 O2 .

B. Metal-Binding Proteins The reactivity of prooxidant metals in biological tissues are mainly controlled by proteins. Metal binding proteins in foods include transferrin (blood plasma), phosvitin (egg yolk), lactoferrin (milk), and ferritin (animal tissues). Transferrin, phosvitin, and lactoferrin are structurally similar proteins consisting of a single polypeptide chain with a molecular weight ranging from 76,000–80,000. Transferrin and lactoferrin each bind two ferric ions, whereas phosvitin has been reported to bind three. Ferritin is a multisubunit protein (molecular weight of 450,000) with the capability of chelating up to 4500 ferric ions. Transferrin, phosvitin, lactoferrin, and ferritin inhibit iron-catalyzed lipid oxidation by binding iron in its inactive ferric state and, possibly, by sterically hindering metal/peroxide interactions. Reducing agents (ascorbate, cysteine, and superoxide anion) and low pH can cause the release of iron from many of the iron-binding proteins, resulting in an acceleration of oxidative reactions. Copper reactivity is controlled by binding to serum albumin, ceruloplasmin, and the skeletal muscle dipeptide, carnosine.

B. Catalase Hydroperoxides are important oxidative substrates because they decompose via transition metals, irradiation, and elevated temperatures to form free radicals. Hydrogen peroxide exists in foods due to its direct addition (e.g., aseptic processing operations) and by its formation in biological tissues by mechanisms including the dismutation of superoxide by SOD and the activity of peroxisomes. Lipid hydroperoxides are naturally found in virtually all food lipids. Removal of hydrogen and lipid peroxides from biological tissues is critical to prevent oxidative damage. Therefore, almost all foods originating from biological tissues contain enzymes that decompose peroxides into compounds less susceptible to oxidation. Catalase is a hemecontaining enzyme that decomposes hydrogen peroxide by the following reaction. 2H2 O2 → 2H2 O + O2 . C. Ascorbate Peroxidase Hydrogen peroxide in higher plants and algae may also be decomposed by ascorbate peroxidase. Ascorbate peroxidase inactivates hydrogen peroxide in the cytosol and chloroplasts by the following mechanism. 2 ascorbate + H2 O2 → 2 monodehydroascorbate + 2H2 O.

C. Phytic Acid Phytic acid or myoinositol hexaphophate is one of the primary metal chelators in seeds where it can be found at concentrations ranging from 0.8–5.3% (Fig. 2). Phytic acid is not readily digested in the human gastrointestinal tract but can be digested by dietary plant phytases and by phytases originating from enteric microorganisms. Phytate is highly phosphorylated, thus, allowing it to form strong chelates with iron, with the resulting iron chelates having lower reactivity. The antioxidant properties of phytic acid are thought to help minimize oxidation in legumes and cereal grains as well as in foods that may be susceptible to oxidation in the digestive tract. Phytic acid has been cited as a preventative agent in iron-mediated colon cancer. Although phytate may be beneficial toward colon cancer, it should be noted that it can potentially have deleterious health effects because of its ability to dramatically decrease the bioavailability of minerals including iron, zinc, and calcium.

342 Two ascorbate peroxidase isozymes have been described that differ in molecular weight (57,000 versus 34,000), substrate specificity, pH optimum, and stability. D. Glutathione Peroxidase Many foods also contain glutathione peroxidase. Glutathione peroxidase differs from catalase in that it decomposes both lipid and hydrogen peroxides. GSH-Px is a selenium-containing enzyme that catalyzes hydrogen or lipid (LOOH) peroxide reduction using reduced glutathione (GSH): H2 O2 + 2GSH—2H2 O + GSSG or, LOOH + 2GSH—LOH + H2 O + GSSG, where GSSG is oxidized glutathione and LOH is a fatty acid alcohol. Two types of GSH-Px exist in biological tissues, of which one shows high specificity for phospholipid hydroperoxides. E. Antioxidant Enzymes in Foods Antioxidant enzyme activity in foods can be altered in raw materials and finished products. Antioxidant enzymes differ in different genetic strains and at different stages of development in plant foods. Heat processing and food additives (e.g., salt and acids) can inhibit or inactivate antioxidant enzyme activity. Dietary supplementation of selenium can be used to increase the glutathione peroxidase activity of animal tissues. These factors suggests that technologies could be developed to increase natural levels of antioxidant enzymes in raw materials and/or minimize their loss of activity during food processing operations.

Natural Antioxidants In Foods

of how these endogenous antioxidants protect foods from oxidation is still in its infancy. In addition, how factors that can alter the activity of endogenous food antioxidants (e.g., heat processing, irradiation, and genetic selection of foods high in antioxidants) is still poorly understood. Finally, research is continuing to show that natural food antioxidants in the diet are very important in the modulation of disease. Thus, finding mechanisms to increase natural food antioxidants may be beneficial to both health and food quality.

SEE ALSO THE FOLLOWING ARTICLES
FOOD COLORS • HYDROGEN BOND • LIPOPROTEIN/ CHOLESTEROL METABOLISM

BIBLIOGRAPHY
Buettner, G. R. (1993). “The pecking order of free radicals and antioxidants: Lipid peroxidation, α-tocopherol, and ascorbate,” Arch. Biochem. Biophys. 300, 535–543. Decker, E. A. (1998). “Strategies for manipulating the prooxidative/antioxidative balance of foods to mazimixe oxidative stability,” Trends Food Sci. Technol. 9, 241–248. Decker, E. A., and Clarkson, P. (2000). “Dietary sources and bioavailability of essential and nonessential antioxidants,” In: Exercise and Oxygen Toxicity (C.K. Sen, L. Packer, and O. Hanninen, eds.). pp. 323–358. Elsevier Science, Amsterdam. Frankel, E. N. (1996). “Antioxidants in lipid foods and their impact on food quality,” Food Chem. 57, 51–55. Graf, E., and Eaton, J. W. (1990). “Antioxidant functions of phytic acid,” Free Rad. Biol. Med. 8, 61–69. Halliwell, B. (1999). “Establishing the significance and optimal intake of dietary antioxidants: The biomarker concept,” Nutr. Rev. 57, 104–113. Halliwell, B., Murcia, M. A., Chirico, S., Aruoma, O. I. (1995). “Free radicals and antioxidants in foods and in vivo: What they do and how they work,” Crit. Rev. Food Sci. Nutr. 35, 7–20. Krinsky, N. I. (1992). “Mechanism of action of biological antioxidants,” Proc. Soc. Exp. Biol. Med. 200, 248–254. Liebler, D. C. (1993). “The role of metabolism in the antioxidant function of vitamin E,” Crit. Rev. Toxicol. 23, 147–169. Liebler, D. C. (1992). “Antioxidant reactions of carotenoids,” Ann. N Y Acad. Sci. 691, 20–30. Nawar, W. W. (1996). “Lipids,” In: Food Chemistry (O. Fennema, ed.), 3rd edition, pp. 225–319. Marcel Dekker, New York.

CONCLUSION
The biological tissues from which foods originate contain multicomponent antioxidant systems that include free radical scavengers, metal chelators, singlet oxygen quenchers, and antioxidant enzymes. Our understanding

Nucleic Acid Synthesis
Sankar Mitra Tapas K. Hazra Tadahide Izumi
University of Texas Medical Branch, Galveston

I. II. III. IV. V. VI. VII.

Structure and Function of Nucleic Acids Nucleic Acid Syntheses DNA Replication and Its Regulation Maintenance of Genome Integrity DNA Manipulations and Their Applications Transcriptional Processes Chemical Synthesis of Nucleic Acids (Oligonucleotides)

GLOSSARY
Cell cycle Stages in the life cycle of replicating eukaryotic cells. After cell division (mitosis), a cell goes through the resting G1/Go phase prior to DNA replication in the S phase. Completion of duplication of cellular materials in the G2 phase occurs prior to mitosis (M phase). Chromatin Cellular genome as nucleoprotein which contains DNA, histones, and a variety of nonhistone chromosomal proteins. Chromatin remodeling Alteration in the structure of segment of chromatin which is brought about by histone acetylation/deacetylation and/or mediated by interaction with large protein complexes as a prerequisite for modulation of transcription activity. Chromosomes Discrete and microscopically visible segments of the eukaryotic genome complexed with proteins and capped by telomeres; each normally contains thousands of genes.

Cis element Short, specific DNA sequences, usually in the promoter, that bind cognate trans-acting factors. Deoxyribonucleotides Monomeric units of DNA, including deoxyadenylic (dAMP), deoxyguanylic (dGMP), deoxycytidylic (dCMP), and deoxythymidylic (dTMP) acids. DNA Deoxyribonucleic acid: linear copolymers of monomeric deoxyribonucleotides normally present as a two-stranded intertwined helix; the deoxyribose sugar moiety lacks DNA helicase An enzyme which unwinds the double helical DNA using energy provided by ATP hydrolysis. DNA ligase The enzyme which catalyzes joining of the 5 and 3 termini of two single-stranded DNA fragments in a double-stranded DNA by forming a phosphodiester bond. DNA repair Enzymatic process that maintains sequence integrity by removing both endogenously and exogenously induced DNA damage. Such lesions could be

.

853

854 mutagenic because of misreplication at the damage site. Replication errors are also corrected by DNA repair. Repair involves removal of the DNA damage site in duplex DNA, followed by resynthesis of the damaged strand using the unaffected complementary strand as the template. Enhancer elements DNA sequences which activate the expression of genes in an orientation- and positionindependent fashion. Episome Small extrachromosomal and sometimes selfreplicating DNA molecules, including infecting viral DNA, founded in both prokaryotes and eukaryotes. Error-bypass DNA polymerases A new class of recently discovered DNA polymerases in both prokaryotes and eukaryotes which are more tolerant of improper base pairing and may function in maintaining genomic continuity when damaged DNA bases have not been repaired. Function The intrinsic 3 exonuclease activity of replicative DNA polymerase or polymerase complexes needed to excise incorrect deoxynucleotides inserted at the terminus of a growing DNA chain. Gene Basic functional unit in the genome which is transcribed to produce messenger RNA, which in turn is translated into protein. (Some genes, e.g., those for ribosomal and transfer RNAs, are only transcribed and not translated.) Genome Complete genetic information stored in the nucleotide sequence (usually DNA) of an organism, organelle, or episome. HMG proteins High mobility group (based on gel electrophoresis) proteins which are associated with chromatin; a subset of nonhistone chromosomal (NHC) proteins. Lagging strand Nascent DNA strand synthesized discontinuously by replication of the 5 → 3 template strand. Leading strand Nascent DNA synthesized by continuous replication of the 3 → 5 template strand. Mitochondrial genome Multiple copies of the circular DNA duplex molecule in eukaryotic mitochondria. Believed to be a vestigial prokaryotic genome, it is replicated by a special DNA polymerase (Pol γ ) which, along with other proteins required for mitochondrial DNA replication, is encoded by the nuclear genome. Mutation Change in the genome sequence via the process of mutagenesis, which can occur either spontaneously due to endogenous reactions or after exposure to external mutagens, including radiation and chemicals. Mutations include large-scale sequence alterations, including deletion or insertion of thousands of DNA base pairs and genomic rearrangement which could involve translocation of one chromosomal seg-

Nucleic Acid Synthesis

ment to another. Mutations could also be subtle, including changes of a single base (known as point mutation), which include loss or addition of a single base. Nontranscribed strand The complementary strand (5 3 ) of DNA with the same sequence as the RNA transcribed from the other (transcribed or template) strand. Nucleosome Smallest repeat unit of chromatin nucleoprotein, containing 145 bp of DNA wrapped around a histone octamer core (2 subunits each of histone H2A, H2B, H3, and H4) along with linker DNA of variable length. Mild treatment of chromatin with DNase digests the linker and generates nucleosome fragments of different repeat lengths (“ladder”). Okazaki fragments Nascent DNA fragments generated by discontinuous synthesis of the lagging (5 → 3 ) strand in all organisms. Operator A small, specific, and often palindromic DNA sequence or its repeats cognate to regulated bacterial genes. A repressor (or activator) binds the operators to prevent (or activate) transcription. Ori (origin) Origin of replication in the genome. These are unique sequences which bind the replication initiation complex as a prerequisite for primer synthesis. PCR Polymerase chain reaction. Plasmid Extrachromosomal DNA molecule, usually much smaller than the cell genome. Plasmids are autonomously replicated in the cell, utilizing the cellular replication machinery. Pol DNA or RNA polymerase. Primase Enzyme (sometimes with other accessory proteins) which is a component of the DNA replication machinery and is needed for synthesis of an oligoribonucleotide primer. Promoter Specific DNA sequence usually found at the beginning of a gene, which binds the transcriptional machinery as a prerequisite to transcription initiation from the gene. Replicon Unit of DNA replication in the genome, containing one ori site. Small genomes of bacteria, plasmids, and viruses have single replicons, while larger eukaryotic genomes have hundreds or thousands of replicons which could be simultaneously or sequentially fired for synthesis of different segments of the genome. This is necessary to reduce the overall replication time of a genome which is 103 times larger than a bacterial genome. Repressor Proteins which bind to specific operators and thus negatively regulate gene expression by inhibiting transcription. Reverse transcriptase (RT) Specialized DNA polymerase encoded by retroviruses, including the AIDS virus (HIV), which utilizes both RNA and DNA

Nucleic Acid Synthesis

855 termediate. The enzyme acts as a swivel for rotating DNA strands around each other. Trans-acting factors Proteins that bind to specific DNA sequences (cis elements) in genes and regulate transcription positively or negatively. Transcribed strand The 3 → 5 DNA strand utilized by RNA polymerase as its transcriptional template. Transcriptional activator Trans-acting proteins which enhance transcription and, thus, the level of specific proteins. Transcription unit Discrete segment of DNA, corresponding to one or more genes, which is utilized as a template by RNA polymerase. In prokaryotes, the transcription unit is called an operon. Translation Synthesis of a protein, directed by mRNA molecules on ribosome.

template. It is responsible for propagation of retroviruses via synthesis of a proviral DNA intermediate. Ribonucleotides Monomeric units of RNA, namely, adenylic (AMP), guanylic (GMP), cytidylic (CMP), and uridylic (UMP); the ribose sugar moiety of each contains a 2 -OH. Ribosome Protein synthesis factory consisting of two differently sized subunits of ribonucleoprotein complexes with several active centers. It travels along mRNA and reads triplet codons for individual amino acids which are brought in by transfer RNAs via base pairing with cognate anticodon sequences in these RNAs. Protein synthesis occurs on the ribosome to which the growing polypeptide chain remains attached. RNA Ribonucleic acid: linear copolymers usually of four ribonucleotides. Three major types of RNA are synthesized in the cell: ribosomal RNA (rRNA), the major component of ribosomes; transfer RNA (tRNA), the adaptor for protein synthesis; and messenger RNA (mRNA), which is required for information transfer. Other small RNAs with specialized functions are also synthesized in small amounts in both prokaryotic and eukaryotic cells. RTPCR Reverse transcript polymerase chain reaction. Modification of the PCR method to amplify RNA, which involves generation of a complementary DNA molecule from RNA (by reverse transcriptase) which is then used in PCR. Telomerase A special eukaryotic DNA polymerase that adds a repeat sequence to chromosome termini without a template. Telomere Terminal region of a linear chromosome, containing partial single-stranded DNA and repeat sequences of short oligonucleotides. Its loss could cause chromosome fusion and rearrangement. Template-independent poly(A) polymerase A template-independent RNA polymerase which catalyzes formation of AMP containing homopolymers up to several hundred monomers at the 3 termini of nascent RNA molecules. The poly(A) tail promotes transport of mRNA from the nucleus, enhances its stability, and is necessary for translation. Terminator Specific sequence found at the end of genes for termination of transcription due to release of RNA and RNA polymerase. Topoisomerase Enzymes which alter topologically constrained DNA, including circular DNA, by changing the linking number. Topoisomerase I changes the linking number one at a time and does not require an external energy source. Topoisomerase II changes the linking number two at a time and generally requires ATP. The linking number is changed by transient breakage and rejoining, with an enzyme-DNA covalent bond in-

NUCLEIC ACIDS are involved in the most fundamental processes of life. Their maintenance and production are essential in all living organisms. The hallmark of the biosphere is diversity of biological processes, even among members of the same genera, e.g., bacteria. Each organism may have some unique features in regard to nucleic acid composition, structure, and metabolism. Thus, studies on nucleic acid synthesis constitute a huge topic of research on which thousands of research articles are published each year. Therefore, it is impossible to cover all aspects of nucleic acid synthesis in this short article. Our goal is to present a broad overview of the key and general features of structure, synthesis, and processing of the various types of nucleic acids. We have limited our discussion mostly to bacteria, specifically Escherichia coli, and to mammals, mostly humans and mice. Most of our current knowledge has been derived from the studies of those organisms. We have also provided appropriate references, which are mostly recent reviews. The readers should be able to peruse these for in-depth knowledge of the topics which are covered only superficially here. Finally, we have included a glossary at the beginning of this article which lists common acronyms and short descriptions of key processes and phenomena.

I. STRUCTURE AND FUNCTION OF NUCLEIC ACIDS
A. Basic Chemical Structure The basic information for all activities in living systems, at least on our planet, is stored ultimately in nucleic acids, namely, deoxyribonucleic (DNA) and ribonucleic (RNA) acids. Except for certain viruses, DNA is the universal genetic material (Fig. 1). The chemical structures of basic

856

Nucleic Acid Synthesis

FIGURE 1 Structure of DNA and RNA: (A) structure of deoxyribonucleotides and ribonucleotides and (B) structure of polynucleotide. Each 3 carbon of the sugar residue is linked to the 5 carbon of the sugar residue in the next nucleotide with a phosphate to form the phosphodiester backbone. (C) Base paring of adenine with thymine (uracil) and guanine with cytosine. Dotted lines denote hydrogen bonding between two bases. R, pentose ring of nucleotide. (D) A three-dimensional structure of a DNA helix.

units of RNA and DNA have been elucidated, and both types of nucleic acids are linear polymers of monomeric units called nucleotides. A nucleotide consists of a purine or pyrimidine base linked to C -1 of a pentose (furanose) via an N •C glycosyl bond and contains a phosphate residue attached to the sugar via an ester bond with a CH2 OH group at the 5 position. The linear polymer in both RNA and DNA is generated by a C -3 ester linkage of 5 nucleotides generating a 3 -5 phosphodiester linkage (Fig. 1B). There are several differences in the chemical structures of DNA and RNA. First is the nature of the pentose ring in these macromolecules, i.e., ribofuranose for RNA and 2 -deoxyribofuranose for DNA (Fig. 1A). Because of the presence of deoxyribose in DNA, the monomeric unit is called a deoxyribonucleotide or simply a deoxynucleotide, while the RNA monomer unit is called a ribonucleotide.

The term “nucleotide” is used generically for both RNA and DNA units. The absence of a 2 -OH group in DNA prevents alkali-mediated cleavage of the 3 -5 phosphodiester cleavage observed in RNA and thus makes DNA more resistant to hydrolysis. Both RNA and DNA contain two types of purines, adenine (A) and guanine (G), and two types of pyrimidine bases (Fig. 1C). The second key difference between RNA and DNA is that while cytosine (C) is present in both RNA and DNA, RNA normally contains uracil (U), while DNA contains 5-methyluracil, called thymine (T), as the other pyrimidine base. The difference in chemical structure is reflected in the intrinsic chemical stability of these nucleic acids. The purine N glycosyl bond in DNA is more unstable than in RNA, and as a result, purines are released much more easily from DNA by acid catalysis. Furthermore, cytosine deamination to produce U also occurs at a finite rate in DNA.

Nucleic Acid Synthesis

857 equimolar amounts of A and T and of G and C (Chargaff’s rule), unlike in RNA, which is single stranded (except in some viruses). X-ray diffraction studies also showed that DNA in double helix exists in the B-form, which is right handed and has a wide major groove and a narrow minor groove. Most of the reactive sites in the bases, including C O and NH groups, are exposed in the major groove (Figs. 1C and 1D). One turn of the helix has10 base pairs (bp) with a rise of 34◦ . Thus, each pair is rotated 36◦ relative to its neighbor. Elucidation of the structure of DNA bound to proteins show that one turn of the helix containing 10.5 bp could be significantly bent or distorted. For example, some DNA binding proteins bind to the minor groove, causing its widening accompanied by compression of the major groove. In some special regions of the genomes, e.g., in telomeres and segments with unusual repeated sequences, alternative forms such as triple helical structure and Z-DNA may exist. The Z-DNA has a left-handed, double-helical structure. In these or in torsionally stressed DNA, the bases can be held together by different type of H-bonding called Hoogsteen base pairing. C. Size, Structure, Organization, and Complexity of Genomes Except for certain viruses, DNA is the genetic material for all organisms and self-replicating units, including viruses and such intracellular organelles as chloroplasts (in plants), kinetoplasts (in protozoa), and mitochondria (in most eukaryotes). Genomic DNA is double helical (except for the genomes of certain bacterial viruses), and its size is related to the complexity of the organism (Table I). In subcellular organelles, viruses, and plasmids, the genome often exists as a circular molecule consisting of up to several thousand base pairs. The genome of bacteria, such as that of the widely studied enteric strain E. coli, is present as a single, circular, double-stranded molecule containing about 4.7 million base pairs. By and large, the genome of many small self-replicating entities is circular DNA, without any terminus in the unbranched polymeric chain. In contrast, the large nuclear genomes of more complex organisms (from lower eukaryotes such as unicellular yeast with a genome size only an order of magnitude larger than that of E. coli, to mammals with genomes larger by three orders of magnitude) consist of multiple, distinct, linear subunits organized in chromosomes. Depending on the stage of the cell cycle, the structure of chromosomes (collectively called chromatin) varies from the highly extended and amorphous state occurring in much of the (interphase) nucleus to highly compacted, linear, organized chromosomes (metaphase) after completion of DNA duplication followed by cell division (mitosis). This

Various processes have evolved to maintain the genomic integrity, as discussed later. Finally, two other critical differences between DNA and RNA are in the length and structure of the polymer chains. DNA polymers, as elaborated later, usually exist as a helix consisting of two intertwining chains, while RNA is present mostly as a single chain. Furthermore, DNA could contain up to several billion deoxynucleotide monomeric units in the genomes of higher organisms, although the genomes of smaller self-replicating units such as viruses contain only a few thousand deoxynucleotides. In contrast, RNA chains are never more than a few thousand nucleotides long. B. Base Pairing in Nucleic Acids: Double Helical Structure of DNA The most important discovery in molecular biology was the identification of the right-handed double helical structure of DNA, where two linear chains are held together by base pair complementarity. This discovery by Watson and Crick in 1953 heralded the era of molecular biology, which was preceded by the rapid accumulation of genetic evidence indicating that DNA, as the genetic material of all organisms, is the primary storehouse of all their information. Exceptions to this fundamental principle were found in certain bacterial, plant, and mammalian viruses, in which RNA constitutes the genome. However, the viruses are obligate parasites and are not able to self-propagate as independent species; thus, they have to depend on their hosts, which have DNA as their genetic material. Thus, DNA in all genomes (except some single-stranded DNA viruses) consists of two strands of polydeoxynucleotides which are anti-parallel in respect to the orientation of the 5 -3 phosphodiester bond in the polymers (Fig. 1D). The two strands are held together by H-bonding between a purine in one strand and a pyrimidine in the complementary strand. Normally, adenine (A) pairs with T and G pairs with C; A and T are held together by two H-bonds, and G and C are held together by three H-bonds involving both exocyclic C O and ring NH (Fig. 1C). As a result, G•C pairs are more stable than A•T pairs. Because U is structurally nearly identical to T, except for the C-5 methyl group, U also pairs with A in the common configuration. Although H-bonds are inherently weak, the stacking of bases in two polynucleotide chains makes the duplex structure of DNA quite stable and induces a fibrillar nature in the DNA polymer. X-ray diffraction studies of the DNA fiber, and subsequent crystallographic studies of small (oligonucleotide) DNA pieces, led to the detailed structural elucidation. This was initially aided by chemical analysis showing equivalence of purines and pyrimidines in all double-stranded DNA and

858
TABLE I Genomic DNA Characterized in Biologya Organism Bacteriophage Virus Structure Linear, circular Total size (bp) 5 ∼ 200 × 103 Up to 2 × 105 4.6 × 106 1.4 × 107 1.4 × 108 3 × 109 Number of genes 10∼100 10∼100 ∼4300 ∼6000 1.4 × 104 4 × 104 to 1 × 105 Sequence

Nucleic Acid Synthesis

Completed for many species Completed for many species Completed Completed Partially completed Partially completed

Bacteria E. coli Circular Eukaryote yeast (S. cerevisiae) Linear Drosophila Human Linear Linear

a As of Feb 2001 the data are to be renewed continuously and are available at the website http://ncbi.nlm. nih.org/entrez.

complex organization of eukaryotic genomes is a distinctive feature which separates them from the prokaryotes. D. Information Storage, Processing, and Transfer The central dogma of molecular biology is that information is transferred from DNA to RNA to proteins. The proteins (which include the enzymes and structural components of cells) are directly responsible for most cellular activities and functions. The information needed for all functions of all organisms is stored in the genomic DNA sequence, which contains discrete units defined as genes. Each gene encodes a protein whose function and activity are determined by its primary sequence. The discovery of colinearity of the DNA nucleotide sequence and the amino acid sequence of the encoded polypeptide in prokaryotes and their viruses led to the discovery of the genetic code

which postulates that a three-nucleotide sequence in DNA, called a codon, is responsible for insertion of a specific amino acid in the polypeptide chain during its synthesis. Thus, the information content in the genomic DNA of a cell needs not only to be preserved and passed on to the progeny cells during replication, an essential characteristic and requirement of all living organisms, but also has to be processed and transferred via proteins to the ultimate cellular activities, including the metabolism. Elucidation of the double-helical structure of DNA lends itself to an elegant but simple mechanism of perpetuation of the DNA information during duplication, called semi-conservative replication. In this model (Fig. 2), the two strands of DNA separate, and each then acts as the template for synthesis of a new daughter strand based on base pair complementarity and strand polarity. Thus, the two strands of the DNA double helix, though not identical in sequence, are equivalent in information content.

FIGURE 2 DNA polymerization reaction. (A) According to the base pairing rules, a deoxythymidinetriphosphate (dTTP) is added at the 3 -OH end of the top strand through a transesterification reaction catalyzed by a DNA polymerase. (B) Two units of DNA polymerase form a heterodimer complex to carry out replication in a semi-conservative way. Because the reaction goes only in the 5 → 3 direction, one side (the leading strand) is synthesized continuously, while the other (the lagging strand) consists of short DNA fragments (Okazaki fragment). DNA replication is initiated by an RNA primer (waved line) which is synthesized by a primase. There are a number of accessory but essential proteins besides the polymerase unit.

Nucleic Acid Synthesis

859 or H5. The polymeric chain nucleosomes are then folded in a 30-nm fiber whose structure is stabilized by the interaction among histones and a number of other proteins collectively called nonhistone chromosomal proteins (NHC), including high mobility group (HMG), which are not particularly basic. Eventually, the fibers are condensed into highly compacted metaphase chromosomes. The nature of the interactions present in interphase and metaphase chromosomes is not clear. However, the implications of this compaction are profound. It is absolutely essential to condense the mammalian genome, which in an extended linear form more than 1 m long, to a volume which can be accommodated in the nuclear volume of 10–30 femtoliters. At the same time, the genes will be buried in condensed chromatin, and yet their specific sequences need to be exposed for various processes of information transfer. Thus, for both transcription and replication, the chromatin has to be decondensed. This was evident in early in vitro studies which showed that both these processes are severely inhibited when DNA is complexed with histones. F. DNA Sequence and Chromosome Organization The massive human genome project should achieve its goal of determining the complete sequence of human and mouse genomes in the near future; a “rough draft” has already been obtained. Furthermore, this genome initiative, pursued by both government and private enterprises in the United States and other countries, has already culminated in elucidating the complete sequence of E. coli and other bacteria, as well as yeast, a nematode, and the fruitfly Drosophila melanogaste. Significant progress has been made in elucidating the nucleotide sequences of both human and mouse genomes by using a two-pronged approach. On one hand, the sequences of transcribed regions of the genomes are being deduced from sequences of randomly isolated mRNA segments reverse transcribed into DNAs. At the same time, complete DNA sequences of fragments of whole chromosomes are being directly determined. This has opened up a huge scientific challenge of deciphering the genetic information, identifying unknown genes and their encoded proteins, and the variability of gene sequences with corresponding changes in the protein sequences in individuals. Functional genomics is a newly created discipline which deals with the deterministic prediction of protein functions from the primary sequences. One extension of such analysis is to ascertain the consequences of allelic polymorphisms in the human genome, i.e., minor changes in the sequences of cellular proteins which do not cause an explicit pathological phenotype and yet

FIGURE 3 An RNA polymerase unit (filled circle), which consists of multiple factors, opens DNA helix (shown as a bubble) and synthesizes RNA in the 5 → 3 direction.

The intermediate carrier in the transfer of information from DNA to protein is the messenger RNA (mRNA), which is copied (transcribed) from only one of the two strands (Fig. 3), based on base pair complementarity (except for the presence of U in RNA in the place of T; Fig. 1C). In the synthesis of both DNA (replication) and RNA (transcription), the polynucleotide chains are synthesized by sequential addition of monomeric units (deoxyribonucleotide for DNA and ribonucleotides for RNA) to the 3 end of the growing chain (Fig. 3). The mRNA is read out by ribosomes, the ribonucleoprotein complex which functions as the factory for protein synthesis. The codons are recognized as blocks because they code for specific amino acids. Thus, the linear polypeptide sequence is determined by the linear mRNA sequence. E. Chromosomal DNA Compaction and Its Implications in Replication and Transcription Metaphase chromosomes in cells undergoing mitosis are visible under the light microscope. Their formation requires some 104 - to 105 -fold condensation of uninterrupted linear duplex DNA which has a 2-nm diameter. Such compaction is accomplished in a highly complex and stepwise fashion. Because DNA is a polyelectrolyte with two negative charges per nucleotide, charge neutralization and shielding is required before the polymer can be folded in an ordered, condensed structure. In addition to metal ions and polyamines, the major source of the positive charge in chromatin is the family of highly basic small proteins, called histones, which are rich in the basic amino acid residues lysine and arginine needed to neutralize the charge of the phosphate backbone of DNA. The prokaryotes also have basic proteins (such as HU protein in E. coli) which induce DNA condensation. However, chromatin compaction in eukaryotes is carried out in stages. The simplest folded unit of DNA is the 10-nm nucleosome, consisting of a core histone octamer containing two molecules each of histone H2A, H2B, H3, and H4 around which nearly two turns of the DNA is wrapped. The nucleosome cores are connected by a stretch of linear DNA (linker) of variable length which is covered by histone H1

860 may affect survival and predisposition to specific diseases in the long term. G. Repetitive Sequences: Selfish DNA Even before the precise genome sequences are elucidated, one unique feature of the metazoan DNA sequence has been established from a number of studies. A large fraction (perhaps up to 90% or more) of the total genomic sequence in metazoan cells do not encode any information. Some of these sequences are present as noncoding intervening regions in genes, named “introns,” which do not code for proteins. However, the intron sequences are transcribed but are removed during processing (“splicing”) to generate mature mRNA, as discussed later. Many of the other genomic sequences are not even transcribed, and these may often be present as multimeric repeats of shorter units. These repetitive sequences have no known function in the cell, yet are maintained and replicated as an integrated part of the genome; such DNA is referred to as “selfish DNA.” Metaphase chromosomes are organized in substructures distinguished by their staining with dyes. Euchromatin regions contain transcribed sequences, while heterochromatin regions contain large segments of repetitive sequences. Metaphase chromosomes are also characterized by specific stained sequences (named centromeres) in the middle of the elongated structure, in addition to telomeres at the termini, as discussed earlier. Both centromeres and telomeres have unique repetitive sequences, and in some cases similar sequences have been observed in other regions of chromosomes; these regions are highly condensed and not transcribed. H. Chromatin Remodeling and Histone Acetylation In order to make the DNA template available for both replication and transcription, the chromatin is “remodeled.” One way to accomplish this reversible process is by altering the electrostatic interaction with histone. Acetylation of lysine residues (and to some extent phosphorylation of serine and threonine residues) reduces the binding affinity of histones with DNA in nucleosome cores and may thus allow exposure of free DNA to the transcriptional machinery. Additionally, a more complex energy-driven process involving the proteins SNF1 and SWI causes a major alteration of the chromatin structure, which is necessary for reprogramming of the transcriptional regimen during growth, development, and associated differentiation. DNA replication also requires access of DNA in free form to the replication machinery and, therefore, may also be dependent on the same remodeling process and could even

Nucleic Acid Synthesis

require dissociation and reassociation of the nucleosome core.

II. NUCLEIC ACID SYNTHESES
A. Similarity of DNA and RNA Synthesis All nucleic acids are usually synthesized by DNA template-guided polymerization of nucleotides—ribonucleotides for RNA and deoxy(ribo)nucleotides for DNA. The reactant monomers are 5 ribonucleoside (or deoxyribonucleoside) triphosphates. These can be described in the following chemical equations: → DNA + ndNTP ← DNA + nPPi and → (DNA) + nNTP ← (DNA) + RNA + nPPi . Enzymatic polymerization is carried out by DNA and RNA polymerases, both of which carry out pyrophosphorolysis, i.e., cleavage of a high energy pyrophosphate bond coupled to esterification of 5 phosphate linked to the 3 -OH of the previous residue. The reaction is reversible, although it strongly favors synthesis. Degradation of nucleic acids is not due to reversal of the reaction, but rather a hydrolytic reaction catalyzed by nucleases, namely, RNases and DNases, which generate nucleotides or deoxynucleotides, respectively. Three distinct stages are involved in the biosynthesis of both DNA and RNA: initiation, chain elongation, and termination. Initiation denotes de novo synthesis of a nucleic acid polymer which is generally well regulated by complex processes, as described later. The key difference in initiation of a DNA vs RNA chain is that RNA polymerases can start a new chain, while all DNA polymerases require a “primer,” usually a short RNA or DNA sequence with a 3 -OH terminus, to which the first deoxynucleotide residue is added. Elongation denotes continuing polymerization of the monomeric nucleotides, and termination defines stoppage of nucleotide addition to the growing polymer chain. During synthesis the enzymes catalyzing the polymerization reaction are guided by nucleic acid templates that provide the complementary sequence for the incorporated nucleotides (Fig. 4). The basic catalytic enzyme in such reactions is called DNA or RNA polymerase. In cells the template for both DNA and RNA is genomic DNA. There are some exceptions to these general rules. Some DNA polymerases can synthesize homo- or heteropolymers of deoxynucleotides in vitro in the absence of a template; the substrate is restricted to one or two dNTPs. While it is unlikely that these homo- or heteropolymers, e.g.,

Nucleic Acid Synthesis

861

FIGURE 4 Replication of circular DNA of prokaryotes and viruses, plasmids, and mitochondria. The basic steps of replication are shown. (A) Rolling circle mode of replication for singlestranded circular DNA: single-stranded (ss) DNA is replicated to the replicative form (RF), which then acts as the template for progeny ssDNA synthesis via a rolling circle intermediate. (B) Circular duplex DNA can be replicated at the ori site by formation of a θ intermediate. Replication could be bidirectional (as shown here) or unidirectional. 5 → 3 chain growth dictates that DNA synthesis is continuous on one side of the ori and discontinuous on the other side for each strand; (+) and (−) strands are shown to distinguish the strand types. (C) Replication of a linear genome with multiple origins.

FIGURE 5 Replication of mammalian viral RNA genome. The basic steps of replication are shown for (A) a (+) strand genome, which acts as an mRNA for encoding viral proteins; (B) a (−) viral genome cannot encode protein and first has to be replicated by the RNA replicase (•) which is present in the virus particle. Once the complementary (+) strand which serves as mRNA is synthesized, viral-specific proteins are synthesized, including RNA replicase. (C) Replication of (+) stranded retroviral genomes first involves synthesis of the reverse transcriptase which directs synthesis of duplex DNA in two stages from the RNA template. Circularization of the DNA followed by its genomic integration allows synthesis of progeny viral RNA by the host transcription machinery.

(dA•dT)n or poly(dA)n •poly(dT)n , are formed in vivo, the availability of these polymers significantly advanced our understanding of the properties of DNA, before the age of chemical or enzymatic oligonucleotide synthesis. There are some exceptions to the norm of DNAdependent DNA or RNA synthesis, mostly in lower eukaryotes or viruses (Fig. 5). One example is RNAdependent RNA synthesis in plant, animal, or bacterial viruses. In these cases, a single-stranded RNA template rather than double-stranded DNA guides synthesis of the complementary RNA strand, based on conventional base pairing. The polarity of RNA adds a level of complexity during synthesis. Thus, the RNA genome of a virus that can be directly read and thus provides the mRNA function is called the positive strand, as in polio virus. In this case, the viral genome RNA functions as the mRNA and encodes the RNA polymerase, which is synthesized like other viral proteins in the infected cell. This RNA polymerase subsequently synthesizes the complementary

negative strand, which then serves as the template for synthesis of the progeny positive strand RNA. The progeny RNA is then packaged into mature progeny virus. In contrast, the genomic RNA of negative strand viruses (e.g., vesicular stomatitis virus) cannot function directly as mRNA and thus cannot guide synthesis of proteins, including the RNA replicase, by itself after the infection of host cells. These viruses carry their own RNA replicase within the virion capsids, which carry out (+) mRNA strand synthesis after infection (Fig. 5). Retroviruses comprise diverse groups of viruses, including human immunodeficiency virus (HIV), which share a common mechanism of genome replication. The RNA genomes of these viruses encode an RNAdependent DNA polymerase (reverse transcriptase or RT) which first generates the complementary (c) DNA of the viral genome. RT has also RNaseH (specific nuclease for degrading RNA from RNA–DNA hybrids) and DNAdependent DNA polymerase activities. After copying the RNA template, the enzyme degrades the RNA and is able to convert the resulting single-stranded cDNA to duplex

862 DNA. This is then integrated into the host cell genome as proviral DNA, from which the progeny viral RNA is eventually transcribed. Thus, the reverse transcriptase is an unusual polymerase because it can utilize both RNA and DNA templates (Fig. 5). There is strong evidence that such reverse transcription was involved in synthesis of “retrotransposons,” a special class of mobile genetic elements, during the evolution of mammalian genomes. These mobile genetic elements, also known as transposons, when identified in bacteria and lower eukaryotes, consist of specific DNA sequences which can be relocated randomly in the genome. The transposition is mediated by enzymes called transposase, usually synthesized by a gene in the transposon. During transposition of retransposons, certain mRNAs are reverse transcribed and then integrated into the genome like the proviral sequence. The presence of specific flanking sequences allows these elements to relocate to other sites in the genome. B. DNA Replication vs Transcription: Enzymatic Processes The broad chemical steps in DNA and RNA synthesis are quite similar, in that both processes represent reading of a DNA strand as the template. However, while both strands of DNA have to be copied, transcription is polar because only one strand is normally copied into RNA whose sequence is identical to the other strand (except for replacement of thymidine by uridine). This is achieved by the presence of discrete start and stop signals bracketing “transcription units” corresponding to each gene containing unique sequences, called promoters; their sequences provide the recognition motif for RNA polymerase to bind and start RNA synthesis unidirectionally. Similarly, the stop sequences are recognition motifs for the transcription machinery to stop and fall off the DNA template. As mentioned before, the two strands of a DNA double helix are of opposite polarity, i.e., one strand is in the 5 → 3 orientation and its complementary strand in the 3 → 5 orientation. Furthermore, the fact that all nucleic acid polymerases can polymerize nucleotide monomers only in the 5 → 3 direction as guided by base pairing with a template does not pose a problem for RNA synthesis because only the 3 → 5 strand of the DNA template is copied. However, DNA replication, where both strands have to be copied in the same 5 → 3 direction of the duplex template, introduces a complication situation (Figs. 2 and 5). The 3 → 5 strand is copied like RNA, while the 5 → 3 strand has to be copied in the opposite direction. It has been observed in all cases that simultaneous replication of both strands is accomplished by continuous copying of the 3 → 5 strand, also called the

Nucleic Acid Synthesis

leading strand, while the 5 → 3 strand is copied after a brief delay when separation of the strands occur, so this nascent strand is called the lagging strand (Fig. 2). The leading strand can be synthesized continuously without interruption, while the lagging strand is synthesized discontinuously after the leading strand is synthesized. The discontinuous fragments are also called Okazaki fragments, named after its discoverer. C. Multiplicity of DNA and RNA Polymerases Multiple DNA and RNA polymerases are present in both eukaryotes and prokaryotes, which evolved to fulfill distinct roles in the cell. In E. coli, DNA polymerases I (Pol I), II (Pol II), and III (Pol III) account for most DNA polymerase activity. Pol I has the highest enzymatic activity and was the first DNA polymerase to be discovered by A Kornberg. However, Pol III is responsible for cellular DNA replication, while Pol I is involved in gap filling necessary during normal DNA replication (to fill in the space of degraded RNA primers) and also during repair of DNA damage. Pol II and two other DNA polymerases, Din B and UmuD/C, are responsible for replication of damaged DNA when it remains unrepaired. Eukaryotic cells express five different DNA polymerases, α, β, γ , δ, and ε, for normal DNA replication and repair. Pol α is involved in synthesis of primers for DNA replication; Pol β and possibly Pol ε are involved in repair replication of damaged DNA. Pol δ (and possibly Pol ε) are responsible for replication of the nuclear genome. Pol γ found in the mitochondria is responsible for replication of the mitochondrial genome. Several additional DNA polymerases recently identified and characterized are involved in replication of unrepaired damaged bases, like the E. coli DinB and UmuD/C (Table II). E. coli has only one RNA polymerase, while eukaryotes have three distinct RNA polymerases, Pol I, Pol II, and Pol III, which transcribe different types of genes. RNA Pol I makes only ribosomal RNAs, which constitute the largest fraction of total RNA and, in fact, a significant fraction of the cellular mass. Pol III transcribes small RNAs, including transfer RNAs, which function as carriers of cognate amino acids and are required for protein synthesis. RNA Pol II transcribes all genes to generate mRNA, which encodes all proteins. Thus, this enzyme recognizes the most diverse group of genes. All of these RNA classes are initially synthesized as longer precursors that require extensive, often regulated, processing to yield the mature RNA product. RNA and DNA polymerases encoded by virus and other episomal genomes are, in general, smaller and have fewer subunits than the cellular polymerases. Cellular polymerase holoenzymes are rather complex with multiple

Nucleic Acid Synthesis
TABLE II Cellular DNA Polymerases Prokaryote (E. coli) Pol I Pol II Pol III Din B UmuC Eukaryote Pol α Pol β Pol δ Pol ε Pol ζ Pol η Pol θ Pol ι Pol γ In vivo function Nonreplicative removal of 5 primer of Okazaki fragments Nonreplicative, damage responsive polymerase Replicative synthesis Lesion bypass DNA synthesis Lesion bypass DNA synthesis RNA primer synthesis Repair synthesis Replicative (repair) synthesis Replicative (repair) synthesis Damage bypass synthesis Damage bypass synthesis Damage bypass synthesis Damage bypass synthesis Mitochondrial DNA synthesis

863 form or RF) is the template for rolling circle synthesis and is formed first by replication of the single-stranded form. Such a single-stranded circular DNA template has been exploited in recombinant DNA techniques. Small organisms (e.g., bacteria), as well as plasmids and many viruses, have only one ori sequence per cellular genome (4.7 × 106 nucleotide pairs in E. coli), which is often an uninterrupted DNA molecule (Figs. 4A and 4B). In complex organisms, with a much larger genome size (∼3 × 109 nucleotide pairs for mammals), which is divided into multiple discrete chromosomes, thousands of ori sequence are present (Fig. 4C), although not all of them may be active in all cells; this requires that replication be regulated and coordinated. B. Regulation of DNA Replication Semi-conservative replication of the genome ensures that each daughter cell receives a full complement of the genome prior to cell division. In eukaryotes, this is achieved by the distinct phases of the cell cycle, namely, G1 phase, during which cells prepare for DNA synthesis; S phase, in which DNA replication is carried out; and G2M (mitosis), during which the replicated chromosomes segregate into the two newly divided daughter cells. Unlike in eukaryotes, DNA replication in prokaryotes may occur continuously during growth (in rich medium). Thus, the copy number of genomes could exceed two in rapidly growing cells. In the case of viruses, which multiply by utilizing the host cell synthetic machinery and eventually killing them, genome replication may be not controlled. However, plasmid DNA, as well as the genomes of organelles such as mitochondria and chloroplasts, is replicated with some degree of regulation. In these cases the genomic copy number can vary within limits as a function of growth condition. C. Regulation of Bacterial DNA Replication at the Level of Initiation In all organisms, as well as autonomously replicating DNA molecules of organelles and plasmids, replication is divided into three stages: initiation, chain elongation, and termination. The control of replication occurs primarily at the level of initiation of DNA synthesis at the “origin” (ori site). Because DNA chains cannot be started de novo and requires a primer, the initiation complex contains primase activity for synthesis of an RNA primer. Discontinuous synthesis of Okazaki fragments needs repeated primer synthesis for each fragment as an integral component of chain elongation. Initiation of the primer at the ori sequence rather than elongation of initiated chains is the critical event in DNA replication control.

subunits which may have distinct functions. These will be discussed later.

III. DNA REPLICATION AND ITS REGULATION
A. DNA Replication DNA replication is initiated at discrete sequences called origin (ori) of replication to which DNA polymerase and accessory proteins bind and copy both strands, as predicted by the semi-conservative replication model (Fig. 2B). In contrast to unidirectional RNA synthesis, DNA replication in most genomes occurs bidirectionally (Fig. 2B). This results in both continuous and discontinuous synthesis of the same strand on two sides of the origin of replication. Some circular genomes, such as mitochondrial DNA, are replicated unidirectionally. In these cases, replication starting at the ori proceeds continuously in the 5 → 3 direction, followed by discontinuous synthesis of the complementary strand. Termination occurs at the same site as the ori after the circle is completely traversed. During replication of the mitochondrial genome, elongation of the continuous strand pauses at some distance from the ori, resulting in a bubble (θ structure) structure named a D-(displacement) loop (Fig. 4A). The single-stranded DNA genomes of certain small E. coli viruses (such as M13 and φX174) are replicated in the form of rolling circles in which unidirectional synthesis of one (virus genome) strand occurs by continuous displacement from the template (complementary strand; Fig. 4A). The initial duplex DNA (called the replicative

864 Different replicons of prokaryotes and eukaryotes utilize distinct mechanisms which vary in complexity, depending on the complexity of the organisms. A common feature of replication initiation control in E. coli genomes and plasmids is the presence of repeats of A•T rich sequences which facilitate unwinding of DNA and one or multiple repeats of a “dnaA box” to which the initiator DnaA protein in E. coli or its functional homolog (called Rep in other cases) binds to allow helical unwinding and primer synthesis. The level of DnaA protein regulates the initiation frequency and, in turn, is controlled at the level of transcription of the dnaA gene. Thus, there are complex negative autofeedback loops to control dnaA gene expression. DnaA regulates its own gene, and its steadystate level in the cell is determined by the cellular growth state. The frequency of replicon firing is dependent on the growth rate of the bacteria. As mentioned before, rapidly growing cells can have multiple copies of the genome, while cells with a very low growth rate have only one copy. Furthermore, as expected in cells with multiple genome copies, the genes near the origin will have a higher average copy number than the genes located near the terminus of replication and, therefore, will be more transcriptionally active. In the case of multicopy plasmids, the control of copy number is mediated by the synthesis of anti-sense RNA of the replication initiator protein Rep, which is copied from the nontranscribed DNA strand and is thus complementary to the normal RNA. Anti-sense RNA prevents synthesis of the Rep protein, which is required for initiation of DNA synthesis and whose concentration is the primary mechanism of controlling initiation frequency. Rep proteins encoded by plasmids bind to additional copies of binding sites called “iterons,” often present upstream of the ori sequences in the plasmids. D. DNA Chain Elongation and Termination in Prokaryotes Once initiated, DNA replication proceeds by coordinated copying of both leading and lagging strands. Although both bacteria and eukaryotes have multiple DNA polymerases, only one, named polymerase III (Pol III), is primarily responsible for replicative DNA synthesis in E. coli. In eukaryotes, DNA polymerases δ and ε have both been implicated in this process along with a suggestion that each of these two enzymes may be specific for leading or lagging strand synthesis. Replication involves separation of two DNA strands which are catalyzed by DNA helicases which hydrolyze ATP during this reaction. ATP hydrolysis provides the energy needed for the unwinding process. All cells have multiple DNA helicases for a variety of DNA transactions.

Nucleic Acid Synthesis

DnaB is the key helicase for replication of the genome E. coli. However, other helicases such as Rep and PriA are also involved in replication and interact with other components of the replication complex called the replisome. Replication requires a large number of proteins, including the holoenzyme of Pol III which includes, in addition to the catalytic polymerase cores, ten or more pairs of other subunits. The polymerase complex appears to have a dimeric asymmetric structure in order to replicate simultaneously two strands with opposite polarity. The continuous leading strand synthesis should be processive without interruption, because periodic RNA primer synthesis is not necessary once the leading DNA strand synthesis is initiated. On the other hand, the discontinuous lagging strand synthesis should not be processive, because repeated synthesis of RNA primers is required to initiate synthesis of each Okazaki fragment. The Pol III holoenzyme appears to assemble in a stepwise fashion, with its key β-subunit dimer acting as a sliding clamp based on its X-ray crystallographic structure of a ring surrounding the DNA. This clamp is loaded on DNA by the γ -complex, accompanied by ATP hydrolysis. The dimeric structure of the replication complex is maintained by the dimeric subunit of the holoenzyme. The β-clamp slides on the duplex DNA template and thus promotes processivity. Proliferating cell nuclear antigen (PCNA) is the sliding clamp homolog in eukaryotic cells and is also used in SV40 replication. Much of the information about the composition of the E. coli Pol III holoenzyme, and DNA chain elongation, was generated from studies of the replication of small, single-stranded circular DNAs of bacterial viruses φX174 and M13 and also of laboratory-constructed plasmid DNA containing the ori (ori C) of E. coli. Asymmetric dimeric replication complexes have also been identified for larger E. coli viruses such as T4 with a linear genome and for the mammalian SV40 virus with a double-strand circular genome. In circular genomes, DNA synthesis is terminated at around 180◦ from the origin. In the case of linear genomes, termination occurs halfway between two neighboring replicons. The mechanism of termination is not completely understood. Although, in the E. coli genome, specific termination (ter) sequences are present, which bind to terminator proteins, such proteins act as anti-helicases to prevent strand separation. However, the termination may not be precise and occurs when the replicating forks collide. E. General Features of Eukaryotic DNA Replication Unlike the genomes in bacteria and plasmids (as well as in mitochondria and chloroplasts) which consist of a circular duplex DNA, with a single ori sequence, the genomes of

Nucleic Acid Synthesis

865 ulator proteins, such as geminin, which blocks licensing, is also regulated by some cell cycle-dependent feedback mechanisms. G. Fidelity of DNA Replication The maintenance of genomic integrity in the form of the organism-specific nucleotide sequence of the genome is essential for preservation of the species during propagation. This requires an extremely high fidelity of DNA replication. Errors in RNA synthesis may be tolerated at a significantly higher level because RNAs have a limited half-life, even in nondividing cells, and are redundant. In contrast, any error in DNA sequence is perpetuated in the future, as there is only one or two copies of the genome per cell under most circumstances. Obviously, all organisms have a finite rate of mutation, which may be necessary for evolution. Genetic errors are one likely cause of such mutations. Inactivation of a vital protein function by mutation of its coding sequence will cause cell death. However, mutations that affect nonessential functions could be tolerated. Some of these mutations can still lead to change in the phenotype, which in extreme cases can cause pathological effects. In other cases, these may be responsible for susceptibility to diseases. In many cases, however, such mutations appear to be innocuous and are defined as an allelic polymorphism. The mammalian genome appears to have polymorphism in one out of several hundred base pairs. Such mutations obviously arose during the evolution and subsequent species propagation. The error rate in replication of mammalian genome is about 10−6 to 10−7 per incorporated deoxynucleotide. The catalytic units of the replication machinery, namely, DNA polymerases, have a significantly higher error rate of the order of 10−4 to 10−5 per deoxynucleotide. In fact, some DNA polymerases, notably the reverse transcriptases of retroviruses, including HIV, the etiologic agent for AIDS, are highly error prone and incorporate a wrong nucleotide for every 102 –103 nucleotides. These mistakes result in a high frequency of mutation in the viral protein, which helps the virus escape from immunosurveillance. The overall fidelity of DNA replication is significantly enhanced by several additional means. The editing or proof-reading function of the replication machinery is a 3 → 5 exonuclease (which is either an intrinsic activity of the core DNA polymerase or is present in another subunit protein of the replication complex) which tests for base pair mismatch during DNA replication and removes the misincorporated base. Such an editing function is also present during RNA synthesis. In addition, after replication is completed, the nascent duplex is scanned for the presence of mispaired bases. Once such mispairs are marked by mismatch recognition proteins, a complex

eukaryotes are not only much larger and linear, but also contain multiple ori sequences for DNA replication and thus multiple replicons. Thousands of replicons are simultaneously fired in mammalian genomes, as is needed to complete replication of the genome in a few hours. Mammalian genomes are three orders of magnitudes larger than the E. coli genome for which one round of replication requires about 40 min at 37◦ C. Replication of a mammalian genome, initiated at a single ori, would thus take more than 1 week with the same rate of synthesis. In fact, it would be even longer because the rate of DNA chain elongation is slower in eukaryotes than in E. coli, possibly because of the increased complexity of eukaryotic chromatin. As mentioned earlier, DNA replication in eukaryotes occurs only during the S phase, which can last for several hours but whose duration varies with the organism, the cell type, and also the developmental stage. For example, in a rapidly growing early embryo of the fruitfly D. melanogaster, cellular multiplication with duplication of the complete genome occurs in less than 15 min. The details of temporal regulation of firing of different replicons are not known. However, euchromatin regions are replicated earlier than the heterochromatin regions. The details of initiation of replication at individual replicons have not been elucidated in eukaryotes. Some ori sequences of the yeast genome, known as autonomous replication sequences (ARS), have been determined. Although such sequences in the mammalian genomes have not been isolated, the ori regions of certain genes which could be selectively amplified have been localized by twodimensional electrophoretic separation. Nevertheless, a significant amount of information has been gathered regarding regulation of DNA replication at the global level. F. Licensing of Eukaryotic Genome Replication Unlike in bacteria and plasmids, DNA replication in eukaryotic cells is extremely precise, and replication initiation occurs only once in each cell cycle to ensure genomic stability. “Licensing” is the process of making the chromatin competent for DNA replication in which a collection of proteins called origin recognition complex (ORC) bind to the ori sequences. This binding is necessary for other proteins required for the onset of the S phase to bind to DNA. ORC is present throughout the cell cycle. However, other proteins required for replication initiation and chain elongation are loaded in a stepwise fashion. The onset of the S phase may be controlled by a minichromosome maintenance (MCM) complex of proteins which licenses DNA for replication, presumably by making it accessible to the DNA synthesis machinery. Several protein factors are involved in the loading process, which is regulated both positively and negatively. The level of reg-

866 mismatch repair process is initiated, which causes removal of a stretch of the newly synthesized strand spanning the mismatch, followed by resynthesis of the segment, as described later. H. Replication of Telomeres—The End Game Because DNA synthesis proceeds unidirectionally from 5 → 3 with respect to deoxyribose, by sequential addition of deoxynucleotides to the 3 terminus of the deoxynucleotide added last, chain elongation can proceed to the terminus of the template strand oriented in the 3 to 5 direction. But how about synthesis of the terminus of the complementary strand ? Because synthesis of this discontinuous (lagging) strand occurs in the opposite direction by repeated synthesis of a primer, the terminus could not be replicated. This problem of end replication is eliminated in the circular genomes of bacteria and the small genomes of plasmids and viruses. However, in the case of linear eukaryotic chromosome, the problem is solved by a specialized mechanism of telomere replication. Telomeres are repeats of short G-rich sequences found at both ends of the chromosomes (Fig. 6). In the human genome, the telomere repeat unit is 5 (T/A)m Gn 3 , where n > 1 and 1 < m < 4. Telomerase is a special DNA polymerase (reverse transcriptase) containing an oligoribonucleotide template 5 Cn(A/T)m3 (which is complementary to the telomere repeat sequence) as an integral part of the enzyme (Fig. 6). In the presence of other accessory proteins, telomerase utilizes its own template to generate the telomeric repeat unit and, by “slippage,” utilizes the same oligoribonucleotide template repeatedly to generate thousands of repeats of the same hexanucleotide unit sequence. Because the lagging strand terminal region does not require an external DNA template, the newly synthesized DNA is present in an extended single-stranded region. Telomeres provide a critical protective function to the chromosome by their unique structures and prevent their abnormal fusion. I. Telomere Shortening: Linkage Between Telomere Length and Limited Life Span One profound implication of the specialized telomere structure and its synthesis is that in the absence of telomerase, the repeat length of telomeres could not be maintained. Telomerase is active in neonatal cells and also in some immortal tumor cells, but is barely detectable in diploid, terminally differentiated mammalian cells. Most such diploid cells can multiply in vitro in specialized culture medium, but have a limited life span. Loss of replicative capacity is associated with shortening of telomere repeat lengths. Furthermore, ectopic and stable expression of telomerase in human diploid cells by introduction of its gene confer an indefinite reproductive life on such cells. It

Nucleic Acid Synthesis

FIGURE 6 A schematic description of the role of telomerase in the maintenance of telomeres at chromosome termini. The double lines with break represent one telomere terminus of a chromosome in which the 5 terminal region of the lagging strand is unreplicated (as in Fig. 4), resulting in an overhanging 3 terminal region. In order to avoid shortening of this telomere sequence during successive rounds of replication, DNA template-independent telomerase extends the 3 overhang by adding the telomere repeat sequence TTGGGG as shown in (C). The template for the repeat is an RNA present in the telomerase complex. The extended 3 single-strand region then allows de novo initiation and filling in of the 5 strand (E). Finally, the 3 overhang loops to anneal with an internal sequence mediated by the telomere repeat factor (TRF2) in order to protect the terminus from degradation by nonspecific nucleases (F).

is generally believed that cells will senesce if the telomere length is reduced below a critical level after repeated replication of the genome.

IV. MAINTENANCE OF GENOME INTEGRITY
The integrity of the genome, both in regard to sequence and to size, is essential for perpetuation of species. This integrity can be threatened in two ways. The first is by errors in DNA replication, as discussed earlier. A second inexorable process of DNA alteration occurs due to chemical reactions which can be either endogenous or induced by

Nucleic Acid Synthesis

867 tion of critical oncogenes and/or tumor suppressor genes, thus causing uncontrolled cellular multiplication and prevention of cell death. Prevention of transcription of bulky adducts in active genes triggers nucleotide excision repair, at least in eukaryotes, in a process called “transcriptioncoupled repair.” In fact, the repair complex has co-opted certain proteins of the transcription complex. Although excision repair requires DNA synthesis, it is distinct from normal semi-conservative replication because it occurs throughout the cell cycle and may utilize nonreplicative DNA polymerases in both prokaryotes and eukaryotes. Pol II and Pol I in E.coli and DNA polymerase β have been identified as such repair polymerases. However, replicative polymerases can also be recruited in some cases, e.g., for mismatch repair synthesis. Interestingly, during the last couple of years, a whole family of DNA polymerase have been identified and characterized in E. coli, yeast, and mammals (Table II). These enzymes are unique in their ability to bypass DNA base adducts which have lost the ability to base pair and thus are not utilized by standard DNA polymerases. It has been suggested that these replication bypass polymerases allow cell survival by allowing DNA replication even at the cost of introducing mutations. B. Post-Replication Recombinational Repair In contrast to the excision repair process in which the DNA damage is actually removed, both eukaryotic and prokaryotic cells have a novel mechanism of adapting to persistent, unrepaired damage by utilizing homologous recombination between the replicated progeny genomes. Recombination, the process of exchange between homologous DNA segments, involves unwinding of one duplex DNA and reciprocal strand exchange. When one strand in the parental DNA has a persistent lesion that prevents replication, a complete duplex is generated from the other, undamaged strand. The new strand subsequently acts as the template for the damaged region by strand exchange during replication of the damaged strand. Thus, recombination allows synthesis of the correct DNA sequence opposite the lesion.

external agents, including environmental genotoxic compounds, drugs, and radiation. Contrary to an earlier belief that DNA is a rather inert chemical, it is, in fact, sensitive to certain chemical reactions, e.g., depurination (loss of purine bases) and deamination of C to U, which occurs at a low but significant rate in DNA. It has been estimated that several hundred to several thousand such lesions are generated in the genome of a human cell per day. Both of these changes could be mutagenic. Loss of purines leads to abasic sites in DNA, which could direct misincorporation of wrong bases during DNA replication. Conversion of C into U is definitely mutagenic, because the change of a G•C to a G•U pair will give rise to one G•C pair and one A•T pair after DNA replication because U, like T, pairs with A. Often, C in the mammalian genome is methylated at the C-5 position, as discussed elsewhere, and 5-methyl C is deaminated more readily than C. Its conversion to T induces the same G•C → A•T mutation and, unlike deamination of C → U, does not produce an “abnormal” base in the DNA. A variety of environmental chemicals and both ultraviolet light present in sunlight and ionizing radiation from radioactive sources and X-rays induce a plethora of DNA lesions which include both base damage and sugar damage and are accompanied by DNA strand breaks. Many of these lesions, in particular, strand breaks and bulky base adducts, are toxic to the cells by preventing both replication and transcription. Other types of base damage and adducts can be mutagenic because they will allow DNA replication to proceed, but will direct incorporation of improper bases in the progeny strand. A. Prevention of Toxic and Mutagenic Effects of DNA Damage by Repair Processes Multiple repair processes have evolved to restore genomic integrity in all organisms ranging from bacteria to mammals. Excision repair comprises one class in which the damaged part of a DNA strand is excised enzymatically from the duplex DNA, leaving a single-strand gap. The gap is then filled by DNA polymerases starting at the 3 -OH terminus by utilizing the undamaged complementary strand as the template, followed by ligation of the nascent segment to the 5 phosphate terminus at the other end of the gap with DNA ligase. The excision repair process consists of three subgroups which are utilized for distinct types of damage, although there is some overlap in their activities. Base excision repair is more commonly used for small base adducts, and nucleotide excision repair is used for replication/transcription-blocking bulky adducts. Mismatch repair evolved primarily to remove DNA mispairs that are generated as errors of replication. Both nucleotide excision and mismatch repair deficiencies have been linked to tumorigenesis, which results from muta-

V. DNA MANIPULATIONS AND THEIR APPLICATIONS
A. Episomal DNA and Recombinant DNA Technology Extrachromosomal or episomal DNA, present in prokaryotes and lower eukaryotes, is distinct from the genome of organelles such as mitochondria or chloroplasts and serves many purposes. In bacteria, plasmid DNA can be

868 transmitted to progeny cells, and the genes in these plasmids encode distinct proteins which provide growth advantage or survival to the host bacteria. For example, many proteins which confer drug resistance by a variety of mechanisms are encoded by the plasmids, which are invariably present as double-stranded circular DNA containing several to hundreds of kilobase pairs. The plasmid DNAs are self-replicating genomic units which are completely dependent on the host bacteria or yeast for their replication. These are also critical vehicles for recombinant DNA technology based on cutting and rejoining DNA fragments. Its invention, some three decades ago, revolutionized molecular biology and is at the root of nearly all modern breakthroughs in biology. Restriction endonucleases, which are enzymes characterized by stringent recognition of specific DNA sequences, cleave DNA duplexes and often leave identical terminal sequences in both plasmid DNA and a gene or segment of a genome. The fragments can then be joined by a DNA ligase. Joining heterologous fragments generates recombinant DNA, for example, a circular plasmid molecule containing foreign genes. These DNA molecules can then be introduced into living cells which allow their reproduction, so that a large amount of recombinant plasmid can then be generated. Recombinant plasmids specific for bacteria, yeast, and even mammalian cells have been generated in the laboratory and exploited for a variety of basic and applied research applications. Specifically, recombinant expression plasmids can be constructed in order to express the ectopic protein encoded by the foreign (trans) gene in the appropriate host cell. Recombinant plasmids of mammalian cells are based on viruses, rather than on episomal DNA. Only the DNA replication function of the virus is incorporated into the plasmid, so that the plasmid is replicated without producing the active virus. In the case of human cells, simian virus 40 (SV40) is commonly used to generate recombinant DNA. The circularity of the plasmid is essential for E. coli, but not mammalian or yeast cells. This may be consistent with the circular genome of the bacteria vs linear genomes of eukaryotes. However, plasmid vectors specific for mammalian cells must be propagated, preferably in E. coli. Such “shuttle” vectors are therefore required to have a circular configuration. B. Polymerase Chain Reaction (PCR) A critical advance in molecular biology came with the invention of PCR, based on a remarkably simple principle, and revolutionized many important aspects of biomedical research and medical jurisprudence. The method is based on the rationale that each strand of a piece of DNA se-

Nucleic Acid Synthesis

FIGURE 7 Principle of polymerase chain reaction (PCR). A copy of a relatively short fragment of DNA (0.1–20 kilobase pairs) can be specifically amplified from genomic DNA by PCR. A typical PCR reaction mixture contains genomic DNA; two oligonucleotide (∼ 20 bp) primers, which have same sequences as the two ends of the DNA fragment to be amplified; and a thermostable DNA polymerase. A cycle of PCR reaction consists of three steps, starting with denaturing the genomic DNA at high temperature (e.g., 95◦ C), followed by primer annealing at near Tm (melting temperature for primer-DNA hybridization), followed by DNA synthesis from the primers by the DNA polymerase. Theoretically, the copy number of the DNA of interest (N) can be amplified to 2C × NO , where NO is the original copy number and C is the number of PCR cycles.

quence can be replicated repeatedly by using an oligonucleotide primer and a DNA polymerase (Fig. 7). After a duplex DNA molecule is generated, the next cycle is carried out by separating the two strands by heating and then starting the next cycle of synthesis after annealing oligonucleotide primers to each template strand. Thus, the repeated cycles of synthesis, denaturation, and primer annealing to both strands allow synthesis of a specific DNA sequence at an exponential rate. Thus, a tiny piece of a DNA molecule could be amplified about a millionfold after 20 cycles of this chain reaction (assuming 100% efficiency of the process; Fig. 7). The PCR technology became viable after discovery of thermostable DNA polymerases derived from bacteria, such as Thermobacillus aqualyticus (Taq), which grow at high temperature. The cycles of PCR could then be automatically set in a thermal cycler. PCR does have some limitations. The most important of these are: (1) errors in DNA replication; (2) less than complete efficiency in each step of the reaction; and (3) improper primer annealing when complex DNA is used. Thus, when amplification of a segment of DNA in a complex genome is desired, the

Nucleic Acid Synthesis

869 RNA polymerases of all organisms are complex machines consisting of multiple subunits which alter conformation. A variety of structural analyses show the presence of a 2.5-nm-wide “channel” on the surface of all DNA polymerases which could be the path for DNA. The RNA polymerase holoenzyme binds to a promoterspecific recognition sequence upstream (5 side of the transcribed strand) of the site of synthesis initiation. While the RNA polymerase is normally present as a closed complex with nonspecific DNA, in which DNA base pairs are not broken, a significant conformational change produces the open complex when RNA the enzyme binds the promoter, unwinds the DNA duplex, and is poised to initiate RNA synthesis. As in the replication process, initiation is the first stage in transcription and denotes the formation of first phosphodiester bond. Unlike in the case of DNA synthesis, RNA chains are initiated de novo without the need of a primer. However, when a primer oligonucleotide is present, RNA polymerases can also extend the primer as dictated by the template strand. A purine nucleotide invariably starts the RNA chains in both prokaryotes and eukaryotes, and the overall rate of chain growth is about 40 nucleotides per second at 37◦ C in E. coli. This rate is much slower than that for DNA chain elongation (∼800 base pairs per second at 37◦ for the E. coli genome). RNA synthesis is not monotonic, and RNA polymerases can move backward like DNA polymerases do for their editing function in which an incorrectly inserted deoxynucleotide is removed by 3 exonuclease activity. RNA polymerases stall, back track, and then cleave off multiple newly inserted nucleotides at the 3 terminus. Subsequently, polymerases move forward along the DNA template and resynthesize the cleaved region. Based on the segment of DNA covered by an RNA polymerase as analyzed by DNA footprinting, it has been proposed that the enzyme alternatively compresses and extends in its binding to the DNA template and acts like an inchworm in its transit. RNA polymerases of both prokaryotes and eukaryotes function as complexes consisting of a number of subunits. The E. coli RNA polymerase enzyme with a total molecular mass of about 465 kD contains two α-subunits, one βand one β -subunit each, and a σ -subunit which provides promoter specificity. During chain elongation, a ternary complex of macromolecules among DNA template, RNA polymerase, and nascent RNA is maintained in which most of the nascent RNA molecule is present in a singlestranded unpaired form. The stability of the complex is maintained by about nine base pairs between RNA and the transcribed (noncoding) DNA strand at the growing point. While DNA replication warrants permanent unwinding of the parental duplex DNA, asymmetric copying of only

first requirement is the sequence information for the termini of the segment, based on which the oligonucleotides will be designed for each terminus and then synthesized. However, errors of replication cannot be completely eliminated. Any error in DNA synthesis that occurs early will be perpetuated. Furthermore, if replication is initiated by primers annealed to an incorrect DNA sequence, the wrong PCR product will be generated. Primarily, because it has both sensitivity and specificity, PCR technology has revolutionized many aspects of biomedical research. Several modifications of the basic methodology have provided additional powerful tools. For example, a trace amount of RNA can be quantitated by reverse transcriptase PCR (RTPCR), where a reverse transcriptase synthesizes the complementary DNA strand of the RNA, which then serves as the template for regular PCR. DNA in a very small amount of biological samples can be amplified by PCR. This technique has been exploited in criminal investigations to identify suspects by “fingerprinting” their DNA, which involves determining a characteristic pattern of repeat sequences in the genome after PCR amplification of the total DNA. PCR has also been utilized in the identification of pathogens and other microorganisms, based on certain unique sequences of each organism. PCR has been exploited for a variety of in vitro manipulations of DNA sequences in plasmids, viruses, and synthetic DNA by generating site-specific mutations and a variety of recombinant DNA plasmids.

VI. TRANSCRIPTIONAL PROCESSES
Transcription is a highly complex process because of its defined initiation and termination sites in the genome and the subsequent processing and regulation of its synthesis. The steady-state level of a protein in the cell is the balance of its rate of synthesis and degradation. The synthesis is determined primarily by the steady-state level of its mRNA. Thus, the rate of transcription often determines the level of its gene product in vivo. As mentioned earlier, RNA synthesis is catalyzed by the RNA polymerase in all organisms. Prokaryotes express a single RNA polymerase used for synthesis of all RNAs, while eukaryotes encode multiple RNA polymerases with dedicated functions. RNA polymerase I (Pol I) in eukaryotic cells is responsible for synthesis of ribosomal RNA, which accounts for more than 70% of total RNA in the cell. Pol III catalyzes synthesis of small RNA molecules, including transfer RNAs which bring in appropriate amino acids to the ribosome for protein synthesis by using their “anti-codon” triplet bases. Pol II is responsible for synthesis of all other RNA, specifically mRNA.

870 one strand by RNA polymerase requires localized strand separation which is induced by the polymerase itself, resulting in a transcription bubble. During chain elongation, this bubble moves along the DNA duplex. Initiation of RNA synthesis is enhanced in an in vitro reaction with supercoiled duplex circular DNA template in which base pairs are destabilized due to torsional stress. Unwinding of the helix at the transcription site causes overwinding (positive supercoiling) of the template DNA ahead of the transcription bubble and underwinding (negative supercoiling) behind the bubble. A. Recognition of Prokaryotic Promoters and Role of σ-Factors In prokaryotic RNA polymerases, the σ -factor is required for promoter recognition and binding. It is loosely bound to the core complex and released after the nascent RNA chain becomes 8–9 nucleotides long. The core polymerase with σ -factor has a high affinity for nonspecific DNA. The σ -factor alters the conformation of the holoenzyme so that its affinity for nonspecific DNA is reduced and the specific binding affinity for the promoter is significantly enhanced. More than one type of σ -factor is present in E. coli, and more such factors are present in other bacteria. These different factors may have specialized functions in altered growth conditions, cause a global change in transcriptional initiation due to their recognition of distinct −35 and −10 sequence elements, and have a preference for different promoters. RNA chain termination in bacteria occurs by two mechanisms, one with assistance of a protein factor rho (ρ) and the other without need of a protein. In both cases, termination occurs at a specific terminator sequence in the gene, at which the RNA polymerase stops adding nucleotides to the growing RNA chain, which is then released from the template. The terminator sequence often has a “hairpin” structure which results from intramolecular base pairing in a palindromic sequence. It is likely that such hairpins at the end of RNA promote its dissociation from DNA. Termination can be prevented by an anti-terminator protein, which allows the polymerase to ignore the terminator signal. A unique distinction between prokaryotic and eukaryotic RNA synthesis is the temporal relationship between its synthesis and utilization in information transfer. Prokaryotic transcription of mRNA is linked to its reading on the ribosome for protein synthesis. Thus, even before transcription is terminated, the 5 terminal region of the nascent mRNA is complexed with a ribosome for initiation and propagation of protein synthesis. In the case of eukaryotes, transcription occurs in the nucleus, from which the RNA has to be transported to the endoplasmic reticulum with ribosomes in the cytoplasm. Two sequence motifs

Nucleic Acid Synthesis

that are common constituents of promoters in prokaryotic genomes and are nominally referred to as −35 and −10 sequences signify that the midpoint of these sequences are located 35 and 10 bp 5 of the start site of transcription. However, the exact distance is somewhat variable for different genes. The consensus −35 sequence is TTGACA, and the consensus of −10 is TATAAT. However, both of the sequences are also somewhat variable. The strength of a promoter, i.e., how efficiently it is recognized for transcriptional initiation, depends on the exact sequence of the −35 and −10 sequences and possibly the intervening “spacer” sequences as well. The promoter strength can vary widely among genes, and mutations in the −35 or −10 sequence in a particular gene can dramatically affect its promoter strength. B. Regulation of Transcription in Bacteria Unlike replication of the complete genome, which is essential for cellular propagation, not all genes need to be transcribed in a particular cell for its survival. Synthesis of mRNA is required for generation of proteins. Because not all proteins are required at all times for cellular survival and metabolism, both in prokaryotes and eukaryotes, and many proteins are expressed only in specific stages of development and differentiation in higher eukaryotes, a gene’s transcription is often highly regulated. Furthermore, the stability of mRNAs and the proteins they encode vary over a wide range. Thus, different mRNAs are not made at the same rate. Additionally, the bulk of RNA, and in fact a large fraction of the cell mass, consists of ribosomal and transfer RNAs needed for carrying out protein synthesis. Both ribosomal and transfer RNAs are extremely stable. Regulation of transcription, first investigated in bacterial viruses, primarily in E. coli, an intestinal microbe and its bacteriophage λ, is the foundation of molecular genetics. The ease of generating and manipulating mutants of various genes in E. coli and λ led to the discovery of repressors, which are proteins that bind to operator sequences of genes and turn off transcription. The genes that were originally studied encode enzymes for sugar (lactose and galactose) metabolism. Inactivation of these genes and their expression could be studied because the proteins are not essential for bacterial survival. An activator needed for expression of lactose-metabolizing β-galactosidase was identified; it is downregulated in the presence of glucose (“glucose effect”) and upregulated by binding to 3 -5 cyclic AMP. Significant advances in elucidating the mechanism of transcriptional regulation came from the life cycle studies of the lysogenic λ virus, whose virus-specific proteins are not expressed in the lysogenic state, when its duplex DNA

Nucleic Acid Synthesis

871 act as positioning factors for correct localization of Pol III initiation. Pol II is the most versatile and widely utilized RNA polymerase in vivo and absolutely needs auxiliary, transcription factors (TFII) whose requirement is dependent on the nature of promoters. 3. The nature of eukaryotic promoters is quite different from the prokaryotic promoters. In addition to the bipartite promoter of Pol I, both Pol II and Pol III have a “TATA box” located about 25 bp upstream of the start site in Pol II responsive genes. The 8-bp sequence consists of only A•T base pairs and is surrounded by G•C pair-rich sequences. Interestingly, the TATA box is quite similar to the −10 sequence in E. coli promoters. There is a second element called a CAAT box, usually about −15 bp 5 of the TATA box. Alternatively a G•C rich sequence is present in some promoters, often at position −90. The consensus GC box sequence is GGGCGG, of which multiple copies are often present and occur in both orientations. These elements are not all present in all promoters; it appears that they work in a “mix and match” fashion. These boxes, and also a octamer box, bind to specific trans-acting factors and are engaged in multiple protein interactions among themselves as well as with components of the RNA Pol II holoenzyme. There is no significant homology among transcription start sites of various genes, except for the tendency for the first base in the transcript to be an A flanked on either side by pyrimidines. This region is defined as the initiator. The first step in transcriptional initiation of a TATAcontaining promoter is the binding of the factor TFIID to the region upstream of the TATA site. The TATA-binding protein, TBP, which specifically binds to the TATA box, is a component of the TFIID complex, along with other proteins collectively called TAFs (TBP-associated factors). TAFs can be variable in the TFIID complex, both in species and amounts, and provide the promoter specificity for initiation. Some TAFs are tissue specific. TFIID has a molecular mass of 800 kD, containing 1 TBP and 11 TAFs. TBP acts as a positioning factor and is able to interact with a wide variety of proteins, including Pol II and Pol III. It binds to the minor groove of the DNA double helix and makes contact with other factors which mostly bind to the major groove and can make multiple contacts. By bending the DNA at the binding site, it appears to bring the factors and RNA polymerase into closer proximity. Although TBP is utilized by both Pol II and Pol III, TFIID is the specific complex for Pol II recognition of a promoter. Other transcription factors (e.g., TFIIA) bind to the TFIID promoter complex and cover increasing segments of DNA. In addition to TFIIA, these include TFIIE, TFIIF, TFIIH, and TFIIJ. Most of the TFII factors

genome is linearly integrated in the host chromosome. Here again, both positive and negative regulatory mechanisms are in play to fine tune the expression of genes from a low maintenance level during lysogeny to large-scale expression of the viral genome when the lysogenic virus enters the lytic phase of growth and exploits the host cell synthetic machinery for replication of its own viral DNA, RNA, and proteins. C. Eukaryotic Transcription The fundamental process is identical in prokaryotes and eukaryotes, in that an RNA polymerase complex binds to the promoter and initiates transcription at a start site downstream to the promoter. De novo initiation of an RNA chain occurs with a purine nucleotide and creation of a transcription bubble with the open complex. The transcription complex can slide back along the nascent chain and endonucleolytically cleave off the 3 segment, then moves forward along the DNA template chain; termination occurs at specific regions in the genes. In spite of this similarity, however, the details are very different in eukaryotic cells and are summarized as follows. 1. Eukaryotic RNA polymerases contain many more subunits, located in the different regions of the nucleus. Pol I, specific for synthesizing rRNA, is located in the nucleolus, a specialized structure within the nucleus, while Pol II and Pol III are in the nucleoplasm. These enzymes have 8–14 subunits with a total molecular mass >500 kD. The large subunits have some sequence similarity with the bacterial RNA polymerases. RNA polymerases of mitochondria and chloroplasts are phylogenetically closer to bacterial RNA polymerase, commensurate with the fact that the target genes of these enzymes are fewer and much smaller in organelles, which are thought to have arisen by symbiotic acquisition of bacteria by primitive eucaryotes. 2. The promoter composition and organization of eukaryotic polymerases are quite specific for each polymerase. The promoters of rRNA genes contain a core and an upstream control element which is needed for high promoter activity. Two ancillary factors, UBFl and SLl, bind to these sequences. Although SLl binds only after UBFl in a cooperative fashion, SL1 is a σ -factor with four proteins among which TBP is also required for initiation by the other polymerases. Pol I is akin to Pol III in that it utilizes both upstream and downstream promoters. There are two types of internal promoters with distinct sequence boxes. One transcription factor (TFIII B) is required for initiation of RNA synthesis by Pol III. Other factors (TFIII A and TFIII C) help TFIII B bind to the right location and

872 are released from the transcription complex before Pol II leaves the promoter and carries out chain elongation. Interestingly, the same general transcription factors, including TFIID, bind to the TATA-less promoter, even though TATA binding by TBP is not available. There is an important contrast in the assembly of RNA polymerase complexes in eukaryotes and prokaryotes. E. coli RNA polymerase binds to the promoter as a complex with the σ -factor, providing the specificity for initiation but not elongation. Eukaryotic Pol II, on the other hand, goes through a much more complex choreography because of the prerequisite for binding to the promoter by other transcription factors. This dichotomy reflects the complex structural organization of the eukaryotic genome and the presence of a much larger number of genes with their complex regulation. Such regulation is not only dependent on the environment, but also on the stage of development and differentiation, at least in the metazoans. 4. A unique difference between prokaryotic and eukaryotic transcription is that in prokaryotes a single mRNA containing many genes can be transcribed from the DNA template as a single transcription unit, coupled with their direct translation on ribosomes into discrete polypeptides. This process reflects the fact that genes which encode enzymes in a given pathway are often clustered in an operon and are co-ordinately regulated. In contrast to the synthesis of polycistronic mRNA in E. coli and other bacteria, eukaryotic transcription units usually consists of single genes. This characteristic may also reflect uncoupled transcription and translation in these organisms. Thus, heterogeneous nuclear RNA (hnRNA) is synthesized in the nucleus and then transported to the cytoplasm along with its processing into mature mRNA including splicing, addition of poly(A) tail at the 3 end, and capping at the 5 end. Subsequently, the RNA is translated on ribosomes (endoplasmic reticulum). Thus, synthesis and utilization of mRNA are temporally and spatially separated.

Nucleic Acid Synthesis

FIGURE 8 A schematic representation of RNA splicing. The coding sequence in metazoan genomes is usually present in segments (exons; indicated by boxes) interspersed between noncoding introns. After synthesis of the primary RNA transcript (called heterogeneous nuclear RNA or hnRNA), the intron sequences are removed by precise cleavage and rejoining is mediated by the spliceosome complex, so that the resulting mature mRNA contains a correctly juxtaposed coding sequence for the polypeptide. The mRNA is also “capped” by 5 -5 linkage with GMP, and a tail of poly(A) is added at the 3 terminus to increase the stability of mRNA and to enhance its efficiency in directing protein synthesis when the mRNA is transported from the nucleus to the cytoplasm.

D. RNA Splicing in Metazoans The central dogma of molecular biology that the information flow from DNA to RNA to protein involves colinearity of the sequences of the monomer units is somewhat violated in metazoans because of the presence of interrupted or fragmented genes (Fig. 8). Thus, while the polypeptide sequence is colinear with the codons of the coding sequence in the mRNA, the RNA itself is not collinear with the gene from which it is transcribed. In other words, the gene contains additional intervening sequences called introns, which are transcribed but whose RNA sequence is subsequently removed from the final mRNA contain-

ing the coding sequence. The primary gene transcripts of nuclear genomes, called heterogeneous nuclear RNA (hnRNAs), are present in a form of protein-bound particles (ribonucleoprotein particles, or hnRNP). RNA splicing is the process of excising introns from hnRNAs, and contiguous exons are then joined to form mature mRNAs, which are subsequently translocated to cytoplasm and are used as templates for translation (Fig. 8). The cleavage and rejoining occur at specific junctions between exons and introns, so that there are no errors in mature mRNA. First, two adjacent exons are aligned, while the intervening intron is extruded, forming a loop (“lariat”) structure. Then the upstream exon is cleaved and joined to the downstream exon via a transesterification reaction. In most cases, two factors are essential for this process. One, the cis-elements in introns and exons, is the signaling sequences for the exact junction sites. The other is the splicing machinery, consisting of several small ribonucleoprotein particles (snRNP; U1, U2, and U4–U6), each of which contains small RNA molecules and proteins. The U1 and U2 snRNPs contain RNA complementary to the intron ciselement and catalyze the formation of the intron lariat, while two adjacent exons are aligned together. With other snRNPs forming an intermediate complex (spliceosome), U6 catalyzes the transesterfication. It should be noted that introns in RNA of some lower eukaryotic species are autospliced and therefore do not require snRNPs.

Nucleic Acid Synthesis

873 signaling cascade is initiated in response to the first signal. The external ligand first binds to its receptor on the cell surface, followed by internalization of the receptor ligand complex. A series of reversible chemical modification (mostly phosphorylation of the regulatory proteins) finally activates the ultimate transcription factors, which then trigger transcription of target genes. The unique difference between the eukaryotes and prokaryotes is in the utilization of transcription factors. In bacteria, one factor is usually specific for one gene or one regulatory unit. In eukaryotes, on the other hand, a single factor activates multiple target genes. Prokaryotic regulatory processes have been elucidated in remarkable detail by utilizing the power of molecular genetics, including “reverse genetics” by which the chromosomal genes in the organism could be mutated at specific sites and the mutant gene products purified and characterized. Furthermore, these genes can be expressed in the episomal state by introducing them into autonomously replicating recombinant plasmids. Commensurate with the significantly higher complexity and size of the genome and differentiation and developmental stages in metazoans, gene regulation in these organisms is very complex and occurs at many levels. Sets of genes are activated at distinct stages of differentiation and development of multicellular organisms in order to encode proteins which are required for specialized functions of the cells in these stages. In contrast, certain “housekeeping” proteins, including enzymes for metabolism and synthesis of all cellular components (i.e., RNA, DNA, structural proteins, and lipids), as well as enzymes for biosynthetic and degradative pathways, are needed in all cell types and developmental stages. Most somatic cells in adult mammals are nondividing and therefore do not require DNA synthesis machinery. However, all cells require transcription for generating proteins for other cellular functions. Unraveling the molecular mechanisms of regulation is the major focus of current research in molecular biology. The regulatory process is affected by multiple parameters. Many genes are activated due to external stimuli, e.g., exposure to hormones and growth factors. In these cases the extracellular signal often acts as a ligand to bind to cell surface receptors which activate the trans-acting factor(s) via multiple steps of signal transduction. 1. Regulation of Transcription via Chromatin Structure Modulation in Eukaryotes The eukaryotic genome is organized at multiple levels, starting with the nucleosome core as described earlier. The nucleosomes are organized in a higher order chromatin structure due to increasing compaction of DNA: from

Termination of eukaryotic transcription is coupled with processing. The mature rRNA is obtained by cleavage of a larger primary transcript synthesized by Pol I. Termination of Pol II transcription occurs at a repeat sequence of U, as in the case of E. coli RNA polymerase, but without the presence of a hairpin structure. More importantly, the 3 termini of mRNAs are generated by cleavage of primary precursor transcripts followed by addition of a tail of poly(A), a homopolymer of up to several hundred AMP residues synthesized by poly(A) polymerase in a template-independent reaction. E. Regulation of Transcription in Eukaryotes While both prokaryotic and eukaryotic genes are regulated by activators and repressors, enhancer elements are unique to eukaryotic genes and can profoundly increase the rate of transcription. These elements are located at a variable distance from the basic promoter itself, can be present both upstream or downstream to the promoter, and, in fact, can even be within the transcription unit. One unexpected feature is that they can function in either orientation and can activate any promoter located in the vicinity. Upstream activating sequences (UAS) have been identified in yeast and are analogous to enhancers in the mammalian genes. Based on the known properties of enhancers, it appears that the presence of these sequences affects chromatin structure and/or the helical structure of the DNA template itself. Further studies are needed to test other possibilities as well, e.g., whether the enhancer provides an entry point for the transcription complex or is needed to place the template at the nuclear matrix where transcription takes place. Positive and negative regulation of prokaryotic genes is achieved by binding of activators and repressors, respectively, to their cognate binding sites in the genes. Downregulation is more common, at least in E. coli, than positive regulation. In fact, the same protein can provide dual functions in a few cases, depending on the location of the sequence motif. In contrast, because of the complexity of chromatin structure and genomic organization development, differentiation, and cell cycle-specific synthesis of proteins, regulation of eukaryotic genes is extremely complex. This is evident from the large number of families of regulatory trans-acting factors which recognize similar if not identical sequence motifs in different genes. Sometimes, these factors have a distinct modular structure—one module for binding to target DNA sequence and another for interaction with components of the transcription apparatus. On top of these complexities, the signal for initiation of transcription may be extracellular, e.g., a growth factor which induces cell proliferation. A highly complex

874 2-nm-wide naked DNA fiber to metaphase chromosomes of microscopic width. The DNA template has to be accessible to transcription machinery containing RNA polymerase; transcriptionally inactive, highly compacted chromatin maintains its structure by multiple protein–protein and protein–DNA interactions, which are yet to be elucidated. However, it is now clear that at the nucleosome level, it is the strength of interaction between histones and DNA which regulates accessibility of the DNA to the transcription machinery, a process controlled by acetylation and phosphorylation of core histones. Multiple histone acetylases and deacetylases, which are themselves regulated, modulate chromatin structure. As stated previously, large protein complexes named SWI and SNF modulate chromatin structure in an energy-dependent process which may be responsible for the differentiation/ development-dependent turning on or off of specific sets of genes. 2. CpG Methylation-Dependent Negative Regulation of Genes In addition to histone modification, DNA itself was found to be modified, most commonly by methylation at the C-5 position of cytosine, but only when it is present as a CpG dinucleotide. Such methylation, catalyzed by specific methyltransferases, invariably inhibits gene expression, which was unequivocally established in the genomes during embryonic development. Sets of genes are selectively methylated or demethylated in the CpG sequences, most commonly in the genes’ promoter regions, leading to their activation or repression. Proteins that bind to methylated CpG sequences have been implicated in the control of histone deacetylation, thereby leading to closing of the promoter. F. Fidelity of Transcription (RNA Editing) The informational content of gene transcripts can be altered during or after transcription by a process collectively called RNA editing. The information changes are carried out at the level of mRNA. RNA editing appears to be a widespread phenomenon for both normal and aberrant RNA processing in organelles and nuclei. It was first discovered in the mitochondria of kinetoplasts in protozoa. Two types of RNA editing have been observed: (1) alteration of coding sequence by nucleotide insertion and/ or deletion and (2) base substitution. In mammalian cells, editing of an individual base in mRNA can cause a change in the sequence of the protein. Such changes can occur by enzymatic deamination in which C is converted to U or A is converted to hypoxanthine. Change of U to C has also been observed in many plants. The (mitochondrial)

Nucleic Acid Synthesis

mRNAs of several kinetoplastid species (Crithidia, Trypanosoma, etc.) were found to be edited by the insertion and deletion of U’s at many sites in mRNAs. The editing process uses a template consisting of a guide RNA (gRNA) whose genes function as independent transcription units. The gRNAs are generally 55–70 nucleotides in length and complementary to the mRNA for a significant distance including and surrounding the edited region. The gRNA dictates the specificity of uridine insertions by its pairing with the pre-edited RNA, but also provides the U residues that are inserted into the target RNA by transesterification reactions; the reaction proceeds along the pre-edited RNA in the 3 -5 direction. The RNA editing process reveals the existence of a previously unrecognized level for the control of gene expression. Recognition of this process has resulted in an expansion of the central dogma. Multiple RNA editing processes play a significant role in normal physiological processes, as well as being responsible for some disease.

VII. CHEMICAL SYNTHESIS OF NUCLEIC ACIDS (OLIGONUCLEOTIDES)
Development of strategies for chemical synthesis of nucleic acids represented a major breakthrough in molecular biology, because most of the current approaches involving PCR, manipulation of recombinant DNA, studies of gene regulation, etc. require synthetic DNA and RNA oligonucleotides with defined sequences. The difficulty of synthesizing RNA and DNA polynucleotide chains from mononucleotide units lies in the reactivity of the side chains of the bases and the susceptibility of the sugar glycosyl bond to cleavage under the harsh conditions needed for condensation reactions to generate phosphodiester bonds. An additional problem in RNA synthesis is the presence of the C 2-OH group in ribose. H. Khorana’s group was the first to solve the problem by blocking all reactive side chains of the bases with reversible blocking groups; a phosphodiester bond between C 3-OH of one nucleotide and the C 5-phosphate of another was generated by condensation in the presence of dicyclohexyl carbodiimide (DCC) under mild conditions. Repeating the process in a cyclic fashion generated oligonucleotides of a defined sequence. While the DCC condensation was efficient, the whole process was extremely laborious, because the products of each reaction had to be purified free of the side products and the blocking groups had to be removed after each cycle. Furthermore, the efficiency of the synthetic reaction fell off rapidly with increasing size of the oligonucleotide. A major advance occurred in the 1970s when two distinct types of chemistries were invented for synthesis of

Nucleic Acid Synthesis

875 between each reaction. The complete procedure has been automated in several commercial instruments. After synthesis is completed, the oligonucleotide product is released from the glass matrix by alkaline treatment, and then the last protective trityl group is removed. The quality and efficiency of polymer synthesis is determined by the efficiency of the individual reactions. The major advantage of phosphoramidite-based synthesis is very high efficiency (99%) of both the condensation and the deprotection reactions. Nonetheless, it is obvious that because the final yield of the oligonucleotide is the product of the yields of each individual cycle, very long oligonucleotides cannot be synthesized at a significant level. In practical terms, the current size limit of an oligonucleotide is usually up to about 120 monomer units. Even then the product has to be purified (usually by gel electrophoresis) from the contaminants, mostly composed of failed synthesis material. A major problem in therapeutic use of oligonucleotides is their degradation by nonspecific nucleases, once delivered inside the tissues and cells. One of several approaches to counter this problem is to synthesize artificial nucleic acids in which phosphate oxygen is replaced with sulfur. In a phosphorothioate oligo (S-oligo), some or all of the internucleotide phosphate groups are replaced by a phosphothioate group. These S-oligos are widely used

deoxyoligonucleotides with the possibility of automating the cyclic procedure. One was based on phosphodiesters of deoxynucleotides as the starting material, which had been utilized early on for synthesis of oligodeoxynucleotides. However, the phosphoramidite method invented later has become the exclusive method of choice for synthesis of both RNA and DNA sequences. The advantages of this method are (1) the relatively high stability of the starting compounds and (2) the mild reaction conditions for removal of the protective groups. Automated procedures have been developed for solidstate synthesis of polymers (Fig. 9), which is initiated by covalent attachment of the first monomer phosphoramidite unit to a glass matrix in the reaction vial; the phosphodiester condensation reaction is carried out by addition of monomer units in the 3 → 5 direction, which is opposite to the direction of enzymatic synthesis. Each cycle of synthesis involves removal of the protective groups after the condensation reaction. Fixed amounts of phosphoramidites of four nucleotides, as well as other modified nucleotides, are added to the reaction vial in predetermined order and amounts. The chemical treatments involving acidic and alkaline solvents are carried out in a preprogrammed sequential order, and the glass matrix containing the oligonucleotide is washed with solvent in

FIGURE 9 An outline of the chemical synthesis of nucleic acids.

876 in anti-sense applications because of their enhanced stability. The modified backbone of an S-oligo is resistant to the action of most nucleases and endonucleases, but they also tend to be subject to more nonspecific interactions due to “stickiness.” A. Peptide Nucleic Acids (PNA) Peptide nucleic acids (PNA; Fig. 10) are synthetic polynucleobase molecules which bind to DNA and RNA with high affinity and specificity. PNA was constructed with a charge-neutral, achiral, pseudopeptide backbone and is therefore chemically more closely related to peptides than to nucleic acids. Thus, PNAs, because of their backbone properties, show extremely good nucleic acid hybridization properties. In fact, PNA–DNA and PNA–RNA duplexes are, in general, thermally more stable than the corresponding DNA(RNA)–DNA(RNA) duplexes. PNAs are relatively easy to synthesize and are stable (especially biologically). These make PNA an attractive candidate for developing effective anti-sense and anti-gene reagents and drugs. PNAs have been found to inhibit RNA polymerase, human telomerase, HIV reverse transcriptase, and many more. Such PNAs are candidates for anti-cancer drugs and also as a means of developing novel drugs to treat HIV infections (AIDS). Despite these encouraging results, further progress is very much impeded by the in-

Nucleic Acid Synthesis

efficient uptake of PNA by living cells and the lack of efficient delivery systems.

SEE ALSO THE FOLLOWING ARTICLES
BIOCONJUGATE CHEMISTRY • DNA TESTING IN FORENSIC SCIENCE • FIBER-OPTIC CHEMICAL SENSORS • GENE EXPRESSION, REGULATION OF • HYBRIDOMAS, GENETIC ENGINEERING OF • ION TRANSPORT ACROSS BIOLOGICAL MEMBRANES • PROTEIN FOLDING • PROTEIN STRUCTURE • PROTEIN SYNTHESIS • TRANSLATION OF RNA TO PROTEIN

BIBLIOGRAPHY
Eckstein, F. (2000). “Phosphorothioate oligodeoxynucleotides: What is their origin and what is unique about them?” Antisense Nucl. Acid Drug Dev. 10, 117–121. Efimov, V. A., Buryakova, A. A., and Chakhmakhcheva, O. G. (1999). “Synthesis of polyacrylamides N-substituted with PNA-like oligonucleotide mimics for molecular diagnostic applications,” Nucl. Acids Res. 27, 4416–4426. Kornberg, A., and Baker, T. A. (1992). “DNA Replication,” Freeman, New York. Lewin, B. (2000). “Genes VII,” Oxford Univ. Press, New York. Malik, S., and Roeder, R. G. (2000). “Transcriptional regulation through mediator-like coactivators in yeast and metazoan cells,” Trends Biochem. Sci. 25, 277–283. Mooney, R. A., Artsimovitch, I., and Landick, R. (1998). “Information processing by RNA polymerase: Recognition of regulatory signals during RNA chain elongation,” J. Bacteriol. 180, 3265–3275. Nielsen, P. E. (2000). “Peptide nucleic acids: On the road to new gene therapeutic drugs,” Pharmacol. Toxicol. 86, 3–7. Peterson, C. L., and Logie, C. (2000). “Recruitment of chromatin remodeling machines,” J. Cell. Biochem. 78, 179–185. Ray, A., and Norden, B. (2000). “Peptide nucleic acid (PNA): Its medical and biotechnical applications and promise for the future,” FASEB J. 14, 1041–1060. Stuart, K., Allen, T. E., Heidmann, S., and Seiwert, S. D. (1997). “RNA editing in kinetoplastid protozoa,” Microbiol. Mol. Biol. Rev. 61, 105– 120. Sudarsanam, P., and Winston, F. (2000). “The Swi/Snf family— nucleosome-remodeling complexes and transcriptional control,” Trends Genet. 16, 345–351. Summers, D. K. (1996). “The Biology of Plasmids,” pp. 39–62, Blackwell Sci., Oxford. Uhlmann, E. (1998). “Peptide nucleic acids (PNA) and PNA-DNA chimeras: From high binding affinity towards biological function,” Biol. Chem. 379, 1045–1052.

FIGURE 10 Structure of peptide nucleic acid (PNA). An artificial oligomer produced by chemical synthesis retains the ability to pair with bases, but is resistant to degradation by nucleases because its backbone does not contain the normal phosphodiester linkage.

Protein Folding
Maurice Eftink Susan Pedigo
University of Mississippi

I. II. III. IV. V.

Introduction Stability of the Tertiary Fold Folding Pathways Empirical Approaches Closing Comments

GLOSSARY
Absorbance spectroscopy Monitors conformational transitions in macromolecules by measuring absorbance changes, usually in the aromatic region of the ultraviolet (UV) spectrum. Circular dichroism A very commonly used method for studying protein conformational changes. Fluorescence The most sensitive of the commonly used optical methods for studying protein unfolding transitions. Nuclear magnetic resonance A powerful method for studies with proteins, as there is such a large number of resolved signals. Scanning calorimetry Measures the variation in the specific heat of a protein containing solution as a protein is thermally unfolded.

this folded structure is achieved remains one of the most important questions in biochemistry and molecular biology. Current efforts toward sequencing the human genome and the genomes of other organisms have led to a large number of putative protein sequences. Much as the genetic code is the Rosetta stone that gave the link between DNA and protein sequences, we now need to find such a link between protein sequences and final structure and function. Without knowledge of the rules of protein folding, there can be little understanding of function from protein sequence information alone.

I. INTRODUCTION
We have interest in protein structure and function at both a fundamental and a practical level. There is astounding beauty in the mastery with which nature has tailored molecules for specific functions, activity levels, regulatory properties, and integration into complex macromolecular assemblies. As will be discussed, in most cases, these molecules assume a final stably folded structure

PROTEINS are only functional as enzymes, transport agents, receptors, and so forth when they exist in a folded, three-dimensional, native structure. The means by which

179

180 spontaneously. Thus, all of the information necessary for biological activity is contained in the simple sequence of amino acids as encoded by the DNA. Practically speaking, predicting protein structure, stability, and function from the primary sequence will open myriad opportunities in the areas of medicine (e.g., drug discovery and understanding molecular basis of disease), industry and manufacturing (e.g., biocatalysis and bioprocessing), and the environment (e.g., bioremediation). Proteins are linear polymers of amino acids that are linked through amide linkages, commonly called the peptide bond. The “backbone” atoms include the amide linkages separated by a carbon that is derivatized by any one of 20 common side chains. The side chains may be grouped at neutral pH as acidic, basic, hydrophobic, and uncharged hydrophilic according to their chemical nature. Thus, although the backbone of the peptide polymer is a repeating identical unit, the side chains and their distinct properties dictate the nature of the protein. Because a subset of the amino acid side chains is charged at neutral pH (acidics are negative and basics are positive), the protein polymer is a polyelectrolyte. The linear sequence of amino acids is called the primary structure of the protein (Fig. 1). The primary structure dictates the way in which the polypeptide folds into a functional protein, in most cases without instructions from other sources. Protein families are proteins related by structure or function. A protein family may be structurally diverse but have a particular cluster of amino acids at the active site that defines the class according to some catalytic function (e.g., dehydrogenases and kinases). Alternatively, proteins may have a structural motif that defines the class (e.g., helix–loop–helix motif of the EF-hand calcium-binding proteins). Proteins with identical function in different organisms often have slightly different primary structures (see below). The presence of certain amino acids relative to others in primary sequences allows putative protein sequences from the Human Genome Project, for example, to be classified into general protein families. Whether this initial classification is valid remains to be seen. To discover the rules of protein folding, two major approaches have emerged: computational and empirical approaches. The computational approach, often termed proteonomics, attempts to predict the structure of a protein based on its sequence by defining a set of rules and criterion for their application. This topic is covered elsewhere in this series. The empirical approach to discovering the rules of protein folding defines global rules for folding based on lessons learned from particular proteins. These two methods are distinctly interwoven.3 Hypotheses derived from one are testable through the other. In this paper, we will discuss the empirical approach to studying protein folding.

Protein Folding

The empirical approach to understanding protein folding has relied heavily on mutational analysis. As mentioned earlier, proteins from different species with identical functions may have slightly different amino acid sequences, or mutations. Often the mutations are conservative, particularly in amino acids that are critical to the structure or function of the protein. Scientists study the different physical properties of these related proteins to gain insight into the role of amino acids in local or global structure and function of the protein. Often mutations are purposely engineered into protein sequences using molecular biological techniques to test hypotheses about roles of certain amino acids in structure or function. Selective substitution of tryptophan into a sequence allows placement of a convenient spectroscopic probe (see below). Although proteins are very diverse, the one thing that almost all have in common is that they adopt spontaneously a unique and stable tertiary structure. This is an utter miracle of nature given the complexity of these heterogeneous polymers. The study of protein folding is focused on understanding the rules that govern the transition into and the stability of this unique fold. The transition into the tertiary structure is studied by kinetic methods. Thus, kinetic studies ask the question, “By what pathway is the final tertiary structure folded?” Alternatively, equilibrium thermodynamic methods ask “How stable is the final fold and why?” Each of these approaches will be discussed individually.

II. STABILITY OF THE TERTIARY FOLD
Stability of a protein is usually studied by observing the energetics of unfolding transitions given by the equations below: N ↔U K un = [U ]/[N ] G o = –RT ln K un un (1) (2) (3)

These equations apply to a simple two-state transition between the native (N ) and the unfolded (U ) state given by the equilibrium constant K un . This is, by definition, a cooperative process without a detectable intermediate species. The denatured or unfolded state of a protein is generally considered to be an ensemble of conformations in which all parts of the protein are exposed to the solvent with a minimum of intramolecular interactions. The denatured state has high conformational entropy and is biologically inactive. The unfolding transition (Eq. (1) and Fig. 2) can be induced by pressure, temperature, extreme pH, and denaturants such as urea and guanidine

Protein Folding

181

FIGURE 1 Diagram of the levels of protein structure. (A) Amino acids are the basic building blocks (monomers) of proteins (polymers). All amino acids contain a carboxylic acid group and an amine group connected by a central carbon called the α-carbon. Each of the 20 common amino acids, designated by a three-letter code, has a unique side chain (R) that is also bonded to the α-carbon. (B) Through a dehydrolysis reaction, an amide bond is formed (boxed region) that links the amino group of one amino acid to the carboxylic acid group of the next amino acid. (C) The primary structure of proteins (1◦ ) is the linear sequence of amino acids written from the amino-terminal end (left) to the carboxy-terminal end (right), by convention. Secondary structure of proteins (2◦ ) is classified into three major categories. In the α-helix structure, the backbone atoms (amide linkage and α-carbon) coil into a right-handed helical shape (residues 55 to 67 from staphylococcal nuclease1 are shown). The α-helix is held together through a series of hydrogen bonds between the amide hydrogen and the carboxyl oxygen of the backbone atoms from amino acids further up the chain. The side chains (not shown) protrude from the central core structure like the spokes of a wheel. Another important secondary structure type is the turn (residues 76 to 88 from Staphyloccocal nuclease are shown). We use the term turn loosely here to represent the regions of proteins that turn corners, thereby allowing interactions between different and often distant (in terms of primary structure) substructures. The β-sheet structure is the third common secondary structure type (residues 9 to 12 and 72 to 76 from Staphylococcal nuclease are shown). It is similar to the α-helical structure in that hydrogen bonds between backbone atoms hold the structure together and the side chains (not shown) protrude from the structure above and below the plane of the sheet. In contrast to the α-helix, the β-sheet can be formed from segments of protein that are far apart in the primary sequence. The tertiary structure (3◦ ) is the three-dimensional association of secondary structures into a unique and stable final fold. A ribbon tracing the backbone atoms of Staphylococcal nuclease is shown.1,2 . The N-terminus of the protein is in the bottom right and the C-terminus is in the top left of the figure. No side chains are drawn except that of residues tryptophan 140.

HCl, as will discussed in a subsequent section. These perturbants disrupt the intramolecular interactions that hold proteins together. One can imagine that the ensemble of unfolded states could be influenced by the means used to unfold.

The native structure of proteins is stabilized by intramolecular, noncovalent interactions including hydrogen bonding, ionic, and van der Waals interactions, and covalent cross-links (disulfide bridges between cysteine residues) according to Eq. (4):

182

Protein Folding

the aqueous solvent. Solvation of nonpolar side chains by aqueous solutions causes a decrease in the entropy of solution. To avoid this entropic penalty, proteins typically bury their nonpolar residues in the interior of a protein.4

III. FOLDING PATHWAYS
The intramolecular interactions discussed above stabilize the final folded structure of a protein. However, knowledge of the end states, N and U , tells us nothing of the path taken between them. Proteins fold on the time scale of microseconds to hundreds of seconds. It is impossible to sample all possible conformations during this time and it is clear that there is a preferred order of events leading to the final tertiary fold. Determining this order of events is an area of active inquiry. The questions that experimentalists are attempting to answer are “Do autonomously folding substructures nucleate the folding of other regions of the protein?” or “Do neighboring substructures fold and then collide to make the tertiary structure?” There is experimental evidence that hydrophobic amino acid residues collapse into a “hydrophobic core” and then the secondary structural units form around the core. It is likely that a combination of these scenarios leads to a correctly folded protein. It is clear that the kinetics of protein folding is protein dependent. Some fold in a distinctly cooperative fashion, such that one can detect only the unfolded and native end states (U ↔ N ), being two-state in a kinetic as well as equilibrium sense. This is equivalent to saying that there is a single rate-limiting step, and intermediate species are not populated. Alternatively, some proteins fold by populating one or more distinct intermediate species (e.g., U ↔ I ↔ N ; see Fig. 2). Thus, formation of the intermediate species is fast, often formed in the dead-time of the instrument, and formation of the native species from the intermediate is relatively slow and easily monitored experimentally. It has been shown that this slow phase in some cases may be due to proline isomerization.5

FIGURE 2 Illustration of cooperative vs. noncooperative unfolding transitions. If the native state of a protein (N) is denatured into the unfolded state (U ) in a single transition (pathway 1), then it is a two-state or cooperative unfolding transition. Alternatively, the native state may be converted into one or more intermediate states (pathway 2). For example, if a protein is comprised of multiple domains, one of the domains may be unfolded first. It is also possible to form a completely different intermediate before unfolding completely. The presence of intermediate species may be observed using kinetic or equilibrium techniques. However, intermediates detectable by kinetic methods may or may not be observable by equilibrium methods.

Go = un

G H-bond + + G S −S +

G ionic + GH phob G vdW (4)

Each term in Eq. (4) will be discussed separately. As mentioned earlier, an important stabilizing factor for the tertiary fold of a protein is its intramolecular hydrogen bonds ( G H-bond ). Secondary structures are stabilized by hydrogen bonds between backbone amide atoms (Fig. 1). The side chains of neighboring secondary structural units can interact through hydrogen bonding. Ionic interactions ( G ionic ) between acidic and basic side chains may stabilize the tertiary structure of proteins and are pH dependent. The actual pKa of an ionizable side chain is influenced by the microenvironment in which it resides. Nonpolar and polar, but uncharged, amino acids interact through van der Waals interactions ( G vd W ). In some proteins, cysteine residues (side chain is a sulfhydryl) form disulfide linkages that can increase the overall stability of the protein ( G S – S ). Other possible factors not considered explicitly here are the effects of metals, nucleotides, prosthetic groups, and cofactors on protein structure and stability. By far the most important noncovalent factor that determines protein stability is hydrophobic interactions ( G H phob ). In globular proteins, hydrophobic amino acids are buried in the interior where they create a “hydrophobic core.” Although these nonpolar residues participate in van der Waals interactions, the primary driving force for the formation of the hydrophobic core is to avoid

IV. EMPIRICAL APPROACHES
A. General Experimental Strategies As discussed above, experimental studies of protein folding reactions fall into the category of either equilibrium or kinetics studies, with the former yielding thermodynamic information about the energy differences between the native and denatured structural states and the latter studies providing information about the folding pathway and

Protein Folding

183 where X i is the mole fraction of each species i and Si is the intrinsic signal of species i. This relationship applies to most solution optical spectroscopic methods. Clearly, for a particular spectroscopic signal to be useful for tracking a N ↔ U transition, the signal of the N and U states must be sufficiently different. The native (X N ) and unfolded (X U ) mole fractions are directly related to the equilibrium constant in Eq. (2), as: X N = 1/(1 + K un ); X U = K un /(1 + K un ). (6)

the height of energy barriers between important species on this pathway. In general, to perform either an equilibrium (thermodynamics) or time-dependent (kinetics) study, one must be able to experimentally monitor a signal that tracks the population of the structural states of the protein. There are a number of ways this can be done. The most convenient experimental methods involve solution-phase spectroscopic measurements; among these methods are absorption spectroscopy, fluorescence, circular dichroism, and nuclear magnetic resonance. Other methods include differential scanning calorimetry, light scattering, electrophoresis, and chromatography. This section gives a brief description of the advantages and disadvantages of some of the above methods. These methods are not equally applicable to equilibrium and time-dependent studies of protein unfolding, as some methods have a rapid response and some have a slow response. Methods also differ in their intrinsic sensitivity, which is related to the concentration of protein necessary to perform the measurement, their ease and economy of use, and whether they provide auxiliary information about the structure of the protein in its native and denatured states. What is meant by the last statement is that some of the spectroscopic signals can provide information about the secondary or tertiary structure of the protein species. For most types of spectroscopy, the signal arises from particular amino acid residues (e.g., aromatic side chains or peptide bond), thus differences in the signals for the conformational states can be related to differences in the local environment of these amino acid residues (e.g., tryptophan residue 140 in staphylococcal nuclease; see Fig. 1). If there are only a very few of such signal origination sites, then site-specific information can be obtained. If there are many probe sites and they are distributed throughout the protein’s structure, then the method yields global information (e.g., signal from the amide linkage in the peptide backbone; see Fig. 1). It goes without saying that the protein sample to be studied must be well defined with regard to purity, and solution conditions must be selected and controlled to be relevant to other functional studies and studies with other proteins. Neutral pH, 20◦ C, and an ionic strength of 0.1 to 0.2 are the most commonly employed solution conditions. A key to most of these methods and their use in protein unfolding studies is that the signal is a mole-fraction weighted average of the signals of each protein species. That is, for the simplest case of a thermodynamics study of the transition between a native, N , and unfolded, U , state of a protein, the observed signal, S, can be expressed as: S= Si X i (5)

The transition from the native state to the unfolded state, or vice versa, can be induced in several ways, essentially by varying the solution conditions in a way that changes the equilibrium between the native and unfolded state. The transition may be induced by varying temperature, adding chemical (chaotropic agent) denaturant, adding acid or base, or increasing pressure. In the case of multimeric proteins, subunit dissociation, which may be accompanied by denaturation of the subunits, can be induced by dilution of the protein. Before discussing the various spectroscopic methods, some thermodynamic relationships are presented for describing the transitions induced in the above ways. B. Basic Thermodynamic Relationships Table I gives some widely accepted relationships for describing the variation of G o for a two-state N ↔ U un transition with temperature, chemical denaturant, pH, or pressure as the perturbations. One of the equations in Table I, when combined with those above and Eqs. (1– 3), can be used to describe data as a function of the denaturing condition. The thermodynamic parameters related to the relationships in Table I are briefly described below. o o 1. Thermal unfolding: Hun and Sun are the enthalpy and entropy changes for a two-state unfolding reaction. o o Both Hun and Sun may be temperature dependent, when the heat capacity change, C p , has a nonzero value. In this case, Eq. (7b) in Table I (the Gibbs-Helmholtz o o equation) should be used, where the Ho,un and So,un are values at some defined reference temperature, To (e.g., 0◦ or 20◦ C).6,7 The heat capacity change for unfolding of proteins is typically found to be positive and to be related to the increase in solvent exposure of apolar side chains upon unfolding. That is, a positive C p is a result of the hydrophobic effect. A consequence is that the G o (T ) un for unfolding of a protein will have a parabolic dependence on temperature and will show both high-temperature and low-temperature induced unfolding.8 2. Denaturant-induced unfolding: The empirical relationship in Table I for chemical denaturation includes

184
TABLE I Relationships Describing Two-State Transitions in Proteins Temperature o o G un (T ) = Hun − T Sun o o G un (T ) = Ho,un + C p (T − To ) − T [ So,un + C p ln(T /To )] where o Ho,un is the enthalpy change at T = To . o is the entropy change at T = T . Sun o C p is the change in heat capacity upon unfolding. Chemical Denaturants G un ([d]) = G o − m[d] o,un where G o is the free energy change in the absence of d. o,un m = δ G un /δ[d]. pH  +   1 + [H ]    K a,U
+    1 + [H ]   K a,N n

Protein Folding

(7a) (7b)

(linear extrapolation model)

(8)

G un ( p H ) = G o − RT o,un

•

           (9)

ln

n

where G o is the free energy change at neutral pH. o,un K a,U is the acid dissociation constant of a residue in the unfolded state. K a,N is the acid dissociation constant of a residue in the native state. Pressure G un (P) = G o – Vun (Po –P) o,un where Vun = volume change for N ↔ U transition. Po = reference pressure. (10)

For a two-state transition, A ↔ B (or N ↔ U for the unfolding of a native, N , to an unfolded, U , state of a protein) the mole fractions of the N and U states are given as X N = 1/Q, X U = exp(– G un /RT )/Q, where Q = 1 + exp(– G un /RT ) and the function for G un is taken from above the average fluorescence signal, Fcalc = X i (Fi + xδ Fi /δx ), where x is a generalized perturbant.

G o , the free energy change for unfolding in the abo,un sence of denaturant, and m, the denaturant susceptibility parameter (= –δ G un /δ[d]), where [d] is the molar concentration of added chemical denaturant.9,10 Through an empirical relationship, the given equation appears to adequately describe the pattern for denaturant-induced unfolding of a number of proteins. The G o value is a o,un direct measure of the stability of a protein at the ambient solvent conditions, which can be moderate temperature and pH (e.g., 20o C and pH 7). The m value also provides structural insights, as m values have been suggested to correlate with the change in solvent accessible apolar surface area upon unfolding of a protein.11 For example, a relatively large m value (i.e., a high susceptibility of the unfolding reaction to denaturant concentration) indicates that there is a large change in the exposure of apolar side chains on unfolding, which might be the case for a protein that has an extensive core of apolar side chains that are exposed upon denaturation.

3. Acid-induced unfolding: The relationship for acidinduced unfolding assumes that there are n equivalent acid dissociating groups on a protein that all have the same pK a,U in the unfolded state and that they are all perturbed to have a pK a,N in the N state. If the pK a,N is more than 2 pH units lower than pK a,U , then the equation simplifies with the denominator of the right term going to unity. The simplest relationship for acid-induced unfolding includes G o , the free energy of unfolding at neutral pH; n, the o,un number of perturbed acid dissociating residues; and their pK a,U in the unfolded state. Presumably, n should be an integer and pK a,U should be approximately equal to the values for such amino acids as glutamate, aspartate (e.g., pK a,U should be about 4 to 4.3) or histidine (e.g., pK a,U should be around 6.5). 4. Pressure-induced unfolding: In the relationship for pressure, P, induced unfolding of proteins, G o o,un is again the value of the free energy change at 1 atmosphere pressure and Vun = VU − VN is the difference in volume

Protein Folding

185 aromatic side chains to water.14 Tryptophan’s absorbance is also sensitive to the local electrostatic field; changes in indole-charge interactions can cause either red or blue shifts upon protein unfolding.15 Table II gives the typical concentration range used for unfolding studies with proteins using this and other methods. The sensitivity of difference absorbance measurements will depend on the molar extinction coefficient of the chromophore and their number, but a concentration range of 0.01 to 0.1 mM protein is usually needed for reasonable signal to noise with a 1-cm pathlength cell. Thermal scans, to induce the unfolding transition, are easy to perform with accessories available for most absorbance spectrophotometers. Chemical denaturant- or pH-induced transitions can be less convenient (unless one has automated titration equipment), since a series of solutions with equal protein concentration and varying denaturant must be prepared. With any of these perturbing conditions, it is important to realize that the variation in the conditions itself (i.e., varying temperature, pH, chemical composition) can lead to a “baseline” change in the absorbance signal from the native and unfolded species.16 So long as these baseline trends are linear and not as large as the absorbance change associated with the conformational transition, the baseline trends can be corrected for in the data analysis. The advantages of absorbance measurements are the ready availability, ease of use, and low cost of the instrumentation. The biggest disadvantage is that it is less sensitive than some other methods. 2. Circular Dichroism Circular dichroism (CD) is a very commonly used method for studying protein conformational changes. The far UV spectral region (180 to 250 nm) is dominated by absorbance by peptide bonds, and there are signature spectra for α-helix and other types of secondary structure in a protein. Additionally, the aromatic CD spectral region of 250 to 300 nm senses the chirality around the aromatic amino acid side chains and there is usually a structured aromatic CD spectrum for the native state of a protein.14,17,18 The effective sensitivity of CD is comparable to or slightly better than that of difference UV absorbance spectroscopy. CD instruments can be purchased with thermoelectric cell holders for thermal scans and with automated titrator syringe pumps for chemical denaturant titrations. Since the far-UV spectral regions is important in protein unfolding studies, it is necessary to work with salts and buffers that have minimal absorbance in this region. When performing CD measurements, it is necessary to pay attention to the buffer and salts and other solution components (e.g., chemical denaturants) being

of the unfolded and native states. Pressure-induced unfolding studies require a specialized high pressure cell.12,13 5. Dissociation/unfolding of oligomeric proteins: Oligomeric proteins are interesting as models for understanding intermolecular protein-protein interactions. A general question for oligomeric proteins, including the simplest dimeric (D) proteins, is whether the protein unfolds in a two-state manner, D ↔ 2U , or whether there is an intermediate state, which might be either an altered dimeric state, D , or a folded (or partially folded) monomer species, M. Models for these two situations are as follows: D ↔ D ↔ 2U D ↔ 2M ↔ 2U (11a) (11b)

For a D ↔ 2U model, the relationships between the observed spectroscopic signal, Sexp ; the mole fraction of dimer, X D , and unfolded monomer, X U ; and the unfolding equilibrium constant (K un = [U ]2 /[D]) will be given by Eq. (5) and XU =
2 K un + 8K un [P]0 − K un ; X D = 1 − X U (12) 4[P]0 1/2

where [P]0 is the total protein concentration (expressed as monomeric form), where Si is the relative signal of species i and where K un will depend on the perturbant as given by one of the above equations. That is, the transition should depend on the total subunit concentration, [P]0 , and on any other perturbation axis. C. Experimental Signals 1. Absorbance Spectroscopy Absorbance spectroscopy (difference spectroscopy) monitors conformational transitions in macromolecules by measuring absorbance changes, usually in the aromatic region of the ultraviolet (UV) spectrum. The amino acids tryptophan and tyrosine are the most important chromophores in the UV region for proteins. As mentioned earlier, tryptophan residues are often engineered into proteins as reporters of local and/or global environment. The indole ring of tryptophan and the phenol ring of tyrosine show sensitivity of their absorbance spectrum to solvent polarity. There is a blue shift in the absorbance of indole and phenol upon increasing solvent polarity. As a result, there will often be a blue shift in the absorbance of tryptophan (typically monitored as a decrease in absorbance in the 291- to 294-nm region of the spectrum) or tyrosine (at 285 to 288 nm) upon unfolding of a protein and a consequent increase in the exposure of these

186
TABLE II Solution Methods for Monitoring the Progress of Protein Unfolding Transitions Method Absorbance Circular dichroism Fluorescence FTIR Light scattering NMR DSC Activity/binding Chemical reactivity Chromatography Electrophoresis Potentiometry a Protein Folding

Conc. Range (mM)a 0.01–1 0.01–0.1 0.0001–0.01 0.5–2 0.1–1 1–10 0.02–0.2 —b variable —c —c 0.1–1

Scanning or Titrationsd TS/AT TS/AT TS/AT TS No No TS P P No Gradients No

Structure Sensed Local Secondary Local/tertiary Secondary Size and shape Local/tertiary Tertiary Tertiary Local/tertiary Size and shape Size, shape, charge Local

Kinetic Applications *** *** *** * * * —e * * — — —

Concentration ranges are for typical experiments with a 20-kDa protein. The concentration range will depend on the method being used to measure enzymatic activity or ligand binding. c The concentration of protein varies during the course of the experiment as the sample flows through the column, gel, or capillary. Initial concentrations are usually in the range of 1 mg/mL. d “TS” refers to the ability to perform thermal scans to unfold a protein; “AT” refers to the ability to perform automated titrations of a protein sample with chemical denaturant, acid, or base while the sample is loaded in the instrument. The label “P” indicates that an automated thermal scan or titration may be possible for certain applications, though this is not commonly done. The “Structure Sensed” column lists the features of the protein structure (e.g., secondary and tertiary structure, local interactions, etc.) that are sensed by the method. Some of these entries are judgment calls. The “Kinetic Applications” column indicates the amenability of the method to protein folding/unfolding kinetics experiments. A label “***” indicates that transient mixing or other means are available for the rapid initiation of the reaction. A label “*” indicates that the method is amenable to study relatively slow reactions (i.e., by a hand-mixing experiment). e Through variation of thermal scan rate or a frequency domain application of DSC, it is possible to obtain kinetics information. b used, particularly if one wishes to make measurements below 200 nm, as various buffers, salts, and denaturants can absorb a significant amount of light in the far-UV. Schmid14 has provided a number of practical tips regarding the application of CD for studies with proteins. There is less interference by buffer, salts, etc. in the aromatic UV spectral region. Whereas the aromatic CD signals can sense the loss of tertiary structure in a protein as it denatures, the CD signals in this region are much smaller than those in the far-UV CD region, giving a lower signal-to-noise ratio. Baseline slopes, as one varies temperature or chemical denaturant, also must be considered in CD measurements in both the far-UV and aromatic spectral region; however, the baselines trends are usually not large. A difference between far-UV CD and other optical methods is that CD signals observe changes throughout the structure of the protein (i.e., its secondary structure) and the magnitude and direction of the signal changes can be more directly related to changes in structure (e.g., a

loss of ellipticity at 222 nm can be related to a loss of α-helix). 3. Fluorescence Fluorescence is the most sensitive of the commonly used optical methods for studying protein unfolding transitions.14,19−21 The absolute sensitivity depends on a number of factors (e.g., lamp or laser intensity, cell pathlength, chromophore extinction coefficient, and quantum yield), of course, but commercial fluorometers can usually detect signals down to the 10-nM range. Either intrinsic or extrinsic fluorophores can be used. The most commonly used intrinsic fluorophores are the tryptophan and tyrosine residues, with the former being the most important due to its larger molar extinction coefficient and a redder absorbance and emission. The fluorescence of tryptophan residues is very dependent on the local microenvironment of its indole side chain, making tryptophan fluorescence responsive to the structure

Protein Folding

187

of a protein. This spectral responsiveness is in terms of its emission maximum and its quantum yield. For example, the emission maximum of tryptophan almost always shifts to longer wavelengths (red shifts) upon unfolding a protein and increasing the solvent exposure of this amino acid side chain. There is a large literature about the fluorescence of tryptophan residues in proteins and its use to study changes in the structure of proteins.19 A variety of extrinsic fluorophores can be attached to proteins to serve as fluorescence probes. These can be selected to maximize sensitivity and to avoid contamination (i.e., by moving to longer absorption and emission wavelengths) from other absorbing components.22 With both intrinsic and extrinsic fluorescence probes, the method focuses only on these probes sites, which might be as few as a single site on a protein. Like the signals from absorption spectroscopy and CD, fluorescence intensity signals (either at a single wavelength or integrated over the emission envelope) follows Eq. (5) and can be used to extract thermodynamic information. However, there are other easily measured fluorescence signals (emission maximum and anisotropy) that do not follow the mole fraction averaging of Eq. (5).19 The apparent emission maximum of a protein will be dominated by the structural state, native or unfolded, which has the higher quantum yield. Consequently, the apparent emission maximum will frequently not give a true reflection of the population of native and unfolded states, thus limiting the value of this type of fluorescence measurement for use in recovering thermodynamic parameters. (Rather than use the apparent emission maximum, it is better to perform curve fitting with composite spectra of the native and unfolded states.) Fluorescence anisotropy values for the fluorescence of a fluorophore on a protein will depend on the fluorophore’s rotational freedom and fluorescence lifetime. Because the motional freedom of intrinsic or extrinsic fluorophores will usually increase when a protein unfolds, a change in a protein’s fluorescence anisotropy is expected upon unfolding. However, to properly use anisotropy to analyze the thermodynamics (or kinetics) of an unfolding transition, Eq. (1) should be replaced with one that includes the fluorescence quantum yield of the protein’s structural states (see Reference 19). As with the above-listed optical methods, fluorescence instruments are designed to allow automated thermal scans and/or titrations. The baseline problem can be more significant with fluorescence than the other methods and should not be ignored. In particular, it is well known that the fluorescence intensity of fluorophores will decrease with increasing temperature, regardless of whether there is a conformational transition. While baseline trends may not

FIGURE 3 Simulation of denaturant-induced unfolding of a protein in a two-state manner. A simulated fluorescence signal ( ) is plotted vs. denaturant concentration for a protein, using Eq. (8). The simulated fluorescence signal decreases with addition of denaturant because the unfolded species has a smaller fluorescence signal (fluorescence is quenched on unfolding). The pre- and posttransition baselines may slope, as shown. The fraction of unfolded species (XU ) increases from left to right as the fraction of native species (X N ) decreases.

be linear over extensive ranges of the perturbing variable (e.g., temperature or chemical denaturant), it is usually adequate to assume linear slopes over a limited range of the variable. The advantages of fluorescence for studying protein unfolding reactions are the wide concentration range that can be measured and the responsiveness of the signal to the microenvironment of the fluorophore. Additionally, fluorescence signals of the native and unfolded state can provide a modicum of structural information about these states (at least with respect to the microenvironment of the fluorophores). Figure 3 shows simulated data for the denaturant-induced unfolding of a protein, as would be monitored by fluorescence intensity measurements. 4. Differential Scanning Calorimetry Another frequently used method is differential scanning calorimetry (DSC), which measures the variation in the specific heat of a protein containing solution as a protein is thermally unfolded.23−25 As opposed to the above optical techniques, where photons are being measured, calorimetry measures the transfer of heat associated with the thermally induced conformational transition. DSC and related types of calorimetry are intrinsically less sensitive than the optical methods. Nonetheless, advances in the technique have made it possible to perform DSC studies with samples as low as 0.1 mg/mL. Temperature is scanned in DSC measurements, so it is the variable that causes the structural transition of a protein. DSC data are typically presented as thermograms that

188 yield a heat-capacity maximum corresponding to the thermal transition temperature, TG , and an enthalpy change, Hun , for the transition. The Hun value can be determined either by integration of the thermogram or by curve fitting (i.e., fitting a van’t Hoff equation to the shape of the thermogram). Referring to these two Hun estimates as the calorimetric and van’t Hoff values, the ratio of the calorimetric and van’t Hoff Hun values can be used to determine whether the transition is best described as a two-state process. That is, a ratio of 1.0 (indicating that two Hun estimates are essentially the same) means that the structural transition is two state. 5. Nuclear Magnetic Resonance Nuclear magnetic resonance (NMR) spectroscopy is a powerful method for studies with proteins, as there is such a large number of resolved signals (due to the individual nuclei, such as the 1 H and 13 C atoms in the backbone and/or side chains of the amino acids).26,27 This gives the potential to track conformational transitions by observing changes at a large number of individual sites on the protein. This is further made possible by the fact that the signals (peaks having various chemical shifts) are usually widely dispersed in the native state of a protein, as a consequence of the sensitivity of the resonance peak for individual nuclei to the local magnetic field, which in turn is related to the three-dimensional structure of the protein. Unfolded proteins, by comparison, usually have a much narrower range of resonance peaks for similar amino acid components. Tracking any of the individual resonance signals, such as those assigned to histidine or tryptophan residues, as a function of denaturing condition (e.g., temperature, pH, or added chemical denaturant) provides a way to study the unfolding process, as the signal is transformed from that of the native state to that of the unfolded state. An important difference between NMR and the above optical methods is that NMR signals can be dynamically averaged signals or individual signals can appear for the native and unfolded states. The latter results if the rate of interconversion of the conformational states is relatively slow in comparison to the difference in resonance frequencies of the signals for the two states. The use of NMR for proteins studies is usually limited to proteins having molecular weight of about 25 kDa or less. The method requires a relatively high concentration of protein, compared to other methods. Besides the above application of NMR to track the population of native and unfolded states, NMR also can provide very high-quality information about the tertiary and secondary structure of proteins. In addition, pulsed isotope labeling experiments can provide information about

Protein Folding

the pathway for protein folding reactions and can provide estimates of the unfolding equilibrium constant at individual sites on the protein.26,27 6. Other Experimental Methods Fourier transform infrared (FTIR) vibrational spectroscopy senses the hydrogen bonding pattern of the peptide bonds of a protein and can detect unfolding transitions in terms of changes in the secondary structure patterns.28 As compared to CD, which also senses secondary structure, FTIR is relatively more responsive to β-sheet structures. A disadvantage of FTIR is that it requires a higher protein concentration and that it is more difficult to automate for titration experiments. Light-scattering methods, such as small-angle X-ray scattering, or quasi-elastic light scattering, can provide information about the size of a protein, in terms of its radius of gyration. Unfolding or aggregation reactions are detected as increases in the hydrodynamic radius.29 These scattering methods are also relatively difficult to adapt to temperature or titration experiments. The ability of a protein to bind a specific ligand or to have catalytic activity can be used to determine the population of native species. The possibilities are numerous, depending on the way that activity and binding are measured. These activity/binding assays should be easy to automate for a series of denaturant concentrations or pH values. Size exclusion chromatography and gel or capillary electrophoresis are methods that separate protein molecules based on size (or size and charge).30–32 In these methods, the protein sample travels down the column, gel slab, or capillary and, for a pure protein, should exit as a single peak traveling past the detector. Denatured proteins should appear to have a larger hydrodynamic radius and should travel more slowly. If the kinetics of interconversion of the native and unfolded species is slower than the time needed to travel through the column (gel or capillary), then it is possible to detect individual peaks for the native and unfolded species. If the interconversion is rapid, a kinetically averaged peak position will be observed. An example of a potentiometric measurement is one in which the pH (or number of protons bound versus those bound at some reference condition) is measured as a function of the denaturing condition. Such an approach would require a difference in the pK a of one or more amino acid side chains in the native and unfolded state. Usually, several such amino acid side chains are in a protein. However, the potentiometric approach requires technical skill, and it is difficult to use in combination with high concentrations of chemical denaturants or temperatures far from ambient.

Protein Folding

189 [d], S∞,[d] , is the signal at equilibrium (infinite time after initiation of reaction) at the denaturing condition, and Si,[d] is the signal amplitude associated with relaxation time constant, τi . If there is only a single step in the process (i = 1), the reaction will be a mono-exponential. With more steps, the reaction becomes bi-, tri-, . . . exponential, and fitting the above equation to the data can be difficult. Even with the number of terms and the corresponding values of τi determined, it is still a challenge to relate the τi and the associated amplitudes to a reaction mechanism. How this can be done is beyond the scope of this entry and we refer readers to References 35 and 36. To experimentally apply the above equation, it is necessary to monitor some signal that responds to the structural state of the protein, just as is the case in equilibrium studies. For very rapid folding/unfolding reactions, the monitoring method must itself be able to respond more rapidly (with adequate signal to noise) than the chemical reaction. Fluorescence, absorbance, and circular dichroism are fairly rapidly responding optical methods and enjoy widespread use in combination with stopped-flow mixers for this purpose. As indicated in Table II, some of the other methods are not easily adapted for the rapid initiation and continuous monitoring of a reaction’s progress. Obviously, when a reaction occurs on the time scale of minutes to hours, then a hand-mixing experiment can suffice.

D. Kinetics Experiments Studying the kinetics of folding or unfolding a protein involves the rapid initiation of the reaction by either removal of the denaturing condition (to initiate folding) or addition of the denaturing condition (unfolding). The most common way in which this is done is by a stopped-flow mixing device, in which a solution of the protein in one solution condition (e.g., neutral buffer with no chemical denaturant) is rapidly mixed with a second solution (e.g., concentrated chemical denaturant or acid). Temperature or pressure can also be used as the perturbing condition. Upward temperature jumps can be initiated by high-powered laser pulses or electrical discharges in the solution. Downward pressure jumps can be initiated by releasing some type of valve after high pressure has been established. Other ways to rapidly initiate a protein folding or unfolding reaction include such things as laser flash-induced chemical reactions, which can dissociate heme-carbon monoxide bonds of heme proteins. The remaining discussion will emphasize stopped-flow mixing reactions, since these are the most widely available approaches. For small, globular proteins, the kinetics of denaturantinduced or acid-induced unfolding reactions are often found to be described as a mono-exponential process, indicating that there is a single energy barrier between the native and unfolded species. For larger proteins, particularly those with multiple domain structures, one often finds the kinetics to be more complicated. In those cases, the folding or unfolding reaction may be described by more than one exponential decay term. This can be an indication of the existence of multiple, slowly interconverting unfolded states, the existence of multiple energy barriers along the folding/unfolding pathway, the existence of more than one pathway, or the existence of some off-path (or dead-end) species. The challenge in kinetics experiments is to first determine the minimum number of decay terms needed to describe the reaction and to then determine a reaction mechanism consistent with the kinetics data. Often, unique mechanisms cannot be determined and it is the art of the scientist to establish which mechanism is most reasonable for the system being studied. In many cases research will focus on determining the number of intermediates on the folding pathway and in trying to gain structural information about these intermediates. Whether studying folding or unfolding, the reaction should be described with the following general empirical relationship:33,34 S(t)[d] = S∞,[d] + i V. CLOSING COMMENTS
The empirical approaches mentioned above have afforded great insight into the transition from a linear sequence of amino acids into a final tertiary structure. Researchers in the field continue to pioneer new techniques that give insight into the complex inter- and intramolecular interactions that dictate structure and function of proteins. These efforts coupled with computational approaches are helping to reveal the rules that nature uses to create the complex and unique structure of proteins.

SEE ALSO THE FOLLOWING ARTICLES
BIOPOLYMERS • CHEMICAL THERMODYNAMICS • ELECTROPHORESIS • NUCLEAR MAGNETIC RESONANCE • PROTEIN STRUCTURE • TRANSLATION OF RNA TO PROTEIN

Si,[d] · exp(−t /τi )

(13)

BIBLIOGRAPHY
Hynes, T. R., and Fox, R. O. (1991). “The crystal structure of staphylococcal nuclease refined at 1.7 A resolution,” Proteins Struct. Funct. Genet. 10, 92–105.

where S(t)[d] is a time-dependent change in some generalized signal at denaturing condition described by

190
Kraulis , P. J. (1991). “MOLSCRIPT: a program to produce both detailed and schematic plots of protein structures,” J. Appl. Crystallogr. 24, 946–950. Laurents, D. V., and Baldwin, R. L. (1998). “Protein Folding: Matching Theory and Experiment” Biophysical J. 75, 428–434. Privalov, P. L., and Gill, S. J. (1988). “Stability of protein structure and hydrophobic interaction,” Adv. Protein Chem. 39, 191–234. Lin, L. N., and Brandts, J. F. (1988). “Separation of the nativelike intermediate from unfolded forms during refolding of ribonuclease A,” Biochemistry 27, 9037–9042. Schellman, J. A. (1987). “The thermodynamic stability of proteins,” Ann. Rev. Biophys. Biophys. Chem. 16, 115–137. Becktel, W. J., and Schellman, J. A. (1987). “Protein stability curves,” Biopolymers 26, 1859–1876. Privalov, P. L. (1990). “Cold denaturation of proteins,” Crit. Rev. Biochem. 25, 281–305. Schellman, J. A. (1978). “Solvent denaturation,” Biopolymers 17, 1305–1322. Pace, C. N. (1986). “Determination and analysis of urea and guanidine hydrochloride denaturation curves,” Methods Enzymol. 131, 266–280. Santoro, M. M., and Bolen, W. D. (1992). “Unfolding free energy changes determined by the linear extrapolation method. 1. Unfolding of phenylmethanesulfonyl alpha-chymotrypsin using different denaturants,” Biochemistry 31, 8063–8068. Palidini, A. A., and Weber, G. (1981). “Pressure induced reversible dissociation of enolase,” Biochemistry 20, 2587–2593. Silva, J. L., and Weber, G. (1993). “Pressure stability of proteins,” Annu. Rev. Phys. Chem. 44, 89–113. Schmid, F. X. (1989). “Spectral methods of characterizing protein conformation and conformational changes,” In “Protein Structure: A Practical Approach” (T. E. Creighton, ed.), pp. 251–285, IRL Press, Oxford. Andrews, L. J., and Forster, L. S. (1972). “Protein difference spectra. Effect of solvent and charge on tryptophan,” Biochemistry 11, 1875– 1879. Eftink, M. R. (1995). “Use of multiple spectroscopic methods to monitor the equilibrium unfolding of proteins,” Methods Enzymol. 259, 487– 512. Johnson, W. C. (1988). “Secondary structure of proteins through circular dichroism spectroscopy,” Ann. Rev. Biophys. Biophys. Chem. 17, 145– 166. Yu, S., Venyaminiov, S. Y., and Yang, J. T. (1996). “Determination of protein secondary structure,” In “Circular Dichroism and the Conformational Analysis of Biomolecules” (G. D. Fasman, ed.), pp. 69–107, Plenum Press, New York. Eftink, M. R. (1996). “Fluorescence techniques for studying protein structure,” Methods Biochem. Anal. 35, 127–205.

Protein Folding
Eftink, M. R. (1994). “The use of fluorescence to monitor unfolding transitions in proteins,” Biophys. J. 66, 482–501. Lakowicz, J. R. (200). “Protein fluorescence,” In “Topics in Fluorescence Spectroscopy,” Vol. 6, Plenum Press, New York. Haugland, R. P. (1983). “Covalent fluorescent probes,” In “Excited States of Biopolymers” (R. F. Steiner, ed.) pp. 29–58, Plenum Press, New York. Sturtevant., J. M. (1897). “Biochemical applications of differential scanning calorimetry,” Ann. Rev. Phys. Chem. 38, 463–488. Biltonen, R. L., and Freire, E. (1978). “Thermodynamic characterization of conformational states of biological macromolecules using differential scanning calorimetry,” CRC Critical Rev. Biochem. 5, 85–124. Cooper, A., and Johnson, C. M. (1994). “Differential scanning microcalorimetry,” Methods Molec. Biol. 22, 125–136. Roder, H. (1989). “Structural characterization of protein folding intermediates by proton magnetic resonance and hydrogen exchange,” Methods Enzymol. 176, 446–473. Englander, W., and Mayne, L. (1992). “Protein folding studied using hydrogen-exchange labeling and two-dimensional NMR,” Ann. Rev. Biochem. Biophys. Chem. 21, 243–265. Muga, A., Mantsch, H. H., and Surewicz, W. K. (1991). “Membrane binding induces destabilization of cytochrome c structure,” Biochemistry 30, 7219–7224. Nicoli, D. F., and Benedek, G. B. (1976). “Study of thermal denaturation of lysozyme and other glubular proteins by light-scattering spectroscopy,” Biopolymers 15, 2421–2437. Corbett, R. T., and Roche, R. S. (1984). “Use of high-speed sizeexclusion chromatography for the study of protein folding and stability,” Biochemistry 23, 1888–1894. Creighton, T. E. (1980). “Kinetic study of protein unfolding and refolding using urea gradient electrophoresis,” J. Mol. Biol. 137, 61–80. Hilser, V. J., Worosila, G. D., and Freire, E. (1993). “Analysis of thermally induced protein folding/unfolding transitions using free solution capillary electrophoresis,” Anal. Biochem. 208, 125–131. Dandliker, W. B., Dandliker, J., Levison, S. A., Kelley, R. J., Hicks, A. N., and White, J. U. (1978). “Fluorescence methods for measuring reaction equilibria and kinetics,” Methods Enzymol. 48, 380–415. Eftink, M. R., and Shastry, M. R. C. (1997). “Fluorescence methods for studying kinetics of protein folding reactions,” Methods Enzymol. 278, 258–287. Utiyama, H., and Baldwin, R. L. (1986). “Kinetic mechanisms of protein folding,” Methods Enzymol. 131, 51–70. Schmid, F. X. (1992). “Kinetics of unfolding and refolding of singledomain proteins,” In “Protein Folding” (T. E. Creighton, ed.), pp. 197–241, W. H. Freeman, New York.

Protein Structure
Ivan Rayment
University of Wisconsin, Madison I. II. III. IV. V. VI. VII. VIII. IX. X. XI. Introduction Amino Acids Protein Structure Determination Structural Hierachy Secondary Structural Motifs Protein Stability Tertiary Structure Membrane Proteins Superfamilies and Structural Evolution Quaternary Structure Conclusions

GLOSSARY
Primary structure Linear sequence of amino acids in a polypeptide. Quaternary structure Arrangement of polypeptides in macromolecular assembly. Secondary structure A description of the three-dimensional structure adopted by a localized sections of the polypeptide chain. Tertiary structure A description of the arrangement of secondary structural elements within the protein. ˚ Units of length 1 A = 0.1 nm.

mer. Most proteins adopt a well-defined three-dimensional structure to fulfil their biological role. Thus, knowledge of protein structure is central to understanding the molecular basis of life.

I. INTRODUCTION
Proteins mediate the majority of biological processes. All proteins share the common feature that they are condensation polymers of amino acids whose sequence is specified by the genetic information contained within the genome of the organism. Complete DNA sequences for organisms ranging from Escherichia coli to humans suggest that the total number of proteins necessary for life lies in the range of 4200–50,000, although the number of genes in higher organisms is still under debate. Most of these proteins

PROTEINS are linear polymers of amino acids linked by amide bonds where the biological function is dictated by the sequence of amino acid residues within the poly-

191

192 adopt a well-defined three-dimensional structure in solution that is essential for protein function. Indeed unfolding or denaturation of a protein typically leads to a loss of biological activity. The amino acid sequence of a protein contains all of the information necessary to dictate its final threedimensional structure or fold. In many cases small proteins can be unfolded and refolded in vitro without loss of activity. In more complex proteins, chaparones are frequently necessary to allow a protein to reach its properly folded or correct three-dimensional state. Chaparones in these instances recognize an incorrectly unfolded protein and provide an energetically favorable pathway, through the hydrolysis of ATP, for the protein to unfold and refold to reach its functional state. Even in these cases, the structure of the protein is dictated by its amino acid sequence. In principle, it should be possible to deduce the structure of a protein from its amino acid sequence. At this time, it is not possible to perform ab initio structure prediction with any great success. As such protein structure prediction remains one of the major problems in biology. Progress in structure prediction has been made through the combination of sequence and structural similarities. This offers hope that with the knowledge of sufficient structures across a wide range of organisms it should be possible to generate the structure of all unknown proteins. Although there is still much to be learned about protein structure, a series of fundamental features, folding rules, and structural motifs have been observed in many of the threedimensional structures determined to date. These common features arise as a consequence of the amino acids used to build the proteins, the peptide bonds that join the amino acids, and the thermodynamic factors that control protein stability. These common threads in protein structure are described in the following.

Protein Structure

II. AMINO ACIDS
All proteins are synthesized from the 20 α-amino acids specified by the genetic code as shown in Fig. 1. The nature of an amino acid is determined by the “sidechain” attached to the α-carbon (Table I). All of these amino acids, except for glycine which carries two hydrogens on its α-carbon, have a chiral center located at the α-carbon. Thus the amino acids exist as either the L- or D-isomers. Only the L-stereoisomer is utilized in protein biosynthesis (Fig. 2). This introduces chirality into all protein molecules that is the source of most of the asymmetric features found in protein structures. The use of only one of the two stereoisomers of the amino acids also establishes a structural uniqueness that is essential for biochemical specificity.

There are four classes of amino acids specified by the genetic code: (1) aliphatic amino acids, (2) aromatic amino acids, (3) polar amino acids, and (4) charged amino acids. These groups of amino acids provide the range of properties necessary to create a stable, functional folded protein. As discussed elsewhere the primary driving force in protein folding and protein structure is the hydrophobic effect. This serves to sequester the hydrophobic side chains away from the bulk solvent. Once folded a typical protein is a densely packed entity that contains few holes larger than a water molecule. The aliphatic amino acids which include glycine, alanine, valine, leucine, isoleucine, and proline provide the range of small hydrophobic amino acids necessary to fill the gaps in the interior of the protein. Glycine and proline serve special roles in protein structure. Glycine is the smallest amino acid and is unique because it lacks a side chain. This gives it more conformational freedom than any other amino acid. Glycine is often found in turns and loops where other amino acids would be sterically unacceptable. It is also found where secondary structural elements intersect and other side chains would introduce molecular collisions. In contrast proline is unusual because it is conformationally restricted. As such it is often found in turns since it introduces an inherent kink in the polypeptide chain without any entropic cost to protein folding. Proline is also unique in that it is the only amino acid (or technically an “imino acid”) that is commonly found to form a cis peptide bond between itself and the residue that precedes it in the polypeptide chain. In this instance the energy barrier to rotation is considerably less than all other peptide bonds (13 kcal/mol vs ∼20 kcal/mol). This post-translational conformational modification often represents a slow step in protein folding. Phenylalanine, tyrosine, and tryptophan are large aromatic residues that are normally found buried in the interior of a protein and are important for protein stability. Tyrosine has special properties since its hydroxyl side chain may function as a powerful nucleophile in an enzyme active site (when ionized) and is a common site for phosphorylation in cell signaling cascades. Tryptophan has the largest side chain and is the least common amino acid in proteins. It has spectral properties that make it the best inherent probe for following protein folding and conformational changes associated with biochemical processes. The polar amino acids include serine, threonine, cysteine, methionine, asparagine, and glutamine. These are an important class of amino acids since they provide many of the functional groups found in proteins. Serine often serves as a nucleophile in many enzyme active sites, and is best known for its role in the serine proteases. Both serine and threonine are sites of phosphorylation and glycosylation which are important for enzyme regulation and

Protein Structure

193

Nonpolar Aliphatic Amino Acids Side chains
Glycine Alanine Valine H Aspartate CH3 CH3 CH CH3 CH3 Leucine CH2 CH CH3 Isoleucine CH CH3 Proline + H 2N CH2 CH2 Arginine CH CH2 COOHistidine CH2 CH3 Lysine Glutamate

Negatively Charged
O CH2 C O O CH2 CH2 C O

Positively Charged + NH3 NH2 CH2 CH2 CH2 H NH C + NH2

CH2

CH2

CH2

CH2

CH2 N H

+

N

Polar Uncharged
Serine CH2 OH CH3 Threonine CH OH Cysteine CH2 SH Phenylalanine CH2

Aromatic Amino Acids Side chains

Methionine

CH2

CH2 O

S

CH3

Tryosine

CH2

OH

Asparagine

CH2

C NH2 O NH2

Tryptophan

CH2

NH

Glutamine

CH2

CH2

C

FIGURE 1 The 20 amino acid side chains specified by the genetic code. All except glycine have a β-carbon. Proline is technically an imino acid since it is a secondary amine.

cell signaling. Cysteine is the most reactive amino acid side chain. It serves as a potent nucleophile and metal ligand (particularly for iron and zinc), but is best known for its ability to form disulfide bonds, which often make an important contribution to the stability of extracellular proteins. Methionine is a fairly hydrophobic amino acid and typically found buried within the interior of a pro-

tein. It can form stacking interactions with the aromatic moieties of tryptophan, phenylalanine, and tyrosine. Asparagine and glutamine are close relatives of aspartate and glutamate but differ in the lack of charge and altered hydrogen bonding characteristics. In general these are not very reactive residues; however, asparagine is a common site for glycosylation.

194
TABLE I Properties of the Amino Acids Amino acid Aliphatic Glycine Alanine Valine Leucine Isoleucine Proline Aromatic Phenylalanine Tyrosine Tryptophan Polar uncharged Serine Threonine Cysteine Methionine Asparagine Glutamine Positively charged Lysine Arginine Histidine Negatively charged Aspartate Glutamate Side chain pK a Occurrence

Protein Structure

Gly Ala Val Leu Ile Pro Phe Tyr Trp Ser Thr Cys Met Asn Gln Lys Arg His Asp Glu

G A V L I P F Y W S T C M N Q K R H D E

7.2 7.8 6.6 9.1 5.3 5.2 3.9 3.2 1.4 6.8 5.9 1.9 2.2 4.3 4.3 5.9 5.1 2.3 5.3 6.3

Lysine, arginine, and histidine can carry a positive charge. Of these, arginine is constitutively positively charged since its pK a lies around 12.5. Lysine also plays an important role in coordinating negatively charged ligands; however, it functions as a nucleophile in some enzyme catalyzed reactions. Histidine is perhaps the most common and versatile catalytic residue in proteins. Its pK a of ∼6.0 allows it to function both as a catalytic acid or base at physiological pH depending on its local environment. Histidine also has the ability to form covalent intermediates during catalysis such as phosphohistidine. In addition, it is often a ligand for transition metal ions such as iron and zinc. A. Post-Translational Modifications Once synthesized and folded, many proteins undergo post-translational modifications before they reach a functional state. Over 200 variant amino acid residues have been identified in proteins thus far. These changes are almost always achieved through an enzymatic pathway. The simplest changes include the formation of disulfide bonds (discussed later) and proteolytic processing of the polypeptide chain to yield a functional protein. Examples of proteolytic processing include the removal of signal peptides, the activation of zymogens to generate active forms of many proteolytic enzymes, and the maturation of viral proteins. Additionally proteolytic processing occurs in the biosynthetic pathway of many hormones. Other simple changes include the glycosylation of asparagine, serine, threonine, and phosphorylation of serine and tyrosine. It is noteworthy that many post-translation modifications are associated with a sequence motif such that it is frequently possible to identify potential sites directly from the amino acid sequence. This arises because most post-translational modifications are the result of enzymatic pathways, which are usually highly specific. Thus protein sequences inferred from DNA sequence are often annotated with sites for post-translational modification. These sites should be viewed with caution since proof of

10.5

∼13 ∼13 8.4

10.5 12.5 6.0 3.9 4.1

The charged amino acids include aspartate, glutamate, lysine, arginine and histidine. As a group these amino acids are relatively abundant and are important for making proteins soluble. Thus, these residues are generally located on the surface of the protein unless they play a specific biological role. Aspartate and glutamate are negatively charged amino acids. Both of these residues can function as general acids or bases in enzyme catalyzed reactions. Likewise they are important metal ion ligands.

COO+

COO+

H3N

C CH3

H

H

C CH3

NH3

L-Alanine

D-Alanine

FIGURE 2 Stereoisomers of L-alanine and D-alanine.

Protein Structure

195 structure prediction. Although considerable progress has been made in recent years toward establishment of a comprehensive structural database many more protein models are needed before structures can be predicted with a high degree of confidence. There are two methods by which protein structures can be determined: X-ray crystallography and NMR. These techniques are complementary, with each having its advantages for providing information about specific aspects of protein structure. A detailed description of these methods is beyond the scope of this summary, but a few comments are noteworthy A. X-Ray Crystallography The first structure of a protein, myoglobin, was determined by X-ray crystallography in 1958 and was followed soon thereafter by the structure of hemoglobin. At that time protein structure determination was a daunting undertaking and few structures were determined in the ensuing years. Fortunately continual developments in the fundamental understanding of X-ray crystallographic theory, data collection, and computational methods have made the determination of protein structure routine. The result of this approach is an electron density map, which is interpreted in terms of a molecular model. The strength of this technique is that it can be applied to any macromolecular assembly that can be crystallized. The overwhelming majority of structures in the protein databank have been determined by X-ray crystallography. The limiting factor in a successful X-ray structure determination is the growth of high quality crystals. In general if suitable crystals can be obtained a three-dimensional structure will be determined. The final quality of an X-ray structure is directly dependent on the three-dimensional order of the crystals since X-ray crystallography is an imaging technique. This is usually indicated by the “resolution” of the data. Resolution refers to the minimum diffraction spacing included in the structural determination where a smaller the number corresponds to a better ˚ structure. Typically a structure at 2.8 A resolution is satisfactory to determine the path of the polypeptide chain, but ˚ data better than 2.5 A are required to define the hydrogen bonding pattern in a protein with great confidence. The one concern leveled at X-ray structures is the influence of the crystalline lattice on the observed conformation of the protein. Fortunately it has been demonstrated repeatedly that the structures of proteins observed in crystalline lattice are consistent with most of the biochemical measurements on the same protein. This arises because protein crystals typically contain about 50% solvent such that very little of a protein molecule is in contact with its neighbors in the crystal lattice and the packing forces are thermodynamically small. In some cases proteins are

modification can only be obtained through chemical or physical experimentation. B. Cofactors Although many proteins derive all of their function from their constitutive amino acids, a large number of proteins require additional cofactors in order to fulfil their biological role. These cofactors provide chemical properties that are not present in the 20 amino acid residues. For example, none of the amino acids are capable of facilitating an oxidation/reduction reaction. A wide range of cofactors are utilized including inorganic ions such as Fe2+ , Mg2+ , and Zn2+ , or complex organic molecules that are normally described as coenzymes such as flavin adenine dinucleotide or nicotinamide adenine dinucleotide. If the coenzyme is covalently bound to the enzyme it is often called a prosthetic group. The complete enzyme is called a holoenzyme, whereas the protein in the absence of its cofactors is called an apoenzyme or apoprotein. Many apoproteins are considerably less stable in the absence of their cofactor. This suggests that although the amino acid sequence dictates the overall threedimensional structure, the cofactor is an integral part of the protein. C. Context Determines Function Typically proteins and especially enzymes contain only a few residues that are absolutely vital for function. In contrast there are usually many other residues of the same type in the protein that do not fulfill any special role. In many cases the catalytic residues have different chemical and physical properties from the same amino acid in solution; for example, the pK a of the side chain might be several pH units higher or lower than the free amino acid. It is generally found that the behavior of an amino acid is profoundly influenced by the context of that amino acid within the protein. Altering the chemical properties of a functional group is one of the major attributes of protein structure and appears to be essential for the activity of most enzymes. There are many ways in which this is achieved; however, a simple example is placement of a charged residue in the interior of a protein such that the deionized state is favored. This serves to raise the pK a of aspartate and glutamate and lower the pK a of lysine.

III. PROTEIN STRUCTURE DETERMINATION
A major requirement for understanding protein structure is a large database of three-dimensional structures. This is particularly important for the comparative method of

196 enzymatically active in the lattice. In others conformational changes are observed between the substrate-free and substrate-bound forms of the enzyme. Typically this requires the crystallization of site-directed mutant proteins complexed with the substrate(s) or the study of complexes with substrate analogs. Except for the use of Laue techniques, protein crystallography yields a time-averaged view of the protein structure. Careful analysis of accurate X-ray diffraction data may provide some indication of conformational flexibility, but that aspect of protein structure is best suited to spectroscopic techniques such as NMR. B. NMR The use of NMR to determine protein structures is a more recent development than X-ray diffraction. It has the advantage that the analysis can be performed in the solution state of the protein which removes any artifacts introduced by crystallization. Its major disadvantage is the size limitation, which restricts most analyses to smaller proteins (