The extraordinary ability of an enzyme to catalyze only one particular reaction is a quality known as specificity (Chapter 14). Specificity means an enzyme acts only on a specific substance, its substrate, invariably transforming it into a specific product. That is, an enzyme binds only certain compounds, and then, only a specific reaction ensues. Some enzymes show absolute specificity, catalyzing the transformation of only one specific substrate to yield a unique product. Other enzymes carry out a particular reaction but act on a class of compounds. For example, hexokinase (ATP : hexose-6-phosphotransferase) will carry out the ATP-dependent phosphorylation of a number of hexoses at the 6-position, including glucose.

15.1 • Specificity Is the Result of Molecular Recognition

An enzyme molecule is typically orders of magnitude larger than its substrate. Its active site comprises only a small portion of the overall enzyme structure. The active site is part of the conformation of the enzyme molecule arranged to create a special pocket or cleft whose three-dimensional structure is complementary to the structure of the substrate. The enzyme and the substrate molecules “recognize” each other through this structural complementarity. The substrate binds to the enzyme through relatively weak forces — H bonds, ionic bonds (salt bridges), and van der Waals interactions between sterically complementary clusters of atoms. Specificity studies on enzymes entail an examination of the rates of the enzymatic reaction obtained with various structural analogs of the substrate. By determining which functional and structural groups within the substrate affect binding or catalysis, enzymologists can map the properties of the active site, analyzing questions such as: Can it accommodate sterically bulky groups? Are ionic interactions between E and S important? Are H bonds formed?

The “Lock and Key” Hypothesis

Pioneering enzyme specificity studies at the turn of the century by the great organic chemist Emil Fischer led to the notion of an enzyme resembling a “lock” and its particular substrate the “key.” This analogy captures the essence of the specificity that exists between an enzyme and its substrate, but enzymes are not rigid templates like locks.

The “Induced Fit” Hypothesis

Enzymes are highly flexible, conformationally dynamic molecules, and many of their remarkable properties, including substrate binding and catalysis, are due to their structural pliancy. Realization of the conformational flexibility of proteins led Daniel Koshland to hypothesize that the binding of a substrate (S) by an enzyme is an interactive process. That is, the shape of the enzyme’s active site is actually modified upon binding S, in a process of dynamic recognition between enzyme and substrate aptly called induced fit. In essence, substrate binding alters the conformation of the protein, so that the protein and the substrate “fit” each other more precisely. The process is truly interactive in that the conformation of the substrate also changes as it adapts to the conformation of the enzyme.
            This idea also helps to explain some of the mystery surrounding the enormous catalytic power of enzymes: In enzyme catalysis, precise orientation of catalytic residues comprising the active site is necessary for the reaction to occur; substrate binding induces this precise orientation by the changes it causes in the protein’s conformation.

“Induced Fit” and the Transition-State Intermediate

The catalytically active enzyme:substrate complex is an interactive structure in which the enzyme causes the substrate to adopt a form that mimics the transition-state intermediate of the reaction. Thus, a poor substrate would be one that was less effective in directing the formation of an optimally active enzyme:transition -state intermediate conformation. This active conformation of the enzyme molecule is thought to be relatively unstable in the absence of substrate, and free enzyme thus reverts to a conformationally different state.

Specificity and Reactivity

Consider, for example, why hexokinase catalyzes the ATP-dependent phosphorylation of hexoses but not smaller phosphoryl-group acceptors such as glycerol, ethanol, or even water. Surely these smaller compounds are not sterically forbidden from approaching the active site of hexokinase (Figure 15.1). Indeed, water should penetrate the active site easily and serve as a highly effective phosphoryl-group acceptor. Accordingly, hexokinase should display high ATPase activity. It does not. Only the binding of hexoses induces hexokinase to assume its fully active conformation.

Figure 15.1  •  A drawing, roughly to scale, of H₂O, glycerol, glucose, and an idealized hexokinase molecule. Note the two domains of structure in hexokinase, (a), between which the active site is located. Binding of glucose induces a conformational change in hexokinase. The two domains close together, creating the catalytic site (b). The shaded area in (b) represents solvent-inaccessible surface area in the active site cleft that results when the enzyme binds substrate.

            In Chapter 16, we explore in greater detail the factors that contribute to the remarkable catalytic power of enzymes and examine specific examples of enzyme reaction mechanisms. Here we focus on another essential feature of enzymes: the regulation of their activity.

15.2 • Controls Over Enzymatic Activity—General Considerations

The activity displayed by enzymes is affected by a variety of factors, some of which are essential to the harmony of metabolism.

1.  The enzymatic rate, v = d[ P]/dt, “slows down” as product accumulates and equilibrium is approached. The apparent decrease in rate is due to the conversion of P to S by the reverse reaction as [P] rises. Once [P]/[S] = K_eq, no further reaction is apparent. K_eq defines thermodynamic equilibrium. Enzymes have no influence on the thermodynamics of a reaction. Also, product inhibition can be a kinetically valid phenomenon: Some enzymes are actually inhibited by the products of their action.

2.  The availability of substrates and cofactors will determine the enzymatic reaction rate. In general, enzymes have evolved such that their K_m values approximate the prevailing in vivo concentration of their substrates. (It is also true that the concentration of some enzymes in cells is within an order of magnitude or so of the concentrations of their substrates.)

3.  There are genetic controls over the amounts of enzyme synthesized (or degraded) by cells. If the gene encoding a particular enzyme protein is turned on or off, changes in the amount of enzyme activity soon follow. Induction, which is the activation of enzyme synthesis, and repression, which is the shutdown of enzyme synthesis, are important mechanisms for the regulation of metabolism. By controlling the amount of an enzyme that is present at any moment, cells can either activate or terminate various metabolic routes. Genetic controls over enzyme levels have a response time ranging from minutes in rapidly dividing bacteria to hours (or longer) in higher eukaryotes.

4.  Enzymes can be regulated by covalent modification, the reversible covalent attachment of a chemical group. For example, a fully active enzyme can be converted into an inactive form simply by the covalent attachment of a functional group, such as a phosphoryl moiety (Figure 15.2). Alternatively, some enzymes exist in an inactive state unless specifically converted into the active form through covalent addition of a functional group. Covalent modification reactions are catalyzed by special converter enzymes, which are themselves subject to metabolic regulation. Although covalent modification represents a stable alteration of the enzyme, a different converter enzyme operates to remove the modification, so that when the conditions that favored modification of the enzyme are no longer present, the process can be reversed, restoring the enzyme to its unmodified state. Many examples of covalent modification at important metabolic junctions will be encountered in our discussions of metabolic pathways. Because covalent modification events are enzyme-catalyzed, they occur very quickly, with response times of seconds or even less for significant changes in metabolic activity. The 1992 Nobel Prize in physiology or medicine was awarded to Edmond Fischer and Edwin Krebs for their pioneering studies of reversible protein phosphorylation as an important means of cellular regulation.

Figure 15.2  •  Enzymes regulated by covalent modification are called interconvertible enzymes. The enzymes (protein kinase and protein phosphatase, in the example shown here) catalyzing the conversion of the interconvertible enzyme between its two forms are called converter enzymes. In this example, the free enzyme form is catalytically active, whereas the phosphoryl-enzyme form represents an inactive state. The -OH on the interconvertible enzyme represents an -OH group on a specific amino acid side chain in the protein (for example, a particular Ser residue) capable of accepting the phosphoryl group.

5.  Enzymatic activity can also be activated or inhibited through noncovalent interaction of the enzyme with small molecules (metabolites) other than the substrate. This form of control is termed allosteric regulation, because the activator or inhibitor binds to the enzyme at a site other than (allo means “other”) the active site. Further, such allosteric regulators, or effector molecules, are often quite different sterically from the substrate. Because this form of regulation results simply from reversible binding of regulatory ligands to the enzyme, the cellular response time can be virtually instantaneous.

6.  Specialized controls: Enzyme regulation is an important matter to cells, and evolution has provided a variety of additional options, including zymogens, isozymes, and modulator proteins.

Zymogens

Most proteins become fully active as their synthesis is completed and they spontaneously fold into their native, three-dimensional conformations. Some proteins, however, are synthesized as inactive precursors, called zymogens or proenzymes, that only acquire full activity upon specific proteolytic cleavage of one or several of their peptide bonds. Unlike allosteric regulation or covalent modification, zymogen activation by specific proteolysis is an irreversible process. Activation of enzymes and other physiologically important proteins by specific proteolysis is a strategy frequently exploited by biological systems to switch on processes at the appropriate time and place, as the following examples illustrate.

Insulin.   Some protein hormones are synthesized in the form of inactive precursor molecules, from which the active hormone is derived by proteolysis. For instance, insulin, an important metabolic regulator, is generated by proteolytic excision of a specific peptide from proinsulin (Figure 15.3).

Figure 15.3  •  Proinsulin is an 86-residue precursor to insulin (the sequence shown here is human proinsulin). Proteolytic removal of residues 31 to 65 yields insulin. Residues 1 through 30 (the B chain) remain linked to residues 66 through 87 (the A chain) by a pair of interchain disulfide bridges.

Proteolytic Enzymes of the Digestive Tract.   Enzymes of the digestive tract that serve to hydrolyze dietary proteins are synthesized in the stomach and pancreas as zymogens (Table 15.1).

Only upon proteolytic activation are these enzymes able to form a catalytically active substrate-binding site. The activation of chymotrypsinogen is an interesting example (Figure 15.4). Chymotrypsino-gen is a 245-residue polypeptide chain cross-linked by five disulfide bonds. Chymotrypsinogen is converted to an enzymatically active form called p-chymotrypsin when trypsin cleaves the peptide bond joining Arg¹⁵ and Ile¹⁶. The enzymatically active p-chymotrypsin acts upon other p-chymotrypsin molecules, excising two dipeptides, Ser¹⁴-Arg¹⁵ and Thr¹⁴⁷-Asn¹⁴⁸. The end product of this processing pathway is the mature protease a -chymotrypsin, in which the three peptide chains, A (residues 1 through 13), B (residues 16 through 146), and C (residues 149 through 245), remain together because they are linked by two disulfide bonds, one from A to B, and one from B to C. It is interesting to note that the transformation of inactive chymotrypsinogen to active p-chymotrypsin requires the cleavage of just one particular peptide bond.

Figure 15.4  •  The proteolytic activation of chymotrypsinogen.

Blood Clotting.   The formation of blood clots is the result of a series of zymogen activations (Figure 15.5). The amplification achieved by this cascade of enzymatic activations allows blood clotting to occur rapidly in response to injury. Seven of the clotting factors in their active form are serine proteases: kallikrein, XII_a, XI_a, IX_a, VII_a, X_a, and thrombin. Two routes to blood clot formation exist. The intrinsic pathway is instigated when the blood comes into physical contact with abnormal surfaces caused by injury; the extrinsic pathway is initiated by factors released from injured tissues. The pathways merge at Factor X and culminate in clot formation. Thrombin excises peptides rich in negative charge from fibrinogen, converting it to fibrin, a molecule with a different surface charge distribution. Fibrin readily aggregates into ordered fibrous arrays that are subsequently stabilized by covalent cross-links. Thrombin specifically cleaves Arg-Gly peptide bonds and is homologous to trypsin, which is also a serine protease (recall that trypsin acts only at Arg and Lys residues).

Figure 15.5  •  The cascade of activation steps leading to blood clotting. The intrinsic and extrinsic pathways converge at Factor X, and the final common pathway involves the activation of thrombin and its conversion of fibrinogen into fibrin, which aggregates into ordered filamentous arrays that become cross-linked to form the clot.

Isozymes

A number of enzymes exist in more than one quaternary form, differing in their relative proportions of structurally equivalent but catalytically distinct polypeptide subunits. A classic example is mammalian lactate dehydrogenase (LDH), which exists as five different isozymes, depending on the tetrameric association of two different subunits, A and B: A₄, A₃B, A₂B₂, AB₃, and B₄ (Figure 15.6). The kinetic properties of the various LDH isozymes differ in terms of their relative affinities for the various substrates and their sensitivity to inhibition by product. Different tissues express different isozyme forms, as appropriate to their particular metabolic needs. By regulating the relative amounts of A and B subunits they synthesize, the cells of various tissues control which isozymic forms are likely to assemble, and, thus, which kinetic parameters prevail.

Figure 15.6  •  The isozymes of lactate dehydrogenase (LDH). Active muscle tissue becomes anaerobic and produces pyruvate from glucose via glycolysis (Chapter 19). It needs LDH to regenerate NAD⁺ from NADH so glycolysis can continue. The lactate produced is released into the blood. The muscle LDH isozyme (A₄) works best in the NAD⁺-regenerating direction. Heart tissue is aerobic and uses lactate as a fuel, converting it to pyruvate via LDH and using the pyruvate to fuel the citric acid cycle to obtain energy. The heart LDH isozyme (B₄) is inhibited by excess pyruvate so the fuel won’t be wasted.

Modulator Proteins

Modulator proteins are yet another way that cells mediate metabolic activity. Modulator proteins are proteins that bind to enzymes, and by binding, influence the activity of the enzyme. For example, some enzymes, such as cAMP-dependent protein kinase (Chapter 23), exist as dimers of catalytic subunits and regulatory subunits. These regulatory subunits are modulator proteins that suppress the activity of the catalytic subunits. Dissociation of the regulatory subunits (modulator proteins) activates the catalytic subunits; reassociation once again suppresses activity  (Figure 15.7). Phosphoprotein phosphatase inhibitor-1 (PPI-1) is another example of a modulator protein. When PPI-1 is phosphorylated on one of its serine residues, it binds to phosphoprotein phosphatase (Figure 15.2), inhibiting its phosphatase activity. The result is an increased phosphorylation of the interconvertible enzyme targeted by the protein kinase/phosphoprotein phosphatase cycle (Figure 15.2). We will meet other important representatives of this class as the processes of metabolism unfold in subsequent chapters. For now, let us focus our attention on the fascinating kinetics of allosteric enzymes.

Figure 15.7  •  Cyclic AMp- dependent protein kinase (also known as PKA) is a 150- to 170-kD R₂C₂ tetramer in mammalian cells. The two R (regulatory) subunits bind cAMP (K_D = 3 x 10^-8 M); cAMP binding releases the R subunits from the C (catalytic) subunits. C subunits are enzymatically active as monomers.

15.3 • The Allosteric Regulation of Enzyme Activity

Allosteric regulation acts to modulate enzymes situated at key steps in metabolic pathways. Consider as an illustration the following pathway, where A is the precursor for formation of an end product, F, in a sequence of five enzyme-catalyzed reactions:

               enz 1      enz 2      enz 3    enz 4     enz 5
            A   →    B   →    C   →    D  →    E   →    F

In this scheme, F symbolizes an essential metabolite, such as an amino acid or a nucleotide. In such systems, F, the essential end product, inhibits enzyme 1, the first step in the pathway. Therefore, when sufficient F is synthesized, it blocks further synthesis of itself. This phenomenon is called feedback inhibition or feedback regulation.

General Properties of Regulatory Enzymes

Enzymes such as enzyme 1, which are subject to feedback regulation, represent a distinct class of enzymes, the regulatory enzymes. As a class, these enzymes have certain exceptional properties:

1.  Their kinetics do not obey the Michaelis - Menten equation. Their v versus [S] plots yield sigmoid- or S-shaped curves rather than rectangular hyperbolas (Figure 15.8). Such curves suggest a second-order (or higher) relationship between v and [S]; that is, v is proportional to [S]ⁿ, where n > 1. A qualitative description of the mechanism responsible for the S-shaped curves is that binding of one S to a protein molecule makes it easier for additional substrate molecules to bind to the same protein molecule. In the jargon of allostery, substrate binding is cooperative.

Figure 15.8  •  Sigmoid v versus [S] plot. The dotted line represents the hyperbolic plot characteristic of normal Michaelis - Menten-type enzyme kinetics.

2.  Inhibition of a regulatory enzyme by a feedback inhibitor does not conform to any normal inhibition pattern, and the feedback inhibitor F bears little structural similarity to A, the substrate for the regulatory enzyme. F apparently acts at a binding site distinct from the substrate-binding site. The term allosteric is apt, because F is sterically dissimilar and, moreover, acts at a site other than the site for S. Its effect is called allosteric inhibition.

3.  Regulatory or allosteric enzymes like enzyme 1 are, in some instances, regulated by activation. That is, whereas some effector molecules such as F exert negative effects on enzyme activity, other effectors show stimulatory, or positive, influences on activity.

4.  Allosteric enzymes have an oligomeric organization. They are composed of more than one polypeptide chain (subunit) and have more than one S-binding site per enzyme molecule.

5.  The working hypothesis is that, by some means, interaction of an allosteric enzyme with effectors alters the distribution of conformational possibilities or subunit interactions available to the enzyme. That is, the regulatory effects exerted on the enzyme’s activity are achieved by conformational changes occurring in the protein when effector metabolites bind.

            In addition to enzymes, noncatalytic proteins may exhibit many of these properties; hemoglobin is the classic example. The allosteric properties of hemoglobin are the subject of a Special Focus beginning on page 480.

15.4 • Models for the Allosteric Behavior of Proteins

The Symmetry Model of Monod, Wyman, and Changeux

In 1965, Jacques Monod, Jeffries Wyman, and Jean-Pierre Changeux proposed a theoretical model of allosteric transitions based on the observation that allosteric proteins are oligomers. They suggested that allosteric proteins can exist in (at least) two conformational states, designated R, signifying “relaxed,” and T, or “taut,” and that, in each protein molecule, all of the subunits have the same conformation (either R or T). That is, molecular symmetry is conserved. Molecules of mixed conformation (having subunits of both R and T states) are not allowed by this model.

            In the absence of ligand, the two states of the allosteric protein are in equilibrium:

R₀ ⇌ T₀

(Note that the subscript “0” signifies “in the absence of ligand.”) The equilibrium constant is termed L: L = T₀/R₀. L is assumed to be large; that is, the amount of the protein in the T conformational state is much greater than the amount in the R conformation. Let us suppose that L = 10⁴.

            The affinities of the two states for substrate, S, are characterized by the respective dissociation constants, K_R and K_T. The model supposes that K_T >> K_R. That is, the affinity of R₀ for S is much greater than the affinity of T₀ for S. Let us choose the extreme where K_R/K_T = 0 (that is, K_T is infinitely greater than K_R). In effect, we are picking conditions where S binds only to R. (If K_T is infinite, T does not bind S.)

            Given these parameters, consider what happens when S is added to a solution of the allosteric protein at conformational equilibrium (Figure 15.9). Although the relative [R₀] concentration is small, S will bind “only” to R₀, forming R₁. This depletes the concentration of R₀, perturbing the T₀/R₀ equilibrium. To restore equilibrium, molecules in the T₀ conformation undergo a transition to R₀. This shift renders more R₀ available to bind S, yielding R1, diminishing [R₀], perturbing the T₀/R₀ equilibrium, and so on. Thus, these linked equilibria (Figure 15.9) are such that S-binding by the R₀ state of the allosteric protein perturbs the T₀/R₀ equilibrium with the result that S-binding drives the conformational transition, T₀ → R₀.

Figure 15.9  •  Monod - Wyman - Changeux (MWC) model for allosteric transitions. Consider a dimeric protein that can exist in either of two conformational states, R or T. Each subunit in the dimer has a binding site for substrate S and an allosteric effector site, F. The promoters are symmetrically related to one another in the protein, and symmetry is conserved regardless of the conformational state of the protein. The different states of the protein, with or without bound ligand, are linked to one another through the various equilibria. Thus, the relative population of protein molecules in the R or T state is a function of these equilibria and the concentration of the various ligands, substrate (S), and effectors (which bind at F_R or F_T). As [S] is increased, the T/R equilibrium shifts in favor of an increased proportion of R-conformers in the total population (that is, more protein molecules in the R conformational state).

            In just this simple system, cooperativity is achieved because each subunit has a binding site for S, and thus, each protein molecule has more than one binding site for S. Therefore, the increase in the population of R conformers gives a progressive increase in the number of sites available for S. The extent of cooperativity depends on the relative T₀/R₀ ratio and the relative affinities of R and T for S. If L is large (that is, the equilibrium lies strongly in favor of T₀) and if K_T >> K_R, as in the example we have chosen, cooperativity is great (Figure 15.10). Ligands such as S here that bind in a cooperative manner, so that binding of one equivalent enhances the binding of additional equivalents of S to the same protein molecule, are termed positive homotropic effectors. (The prefix “homo” indicates that the ligand influences the binding of like molecules.)

Figure 15.10  •  The Monod - Wyman - Changeux model. Graphs of allosteric effects for a tetramer (n = 4) in terms of Y, the saturation function, versus [S]. Y is defined as [ligand-binding sites that are occupied by ligand]/[ total ligand-binding sites]. (a) A plot of Y as a function of [S], at various L values. (b) Y as a function of [S], at different c, where c = K_R/K_T. (When c = 0, K_T is infinite.) (Adapted from Monod, J., Wyman, J., and Changeux, J.-P., 1965. On the nature of allosteric transitions: A plausible model. Journal of Molecular Biology 12:92.)

Heterotropic Effectors

This simple system also provides an explanation for the more complex substrate-binding responses to positive and negative effectors. Effectors that influence the binding of something other than themselves are termed heterotropic effectors. For example, effectors that promote S binding are termed positive heterotropic effectors or allosteric activators. Effectors that diminish S binding are negative heterotropic effectors or allosteric inhibitors. Feedback inhibitors fit this class. Consider a protein composed of two subunits, each of which has two binding sites: one for the substrate, S, and one to which allosteric effectors bind, the allosteric site. Assume that S binds preferentially (“only”) to the R conformer; further assume that the positive heterotropic effector, A, binds to the allosteric site only when the protein is in the R conformation, and the negative allosteric effector, I, binds at the allosteric site only if the protein is in the T conformation. Thus, with respect to binding at the allosteric site, A and I are competitive with each other.

Positive Effectors

If A binds to R₀, forming the new species R_{1( A)}, the relative concentration of R₀ is decreased and the T₀/R₀ equilibrium is perturbed (Figure 15.11). As a consequence, a relative T₀ → R₀ shift occurs in order to restore equilibrium. The net effect is an increase in the number of R conformers in the presence of A, meaning that more binding sites for S are available. For this reason, A leads to a decrease in the cooperativity of the substrate saturation curve, as seen by a shift of this curve to the left (Figure 15.11). Effectively, the presence of A lowers the apparent value of L.

Figure 15.11  •  Heterotropic allosteric effects: A and I binding to R and T, respectively. The linked equilibria lead to changes in the relative amounts of R and T and, therefore, shifts in the substrate saturation curve. This behavior, depicted by the graph, defines an allosteric “K” system. The parameters of such a system are: (1) S and A (or I) have different affinities for R and T and (2) A (or I) modifies the apparent K_0.5 for S by shifting the relative R versus T population.

Negative Effectors

The converse situation applies in the presence of I, which binds “only” to T. I-binding will lead to an increase in the population of T conformers, at the expense of R₀ (Figure 15.11). The decline in [R₀] means that it is less likely for S (or A) to bind. Consequently, the presence of I increases the cooperativity (that is, the sigmoidicity) of the substrate saturation curve, as evidenced by the shift of this curve to the right (Figure 15.11). The presence of I raises the apparent value of L.

K Systems and V Systems

The allosteric model just presented is called a K system because the concentration of substrate giving half-maximal velocity, defined as K_0.5, changes in response to effectors (Figure 15.11). Note that V_max is constant in this system.
            An allosteric situation where K_0.5 is constant but the apparent V_max changes in response to effectors is termed a V system. In a V system, all v versus S plots are hyperbolic rather than sigmoid (Figure 15.12). The positive heterotropic effector A activates by raising V_max, whereas I, the negative heterotropic effector, decreases it. Note that neither A nor I affects K_0.5. This situation arises if R and T have the same affinity for the substrate, S, but differ in their catalytic ability and their affinities for A and I. A and I thus can shift the relative T/R distribution. Acetyl-coenzyme A carboxylase, the enzyme catalyzing the committed step in the fatty acid biosynthetic pathway, behaves as a V system in response to its allosteric activator, citrate (see Chapter 25).

Figure 15.12  •  v versus [S] curves for an allosteric “V” system. The V system fits the model of Monod, Wyman, and Changeux, given the following conditions: (1) R and T have the same affinity for the substrate, S. (2) The effectors A and I have different affinities for R and T and thus can shift the relative T/R distribution. (That is, A and I change the apparent value of L.) Assume as before that A binds “only” to the R state and I binds “only” to the T state. (3) R and T differ in their catalytic ability. Assume that R is the enzymatically active form, whereas T is inactive. Because A perturbs the T/R equilibrium in favor of more R, A increases the apparent V_max. I favors transition to the inactive T state.

K Systems and V Systems Fill Different Biological Roles

The K and V systems have design features that mean they work best under different physiological situations. “K system” enzymes are adapted to conditions in which the prevailing substrate concentration is rate-limiting, as when [S] in vivo » K_0.5. On the other hand, when the physiological conditions are such that [S] is usually saturating for the regulatory enzyme of interest, the enzyme conforms to the “V system” mode in order to have an effective regulatory response.

15.5 • Glycogen Phosphorylase: Allosteric Regulation and Covalent Modification

The Glycogen Phosphorylase Reaction

The cleavage of glucose units from the nonreducing ends of glycogen molecules is catalyzed by glycogen phosphorylase, an allosteric enzyme. The enzymatic reaction involves phosphorolysis of the bond between C-1 of the departing glucose unit and the glycosidic oxygen, to yield glucose-1-phosphate and a glycogen molecule that is shortened by one residue (Figure 15.13).

Figure 15.13  •  The glycogen phosphorylase reaction.

(Because the reaction involves attack by phosphate instead of H₂O, it is referred to as a phosphorolysis rather than a hydrolysis.) Phosphorolysis produces a phosphorylated sugar product, glucose-1-P, which is converted to the glycolytic substrate, glucose-6-P, by phosphoglucomutase (Figure 15.14). In muscle, glucose-6-P proceeds into glycolysis, providing needed energy for muscle contraction. In liver, hydrolysis of glucose-6-P yields glucose, which is exported to other tissues via the circulatory system.

Figure 15.14  •  The phosphoglucomutase reaction.

The Structure of Glycogen Phosphorylase

Muscle glycogen phosphorylase is a dimer of two identical subunits (842 residues, 97.44 kD ). Each subunit contains a pyridoxal phosphate cofactor, covalently linked as a Schiff base to Lys⁶⁸⁰. Each subunit contains an active site (at the center of the subunit) and an allosteric effector site near the subunit interface (Figure 15.15). In addition, a regulatory phosphorylation site is located at Ser¹⁴ on each subunit. A glycogen-binding site on each subunit facilitates prior association of glycogen phosphorylase with its substrate and also exerts regulatory control on the enzymatic reaction.

Figure 15.15  •  (a) The structure of a glycogen phosphorylase monomer, showing the locations of the catalytic site, the PLP cofactor site, the allosteric effector site, the glycogen storage site, the tower helix (residues 262 through 278), and the subunit interface. (b) Glycogen phosphorylase dimer.

            Each subunit contributes a tower helix (residues 262 to 278) to the subunit - subunit contact interface in glycogen phosphorylase. In the phosphorylase dimer, the tower helices extend from their respective subunits and pack against each other in an antiparallel manner.

Regulation of Glycogen Phosphorylase by Allosteric Effectors

Muscle Glycogen Phosphorylase Shows Cooperativity in Substrate Binding

The binding of the substrate inorganic phosphate (P_i) to muscle glycogen phosphorylase is highly cooperative (Figure 15.16a), which allows the enzyme activity to increase markedly over a rather narrow range of substrate concentration. P_i is a positive homotropic effector with regard to its interaction with glycogen phosphorylase.

Figure 15.16  •  v versus S curves for glycogen phosphorylase. (a) The sigmoid response of glycogen phosphorylase to the concentration of the substrate phosphate (P_i) shows strong positive cooperativity. (b) ATP is a feedback inhibitor that affects the affinity of glycogen phosphorylase for its substrates but does not affect V_max. (Glucose-6-P shows similar effects on glycogen phosphorylase.) (c) AMP is a positive heterotropic effector for glycogen phosphorylase. It binds at the same site as ATP. AMP and ATP are competitive. Like ATP, AMP affects the affinity of glycogen phosphorylase for its substrates, but does not affect V_max.

ATP and Glucose-6-P Are Allosteric Inhibitors of Glycogen Phosphorylase

ATP can be viewed as the “end product” of glycogen phosphorylase action, in that the glucose-1-P liberated by glycogen phosphorylase is degraded in muscle via metabolic pathways whose purpose is energy (ATP) production. Glucose-1-P is readily converted into glucose-6-P to feed such pathways. (In the liver, glucose-1-P from glycogen is converted to glucose and released into the bloodstream to raise blood glucose levels.) Thus, feedback inhibition of glycogen phosphorylase by ATP and glucose-6-P provides a very effective way to regulate glycogen breakdown. Both ATP and glucose-6-P act by decreasing the affinity of glycogen phosphorylase for its substrate P_i (Figure 15.16b). Because the binding of ATP or glucose-6-P has a negative effect on substrate binding, these substances act as negative heterotropic effectors. Note in Figure 15.16b that the substrate saturation curve is displaced to the right in the presence of ATP or glucose-6-P, and a higher substrate concentration is needed to achieve half-maximal velocity (V_max/2). When concentrations of ATP or glucose-6-P accumulate to high levels, glycogen phosphorylase is inhibited; when [ATP] and [glucose-6-P] are low, the activity of glycogen phosphorylase is regulated by availability of its substrate, P_i.

AMP Is an Allosteric Activator of Glycogen Phosphorylase

AMP also provides a regulatory signal to glycogen phosphorylase. It binds to the same site as ATP, but it stimulates glycogen phosphorylase rather than inhibiting it (Figure 15.16c). AMP acts as a positive heterotropic effector, meaning that it enhances the binding of substrate to glycogen phosphorylase. Significant levels of AMP indicate that the energy status of the cell is low and that more energy (ATP) should be produced. Reciprocal changes in the cellular concentrations of ATP and AMP and their competition for binding to the same site (the allosteric site) on glycogen phosphorylase, with opposite effects, allow these two nucleotides to exert rapid and reversible control over glycogen phosphorylase activity. Such reciprocal regulation ensures that the production of energy (ATP) is commensurate with cellular needs.
            To summarize, muscle glycogen phosphorylase is allosterically activated by AMP and inhibited by ATP and glucose-6-P; caffeine can also act as an allosteric inhibitor (Figure 15.17). When ATP and glucose-6-P are abundant, glycogen breakdown is inhibited. When cellular energy reserves are low (i.e., high [AMP] and low [ATP] and [G-6-P]), glycogen catabolism is stimulated.

Figure 15.17  •  The mechanism of covalent modification and allosteric regulation of glycogen phosphorylase. The T states are blue and the R states blue-green.

            Glycogen phosphorylase conforms to the Monod - Wyman - Changeux model of allosteric transitions, with the active form of the enzyme designated the R state and the inactive form denoted as the T state (Figure 15.17). Thus, AMP promotes the conversion to the active R state, whereas ATP, glucose-6-P, and caffeine favor conversion to the inactive T state.
                        X-ray diffraction studies of glycogen phosphorylase in the presence of allosteric effectors have revealed the molecular basis for the T ⇌ R conversion. Although the structure of the central core of the phosphorylase subunits is identical in the T and R states, a significant change occurs at the subunit interface between the T and R states. This conformation change at the subunit interface is linked to a structural change at the active site that is important for catalysis. In the T state, the negatively charged carboxyl group of Asp²⁸³ faces the active site, so that binding of the anionic substrate phosphate is unfavorable. In the conversion to the R state, Asp²⁸³ is displaced from the active site and replaced by Arg⁵⁶⁹. The exchange of negatively charged aspartate for positively charged arginine at the active site provides a favorable binding site for phosphate. These allosteric controls serve as a mechanism for adjusting the activity of glycogen phosphorylase to meet normal metabolic demands. However, in crisis situations in which abundant energy (ATP) is needed immediately, these controls can be overridden by covalent modification of glycogen phosphorylase. Covalent modification through phosphorylation of Ser¹⁴ in glycogen phosphorylase converts the enzyme from a less active, allosterically regulated form (the b form) to a more active, allosterically unresponsive form (the a form). Covalent modification is like a “permanent” allosteric transition that is independent of [allosteric effector], such as AMP.

Regulation of Glycogen Phosphorylase by Covalent Modification

As early as 1938, it was known that glycogen phosphorylase existed in two forms: the less active phosphorylase b and the more active phosphorylase a. In 1956, Edwin Krebs and Edmond Fischer reported that a “converting enzyme” could convert phosphorylase b to phosphorylase a. Three years later, Krebs and Fischer demonstrated that the conversion of phosphorylase b
            Phosphorylation of Ser¹⁴ causes a dramatic conformation change in phosphorylase. Upon phosphorylation, the amino-terminal end of the protein (including residues 10 through 22) swings through an arc of 120°, moving into the subunit interface (Figure 15.18). This conformation change moves Ser¹⁴ by more than 3.6 nm.

Figure 15.18  •  In this diagram of the glycogen phosphorylase dimer, the phosphorylation site (Ser14) and the allosteric (AMP) site face the viewer. Access to the catalytic site is from the opposite side of the protein. The diagram shows the major conformational change that occurs in the N-terminal residues upon phosphorylation of Ser¹⁴. The solid black line shows the conformation of residues 10 to 23 in the b, or unphosphorylated, form of glycogen phosphorylase. The conformational change in the location of residues 10 to 23 upon phosphorylation of Ser¹⁴ to give the a (phosphorylated) form of glycogen phosphorylase is shown in yellow. Note that these residues move from intrasubunit contacts into intersubunit contacts at the subunit interface. (Sites on the two respective subunits are denoted, with those of the upper subunit designated by primes (9).)

(Adapted from Johnson, L. N., and Barford, D., 1993. The effects of phosphorylation on the structure and function of proteins. Annual Review of Biophysics and Biomolecular Structure 22:199-232.)

            Dephosphorylation of glycogen phosphorylase is carried out by phosphoprotein phosphatase 1. The action of phosphoprotein phosphatase 1 inactivates glycogen phosphorylase.

Enzyme Cascades Regulate Glycogen Phosphorylase

Figure 15.19  •  The hormone-activated enzymatic cascade that leads to activation of glycogen phosphorylase.

The phosphorylation reaction that activates glycogen phosphorylase is mediated by an enzyme cascade (Figure 15.19). The first part of the cascade leads to hormonal stimulation (described in the next section) of adenylyl cyclase, a membrane-bound enzyme that converts ATP to adenosine-3',5' -cyclic monophosphate, denoted as cyclic AMP or simply cAMP (Figure 15.20). This regulatory molecule is found in all eukaryotic cells and acts as an intracellular messenger molecule, controlling a wide variety of processes. Cyclic AMP is known as a second messenger because it is the intracellular agent of a hormone (the “first messenger”). (The myriad cellular roles of cyclic AMP are described in detail in Chapter 34.)

Figure 15.20  •  The adenylyl cyclase reaction yields 3',5' -cyclic AMP and pyrophosphate. The reaction is driven forward by subsequent hydrolysis of pyrophosphate by the enzyme inorganic pyrophosphatase.

            The hormonal stimulation of adenylyl cyclase is effected by a transmembrane signaling pathway consisting of three components, all membrane-associated . Binding of hormone to the external surface of a hormone receptor causes a conformational change in this transmembrane protein, which in turn stimulates a GTP-binding protein (abbreviated G protein). G proteins are heterotrimeric proteins consisting of a - (45 - 47 kD ), b- (35 kD), and g - (7 - 9 kD) subunits. The a -subunit binds GDP or GTP and has an intrinsic, slow GTPase activity. In the inactive state, the G _{a b g} complex has GDP at the nucleotide site. When a G protein is stimulated by a hormone-receptor complex, GDP dissociates and GTP binds to G _a , causing it to dissociate from G_{b g} and to associate with adenylyl cyclase (Figure 15.21). Binding of G _a (GTP) activates adenylyl cyclase to form cAMP from ATP. However, the intrinsic GTPase activity of G _a eventually hydrolyzes GTP to GDP, leading to dissociation of G _a (GDP) from adenylyl cyclase and reassociation with G _{a b g} to form the inactive G _{a b g} complex. This cascade amplifies the hormonal signal because a single hormone-receptor complex can activate many G proteins before the hormone dissociates from the receptor, and because the G _a -activated adenylyl cyclase can synthesize many cAMP molecules before bound GTP is hydrolyzed by G _a . More than 100 different G protein - coupled receptors and at least 21 distinct G _a proteins are known (Chapter 34).

Figure 15.21  •  Hormone (H) binding to its receptor (R) creates a hormone;receptor complex (H:R) that catalyzes GDP-GTP exchange on the a -subunit of the heterotrimer G protein (G a b g ), replacing GDP with GTP. The G a -subunit with GTP bound dissociates from the b g -subunits and binds to adenylyl cyclase (AC). AC becomes active upon association with G a :GTP and catalyzes the formation of cAMP from ATP. With time, the intrinsic GTPase activity of the G a -subunit hydrolyzes the bound GTP, forming GDP; this leads to dissociation of G a :GDP from AC, reassociation of G a with the b g subunits, and cessation of AC activity. AC and the hormone receptor H are integral plasma membrane proteins; G a and Gb g are membrane-anchored proteins.

            Cyclic AMP is an essential activator of cAMP-dependent protein kinase (PKA). This enzyme is normally inactive because its two catalytic subunits (C) are strongly associated with a pair of regulatory subunits (R), which serve to block activity. Binding of cyclic AMP to the regulatory subunits induces a conformation change that causes the dissociation of the C monomers from the R dimer (Figure 15.7). The free C subunits are active and can phosphorylate other proteins. One of the many proteins phosphorylated by PKA is phosphorylase kinase (Figure 15.19). Phosphorylase kinase is inactive in the unphosphorylated state and active in the phosphorylated form. As its name implies, phosphorylase kinase functions to phosphorylate (and activate) glycogen phosphorylase. Thus, stimulation of adenylyl cyclase leads to activation of glycogen breakdown.

Special Focus:
Hemoglobin and Myoglobin — Paradigms of Protein Structure and Function

Ancient life forms evolved in the absence of oxygen and were capable only of anaerobic metabolism. As the earth’s atmosphere changed over time, so too did living things. Indeed, the production of O₂ by photosynthesis was a major factor in altering the atmosphere. Evolution to an oxygen-based metabolism was highly beneficial. Aerobic metabolism of sugars, for example, yields far more energy than corresponding anaerobic processes. Two important oxygen-binding proteins appeared in the course of evolution so that aerobic metabolic processes were no longer limited by the solubility of O₂ in water. These proteins are represented in animals as hemoglobin (Hb) in blood and myoglobin (Mb) in muscle. Because hemoglobin and myoglobin are two of the most-studied proteins in nature, they have become paradigms of protein structure and function. Moreover, hemoglobin is a model for protein quaternary structure and allosteric function. The binding of O₂ by hemoglobin, and its modulation by effectors such as protons, CO₂, and 2,3 -bisphosphoglycerate, depend on interactions between subunits in the Hb tetramer. Subunit  - subunit interactions in Hb reveal much about the functional significance of quaternary associations and allosteric regulation.

The Comparative Biochemistry of Myoglobin and Hemoglobin

A comparison of the properties of hemoglobin and myoglobin offers insights into allosteric phenomena, even though these proteins are not enzymes. Hemoglobin displays sigmoid-shaped O₂-binding curves (Figure 15.22). The unusual shape of these curves was once a great enigma in biochemistry. Such curves closely resemble allosteric enzyme:substrate saturation graphs (see Figure 15.8). In contrast, myoglobin’s interaction with oxygen obeys classical Michaelis - Menten-type substrate saturation behavior.

Figure 15.22  •   O₂-binding curves for hemoglobin and myoglobin.

            Before examining myoglobin and hemoglobin in detail, let us first encapsulate the lesson: Myoglobin is a compact globular protein composed of a single polypeptide chain 153 amino acids in length; its molecular mass is 17.2 kD (Figure 15.23). It contains heme, a porphyrin ring system complexing an iron ion, as its prosthetic group (see Figure 5.15). Oxygen binds to Mb via its heme. Hemoglobin (Hb) is also a compact globular protein, but Hb is a tetramer. It consists of four polypeptide chains, each of which is very similar structurally to the myoglobin polypeptide chain, and each bears a heme group. Thus, a hemoglobin molecule can bind four O₂ molecules. In adult human Hb, there are two identical chains of 141 amino acids, the a-chains, and two identical b-chains, each of 146 residues. The human Hb molecule is an a₂b₂-type tetramer of molecular mass 64.45 kD. The tetrameric nature of Hb is crucial to its biological function: When a molecule of O₂ binds to a heme in Hb, the heme Fe ion is drawn into the plane of the porphyrin ring. This slight movement sets off a chain of conformational events that are transmitted to adjacent subunits, dramatically enhancing the affinity of their heme groups for O₂. That is, the binding of O₂ to one heme of Hb makes it easier for the Hb molecule to bind additional equivalents of O₂. Hemoglobin is a marvelously constructed molecular machine. Let us dissect its mechanism, beginning with its monomeric counterpart, the myoglobin molecule.

Figure 15.23  •  The myoglobin and hemoglobin molecules. Myoglobin (sperm whale): one polypeptide chain of 153 aa residues (mass 5 17.2 kD) has one heme (mass = 652 D) and binds one O₂. Hemoglobin (human): four polypeptide chains, two of 141 aa residues ( a ) and two of 146 residues (b); mass = 64.45 kD. Each polypeptide has a heme; the Hb tetramer binds four O₂. (Irving Geis)

Myoglobin

Myoglobin is the oxygen-storage protein of muscle. The muscles of diving mammals such as seals and whales are especially rich in this protein, which serves as a store for O₂ during the animal’s prolonged periods underwater. Myoglobin is abundant in skeletal and cardiac muscle of nondiving animals as well. Myoglobin is the cause of the characteristic red color of muscle.

The Mb Polypeptide Cradles the Heme Group

The myoglobin polypeptide chain is folded to form a cradle (4.4 x 4.4 x 2.5 nm) that nestles the heme prosthetic group (Figure 15.24). O₂ binding depends on the heme’s oxidation state. The iron ion in the heme of myoglobin is in the +2 oxidation state, that is, the ferrous form. This is the form that binds O₂. Oxidation of the ferrous form to the +3 ferric form yields metmyoglobin, which will not bind O₂. It is interesting to note that free heme in solution will readily interact with O₂ also, but the oxygen quickly oxidizes the iron atom to the ferric state. Fe³⁺:protoporphyrin IX is referred to as hematin. Thus, the polypeptide of myoglobin may be viewed as serving three critical functions: it cradles the heme group, it protects the heme iron atom from oxidation, and it provides a pocket into which the O₂ can fit.

Figure 15.24  •  Detailed structure of the myoglobin molecule. The myoglobin polypeptide chain consists of eight helical segments, designated by the letters A through H, counting from the N-terminus. These helices, ranging in length from 7 to 26 residues, are linked by short, unordered regions that are named for the helices they connect, as in the AB region or the EF region. The individual amino acids in the polypeptide are indicated according to their position within the various segments, as in His F8, the eighth residue in helix F, or Phe CD1, the first amino acid in the interhelical CD region. Occasionally, amino acids are specified in the conventional way, that is, by the relative position in the chain, as in Gly¹⁵³. The heme group is cradled within the folded polypeptide chain. (Irving Geis)

O₂ Binding to Mb

Iron ions, whether ferrous or ferric, prefer to interact with six ligands, four of which share a common plane. The fifth and sixth ligands lie above and below this plane (see Figure 15.25). In heme, four of the ligands are provided by the nitrogen atoms of the four pyrroles. A fifth ligand is donated by the imidazole side chain of amino acid residue His F8. When myoglobin binds O₂ to become oxymyoglobin, the O₂ molecule adds to the heme iron ion as the sixth ligand (Figure 15.25). O₂ adds end on to the heme iron, but it is not oriented perpendicular to the plane of the heme. Rather, it is tilted about 60° with respect to the perpendicular.

Figure 15.25  •  The six liganding positions of an iron ion. Four ligands lie in the same plane; the remaining two are, respectively, above and below this plane. In myoglobin, His F8 is the fifth ligand; in oxymyoglobin, O₂ becomes the sixth.

In deoxymyoglobin, the sixth ligand position is vacant, and in metmyoglobin, a water molecule fills the O₂ site and becomes the sixth ligand for the ferric atom. On the oxygen-binding side of the heme lies another histidine residue, His E7. While its imidazole function lies too far away to interact with the Fe atom, it is close enough to contact the O₂. Therefore, the O₂-binding site is a sterically hindered region. Biologically important properties stem from this hindrance. For example, the affinity of free heme in solution for carbon monoxide (CO) is 25,000 times greater than its affinity for O₂. But CO only binds 250 times more tightly than O₂ to the heme of myoglobin, because His E7 forces the CO molecule to tilt away from a preferred perpendicular alignment with the plane of the heme (Figure 15.26). This diminished affinity of myoglobin for CO guards against the possibility that traces of CO produced during metabolism might occupy all of the heme sites, effectively preventing O₂ from binding. Nevertheless, CO is a potent poison and can cause death by asphyxiation.

Figure 15.26  •  Oxygen and carbon monoxide binding to the heme group of myoglobin.

O₂ Binding Alters Mb Conformation

What happens when the heme group of myoglobin binds oxygen? X-ray crystallography has revealed that a crucial change occurs in the position of the iron atom relative to the plane of the heme. In deoxymyoglobin, the ferrous ion has but five ligands, and it lies 0.055 nm above the plane of the heme, in the direction of His F8. The iron:porphyrin complex is therefore dome-shaped. When O₂ binds, the iron atom is pulled back toward the porphyrin plane and is now displaced from it by only 0.026 nm (Figure 15.27). The consequences of this small motion are trivial as far as the biological role of myoglobin is concerned. However, as we shall soon see, this slight movement profoundly affects the properties of hemoglobin. Its action on His F8 is magnified through changes in polypeptide conformation that alter subunit interactions in the Hb tetramer. These changes in subunit relationships are the fundamental cause of the allosteric properties of hemoglobin.

Figure 15.27  •  The displacement of the Fe ion of the heme of deoxymyoglobin from the plane of the porphyrin ring system by the pull of His F8. In oxymyoglobin, the bound O₂ counteracts this effect.

The Physiological Significance of Cooperative

Binding of Oxygen by Hemoglobin

The oxygen-binding equations for myoglobin and hemoglobin are described in detail in the Appendix at the end of this chapter. The relative oxygen affinities of hemoglobin and myoglobin reflect their respective physiological roles (see Figure 15.22). Myoglobin, as an oxygen storage protein, has a greater affinity for O₂ than hemoglobin at all oxygen pressures. Hemoglobin, as the oxygen carrier, becomes saturated with O₂ in the lungs, where the partial pressure of O₂ (pO₂) is about 100 torr.¹ In the capillaries of tissues, pO₂ is typically 40 torr, and oxygen is released from Hb. In muscle, some of it can be bound by myoglobin, to be stored for use in times of severe oxygen deprivation, such as during strenuous exercise.

The Structure of the Hemoglobin Molecule

As noted, hemoglobin is an a ₂ b₂ tetramer. Each of the four subunits has a conformation virtually identical to that of myoglobin. Two different types of subunits, a and b, are necessary to achieve cooperative O₂-binding by Hb. The b-chain at 146 amino acid residues is shorter than the myoglobin chain (153 residues), mainly because its final helical segment (the H helix) is shorter. The a -chain (141 residues) also has a shortened H helix and lacks the D helix as well (Figure 15.28).

Figure 15.28  •  Conformational drawings of the a - and b-chains of Hb and the myoglobin chain. (Irving Geis)

Max Perutz, who has devoted his life to elucidating the atomic structure of Hb, noted very early in his studies that the molecule was highly symmetrical. The actual arrangement of the four subunits with respect to one another is shown in Figure 15.29 for horse methemoglobin. All vertebrate hemoglobins show a three-dimensional structure essentially the same as this. The subunits pack in a tetrahedral array, creating a roughly spherical molecule 6.4 x 5.5 x 5.0 nm. The four heme groups, nestled within the easily recognizable cleft formed between the E and F helices of each polypeptide, are exposed at the surface of the molecule. The heme groups are quite far apart; 2.5 nm separates the closest iron ions, those of hemes a ₁ and b₂, and those of hemes a ₂ and b₁. The subunit interactions are mostly between dissimilar chains: each of the a -chains is in contact with both b-chains, but there are few α-α or b-b interactions.

Figure 15.29  •  The arrangement of subunits in horse methemoglobin, the first hemoglobin whose structure was determined by X-ray diffraction. The iron atoms on metHb are in the oxidized, ferric (Fe³⁺) state. (Irving Geis)

Oxygenation Markedly Alters the Quaternary Structure of Hb

Crystals of deoxyhemoglobin shatter when exposed to O₂. Further, X-ray crystallographic analysis reveals that oxy- and deoxyhemoglobin differ markedly in quaternary structure. In particular, specific a b-subunit interactions change. The ab contacts are of two kinds. The a ₁ b₁ and a ₂ b₂ contacts involve helices B, G, and H and the GH corner. These contacts are extensive and important to subunit packing; they remain unchanged when hemoglobin goes from its deoxy to its oxy form. The a ₁ b₂  anda ₂ b₁  contacts are called sliding contacts. They principally involve helices C and G and the FG corner (Figure 15.30).

Figure 15.30  •  Side view of one of the two a b dimers in Hb, with packing contacts indicated in blue. The sliding contacts made with the other dimer are shown in yellow. The changes in these sliding contacts are shown in Figure 15.31. (Irving Geis)

When hemoglobin undergoes a conformational change as a result of ligand binding to the heme, these contacts are altered (Figure 15.31). Hemoglobin, as a conformationally dynamic molecule, consists of two dimeric halves, an a ₁ b₁-subunit pair and an a ₂ b₂-subunit pair. Each a b dimer moves as a rigid body, and the two halves of the molecule slide past each other upon oxygenation of the heme. The two halves rotate some 15° about an imaginary pivot passing through the a b-subunits; some atoms at the interface between a b dimers are relocated by as much as 0.6 nm.

Figure 15.31  •   Subunit motion in hemoglobin when the molecule goes from the (a) deoxy to the (b) oxy form. (Irving Geis)

Movement of the Heme Iron by Less Than 0.04 nm Induces the Conformational Change in Hemoglobin

In deoxyhemoglobin, histidine F8 is liganded to the heme iron ion, but steric constraints force the Fe²⁺:His -N bond to be tilted about 8° from the perpendicular to the plane of the heme. Steric repulsion between histidine F8 and the nitrogen atoms of the porphyrin ring system, combined with electrostatic repulsions between the electrons of Fe²⁺ and the porphyrin p-electrons, forces the iron atom to lie out of the porphyrin plane by about 0.06 nm. Changes in electronic and steric factors upon heme oxygenation allow the Fe²⁺ atom to move about 0.039 nm closer to the plane of the porphyrin, so now it is displaced only 0.021 nm above the plane. It is as if the O₂ were drawing the heme Fe²⁺ into the porphyrin plane (Figure 15.32). This modest displacement of 0.039 nm seems a trivial distance, but its biological consequences are far-reaching. As the iron atom moves, it drags histidine F8 along with it, causing helix F, the EF corner, and the FG corner to follow. These shifts are transmitted to the subunit interfaces, where they trigger conformational readjustments that lead to the rupture of interchain salt links.

Figure 15.32  •  Changes in the position of the heme iron atom upon oxygenation lead to conformational changes in the hemoglobin molecule.

The Oxy and Deoxy Forms of Hemoglobin Represent

Two Different Conformational States

Hemoglobin resists oxygenation (see Figure 15.22) because the deoxy form is stabilized by specific hydrogen bonds and salt bridges (ion-pair bonds). All of these interactions are broken in oxyhemoglobin, as the molecule stabilizes into a new conformation. A crucial H bond in this transition involves a particular tyrosine residue. Both a - and b-subunits have Tyr as the penultimate C-terminal residue (Tyr a 140 = Tyr HC2; Tyr b145 = Tyr HC2, respectively²). The phenolic -OH groups of these Tyr residues form intrachain H bonds to the peptide C=O function contributed by Val FG5 in deoxyhemoglobin. (Val FG5 is a 93 and b98, respectively.) The shift in helix F upon oxygenation leads to rupture of this Tyr HC2:Val FG5 hydrogen bond. Further, eight salt bridges linking the polypeptide chains are broken as hemoglobin goes from the deoxy to the oxy form (Figure 15.33). Six of these salt links are between different subunits. Four of these six involve either carboxyl-terminal or amino-terminal amino acids in the chains; two are between the amino termini and the carboxyl termini of the a -chains, and two join the carboxyl termini of the b-chains to the e -NH31 groups of the two Lys a 140 residues. The other two interchain electrostatic bonds link Arg and Asp residues in the two a -chains. In addition, ionic interactions between Asp b94 and His b146 form an intrachain salt bridge in each b-subunit. In deoxyhemoglobin, with all of these interactions intact, the C-termini of the four subunits are restrained, and this conformational state is termed T, the tense or taut form. In oxyhemoglobin, these C-termini have almost complete freedom of rotation, and the molecule is now in its R, or relaxed, form.

Figure 15.33  •  Salt bridges between different subunits in hemoglobin. These noncovalent, electrostatic interactions are disrupted upon oxygenation. Arga¹⁴¹ and Hisb¹⁴⁶ are the C-termini of the a - and b-polypeptide chains. (a) The various intrachain and interchain salt links formed among the a - and b-chains of deoxyhemoglobin. (b) A focus on those salt bridges and hydrogen bonds involving interactions between N-terminal and C-terminal residues in the a -chains. Note the Cl^- ion, which bridges ionic interactions between the N-terminus of a ₂ and the R group of Arga¹⁴¹. (c) A focus on the salt bridges and hydrogen bonds in which the residues located at the C-termini of b-chains are involved. All of these links are abolished in the deoxy to oxy transition. (Irving Geis)

A Model for the Allosteric Behavior of Hemoglobin

A model for the allosteric behavior of hemoglobin is based on recent observations that oxygen is accessible only to the heme groups of the a -chains when hemoglobin is in the T conformational state. Perutz has pointed out that the heme environment of b-chains in the T state is virtually inaccessible because of steric hindrance by amino acid residues in the E helix. This hindrance disappears when the hemoglobin molecule undergoes transition to the R conformational state. Binding of O₂ to the b-chains is thus dependent on the T to R conformational shift, and this shift is triggered by the subtle changes that occur when O₂ binds to the a -chain heme groups.

H+ Promotes the Dissociation of Oxygen from Hemoglobin

Protons, carbon dioxide, and chloride ions, as well as the metabolite 2,3 -bisphosphoglycerate (or BPG), all affect the binding of O₂ by hemoglobin. Their effects have interesting ramifications, which we shall see as we discuss them in turn. Deoxyhemoglobin has a higher affinity for protons than oxyhemoglobin. Thus, as the pH decreases, dissociation of O₂ from hemoglobin is enhanced. In simple symbolism, ignoring the stoichiometry of O₂ or H⁺ involved:

            HbO₂ + H⁺ ⇌  HbH⁺ + O₂

Expressed another way, H⁺ is an antagonist of oxygen binding by Hb, and the saturation curve of Hb for O₂ is displaced to the right as acidity increases (Figure 15.34). This phenomenon is called the Bohr effect , after its discoverer, the Danish physiologist Christian Bohr (the father of Niels Bohr, the atomic physicist). The effect has important physiological significance because actively metabolizing tissues produce acid, promoting O₂ release where it is most needed. About two protons are taken up by deoxyhemoglobin. The N-termini of the two a -chains and the His b146 residues have been implicated as the major players in the Bohr effect . (The pK_a of a free amino terminus in a protein is about 8.0, but the pK_a of a protein histidine imidazole is around 6.5.) Neighboring carboxylate groups of Asp b94 residues help to stabilize the protonated state of the His b146 imidazoles that occur in deoxyhemoglobin. However, when Hb binds O₂, changes in the conformation of b-chains upon Hb oxygenation move the negative Asp function away, and dissociation of the imidazole protons is favored.

Figure 15.34  •  The oxygen saturation curves for myoglobin and for hemoglobin at five different pH values: 7.6, 7.4, 7.2, 7.0, and 6.8.

CO₂ Also Promotes the Dissociation of O₂ from Hemoglobin

Carbon dioxide has an effect on O₂ binding by Hb that is similar to that of H⁺, partly because it produces H⁺ when it dissolves in the blood:

The enzyme carbonic anhydrase promotes the hydration of CO₂. Many of the protons formed upon ionization of carbonic acid are picked up by Hb as O₂ dissociates. The bicarbonate ions are transported with the blood back to the lungs. When Hb becomes oxygenated again in the lungs, H1 is released and reacts with HCO₃^- to re-form H₂CO₃, from which CO₂ is liberated. The CO₂ is then exhaled as a gas.

            In addition, some CO2 is directly transported by hemoglobin in the form of carbamate (—NHCOO^-). Free a -amino groups of Hb react with CO₂ reversibly:

R—NH₂ + CO₂ ⇌ R—NH—COO^- + H⁺

This reaction is driven to the right in tissues by the high CO₂ concentration; the equilibrium shifts the other way in the lungs where [CO₂] is low. Thus, carbamylation of the N-termini converts them to anionic functions, which then form salt links with the cationic side chains of Arg a 141 that stabilize the deoxy or T state of hemoglobin.

            In addition to CO₂, Cl^- and BPG also bind better to deoxyhemoglobin than to oxyhemoglobin, causing a shift in equilibrium in favor of O₂ release. These various effects are demonstrated by the shift in the oxygen saturation curves for Hb in the presence of one or more of these substances (Figure 15.35). Note that the O₂-binding curve for Hb + BPG + CO₂ fits that of whole blood very well.

2,3 -Bisphosphoglycerate Is an Important

Allosteric Effector for Hemoglobin

Figure 15.35  •  Oxygen-binding curves of blood and of hemoglobin in the absence and presence of CO2 and BPG. From left to right: stripped Hb, Hb + CO₂, Hb + BPG, Hb + BPG + CO₂, and whole blood.

The binding of 2,3 -bisphosphoglycerate (BPG) to Hb promotes the release of O₂ (Figure 15.35). Erythrocytes (red blood cells) normally contain about 4.5 mM BPG, a concentration equivalent  to that of tetrameric hemoglobin molecules. Interestingly, this equivalence is maintained in the Hb:BPG binding stoichiometry because the tetrameric Hb molecule has but one binding site for BPG. This site is situated within the central cavity formed by the association of the four subunits. The strongly negative BPG molecule (Figure 15.36) is electrostatically bound via interactions with the positively charged functional groups of each Lys b82, His b2, His b143, and the NH₃⁺-terminal group of each b-chain.

Figure 15.36  •  The structure, in ionic form, of BPG or 2,3 -bisphosphoglycerate, an important allosteric effector for hemoglobin.

These positively charged residues are arranged to form an electrostatic pocket complementary to the conformation and charge distribution of BPG (Figure 15.37). In effect, BPG cross-links the two b-subunits. The ionic bonds between BPG and the two b-chains aid in stabilizing the conformation of Hb in its deoxy form, thereby favoring the dissociation of oxygen. In oxyhemoglobin, this central cavity is too small for BPG to fit. Or, to put it another way, the conformational changes in the Hb molecule that accompany O₂-binding perturb the BPG-binding site so that BPG can no longer be accommodated. Thus, BPG and O₂ are mutually exclusive allosteric effectors for Hb, even though their binding sites are physically distinct.

Figure 15.37  •  The ionic binding of BPG to the two b-subunits of Hb. (Irving Geis)

The Physiological Significance of BPG Binding

The importance of the BPG effect is evident in Figure 15.35. Hemoglobin stripped of BPG is virtually saturated with O₂ at a pO₂ of only 20 torr, and it cannot release its oxygen within tissues, where the pO₂  is typically 40 torr. BPG shifts the oxygen saturation curve of Hb to the right, making the Hb an O₂ delivery system eminently suited to the needs of the organism. BPG serves this vital function in humans, most primates, and a number of other mammals. However, the hemoglobins of cattle, sheep, goats, deer, and other animals have an intrinsically lower affinity for O₂, and these Hbs are relatively unaffected by BPG. In fish, whose erythrocytes contain mitochondria, the regulatory role of BPG is filled by ATP or GTP. In reptiles and birds, a different organophosphate serves, namely inositol pentaphosphate (IPP) or inositol hexaphosphate (IHP) (Figure 15.38).

Figure 15.38  •  The structures of inositol pentaphosphate and inositol hexaphosphate, the functional analogs of BPG in birds and reptiles.

Fetal Hemoglobin Has a Higher Affinity for O2

Because It Has a Lower Affinity for BPG

The fetus depends on its mother for an adequate supply of oxygen, but its circulatory system is entirely independent. Gas exchange takes place across the placenta. Ideally then, fetal Hb should be able to absorb O₂ better than maternal Hb so that an effective transfer of oxygen can occur. Fetal Hb differs from adult Hb in that the b-chains are replaced by very similar, but not identical, 146-residue subunits called g -chains (gamma chains). Fetal Hb is thus a ₂ g ₂ . Recall that BPG functions through its interaction with the b-chains. BPG binds less effectively with the g -chains of fetal Hb (also called Hb F). (Fetal g -chains have Ser instead of His at position 143, and thus lack two of the positive charges in the central BPG-binding cavity.) Figure 15.39 compares the relative affinities of adult Hb (also known as Hb A) and Hb F for O₂ under similar conditions of pH and [BPG]. Note that Hb F binds O₂ at pO₂ values where most of the oxygen has dissociated from Hb A. Much of the difference can be attributed to the diminished capacity of Hb F to bind BPG (compare Figures 15.35 and 15.39); Hb F thus has an intrinsically greater affinity for O₂, and oxygen transfer from mother to fetus is ensured.