Skip to Main Content
Skip Nav Destination

Enzymes are the all-purpose catalysts that make the Chemistry of Life run smoothly and efficiently. They do the sorts of things that chemists want to do, under the mildest, “greenest” conditions – in aqueous solution near pH 7, at atmospheric pressure and temperatures close to ambient. They are wonderfully efficient catalysts, capable of handling with ease the most unreactive compounds present in biological systems, and their reactions, where necessary, are completely chemo-, regio- and stereoselective. Small wonder that an understanding of how enzymes work has been an ambition for generations of researchers in a whole range of disciplines, from pure enzymology, through almost all of chemistry to X-ray crystallography.

An understanding of the principles of enzyme catalysis is of far more than academic interest. The industrial use of enzymes is widespread and growing.1  The food industry has always used the enzymes present in various organisms such as yeasts, but a growing trend is to use isolated enzymes at key stages to improve quality control. Medical applications have become hugely important, both in diagnostic testing and directly in therapeutic applications. And nowhere is the need to understand the fundamental principles more important than in the development of artificial enzymes, which have far-reaching potential in the fine chemicals and pharmaceutical industries. Thus, asymmetric synthesis is a major activity and growth area for organic and pharmaceutical chemistry, and chiral catalysis the most elegant – and most efficient – way of achieving it. Enzymes are chiral catalysts par excellence, and natural (wild-type) or specifically modified enzymes play increasingly important roles.

As proteins, enzymes do, however, have certain practical disadvantages outside their native organisms: they are often denatured inconveniently fast, by changes in pH, heat or solvent, and by surfactants and many other chemicals. And the typical enzyme works best on just one specific substrate, in water, and at concentrations that are inconveniently low for serious synthesis. Hence the interest in developing synthetic “artificial enzymes”: which can be more robust, can work in a solvent or solvents of choice; and could in principle be designed to catalyze a particular reaction, rather than a particular reaction of a specific substrate. Last but not least, a major advantage of synthetic systems is that they can in principle be designed to catalyze any reaction of interest, including non-natural reactions, for which no natural enzymes exist. Successful design in this context will inevitably be based on developing enzyme models.

Enzymes are far more than just highly evolved catalysts for specific reactions: they may also have to recognize and respond to molecules other than their specific substrate and product, as part of the control mechanisms of the cell. The evolution of artificial enzymes is at a much more primitive stage, with efficient catalysis the primary, and often the sole, objective. Systems are known that model various other functions, including potential control mechanisms. But to be useful as an industrial catalyst an artificial enzyme has no need of sophisticated built-in feedback control mechanisms or high substrate specificity: a stable molecule that is an efficient catalyst for a key target reaction in a chemical reactor will not be required to select its substrate from many hundreds in the same solution, as enzymes routinely must in the cell. So, a rational design strategy is indeed to consider simply those features of enzymes that are essential for catalytic efficiency.

In these first two chapters we discuss enzyme mechanisms in rather general terms, to identify and define these key features. We then go on to discuss the developing range of enzyme models: by which we mean systems designed to test basic ideas on enzyme mechanism by reproducing specific, key features of enzyme reactions; and attempts to develop them into enzyme-like catalysts. We reserve the term enzyme mimics for the most highly developed enzyme models, which combine successfully more than one of these key features, and catalyze reactions by mechanisms that are demonstrably enzyme-like, involving both binding and catalysis. An enzyme mimic that can do all this, and achieve turnover at a reasonable rate, deserves to be called an artificial enzyme.

Enzymes are proteins. The “central dogma” of biological chemistry underlines the pivotal role of enzyme catalysis (and highlights a fascinating problem in biochemical evolution!):

graphic

Enzyme proteins are made up of one (sometimes more than one) polypeptide chain (Figure 1.1), each of which is folded into a flexible, more or less unique active conformation. The preferred 3-dimensional structure is determined by a complex array of physicochemical interactions between side-chains, main-chain amide groups and especially solvent water. Whole books and half a dozen current journals deal specifically with protein chemistry, and the basic ideas are described in many textbooks. So, only those properties of special relevance to catalysis will be introduced in this chapter, and discussed in the necessary detail later in the book. Specific suggestions for further reading are to be found at the end of each chapter.

Figure 1.1

Protein structure and biosynthesis.

Figure 1.1

Protein structure and biosynthesis.

Close modal

The easiest way to a broader understanding of the 3-dimensional structures of proteins is to spend time on your computer “playing” with real structures. A good place to start is with the simple-to-use software available at http://www.umass.edu/microbio/rasmol/ or http://www.pymol.org/). While the structure of practically any enzyme of special interest is likely to be one of the many thousands accessible online from the Protein Data Base (http://www.rcsb.org/pdb/).

Enzymes have evolved to operate under most of the various environments natural to living organisms. The most important of these are the cytoplasm – an aqueous solution containing hundreds of other proteins and small-molecule metabolites – and the surfaces of membranes of various sorts. So enzymes have to be “comfortable” in various operating environments, and “tunable” – to work at different, controlled levels of activity appropriate to the changing requirements of the system. They must also be capable of catalyzing specific reactions of specific substrates at rates (based on values of kcat – see Section 1.2 – typically in the range 1–1000 s−1) high enough to support the immediate demands of the interactive network of local control mechanisms. Substrates range in size from O2 and CO2 to macromolecules, and kcat values between 1–1000 s−1 can represent accelerations of up to 1020 compared with rates of the corresponding uncatalyzed reactions at physiological pH in water.

So, an enzyme has to provide a highly sophisticated single-molecule “reaction vessel,” to support such rapid reactions. Not surprisingly, filling and emptying this “reaction vessel” efficiently poses its own problems, because when reactions become very fast, simple diffusion processes can become rate limiting. So, important additional requirements for the active-site “reaction vessel” are rapid substrate binding and – no less important – rapid release of product. (This last is no trivial requirement, considering that the product is typically almost identical to the substrate, give or take the cleavage or formation of a single covalent bond.) Furthermore, most if not all of any water molecules present in the resting active site have to be removed for the duration of the reaction process. All this makes the simple picture of a (static) cavity complementary in structure to the substrate a highly unconvincing proposition. Fischer's imaginative lock-and-key principle2  remains a valuable starting point, but an enzyme must provide specific binding complementarity not just to its substrate (or substrates), but to all intermediates and transition states on the reaction pathway.

Thus, the processes involved in catalysis are dynamic, and make complex demands on the enzyme protein. So, we should not be surprised that all this should require a large, sometimes a very large, molecule. An additional if less immediate factor is what might be called “protein bloat”. We are all familiar with the way systems like software applications (or government legislation) that are continually being improved and extended, can grow rapidly in size and complexity over just a small number of years. Enzyme evolution is a great deal slower – but it has been going on for millions of years…

What makes one protein different from another is the sequence of amino-acids, and thus the arrangement of the side-chains Rn (Figure 1.1) in the polypeptide chain. The chemical reactions involved in enzyme catalysis are implemented by the functional groups available on the amino-acid side-chains (backbone amide groups are not usually directly involved). Nine of the 20 naturally occurring amino-acids carry one of six different reactive groups (seven if phenol and alcohol OH groups count as different). These are listed in Table 1.1, which also gives the standard one- and three-letter abbreviations for the amino-acids, and approximate values for the pKas of the side-chain groups. The pKa value tells us how much of each ionic form of a group is present at any given pH, for the amino-acid free in solution. Though it gives only an indication of what might be the situation in the controlled environment of an active site, where in the absence of full solvation pKas can be perturbed, sometimes by as much as 4–5 units. (Qualitatively, this perturbation generally favours neutral species, because ionized forms are stabilized more strongly by hydrogen-bonding solvation.)

Table 1.1

Amino-acids with side-chains ionizing in the pH region

Amino acidSide-chain functional groupCatalytic function of the group
Aspartic acid (Asp, D)  The COOH group can act as a general acid, but will be present near pH 7 only in a specially controlled environment. 
Glutamic acid (Glu, E)  The carboxylate anion can act as a general base or as a nucleophile. 
Histidine (His, H)  The imidazole group is about 50% protonated at pH 7, so is the most versatile side-chain group: making available general acid, general base or nucleophilic groups near pH 7. 
Cysteine (Cys, C)  The thiolate group is a powerful nucleophile, and more than 1% may be present in a controlled active-site environment if its pKa is lowered. 
Lysine (Lys, K)  NH2 is a good nucleophile, especially for C=O, and its effective pKa can be lowered in a controlled active-site environment. 
Tyrosine (Tyr, Y)  An oxyanion is a good nucleophile, especially for phosphorus. 
Arginine (Arg, R)  The guanidinium group generally stays protonated. It makes strong, stabilizing double H-bonding interactions with CO2 and PO2 groups. 
Serine (Ser, S) Threonine (Thr, T)  The OH group is not ionized significantly in the pH region, but can act as a nucleophile as the proton is removed by a general base. 
Amino acidSide-chain functional groupCatalytic function of the group
Aspartic acid (Asp, D)  The COOH group can act as a general acid, but will be present near pH 7 only in a specially controlled environment. 
Glutamic acid (Glu, E)  The carboxylate anion can act as a general base or as a nucleophile. 
Histidine (His, H)  The imidazole group is about 50% protonated at pH 7, so is the most versatile side-chain group: making available general acid, general base or nucleophilic groups near pH 7. 
Cysteine (Cys, C)  The thiolate group is a powerful nucleophile, and more than 1% may be present in a controlled active-site environment if its pKa is lowered. 
Lysine (Lys, K)  NH2 is a good nucleophile, especially for C=O, and its effective pKa can be lowered in a controlled active-site environment. 
Tyrosine (Tyr, Y)  An oxyanion is a good nucleophile, especially for phosphorus. 
Arginine (Arg, R)  The guanidinium group generally stays protonated. It makes strong, stabilizing double H-bonding interactions with CO2 and PO2 groups. 
Serine (Ser, S) Threonine (Thr, T)  The OH group is not ionized significantly in the pH region, but can act as a nucleophile as the proton is removed by a general base. 

The functional groups listed in Table 1.1 are by no means an impressive selection of reagents, by the standards of the present-day organic or inorganic chemist. Nor can they be, because they must operate, and have long-term stability, in water: any strong base, acid or electrophile would immediately be neutralized by the solvent.

Most current thinking about enzyme catalysis is based on Pauling's original suggestion that enzymes work by binding and thus selectively stabilizing the transition states for their reactions. Starting from simple transition-state theory: consider the interaction between two reacting molecules A and B. The essential first step is for the two to come together. In the gas phase this involves a simple collision, but in solution molecules are separated by bulk solvent, and since each has its own solvation shell, making contact is a more complicated business. We can allow for this step by introducing into the reaction pathway the “encounter complex” A·B: without defining it in detail. Once in contact the molecules can undergo multiple “collisions” within the encounter complex before either reacting or diffusing apart. Thus, simple geometrical requirements for reaction, e.g. the directionality of approach of the reacting centres on the separate molecules, are not generally critical. If the chemical reaction is faster than diffusional separation, as is typically the case for many proton-transfer reactions, the diffusion step is rate determining and the reaction is diffusion controlled. (For examples, see Section 4.1.)

graphic

A ball-park estimate of the equilibrium constant for the random association of small molecules in aqueous solution is Ka ∼ 0.07 M−1. Making the interaction between A and B “sticky” – i.e. by the various binding interactions of molecular recognition – will increase Ka, and thus, other things being equal, (see Figure 1.4) the overall rate of formation of products. Enzymes work by (i) binding their own particular substrate, usually very specifically from the hundreds available in solution in the cell, (ii) catalyzing a specific reaction of the bound molecule; and (iii) finally releasing the product into solution.

Figure 1.2

Representative pH–rate profiles for three typical reaction types in water.

Figure 1.2

Representative pH–rate profiles for three typical reaction types in water.

Close modal

graphic
This mode of catalysis – whether by an enzyme or a simpler catalyst – is characterized by “saturation” or “Michaelis–Menten” kinetics: whereby a limiting rate is reached when all catalyst molecules are “busy” – i.e. binding reactant, intermediates or product. The defining equation for Michaelis–Menten kinetics introduces two key parameters, kcat and KM:
graphic

In the simplest case, where the chemical step kcat is clearly rate determining (i.e.E.S dissociates faster than it is converted to products), the rate becomes kcat[E]0 at [S] >> KM, defining kcat as the first-order rate constant for the conversion of E.S to products. And at [S] << KM (as [S] becomes very small) the rate → kcat/KM [E]0[S], defining kcat/KM as the second-order rate constant for the overall reaction. (For further detail see Section 2.3.)

We will use the parameters kcat and KM to characterize catalysis by enzyme models as well as enzymes: because they are familiar, and because they allow direct comparisons with enzyme reactions. We will also use the association constant Ka, in discussions of simple binding equilibria. (This corresponds to 1/KM in the simple Michaelis–Menten mechanism.)

The basic mechanistic problem solved by enzyme catalysis is simply illustrated by the schematic pH–rate profiles shown in Figure 1.2. Ionic reactions between nucleophilic and electrophilic reactants, for example most hydrolysis reactions, are typically acid and/or base catalyzed; so slowest (as shown by the minimum in the pH–rate profile I) near pH 7. A sufficiently highly reactive system may react without the need for acid/base catalysis, and show also an uncatalyzed “water reaction” (curve II): which is pH independent over a certain range but still slowest near neutrality. Enzyme reactions, by contrast, are fastest under physiological conditions, often showing bell-shaped pH–rate profiles (curve III), with a rate maximum near pH 7.

Figure 1.3

Distribution as a function of pH of the three ionic forms of a dibasic acid (e.g. an amino-acid) with pKas of 6 and 8. (The “diacid” is a cation in the case of an amino-acid.)

Figure 1.3

Distribution as a function of pH of the three ionic forms of a dibasic acid (e.g. an amino-acid) with pKas of 6 and 8. (The “diacid” is a cation in the case of an amino-acid.)

Close modal

The reactions at high and low pH (curves I and II of Figure 1.2) are not directly relevant to catalysis by enzymes, which operate of necessity near pH 7, and use catalytic groups that are only weak acids, bases and nucleophiles. The great majority of enzyme-catalyzed reactions are ionic, involving heterolytic bond making and breaking, and thus the creation or neutralization of charge. Under conditions of constant pH this commonly requires the transfer of protons. General acid–base catalysis provides mechanisms for bringing about the necessary proton transfers without involving hydrogen or hydroxide ions, which are present in water at concentrations of the order of 10−7 M under physiological conditions. At pHs near neutrality relatively weak acids and bases can compete effectively with lyonium or lyate species because they can be present in much higher concentrations.

Acid–base catalysis is termed specific if the rate of the reaction concerned depends only on the acidity (pH, etc.) of the medium (as in curve I of Figure 1.2). This is the case if the reaction involves as an intermediate a small amount of the conjugate acid or base of the reactant preformed in a rapid equilibrium process – normal behaviour if the reactant is weakly basic or acidic.

General acid–base catalysis is defined experimentally by the appearance in the rate law of acids and/or bases other than lyonium or lyate ions. Thus, the hydrolysis of enol ethers (Scheme 1.1) is general acid catalyzed: the rate depends on pH, but near neutrality depends also on the concentration of the buffer (HA+A) used to maintain the pH. Measurements at different buffer ratios show that the catalytic species is the conjugate acid HA. Any “general acid” can be a catalyst: the most reactive will always be the hydronium ion H3O+, which is (by definition) the strongest acid available in aqueous solution.

Scheme 1.1

The hydrolysis of enol ethers is a typical general acid-catalyzed reaction. The rate law is found to be −d[1]/dt=kH[1] [H3O+]+kHA[1] [HA].

Scheme 1.1

The hydrolysis of enol ethers is a typical general acid-catalyzed reaction. The rate law is found to be −d[1]/dt=kH[1] [H3O+]+kHA[1] [HA].

Close modal

If the measurements at different buffer ratios show that the catalytic species is the conjugate base A the reaction is kinetically general base catalyzed (in which case HA and A will usually subsequently be referred to as BH+ and B). Thus, the enolization of ketones is general base catalyzed (Scheme 1.2).

Scheme 1.2

The enolization of ketones is general base catalyzed. The rate law is found to be −d[2]/dt = kOH[2] [HO]+kB [2] [B].

Scheme 1.2

The enolization of ketones is general base catalyzed. The rate law is found to be −d[2]/dt = kOH[2] [HO]+kB [2] [B].

Close modal

The rate constants kHA and kB depend on the strength of the acid or base, and for a given reaction are correlated by the Brönsted equation. This is written differently for general acid- and general base-catalyzed reactions, respectively:

graphic

The pKas used are those of the conjugate acids, HA and BH+.

To illustrate the fundamental mechanisms involved we use as examples the hydrolysis reactions of carboxylic ester and amide groups: though the treatment applies to any reaction involving the attack of a nucleophile bearing a potentially acidic proton on any polar single bond, or its addition to a multiple bond.

If the substrate is reactive (electrophilic) enough, activation by protonation is not necessary and the attack of water on the neutral molecule is rate determining near pH 7 (see curve II of Figure 1.2). Formally this generates both a positive and a negative charge, and as the reaction proceeds both will be “delocalized” into the surrounding solvent via the network of hydrogen bonds. Scheme 1.3 shows the first step in the hydrolysis of a carboxylic acid derivative, RCOX, where X is a potential leaving group.

Scheme 1.3

Mechanism of spontaneous hydrolysis of a carboxylic acid derivative. The second step (“etc.”) of the reaction is discussed in Scheme 1.6. Note that only the principal reacting solvent molecules are shown. Every lone pair on every O or X atom will also be solvated by hydrogen bonding to a water OH, and every potentially acidic proton solvated by hydrogen bonding to a water oxygen.

Scheme 1.3

Mechanism of spontaneous hydrolysis of a carboxylic acid derivative. The second step (“etc.”) of the reaction is discussed in Scheme 1.6. Note that only the principal reacting solvent molecules are shown. Every lone pair on every O or X atom will also be solvated by hydrogen bonding to a water OH, and every potentially acidic proton solvated by hydrogen bonding to a water oxygen.

Close modal

The nucleophile in a hydrolysis reaction is of course a water molecule (labeled nuc in Scheme 1.3). As the new C–O bond forms, positive charge develops on the nucleophilic oxygen, and the attached OH protons become more and more acidic, until they can be transferred to solvating water, acting formally as a general base (gb). The negative charge developing on the carbonyl oxygen may similarly be transferred via a hydrogen bond to another water molecule, acting this time as a general acid (ga). This mechanism is always available, though it does not always lead to reaction at an observable rate. It accounts, for example, for the rapid hydrolysis of ethyl trifluoroacetate, CF3COOEt, which shows a pH-independent region near neutrality (see curve II of Figure 1.2): but the hydrolysis of ethyl acetate, CH3COOEt is extremely slow (half-life ∼80 years at 25 °C at the pH minimum (see curve I of Figure 1.2), while the neutral reaction of CH3CONHEt is too slow to observe.

The mechanism outlined in Scheme 1.3 is quite general, and can be applied to the ionic cleavage of any σ- or π-bond X–Y or X=Y (Scheme 1.4).

Scheme 1.4

Generalized reaction scheme for ionic bond breaking. Given the right combination of nucleophile, general base and general acid this mechanism can provide major reductions in the enthalpy of activation for the cleavage of the X–Y bond.

Scheme 1.4

Generalized reaction scheme for ionic bond breaking. Given the right combination of nucleophile, general base and general acid this mechanism can provide major reductions in the enthalpy of activation for the cleavage of the X–Y bond.

Close modal

However, H2O is a weak nucleophile, acid and base and it is easy to find stronger examples of all three types of reagent – general base, nucleophile and general acid – among the functional groups available on amino-acid side-chains (Table 1.1). So we are not surprised to learn that the hydrolysis of CF3COOEt, for example, is general base catalyzed by carboxylate anions (Scheme 1.5).

Scheme 1.5

Mechanism of spontaneous hydrolysis of ethyl trifluoroacetate. Nucleophilic catalysis [in brackets] is disfavoured because the tetrahedral addition intermediate T reverts too rapidly to reactants. See the text.

Scheme 1.5

Mechanism of spontaneous hydrolysis of ethyl trifluoroacetate. Nucleophilic catalysis [in brackets] is disfavoured because the tetrahedral addition intermediate T reverts too rapidly to reactants. See the text.

Close modal

The direct reaction (Scheme 1.6) does have the entropic advantage of not involving a third molecule, and a carboxylate anion can – given the right conditions – displace better leaving groups than ethoxide (and poorer leaving groups than itself). As long as the intermediate (in this case an anhydride) is more reactive than the starting material the result is catalysis of hydrolysis.

This mechanism competes equally with general base catalysis (i.e. accounts for some 50% of the reaction) for esters with a leaving group of pKa of ∼7, and becomes dominant for derivatives with better leaving groups. The two mechanisms are kinetically identical (the third participant in the general base-catalysis mechanism being a kinetically invisible solvent molecule). But they can be distinguished in a number of ways, most simply by identifying the intermediate, directly or by trapping, and showing that it is “kinetically competent” (reactive enough to account for the observed rate of reaction). This provides convincing positive evidence for a nucleophilic mechanism: as do irregularities in the fit of data to the Brönsted equation (Section 1.3.1) that can be ascribed to steric effects (not to be expected for general base catalysis, where the attacking nucleophile is always H2O). The simplest positive evidence for general acid or general base catalysis is a significant solvent deuterium isotope effect, kH2O/ kD2O, of ≥2 consistent with a proton transfer in the rate-determining step. Though, as always, an accumulation of several independent pieces of evidence consistent with the suggested mechanism is preferred to support a convincing conclusion.3 

Scheme 1.6

Nucleophilic catalysis in the spontaneous hydrolysis of an acyl derivative with a good leaving group. See the text.

Scheme 1.6

Nucleophilic catalysis in the spontaneous hydrolysis of an acyl derivative with a good leaving group. See the text.

Close modal

If the leaving group is poor it can be made viable by protonation: complete protonation to form the conjugate acid if the group (e.g. amino) is sufficiently basic, but involving partial proton transfer in the case of a weakly basic group like OR or OH. This mechanism (general acid catalysis) is involved, though not easily observed, in the breakdown of the tetrahedral addition intermediates involved in the acyl transfer reactions of esters and amides (Scheme 1.7).

Scheme 1.7

General acid catalysis of the second step in the hydrolysis of carboxylic acid derivatives (Scheme 1.3). Note that general acid catalysis is the microscopic reverse of general base-catalyzed addition (dashed arrows).

Scheme 1.7

General acid catalysis of the second step in the hydrolysis of carboxylic acid derivatives (Scheme 1.3). Note that general acid catalysis is the microscopic reverse of general base-catalyzed addition (dashed arrows).

Close modal

General acid catalysis is most conveniently observed in the reactions of stable tetrahedral species with two or more O, S or N atoms attached to a central atom (though only in special cases for simple acetals, as discussed in Section 4.3.1). That general acid catalysis is observed in the hydrolysis of ortho esters 1.3 (Scheme 1.8) is explained as follows: the C–OR oxygens are very weakly basic and the dioxocarbocation intermediate 1.4 particularly stable, a result of π-donation from the two remaining oxygens. Under these conditions C–OR cleavage can occur before proton transfer is complete, thus becoming concerted with proton transfer (Scheme 1.8).

Scheme 1.8

Classical general acid catalysis involves proton transfer becoming concerted with the cleavage of a covalent bond between heavy atoms.

Scheme 1.8

Classical general acid catalysis involves proton transfer becoming concerted with the cleavage of a covalent bond between heavy atoms.

Close modal

Given a viable electrofuge (e.g. the cation 1.4 in Scheme 1.8 ) proton transfer can be expected to become concerted with the making or breaking of covalent bonds if the reaction is sufficiently thermodynamically favourable. The practical requirements are summarized in Jencks’ so-called “libido rule”.4,5 

Concerted general acid–base catalysis of complex reactions in aqueous solution can occur only (a) at sites that undergo a large change in pKa in the course of the reaction, and (b) when this change in pK converts an unfavorable to a favorable proton transfer with respect to the catalyst; i.e. the pK of the catalyst is intermediate between the initial and final pKa values of the substrate site.

Thus, general acids with pKas of 7 ± 4, of potential interest in biological systems, are well qualified to assist in the cleavage of bonds to oxygen leaving groups: since the pKas of ester oxygens are typically negative, while those of their alcohol cleavage products are usually >14.

These simple examples might suggest that the observation of general acid or general base kinetics is prima facie evidence for the corresponding general acid or general base catalysis mechanisms. This is not the case, for the usual reasons of (a) kinetic equivalence (the proton is a uniquely mobile species), and (b) the possible involvement of the solvent (e.g. the water molecules in Scheme 1.3) in the mechanism. For example, ketone enolization below pH 7 is general acid catalyzed (i.e. is first order in the concentration of general acids HA), but is explained not by the general acid-catalysis mechanism (gac, Scheme 1.9) but by the kinetically equivalent specific-acid general-base catalysis (gbc route in Scheme 1.9): which requires only bimolecular encounters for the rate-determining step of the reaction).

Scheme 1.9

The general-acid catalyzed enolization of ketones involves the kinetically equivalent general-base catalyzed removal of the CH proton from the substrate conjugate acid.

Scheme 1.9

The general-acid catalyzed enolization of ketones involves the kinetically equivalent general-base catalyzed removal of the CH proton from the substrate conjugate acid.

Close modal

In practice, the entropic cost of bringing four molecules together is far too great for the complete mechanism of Scheme 1.4 to be observable in free solution: even a three-molecule collision is unfavourable enough that the mechanism of Scheme 1.5 is observed only when the enthalpy of activation for the reaction is intrinsically low: and even nucleophilic catalysis, with the more modest entropic requirements of a bimolecular encounter, is observed in water only with activated substrates. Since many enzymes catalyze reactions of highly unreactive substrates, it follows that they have evolved ways of reducing the free energies of activation. The simplest of these involves bringing the reacting groups into close proximity in the substrate-binding step, in such a way that the entropy debt is paid off “in advance”, in the binding step. So the simplest models of fundamental enzyme processes are intramolecular reactions, in which reacting groups are brought into close proximity by being positioned close together on the same molecule.

Though the entropic cost of bringing four molecules together is too great for the complete mechanism outlined in Scheme 1.4 to be observable in free solution, this is not a problem for a substrate bound in close proximity to the catalytic groups of an enzyme-active site. A classic example is the mechanism (Scheme 1.10, A) used by many serine proteinases (enzymes that hydrolyze the intrinsically unreactive (at pH 7) peptide (i.e. amide) bonds of proteins, using the OH group of an active-site serine as a nucleophile). The earliest work on these mechanisms involved studying the component parts in simple bimolecular systems, as described above: and then – more appropriately, and also more conveniently – in systems where the reacting functional groups are held in close proximity on the same molecule. Thus, the central nucleophilic reaction can be modeled by the cyclization (lactonization) of an appropriate hydroxyamide (Scheme 1.10, B), which can be much faster than the neutral hydrolysis of the amide group under the same conditions, allowing the use of unactivated substrate groups.

Scheme 1.10

Lactonization model for nucleophilic catalysis by serine proteases.

Scheme 1.10

Lactonization model for nucleophilic catalysis by serine proteases.

Close modal

Intramolecular reactions, and particularly intramolecular nucleophilic reactions (i.e. cyclizations) are the only simple reactions that begin to rival their enzyme-catalyzed counterparts in rate under physiological conditions. Making the reaction of interest part of a thermodynamically favourable cyclization can produce quite simple systems in which the extraordinarily stable groups of structural biology (amides, glycosides and phosphate esters typically have half-lives of tens to millions of years under physiological conditions near pH 7) can be cleaved in fractions of a second. An early example was 1.5 (Scheme 1.11), where an ordinary aliphatic amide is forced into close proximity with a COOH group in such a way that the half-life of 1.5, where R is a simple alkyl group, is less than a second.6 

Scheme 1.11

The extraordinarily rapid cyclization of simple, unactivated amides 1.5 of dimethylmaleic acid involves nucleophilic “catalysis” (see the text) by the neighbouring carboxyl group and catalysis by external general acids.6 

Scheme 1.11

The extraordinarily rapid cyclization of simple, unactivated amides 1.5 of dimethylmaleic acid involves nucleophilic “catalysis” (see the text) by the neighbouring carboxyl group and catalysis by external general acids.6 

Close modal

Given a suitable cyclization reaction going at a useful rate, it is possible to ask relevant questions about the basic reaction between the two interacting functional groups, specifically chosen from those involved in enzyme reactions as substrate or catalytic group. Varying the length and structure of the linker (Scheme 1.10, B) can provide information about preferred geometries of approach. And because amide cleavage can be so fast in this efficient intramolecular reaction, it is possible to observe also external general acid or general base catalysis of the cyclization. This in turn provides the information needed for the rational design of a next generation of models in which the general acid or base is also built in to the reacting system, so that the general acid/base catalysis becomes itself intramolecular. At each stage we interrogate the system about its intrinsic preferences, and then try to improve the efficiency of catalysis in light of the answers.

This is not a trivial exercise, and there can be surprises (positive as well as negative!). For example, the reaction of 1.5 (Scheme 1.11) is general acid catalyzed: the neutral tetrahedral intermediate T0 breaks down rapidly to the reactant acid amide, and goes on to product only after a double proton transfer has converted it to the zwitterion T±, with a neutral amine as the leaving group. The double proton transfer is neatly catalyzed by the CO2H group of a general acid, which can transfer its proton to the N centre of T0 while simultaneously accepting one from the OH group on its other oxygen (Scheme 1.12).

Scheme 1.12

The synchronous mechanism for the general acid-catalyzed conversion of T0 to T± is not geometrically possible (1.6) for the reaction of the β-alanine derivative 1.5a. (1.6 shows the OH of the CO2H group in its preferred z conformation).

Scheme 1.12

The synchronous mechanism for the general acid-catalyzed conversion of T0 to T± is not geometrically possible (1.6) for the reaction of the β-alanine derivative 1.5a. (1.6 shows the OH of the CO2H group in its preferred z conformation).

Close modal

The observation of general acid catalysis suggests that reaction would be faster if the general acid was a built-in CO2H group. This would have to be attached via a linker long enough to support the geometry required for the double proton transfer, but not so long that catalysis would be inefficient. A study of a series of amides derived from amino-acids showed that the most reactive were derivatives of β-amino-acids, e.g. 1.5a (Scheme 1.12) from β-alanine: but the catalytic group involved turned out to be not CO2H but the carboxylate anion, CO2.7  The CO2H-catalyzed reaction might still be most effective in an enzyme-active site, where the geometry can be optimal (as for T0 in Scheme 1.12): but this geometry cannot be achieved in the intramolecular system 1.6. The mechanism of this reaction is discussed in more detail in Section 4.1.3.1.

This result contains several important lessons.

  1. Experimentally the pH–rate profile shows a rate maximum in the region between the pKas of the two carboxyl groups, because one is involved as CO2H and the other as carboxylate. This is the simple, general explanation for the pH–rate optima observed for many enzyme reactions (see Figure 1.2). If the mechanism depends on two (or more) functional groups, one in the basic and one in the acidic form, the concentration of the reactive form (Figure 1.3), and thus the rate of the catalytic reaction, is at a maximum at a pH half-way between the two pKas.

  2. Intramolecular reactions are useful if limited models for enzyme reactions. The neighbouring groups can be held in close proximity, but the linker may place constraints on their geometry of approach.

  3. In reactions involving the making or breaking of bonds between heavy (nonhydrogen) atoms the intramolecular reaction is a one-off event – there is no possibility of turnover. The term intramolecular catalysis can be used correctly, because the catalytic group may be regenerated: but the neighbouring substrate group is not, so no turnover is possible.

Figure 1.4

Energy-profile diagrams compared for a reaction catalyzed by an enzyme or enzyme model E and for the same reaction under the same conditions with no catalysis.

Figure 1.4

Energy-profile diagrams compared for a reaction catalyzed by an enzyme or enzyme model E and for the same reaction under the same conditions with no catalysis.

Close modal

Despite these limitations the study of intramolecular reactions, apart from being an integral part of the broader study of organic reactivity, has informed our thinking on enzyme catalysis by defining, in detail in simple systems, the basic mechanisms involved in most enzyme reactions. (Enzyme evolution has involved a great deal of attention to mechanistic detail!) For example, general acid catalysis of acetal hydrolysis was first observed as an intramolecular reaction, after being suggested as the likely role for one of the two active-site groups of lysozyme: and is now a well-understood process. Furthermore, since intramolecular reactions offer the best prospect of observing reactions of the least reactive naturally occurring functional groups in simple systems, the efficiencies as well as the mechanisms of the reactions are of special interest.

The efficiency of intramolecular catalysis is defined most simply in terms of the effective molarity (EM) of the catalytic group.8  The EM is defined as the ratio of the first-order rate constant for the intramolecular reaction and the second-order rate constant for the corresponding intermolecular reaction (Scheme 1.13) proceeding by the same mechanism under the same conditions. It has the dimensions of molarity and is simply the concentration of the catalytic group that would be required in the equivalent intermolecular process to match the rate of the intramolecular reaction. (This concentration is purely nominal in most cases, since EMs are often greater than 10 M.)

Scheme 1.13

Rate constants defined for intra- vs. intermolecular catalyzed reactions going by the same (general base catalysis) mechanism.

Scheme 1.13

Rate constants defined for intra- vs. intermolecular catalyzed reactions going by the same (general base catalysis) mechanism.

Close modal

Two generalizations are of major importance:

  1. Intramolecular catalysis can be highly efficient in cyclization reactions, which involve the formation of bonds between heavy (nonhydrogen) atoms. EMs of 108–9 can be observed even in flexible systems,8  and higher figures are possible in rigid systems where ground-state strain is relieved as part of the reaction: the EM of the carboxyl group in the cyclization of 1.5 (Scheme 1.11 ), for example, is over 1013 M.8  Enzymes cannot exert the sorts of very strong nonbonding interactions that can be built by synthesis into structures like 1.5, because protein structures are relatively flexible; but studies of such rigid systems can give us access to valuable mechanistic information about the reactions of relevant functional groups going at rates comparable to those of the corresponding enzyme-catalyzed processes. There has been much discussion of the origins of the rate enhancements in these systems, in terms of at least a dozen theories,9  from orbital steering to the spatiotemporal hypothesis.10  Some of these provide useful guidance for systems design, but the fundamental requirement for a high (kinetic) EM is a thermodynamically favourable cyclization with no stereoelectronic impediment to reaction. The entropy term is always more favourable for an intramolecular process (the basis of the description of enzymes as “entropy traps”): but this factor is common to all simple intramolecular processes, and does not vary a great deal for many favourable reactions; whereas the enthalpy term can depend strongly on the way the reacting groups are brought together.11 

  2. Intramolecular reactions involving rate-determining intramolecular proton transfer – i.e. general acid and general base-catalyzed reactions – are typically far less efficient, with EMs generally below about 10 M.8  For example, the EM of the carboxylate group in a series of half-esters 1.7 of dialkylmalonic acids, hydrolyzed by the general base-catalysis mechanism shown in Scheme 1.14 (and observed “by default” – because cyclization, to form a 4-membered ring, is thermodynamically strongly disfavoured) ranges from less than 1 up to no more than 60 M.12 

Scheme 1.14

The neighbouring carboxylate group of malonate monoesters 5 with good leaving groups acts as a general base, rather than a nucleophile: because it is too close.

Scheme 1.14

The neighbouring carboxylate group of malonate monoesters 5 with good leaving groups acts as a general base, rather than a nucleophile: because it is too close.

Close modal

These low numbers would appear to present something of a problem, since proton transfer is the commonest reaction catalyzed by enzymes. There is no doubt that enzymes catalyze such reactions efficiently, so the reasonable conclusion is that the models, not the enzymes, are deficient. More recent work suggesting a solution is discussed in Section 4.3.

The efficiency of enzyme catalysis depends on a large number of factors, no single one of which is uniquely responsible for the accelerations observed in any particular case. The most important of these are discussed in the following pages, together with attempts to identify and to quantify them in model systems. But we must always bear in mind the overall thermodynamics of the process: which can be summarized in the energy-profile diagram of Figure 1.4.

Figure 1.4 compares simple energy-profile diagrams for a general reaction of a substrate S catalyzed by an enzyme E (left to right arrow) with that (right to left arrow) for the same, uncatalyzed, reaction, involving the same functional groups, in water. The picture is oversimplified, greatly so for many enzyme-catalyzed processes, to consider only a single (chemical) rate-determining step in each direction: but has the advantages (i) that it is based on the Michaelis–Menten equation, and thus on standard, readily measurable parameters, kcat and KM: and (ii) that it can be applied equally well to measure the efficiency of an enzyme model or mimic. The key parameter derived from the direct comparison of catalyzed and uncatalyzed reactions is the binding energy of the transition state of the catalyzed reaction, defined in terms of the rate ratio (kcat/KM)/kuncat (for more detailed discussion see Chapter 2).

graphic

In the simple case this is a measure of the stabilization of the rate-determining transition state by the catalyst compared with stabilization by solvent water. (This assumes that substrate binding is thermodynamically neutral under the conditions. Stronger substrate binding stabilizes the effective ground state for the reaction (Figure 1.4), and hence reduces catalytic efficiency.)

A complete description of the mechanism of a reaction must include a description of the transition state, but – as a species with no significant lifetime – its structure is rarely well defined, especially for reactions taking place in the active site of a complex protein enzyme. However, we can draw some general conclusions.

An enzyme has evolved to bind – and thus stabilize specifically – the transition state for its particular reaction. For “binding the transition state” to lead to catalysis the interactions of molecular recognition concerned must be strongest in the transition state. But what exactly do we mean by “binding the transition state” – this dynamic entity with no significant lifetime?

Take the transition state for the serine protease reaction discussed previously (Scheme 1.10, A). The transition state is conventionally drawn as shown in Scheme 1.15. Note that there is no well-defined boundary between enzyme and “transition state,” because enzyme and substrate are linked by partial covalent bonds. The details will depend on the particular reaction, but this partial covalent bonding gives rise to particularly strong, stabilizing, dynamic interactions specific to the short-lived transition-state structure, which can contribute significantly to the “binding of the transition state” that makes catalysis efficient. This contribution is particularly strong for partial covalent bonding between heavy (nonhydrogen) atom centres. Thus, those enzymes that have to be the most efficient, because they catalyze reactions of substrates of very low intrinsic reactivity, commonly do so by mechanisms in which the enzyme acts as a nucleophile (as in Scheme 1.10, A): using one of the groups available on amino-acid side-chains (Table 1.1), perfectly positioned in the active site for the purpose.13,14 

Scheme 1.15

Binding the transition state. The arrows indicate three of the five partial covalent bonds that contribute directly to transition-state binding.

Scheme 1.15

Binding the transition state. The arrows indicate three of the five partial covalent bonds that contribute directly to transition-state binding.

Close modal

Enzymes work by a sophisticated, integrated system of binding and catalysis. The essential first step of any enzyme reaction (Figure 1.4) is substrate binding, by which the specific substrate is selected from many others available after diffusing into contact with the protein, and then into the active-site “reaction vessel”, where it comes into contact with the catalytic apparatus. The rest of the reaction coordinate for the catalyzed reaction is accessible only to the bound substrate. Binding and catalysis are thus complementary, and if we accept that an enzyme has evolved to bind specifically the transition state for its particular reaction both can be regarded as binding phenomena. Discrimination between closely similar substrates – enantiomers are obvious examples1 – depends on the selective binding of both substrate and transition state. Of these, transition-state binding is more important because it affects the rate-determining step directly.

Comparisons with the results of much recent work on molecular recognition show that substrate binding falls quantitatively into the class of typical host–guest interactions involving the binding of small molecules to synthetic hosts, which are characterized by association constants in water in the range 10 to 104 M−1.16  The direct noncovalent intermolecular interactions involved – hydrogen bonding, electrostatic and van der Waals attractions – are fairly well understood, with multiple interactions the main source of stronger binding. They are conveniently studied in nonpolar solvents, because hydrogen-bonding and charge–charge-based associations are generally at their weakest in strongly polar water. Water is a good hydrogen-bonding solvent for individual charged or strongly polar molecules, and the major role in stabilizing host–guest complexes involving neutral systems in aqueous solution is played by the more subtle hydrophobic interactions.

This “hydrophobic effect” refers to the tendency for nonpolar organic compounds – or groups – to avoid water in favour of a nonpolar environment (oil and water don’t mix!). It is responsible for the formation from nonpolar compounds of aggregates of all sizes, from short-lived “encounter complexes” and well-defined host–guest complexes – thus stabilized by “hydrophobic binding” – to (relatively long-lived but less well-defined) micelles and vesicles. It is also one of the most significant factors controlling protein folding: nonpolar (hydrophobic) groups on amino-acid side-chains are typically buried in the interior of a globular protein: the way they pack together, together with the preference for charged groups to lie on the surface in contact with solvent water, controls the folding of the peptide chain. Similarly, small molecules with nonpolar groups aggregate in water to form micelles or vesicles, in which hydrophobic groups pack together and charged groups lie on the surface in contact with the solvent. Last, but not least, because charged species are disfavoured in an apolar environment, the pKa values of ionizable catalytic groups (at least one form is always charged) may be significantly perturbed in a hydrophobic active site (see Section 3.2).

When water molecules come into unavoidable contact with a nonpolar (hydrophobic) surface the solvent's unique, dynamic three-dimensional network of hydrogen bonds is disrupted. These hydrogen-bonding interactions are dominant as the system adjusts to the disruption in two important ways. (i) The local network of solvent hydrogen bonds at the interface becomes close to two-dimensional, with O–H bonds making tangential contact with the nonpolar surface. This enforced ordering process involves strongly unfavourable (negative) entropy effects.2 (ii) The extent of the interaction is minimized by bringing nonpolar molecules, or parts of molecules, together. This hydrophobic binding is reinforced by (much weaker) van der Waals attractions, and depends for small molecules on their nonpolar surface area, though not with great precision. In terms of equilibrium constants at room temperature, from 70 Å2 to as little as 10 Å2 of surface area buried can contribute a factor of 10 to the equilibrium constant for association.16 

Hydrophobic binding is by definition specific to water as a solvent. Direct noncovalent intermolecular interactions – hydrogen bonding, electrostatic and van der Waals attractions – operate in all solvents. We can get some idea of the importance of hydrophobic binding of small neutral molecules to organic hosts from the difference between free energies of binding in water and organic solvents. On average this is about 1.5 kcal mol−1, worth a factor of about 12 in association constant.16 

A useful practical measure of hydrophobicity is log P, based on the partition coefficient P for a given molecule between water and 1-octanol.19  Values are more or less additive, so can be broken down into group contributions and used in linear free-energy relationships for simple binding processes; or for structure–activity relationships involving binding equilibria, from enzyme reactions to anaesthesia. Such relationships do not need to be very precise to allow interesting and sometimes useful predictions from simple extrapolation.

We know that enzymes, to be efficient catalysts, must bind transition states much more strongly than their substrates (see Section 1.5): presumably by a combination of mechanisms including those discussed above; plus, perhaps, additional special factors. Uncertainties arise because as we have seen, binding a transition state is not the same as binding a stable molecule; and because we need to estimate how strongly a protein could bind its substrate, using only a combination of the familiar intermolecular interactions plus hydrophobic binding. We can get a good idea by considering the properties of antibodies, proteins produced by the immune system and evolved specifically to bind stable antigens of various sorts.

Antibodies bind small organic molecule antigens with association constants of 107±2 M−1, corresponding to a binding free energy of −10 ±3 kcal mol−1 – about double that observed for cyclodextrins (see Section 4.1.1.3.1) and typical synthetic hosts.16  It is reasonable to suppose that this approaches the limit for binding small molecules under physiological conditions. As might be expected, bigger natural ligands like proteins and oligosaccharides offering multiple binding interactions can be bound up to two orders of magnitude more strongly in vivo (Ka=109± 2 M−1).

In general, the strongest protein–small molecule binding interactions, matching these values (Ka= 109± 4 M−1), are found for enzymes binding transition-state mimics (or analogues): stable compounds that bind specifically to active sites, and of practical interest because they can act as enzyme inhibitors (see Section 3.3). It is not surprising that enzyme-active sites should bind particularly strongly molecules designed to mimic their specific transition states. However, no stable molecule can be a perfect mimic for the dynamic transition state that the enzyme has evolved to recognize. There remains a substantial gap between Ka values for the binding of the best transition-state mimic inhibitors and “equilibrium constants” for transition-state binding to enzymes. These fall typically in the range 1016.0± 4.0 M−1, equivalent to free energies of binding of 22 ± 6 kcal mol−1, but can reach as high as 32 kcal mol−1 for reactions of common, very unreactive substrates.16 

In a small number of cases proteins are known that bind relatively small molecules with free energies of binding that fall in this elevated range. The best-known example is the formation of the avidin:biotin complex, with a free energy of binding of −21 kcal mol−1. This can be interpreted in terms of multiple complementary contributions, from hydrogen bonding, van der Waals interactions and hydrophobic binding.20  The “substantial gap” referred to above, of the order of 10 kcal mol−1 remains. But it is clear that the very large free energies of binding by enzymes of transition states for reactions of highly unreactive substrates also represent the summation of multiple, complementary contributions from recognized effects. The main effect specific to transition-state binding that is not involved in the binding of ordinary molecules is likely to be the partial covalent bonding between substrate and nucleophilic groups of the enzyme-active site.

Finally, it is important to stress that very strong binding of ordinary molecules inevitably involves slow release of the bound guest: for example, biotin is bound effectively irreversibly to avidin. Tight binding of either substrate or product to an enzyme would inhibit turnover (and slow product release is a common problem for enzyme mimics). Only a high energy, and thus short-lived species – like a transition state – can be bound so strongly without dissociation becoming unacceptably slow. Binding (or association) constants Ka can be dissected into rates of binding (k1) and rates of release (k−1), such that Ka = k1/k−1. k1 has an upper limit, the rate constant for diffusion through water. This is of the order of 108 M−1 s−1 or less for the binding of small molecules to proteins, whose conformations may have to change and binding sites be desolvated. For example, k1 is typically ∼ 106 M−1 s−1 for the binding of small molecules to antibodies, which, combined with an association constant of 107 M−1 or more gives k−1 of the order of 10−1 s−1 or less for antigen release.

The active site of an enzyme is far more than a passive “reaction vessel.” It provides a dynamic environment tailored specifically to the whole of the developing reaction coordinate of the reaction it has evolved to catalyze. In terms of molecular recognition it acts as host to a particularly demanding guest: starting with a cautious welcome, followed – if the guest's “face” fits – by the tightest of embraces: after which the guest – changed perhaps permanently by the experience – can hardly get away fast enough. At each stage the microenvironment of the active site is adjusting to match the electronic profile of the guest in a preprogrammed “vectorial” process.21  The process is only complete when the active site is vacant, and ready for the next guest.

Of course the “vacant” active site in the “resting” enzyme is no less dynamic. Fischer's seminal lock-and-key hypothesis introduced the important concept of enzyme–substrate recognition,2  but we understand now that there is no such thing as an active-site cavity with a fixed shape precisely complementary to that of the substrate. A “vacant” active site must either collapse (reversibly, of course), or be occupied by (more or less rapidly exchanging) water molecules. A specific substrate will replace some or all of these, as the local geometry adjusts to the requirements of the new guest: opening wider to accept it, closing, perhaps all round it in the transition state; and opening again to release the product. All these changes involve reciprocal interactions with the surrounding protein structure, which can play a significant part in the overall process.

The active-site microenvironment is of course not homogeneous: it is formed from individual peptide groups and amino-acid side-chains, and may be largely hydrophobic, largely polar, or – significantly – a specifically “designed” combination of the two: with the functional groups involved in catalysis positioned to optimize the stereoelectronic interactions involved in bond making and breaking. It is sometimes suggested that a particular organic solvent, or dielectric constant, offers a better model than water for an “active-site environment”. There can be an element of truth in this for some simple systems. But no homogeneous medium can model the dynamic three-dimensional arrangement of molecular diversity available in a protein microenvironment.

The functional groups available on the side-chains of the 20 amino-acids that occur naturally in protein enzymes (Table 1.1) may make up an unimpressive selection of reagents, but they perform extraordinarily effectively in a vast number of reactions. Nevertheless, they do have their limitations: for example they include no oxidizing or reducing agent, and no significant electrophilic centre. (All this is for good reasons: proteins contain groups that are relatively sensitive to oxidation, and have to operate, and thus be stable in (relatively nucleophilic) water.)

The missing functionality for these, and some other types of essential chemistry not available from side-chain groups is provided by cofactors, auxiliaries that are bound, usually reversibly, to extend the capabilities of the active sites of large numbers of relevant enzymes. Cofactors include metal ions, which provide precisely positioned electrophilic centres and the possibility of single-electron transfer processes; and the coenzymes, a range of specialized organic systems. Familiar examples are NADH, an organic reducing agent (sometimes called biological borohydride); and pyridoxal 1.8, which acts as a “portable electron sink” (Scheme 1.16).22 

Scheme 1.16

Outline mechanism for the decarboxylation of an α-amino-acid by pyridoxal (1.8). The pair of sigma-bonding electrons forming the C–C bond that is broken can be delocalized into the electron sink provided by the pyridinium system most efficiently in the geometry shown, which optimizes overlap with the (π*-orbital of the) π-system. The same process can lead to cleavage of the C–H or C–R bonds, in conformations with these bonds perpendicular to the plane of the π-system.22 

Scheme 1.16

Outline mechanism for the decarboxylation of an α-amino-acid by pyridoxal (1.8). The pair of sigma-bonding electrons forming the C–C bond that is broken can be delocalized into the electron sink provided by the pyridinium system most efficiently in the geometry shown, which optimizes overlap with the (π*-orbital of the) π-system. The same process can lead to cleavage of the C–H or C–R bonds, in conformations with these bonds perpendicular to the plane of the π-system.22 

Close modal

In both cases the coenzyme acts primarily as an extension of the enzyme. The chemistry can be shown to work perfectly well in the absence of any enzyme, albeit far less efficiently. The enzyme provides the environment for catalysis, just as it does for reactions not involving cofactors. Thus, the principles of catalysis, which enzyme models are designed to test and illustrate, are also no different, and we will not discuss coenzymes (or coenzyme models) as such in this book: beyond a single high-profile example (Section 4.2), chosen to illustrate this powerful way of extending the capabilities of protein enzymes.

There is always one (or a handful of) best – meaning most efficient – mechanisms for the catalysis of a particular reaction by a protein enzyme under physiological conditions, and biochemical evolution is efficient enough – and may be assumed to have had long enough – to have found them. So, it should be no surprise that large numbers of the thousands of enzymes use basically the same mechanism to carry out the relatively small number of particular common tasks, like amide hydrolysis, phosphate transfer or reduction using NADH. This means that the active sites of such enzyme “families” have in common the same set of basic features: including for example recognizable nucleotide binding domains for cofactors like ATP and NADH; and in particular the same side-chain groups involved in catalysis. The need for large numbers of different enzymes to catalyze the same basic reaction comes from various extra-kinetic requirements, such as mechanisms for feedback control of activity, and especially the need for substrate specificity. This economy of mechanism is a major advantage in the design of enzyme models. It is also the basis of fascinating insights into biochemical evolution, because homologies in primary sequence can be used to construct evolutionary “family trees” of groups of such enzymes.

The first such family to be identified was the serine proteases, enzymes like chymotrypsin with a common tertiary structure (and thus thought to be derived from a common ancestor proteinase). These enzymes use an active-site serine (part of a “catalytic triad” that includes a histidine general base) as a nucleophilic catalyst for amide hydrolysis (the first step of the mechanism is outlined in Scheme 1.10). This mechanism has probably inspired more enzyme models and mimics, of various sorts, than any other. Nature has also been involved: the same catalytic triad, in the same 3-dimensional arrangment, is also found in a second, smaller group of serine proteinases (e.g. subtilisin) that have no significant sequence homology with the first group. Evidently, evolution has arrived at the same solution to this particular mechanistic problem by a quite different route: a clear case of convergent evolution. (Divergent evolution has also been identified, for example in the very different activities of the enzymes tyrosinase, hemocyanin, and catechol oxidase, which have almost identical active sites.23 )

This multiplicity of different enzymes all using basically the same chemical mechanism naturally makes the reactions involved attractive targets for enzyme modelling. Apart from the serine proteases one of the biggest enzyme classes is the glycosidases, which have been classified on the basis of similarities of primary sequence into over 100 families (for the current score go to http://www.cazy.org/fam/acc_fam.html). All of these, with the exception of a few special cases, use one of just two different mechanisms. In both cases, the hydrolysis of the glycosidic bond is accomplished by two catalytic residues of the enzyme: a nucleophile/base and a general acid.

The observation that glycosidases “choose” to use a general acid, rather than a general base, clearly makes good chemical sense. We know that acetals (a glycoside is a particular sort of acetal) are stable in base, and hydrolyzed in acid. But when the first glycoside active site (of lysozyme) was shown to contain just two carboxyl groups general acid catalysis of acetal hydrolysis was not a known reaction.24  Given the hint, it did not take long for the first enzyme model to appear, when it was shown that the glucoside 1.9 containing salicylic acid was hydrolyzed some 105 times faster than its isomer with the COOH group in the para position.25  The mechanism, written as classical intramolecular general acid catalysis, is shown in Scheme 1.17. Since 1.9 has a half-life in water of 8 min even at 91.3 °C, and a phenol is a much better leaving group than a sugar OH, this result is only a first indication of what might be happening in a fraction of a second in an enzyme-active site under physiological conditions, and 1.9 is only the first – albeit the key first example – of a series of models designed to address specific questions about the mechanism of the enzyme reaction (see Section 4.3.3).

Scheme 1.17

Intramolecular general acid catalysis can account for the rapid hydrolysis of 2-carboxyphenyl β-D-glucoside 1.9. See the text.

Scheme 1.17

Intramolecular general acid catalysis can account for the rapid hydrolysis of 2-carboxyphenyl β-D-glucoside 1.9. See the text.

Close modal

It is, however, a start. Given a convincing working model – in this case efficient catalysis near pH 4 of the hydrolysis of an aryl β-D-glucoside – a type of substrate normally showing no observable reaction under the conditions in the absence of an enzyme – it becomes possible to ask questions about the detailed mechanism of catalysis. Questions relevant to catalytic efficiency include the identification of significant electronic and stereoelectronic effects on the general acid, the leaving group, and their geometry of approach; and the influence of the medium on the reaction. The deeper understanding that emerges can then be applied – with appropriate reservations – to the same part-reaction going on in the enzyme-active site. It can also inform the design of improved enzyme models, as discussed further in Section 4.3.3.

We reserve the term enzyme mimics for more highly developed enzyme models, which combine (at least) substrate binding with catalysis: and have the evidence to prove it. It is not unusual to come across work that claims to mimic enzyme reactions but provides no evidence for the formation of a catalyst–substrate complex (either directly or in the form of saturation kinetics: see Section 2.3). In the absence of such evidence we file such systems (even if only provisionally) as pretenders. Always, the ultimate target is a properly qualified enzyme mimic that can also achieve turnover at a reasonable rate. This could in many cases be called, not unreasonably, an artificial enzyme. We now prefer (see the discussion in the Preface) the more general term model enzyme.

  • W. Aehle, ed., Enzymes in Industry; 3rd edn., Wiley-VCH, Weinheim, 2007.

  • J.-P. Behr, ed., The Lock and Key Principle, Wiley, Chichester, 1994.

  • R. Breslow, ed., Artificial Enzymes, Wiley-VCH, Weinheim, 2005.

  • A. R. Fersht, Structure and Mechanism in Protein Science, Freeman, New York, 1999.

  • P. A. Frey, A. D. Hegeman, Enzymatic Reaction Mechanisms, OUP, Oxford and New York, 2006.

  • H. Maskill, The Physical Basis of Organic Chemistry, OUP, Oxford and New York, 1985.

  • X. Y. Zhang and K. N. Houk, Why enzymes are proficient catalysts. Accts. Chem. Res., 2005, 38, 379–385.

1

Enzymes are well able to discriminate with remarkably high selectivity between closely similar substrates, from large molecules differing by just one carbon atom to enantiomers. In practice, not all enzymes show perfect selectivity, including enantioselectivity.15 

2

The hydrophobic effect near room temperature is now generally considered to result from entropy effects, resulting from the increased ordering of water molecules in contact with a nonpolar solute, as discussed. However, at high temperatures the unfavorable enthalpy of interaction of water with a nonpolar solute rather than with other water molecules becomes the dominant effect.

1.
Enzymes in Industry
; 3 edn.; Aehle, W., ed.;
Wiley-VCH
:
Weinheim
,
2007
2.
The Lock and Key Principle
; Behr, J.-P., ed.;
Wiley
:
Chichester
,
1994
3.
H.
Maskill
,
The Physical Basis of Organic Chemistry
;
Oxford University Press: Oxford
,
New York
,
1985
4.
Jencks
 
W. P.
J. Am. Chem. Soc.
1972
, vol. 
94
 pg. 
4731
 
5.
Jencks
 
W. P.
Chem. Rev.
1972
, vol. 
72
 (pg. 
705
-
718
)
6.
Kirby
 
A. J.
Lancaster
 
P. W.
J. Chem. Soc. Perkin Trans. 2
1972
, vol. 
1206
 
7.
Kirby
 
A. J.
McDonald
 
R. S.
Smith
 
C. R.
J.C.S. Perkin Trans. 2
1974
(pg. 
1495
-
1504
)
8.
Kirby
 
A. J.
Adv. Phys. Org. Chem.
1980
, vol. 
17
 (pg. 
183
-
278
)
9.
Jencks
 
W. P.
Adv. Enzymol.
1975
, vol. 
43
 (pg. 
219
-
410
)
10.
Menger
 
F. M.
Accts. Chem. Res.
1985
, vol. 
18
 (pg. 
28
-
33
)
11.
Mandolini
 
L.
Adv. Phys. Org. Chem.
1986
, vol. 
22
 pg. 
1
 
12.
Kirby
 
A. J.
Lloyd
 
G. J.
J. Chem. Soc. Perkin Trans. 2
1976
pg. 
1753
 
13.
L. B.
Spector
,
Covalent Catalysis by Enzymes
,
Springer-Verlag
,
New York
,
1982
14.
Zhang
 
X. Y.
Houk
 
K. N.
Accts. Chem. Res.
2005
, vol. 
38
 (pg. 
379
-
385
)
15.
Jaeger
 
K.-E.
Eggert
 
T.
Curr. Opin. Biotechnol.
2004
, vol. 
15
 (pg. 
305
-
313
)
16.
Houk
 
K. N.
Leach
 
A. G.
Kim
 
S. P.
Zhang
 
X. Y.
Angew. Chem., Int. Ed. Engl.
2003
, vol. 
42
 (pg. 
4872
-
4897
)
17.
Blokzijl
 
W.
Engberts
 
J.
Angew. Chem., Int. Ed. Engl.
1993
, vol. 
32
 (pg. 
1545
-
1579
)
18.
Otto
 
S.
Engberts
 
J. B. F.
N. Org. Biomol. Chem.
2003
, vol. 
1
 (pg. 
2809
-
2820
)
19.
Mannhold
 
R.
Rekker
 
R. F.
Perspec. Drug Discov. Des.
2000
, vol. 
18
 (pg. 
1
-
18
)
20.
DeChancie
 
J.
Houk
 
K. N.
J. Am. Chem. Soc.
2007
, vol. 
129
 (pg. 
5419
-
5429
)
21.
Jencks
 
W. P.
Ann. Rev. Biochem.
1997
, vol. 
66
 (pg. 
1
-
18
)
22.
Eliot
 
A. C.
Kirsch
 
J. F.
Ann. Rev Biochem
2004
, vol. 
73
 (pg. 
383
-
415
)
23.
Decker
 
H.
Schweikardt
 
T.
Tuczek
 
F.
Angew. Chem., Int. Ed. Engl.
2006
, vol. 
45
 (pg. 
4546
-
4550
)
24.
T.
Imoto
L. N.
Johnson
A. C. T.
North
D. C.
Phillips
J. A.
Rupley
, In:
The Enzymes
, P. D. Boyer (Ed.),
1972
, Vol.
VII
, p. 665
25.
Capon
 
B.
Smith
 
M. C.
Anderson
 
E.
Dahm
 
R. H.
Sankey
 
G. H.
J. Chem. Soc. B
1969
(pg. 
1038
-
1047
)
Close Modal

or Create an Account

Close Modal
Close Modal