Robert A Copeland Richard R Gontarek and Lusong

CONTENTS

11.1 Introduction 173

11.2 Modes of Inhibitor Interaction with Enzymes 174

11.3 Protein Dynamics in Enzyme Catalysis and Inhibitor Interactions 175

11.4 Mechanism-Based Inhibitor Design 177

11.4.1 Substrate Structure-Based Design 178

11.4.1.1 Nucleoside and Nucleotide Inhibitors of HIV Reverse

Transcriptase 178

11.4.1.2 Human Steroid 5a-Reductase Inhibitors 179

11.4.2 Intermediate State-Based Design 179

11.4.2.1 Inhibitors of Hydroxymethylglutaryl-CoA Reductase

(HMG-CoA Reductase) 179

11.4.2.2 Inhibitors of Purine Nucleoside Phosphorylase 181

11.5 Biostructure-Based Design 182

11.5.1 Structure-Based Design of Protein Kinase Inhibitors 183

11.5.2 Structure-Based Design of HIV Protease Inhibitors 185

11.6 Concluding Remarks 188

Further Readings 188

11.1 INTRODUCTION

Inhibition of disease-associated enzyme targets by small molecular weight drugs is a well-established modality for pharmacologic intervention in human disease. Indeed, a recent survey of the FDA Orange Book showed that more than 300 marketed drugs work through enzyme inhibition. Among orally dosed drugs in clinical use, nearly half of them function by inhibition of specific enzyme targets. Likewise, much of current preclinical drug discovery efforts in biotechnology and pharmaceutical companies—as well as those same efforts in government and academic laboratories—is focused on the identification and optimization of small molecules that function by inhibition of specific enzyme targets. The reasons for the popularity of enzymes as targets for drug discovery have been reviewed a number of times recently (see for example, Copeland (2005)). In brief, enzymes make good drug targets for two significant reasons. First, the catalytic activity of specific enzymes is often critical to the pathophysiology of disease, such that inhibition of catalysis is disease modifying. Second, the binding pockets for natural ligands of enzymes, that play a crucial role in catalytic activity, are often uniquely well-suited for interactions with small molecule drugs. Thus, the very nature of the chemistry of enzyme catalysis makes these proteins highly vulnerable to inactivation by small molecule inhibitors that have the physicochemical characteristics of oral drugs.

Enzyme catalysis involves the conversion of a natural ligand (the substrate) into a different chemical species (the product), most often through a process of chemical bond breaking and formation steps. The chemical transformation of substrate to product almost always involves the formation of a sequential series of intermediate chemical species along the reaction pathway. Paramount in this reaction pathway is the formation of a short-lived, high-energy species referred to as the transition state. To facilitate this sequential process of intermediate species formation, the ligand-binding pocket(s) of enzymes must undergo specific conformational changes that induce strains at correct locations and align molecular orbitals to augment the chemical reactivity of the appropriate functionalities on the substrate molecule(s), at defined moments during the reaction cycle. The bases of mechanistic enzymology include understanding the chemical nature of the various intermediate species formed, and their interactions with those elements of the enzyme-binding pocket that facilitate chemical transformations. When these studies are coupled with structural biology methods, such as x-ray crystallography and multidimensional nuclear magnetic resonance (NMR) spectroscopy, a rich understanding of the structure-activity relationships (SAR) that attend enzyme catalysis can be obtained. What is germane to the present discussion is that this structural and mechanistic understanding can be exploited to discover and design small molecule inhibitors—mimicking key structural features of reaction intermediates—that form high-affinity interactions with specific conformational states of the ligand-binding pocket of the target enzyme. In this chapter, we describe the application of mechanistic and structural enzymology to drug discovery efforts with an emphasis on the evolution of structural changes that attend catalysis and the exploitation of these various conformational forms for high-affinity inhibitor development.

11.2 modes of inhibitor interaction with enzymes

The simplest enzyme-catalyzed reaction one can envisage is that of a single substrate (S) being converted by the enzyme (E) to a single product (P). This reaction can be summarized by the following equation:

As summarized by Equation 11.1, enzyme and substrate combine to form a reversible initial encounter complex (ES) that is governed by a forward rate constant for association (k1) and a reverse rate constant for dissociation (k2). The equilibrium dissociation constant for the ES complex is given the symbol KS and is mathematically equivalent to the ratio of the rate constants k2/kj. Subsequent to initial complex formation, a series of chemical steps ensue that are collectively quantified by the cumulative rate constant kcat. Thus, kcat is not a microscopic rate constant, but rather summarizes all of the intermediate states that must be formed during the chemical transformation of substrate to product (see Section 11.3 for more details on the individual intermediate steps that may contribute to kcat).

Three modes of inhibitor interaction with an enzyme target can be defined, based on their effects on the catalytic steps summarized in Equation 11.1. Competitive inhibitors bind to the free enzyme in a manner that blocks the binding of substrate so that they increase the apparent value of KS, but have no effect on the apparent value of kcat. Noncompetitive inhibitors can bind to both the free enzyme and to the ES complex (or intermediate species that follow the formation of the ES complex). Such inhibitors can have some effect on the value of KS but show the greatest effect on kcat, as they inhibit by blocking catalytic steps subsequent to substrate binding. Finally, uncompetitive inhibitors have no affinity for the free enzyme and only bind subsequent to the formation of the ES complex. These inhibitors decrease the apparent value of KS (i.e., increasing the apparent affinity of the enzyme for substrate) and also decrease the apparent value of kcat (i.e., diminishing the ability of the enzyme to catalyze chemical steps subsequent to substrate binding). Among drugs in current clinical use one finds multiple examples of each of these three modalities of enzyme inhibition.

11.3 PROTEIN DYNAMICS IN ENZYME CATALYSIS AND INHIBITOR INTERACTIONS

The catalytic pathway summarized in Equation 11.1 is a gross oversimplification of even the simplest of enzymatic reactions. At minimum, this reaction requires the formation of two additional forms of enzyme-ligand binary complex, these being the enzyme-transition state complex and the enzyme-product complex. In practice, one often finds that additional intermediate states are accessed during the catalytic cycle of an enzyme. Thus, one can say that the catalytic cycle of an enzyme is a sequential series of protein-ligand complexes, each representing a unique chemical form of the ligand with attendant changes in the protein conformation of the ligand-binding pocket. Each conformational state of the ligand-binding pocket that is accessed during this catalytic cycle is a potential target for small molecule drug interactions. Hence, one can think of the ligand-binding pocket of an enzyme not as a single target for drug intervention, but rather a collection of targets that evolve and interconvert over the time course of catalytic turnover.

To illustrate these concepts, let us consider the reaction cycle of an aspartyl protease. The aspartyl proteases constitute a family of protein/peptide hydrolyzing enzymes that use a pair of aspartic acid residues within the enzyme active site to facilitate peptide bond cleavage. Figure 11.1 provides

Flap closing
H
FIGURE 11.1 The reaction pathway for an aspartyl protease. (Adapted and modified from Copeland, R.A., Evaluation of Enzyme Inhibitors in Drug Discovery: A Guide for Medicinal Chemists and Pharmacologists, Wiley, Hoboken, NJ, 2005.)

a schematic representation of the canonical reaction cycle of an aspartyl protease, illustrating the changes in active site structure that attend catalysis. Before substrate binding the enzyme is in a resting form (E) in which the two active site aspartic acid residues are bridged by a water molecule. One of the aspartates is present in the protonated acid form while the other is present as the conjugate base form, and the two residues share the acid proton through a strong hydrogen bond. Initial substrate binding causes disruption of the hydrogen bonding interactions and displacement of the water molecule in species ES. After initial substrate binding, a flexible loop of the protein, referred to as the "flap" closes down over the active site to occlude the active site groups from bulk solvent. The substrate-bound enzyme in this altered conformation is referred to as form E'S in Figure 11.1. Subsequently, the water of the enzyme active site attacks the carbonyl carbon of the scissile peptide bond to form a dioxy, tetrahedral carbon center on the substrate. This constitutes the enzyme-bound transition state of the reaction and is symbolized as E'S* in the figure. Bond rupture then occurs, leading to a species with both active site aspartates protonated and with both the anionic and cationic peptide products bound (form E'P). After that, the flap retracts from the active site to generate a new conformational state of the active site, referred to as FP. With the flap out of the way, the product peptides can now dissociate from the enzyme, forming enzyme state F. Deprotonation of one of the active site aspartates occurs next to form state G. Finally, addition of a water molecule returns the enzyme back to the original conformational state E, thus completing the catalytic cycle.

It is clear from Figure 11.1 that protein dynamics is an important component of the catalytic cycles of enzyme reactions. As stated earlier, the importance of this concept to drug discovery is that each intermediate state accessed along the reaction pathway provides unique opportunities for inhibitor interactions. For example, in the case of the aspartyl proteases, there are three distinct ligand-free conformational forms of the enzyme (states E, F, and G); small molecule inhibitors are known that preferentially bind to each of these individual states.

The protein structures that are represented along the reaction pathway of an enzyme reflect a collection of conformational microstates that interconvert among themselves through rotational and vibrational excursions. Thus, any particular conformation of the enzyme can be represented as a manifold of conformational substates (microstates), each stabilized to different degrees by specific interactions with ligands. Isomerization of the enzyme from one structure to another therefore involves the potential energy stabilization of certain microstates at the expense of others.

Similarly, high-affinity inhibitor interactions often involve isomerization of the enzyme from an initial structure resembling the unliganded enzyme to a new conformation in which a particular microstate(s) is highly stabilized by interactions with the inhibitor. Kinetically, this requires two distinct steps in overall binding of high-affinity inhibitors. As illustrated in Figure 11.2A, the first step involves the formation of a reversible encounter complex between the enzyme and inhibitor (EI), which often displays only modest affinity. Forward binding of the inhibitor to the enzyme is dictated by the association rate constant k3 and dissociation of the initial EI complex is dictated by the dissociation rate constant k4. Once the EI complex is formed, enzyme isomerization can occur to form the much higher affinity complex E*I. The forward conversion of EI to E*I is dictated by the isomerization rate constant k5 and the reverse conversion of E*I back to EI is dictated by rate constant k6. All three enzyme forms (E, EI, and E*I) can also be represented as potential energy diagrams, as illustrated in Figure 11.2B. It should be noted that the affinity of the enzyme-inhibitor complex is related to the potential energy stabilization of the system, which is reflected in the depth of the potential well in the energy diagram. The deeper the potential energy well for the inhibited form, the more energy that is required to escape this well and thus access the other conformational microstates required for continued catalysis. The potential energy stabilization of the inhibitor-bound form is mediated by productive interactions between the inhibitory ligand and the binding pocket of the protein, in the form of hydrogen bonds, electrostatic interactions, hydrophobic interactions, van der Waal forces, and the like. These concepts can also be conceptualized in terms of an induced fit between the binding pocket and the inhibitor, as illustrated in cartoon form in Figure 11.2C.

FIGURE 11.2 A two-step inhibitor-binding mechanism involving initial binding of the inhibitor to the enzyme in one conformation and a subsequent isomerization of the enzyme to a new conformation.

(A) Reaction sequence illustrating the forward and reverse kinetic steps of binding and enzyme isomerization.

(B) Potential energy diagrams representing the three conformational states of the enzyme: E, EI, and E*I.

(C) Cartoon representation of the inhibitor binding and enzyme isomerization steps in this mechanism.

The state E*I then represents a state of high-affinity interactions between the enzyme and the inhibitor. As long as the inhibitor is bound to the enzyme, either in the form of EI or E*I, the biological activity of the enzyme is blocked. Dissociation of the inhibitor from the enzyme can occur for any reversible inhibition process; once the enzyme is free of inhibitor, catalytic activity is restored. In the case of tight-binding inhibitors that induce enzyme isomerization, the overall rate constant for inhibitor dissociation, koff, must take into account reversal of the isomerization step, reisomerization via k5, and dissociation of the inhibitor from EI via k4. Mathematically, the value of koff is given by koff = ( kf6 k ) (11.2) (k3 + k5 + k6)

For this two-step binding mechanism, it is almost always the case that the reverse isomerization step, mediated by rate constant k6, is by far the slowest step in overall inhibitor dissociation. Thus, the lower the value of k6, the longer the duration of potent inhibition by the drug. There are a large and growing number of examples of highly efficacious drugs that demonstrate tight-binding interactions with their target enzyme through a two-step enzyme isomerization mechanism as described here. In some cases, the slowness of the reverse isomerization step leads to prolonged duration of inhibition that may translate into an extended duration of pharmacodynamic activity in vivo; this concept is considered further in Section 11.6.

11.4 MECHANISM-BASED INHIBITOR DESIGN

Enzymes are designed by nature to catalyze a specific chemical reaction. As described earlier, every enzyme accesses a sequential series of intermediate states along the reaction pathway, thus providing unique opportunities for inhibitor interaction. Consequently, enzyme inhibitors can be effectively designed based on an understanding of the mechanistic and structural details of the catalyzed reaction pathway. The majority of known enzyme inhibitors are structurally related to natural ligands of the enzymatic reaction; a recent survey suggested that more than 60% of marketed drugs that target enzymes are either analogs of substrates or enzyme cofactors, or they undergo catalyzed structural conversion within the active site of an enzyme. Substrate, cofactor, and product mimicry, however, is not the only method for the design of high-affinity, selective enzyme inhibitors. Advances in transition state theory during the past three decades have helped to establish an alternative approach for mechanism-based design: intermediate-state-based design (sometimes also referred to as transition-state-based design). In this latter approach, inhibitors that mimic the steric and electronic features of high-energy reaction intermediate states are designed to capitalize on the specific interaction of active site residues with the reaction intermediate. In the next two sections, cases for substrate structure-based design and intermediate-state-based design will be discussed to exemplify inhibitor design strategies that have led to successfully marketed products or clinical candidates.

11.4.1 Substrate Structure-Based Design

11.4.1.1 Nucleoside and Nucleotide Inhibitors of HIV Reverse Transcriptase

HIV reverse transcriptase (RT) is one of two main targets for antiacquired immunodeficiency syndrome (AIDS) therapy (the second target being the HIV protease; vide infra). The RT enzyme catalyzes the synthesis of double stranded proviral DNA from single stranded genomic HIV RNA. Drugs targeting HIV RT can be divided into two categories: (i) nucleoside and nucleotide RT inhibitors (NRTIs), which are competitive with respect to the natural deoxynucleotide triphosphates (dNTPs) and serve as alternative substrates for catalysis (resulting in chain termination); (ii) nonnucleoside RT inhibitors (NNRTIs), which are allosteric, noncompetitive inhibitors that bind at a site distal to the RT active site. NRTIs were the first class of chemotherapeutic agents to be utilized in the clinic to treat AIDS patients and offer excellent examples of inhibitor design based on substrate mimicry. The first NRTI, Zidovudine (AZT) was approved by the FDA in 1987 (Figure 11.3).

This molecule is a thymidine analog with an azido group in place of the hydroxyl group at the 3' position of the ribose. Since the advent of AZT-based therapy, a number of NRTIs have joined the

FIGURE 11.3 Representative FDA-approved nucleoside/nucleotide RT inhibitors (top panel) that closely mimic the natural deoxynucleotides (bottom panel).

anti-AIDS treatment armamentarium. Most of these NRTIs are nucleoside analogs with the exception of tenofovir disoproxil fumarate (TDF), which is a nucleotide analog of adenosine phosphate. The NRTIs are administered as unphosphorylated prodrugs. Upon entering the host cell, these prodrugs are recognized by cellular kinases and further converted to the tri-phosphorylated form. The tri-phosphorylated NRTIs then bind to the active site of RT and are catalytically incorporated into the growing DNA chain. The incorporated NRTIs block the further extension of the chain since the NRTIs lack the 3' hydroxyl group on their ribose or pseudoribose moiety and thus cannot form the 3'-5' phosphodiester bond needed for DNA extension. NRTIs are one of the major classes of inhibitors used in all combination therapies for the treatment of HIV-infected patients. However, the clinical successes of these agents are limited by viral resistances to NRTIs, arising through mutations in the coding region of RT. These mutations confer viral resistances through improved discrimination of a nucleotide analog relative to the natural substrate, or by increased phosphorolytic cleavage of an analog-blocked primer. To overcome these acquired resistances, the design of the next generation of NRTIs has been mainly focused on two fronts: (i) nucleoside analogs possessing a 3' hydroxyl group that can induce delayed polymerization arrest; (ii) nucleotide analogs that are designed to be incorporated into the viral genome during replication. These nucleotide analogs can introduce mutations into the HIV genome through mispairing and blockade of the replication process.

11.4.1.2 Human Steroid 5a-Reductase Inhibitors

The human enzyme steroid 5a-reductase is responsible for the conversion of testosterone (T) to the more potent androgen, dihydrotesterone (DHT). It has been shown that abnormally high 5a-reductase activity in humans leads to excessively high DHT levels in peripheral tissues. Inhibition of 5a-reductase thus offers a potential treatment for DHT-associated diseases, such as benign prostate hyperplasia, prostate cancer, acne, and androgenic alopecia. In humans, there are two types of steroid 5a-reductase: type I and type II. The type I 5a-reductase is mainly expressed in the sebaceous glands of skin and the liver, while the type II enzyme is most abundant in the prostate, seminal vesicles, liver, and epididymis. The first 5a-reductase inhibitor approved for clinical application in the United States was finasteride; it is currently employed in the treatment of benign prostatic hyperplasia (BPH) in men. This compound is approximately 100-fold more potent toward the type II than the type I isozyme of 5a-reductase. In humans, finasteride decreases prostatic DHT levels by 70%-90%, resulting in reduced prostate size. The detailed biochemical characterization of finasteride inhibition suggested that finasteride is a mechanism-based inhibitor. It is proposed that by closely mimicking the substrate (testosterone), finasteride is accepted as an alternate substrate and forms an NADP-dyhydrofinasteride adduct at the enzyme active site (Figure 11.4). This covalent NADP-dyhydrofinasteride adduct represents a bisubstrate analog with extremely high affinity (K < 1 x 10-13 M) to the type II 5a-reductase. Interestingly, finasteride is also a mechanism-based inhibitor of the human type I 5a-reductase. However, the NADP-dyhydrofinasteride adduct formation rate at the type I 5a-reductase active site is reduced by more than 100-fold compared to that for the type II isozyme. This difference in NADP-dyhydrofinasteride adducts formation rate accounts for the isozyme selectivity of finasteride both in vitro and in vivo. Knowledge of the mechanism of inhibition of 5a-reductase by 4-azasteroids (represented by finasteride) and of the SAR for dual 5a-reductase inhibition, led to the discovery of a potent, dual inhibitor of 5a-reductase, known as dutasteride. Dutasteride, is equipotent versus type I and type II 5a-reductase and demonstrates exceptional in vivo potency. This compound has also been approved for clinical use in the treatment of BPH.

11.4.2 Intermediate State-Based Design

11.4.2.1 Inhibitors of Hydroxymethylglutaryl-CoA Reductase (HMG-CoA Reductase)

The biosynthetic pathway for cholesterol involves more than 25 different enzymes. The enzyme 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase catalyzes the conversion from

NADPH

NADP+

(A) Testosterone

Dihydrote stosterone (DHT)

NADPH

NADP+

(A) Testosterone

Dihydrote stosterone (DHT)

(B) Finasteride

Dutasteride

(C) NADP-dihydrofinasteride adduct

FIGURE 11.4 (A) 5a-Reductase catalyzed conversion of testosterone (T) to dihydrotesterone (DHT); (B) chemical structures for finasteride and dutasteride; and (C) the proposed structure of the NADP-dihydrofinasteride adduct. PADPR, phosphoadenosine diphosphoribose.

(B) Finasteride

Dutasteride

(C) NADP-dihydrofinasteride adduct

FIGURE 11.4 (A) 5a-Reductase catalyzed conversion of testosterone (T) to dihydrotesterone (DHT); (B) chemical structures for finasteride and dutasteride; and (C) the proposed structure of the NADP-dihydrofinasteride adduct. PADPR, phosphoadenosine diphosphoribose.

HMG-CoA to mevalonate, the rate-limiting step of the entire pathway. Inhibition of HMG-CoA reductase provides a very attractive opportunity to inhibit cholesterol biosynthesis because no buildup of potentially toxic precursors occurs upon inhibition. In 1976, Japanese microbiologist Akira Endo isolated a series compounds including ML236B (compactin) from Penicillium citrinum with powerful inhibitory effect on HMG-CoA reductase. Since then, seven HMG-CoA reductase inhibitors have become marketed drugs for lowering cholesterol levels (Figure 11.5).

These HMG-CoA reductase inhibitors, commonly referred to as statins, have accounted for the majority of prescriptions for cholesterol-lowering drugs worldwide. All the statins in clinical use are analogues of the substrate HMG-CoA with an HMG-like moiety, which may be present in an inactive lactone form in the prodrugs (Figure 11.5.). These statins are classified in two groups according to their molecular structures. Type I statins, including lovastatin and simvastatin, are lactone prodrugs originally isolated from fungi. They are enzymatically hydrolyzed in vivo to produce the active drug. Type II statins are all synthetic products with larger groups attached to the HMG-like moiety. All the statins are competitive with respect to HMG-CoA and noncom-petitive with respect to NADPH, a cosubstrate of the reaction. Crystal structures of HMG-CoA reductase complexed with six different statins showed that the statins occupy the HMG-binding region, but do not extend into the NADPH site. The orientation and bonding interactions of the HMG-moiety of the statins resemble those of the substrate complex. However, from combination crystal structures, binding thermodynamics, and SAR studies it is clear that the 5'-hydroxyl group of the acidic side-chain acts as a mimetic of the tetrahedral intermediate of the reduction reaction. The multiple H-bonds between the C-5-OH of the statins and the HMG-CoA reductase active site contribute significantly to the tight binding of the statin inhibitors. Strictly speaking, the HMG-CoA reductase inhibitors are not products of rational design; rather they were identified

S-CoA NADPH NADP+

S-CoA NADPH NADP+

S-CoA Melvaldyl-CoA

HMG-CoA

S-CoA Melvaldyl-CoA

HO H3C

NADPH NADP+

+ CoASH

Melvalonate

HMG-CoA

Melvalonate

Compactin

Lovastatin

Simvastatin

Pravastatin

Compactin

Lovastatin

Simvastatin

Pravastatin

Fluvastatin
Cerivastatin
Atorvastatin

Rosuvastatin

Rosuvastatin

FIGURE 11.5 Structures of HMG-CoA reductase reaction substrate, tetrahedral intermediate, product, and the statin inhibitors. Compactin, Lovastatin, and Simvastatin are type I statins. All other statins are type II statins. The melvaldyl tetrahedral intermediate that is mimicked in all statins is shaded in gray.

through natural product screening and analoging of the natural product hits. Nevertheless, it is quite clear that all statins share a common strategy for inhibiting their target: tetrahedral intermediate state mimicry.

11.4.2.2 Inhibitors of Purine Nucleoside Phosphorylase

Purine nucleoside phosphorylase (PNP) catalyzes the phosphorolysis of 6-oxypurine nucleosides and deoxynucleosides. In humans, the PNP pathway is the only route for deoxyguanosine degradation and genetic deficiency in this enzyme leads to profound T-cell-mediated immunosuppression. Inhibition of PNP has applications in treating aberrant T lymphocyte activity, which is implicated in T-cell leukemia and autoimmune diseases. The challenge to inhibitor design for PNP arises from the abundance of the enzyme in human tissues. It has been shown that near complete inhibition of PNP (>95%) is required for significant reduction in T-cell function. Structural-based inhibitor design produced some inhibitors with Kd values in the nanomolar range. However, clinical evaluations showed that these inhibitors did not produce sufficient inhibition of PNP to be effective anti-T-cell therapies. Much more potent PNP inhibitors were later designed with the aid of transition state analysis. In theory, a perfect transition state inhibitor of PNP should bind with a Kd value of approximately 10-17 M (10 attomolar). The structure of the transition state for human PNP was determined by Schramm and coworkers by measuring kinetic isotope effects. Their studies revealed a transition state with significant ribooxycarbenium character (Figure 11.6). Based on the features of this transition state, compounds with picomolar affinity to PNP were synthesized. Among them, o

OH OH

Was this article helpful?

0 0

Post a comment