key: cord-010203-dt9m596i
authors: Hellen, Christopher U.T.; Wimmer, Eckard
title: Viral proteases as targets for chemotherapeutic intervention
date: 2004-08-26
journal: Curr Opin Biotechnol
DOI: 10.1016/0958-1669(92)90010-g
sha: 
doc_id: 10203
cord_uid: dt9m596i

Many viruses encode proteinases that are essential for infectivity, and are consequently attractive chemotherapeutic targets. The biochemistry and structure of the human immunodeficiency virus proteinase have been characterized extensively, and potent peptide-mimetic inhibitors have been developed. Techniques and strategies used to improve the efficiency of these compounds are likely to be applicable to other viral proteinases.

Many human and animal viruses encode proteinases that play important roles at different stages in the infection cycle, including separation of functionally different domains from a precursor polyprotein (enabling cleavage products to be transported to different cellular compartments) and regulation of a variety of events in viral replication, such as uncoating, activation of replicative enzymes and morphogenesis [1] . These proteinases are essential for virus infectivity, and have therefore come to be seen as attractive targets for chemotherapeutic intervention, particularly because they have unusual cleavage specificities that differ from those of host proteases.

Proteinases are encoded by all retroviruses, including HIV and human T-cell leukemia virus and have been identified in a growing number of DNA and positivesense RNA viruses. These include adenoviruses and herpesviruses [2",3"] as well as picornaviruses, flaviviruses (such as Dengue and yellow fever viruses), pestiviruses and the related hepatitis C virus. A number of these viruses cause diseases of medical or veterinary importance that are not amenable to conventional preventive or prophylactic measures, and proteinase inhibition therefore represents a valid alternative therapeutic approach. Many viral proteinases have only been identified recently, and characterization is consequently in its early stages. To illustrate potential strategies in the analysis of viral proteinases, and in the design and development of inhibitors we shall therefore focus on the picornavirus and retrovirus proteinases, as they have elicited the greatest academic and industrial interest, and as a result, have been characterized in detail.

HIV-1 has the genetic organization 5'-gag-pol-env-3' that is typical of retroviruses. The gag and pol genes encode inner structural, and replicative proteins respectively, and are translated as polyproteins that are cleaved at eight sites by the proteinase (PR) ( Fig. 1 and Table 1 ). These polyproteins are transported to the plasma membrane and cleavage occurs after budding of immature particles, resulting in morphological changes associated with virion maturation. The substrate specificity of HIV-1 PR is puzzling in that PR catalyzes specific cleavage at a small number of polyprotein sites that show no apparent sequence conservation. Analysis of viral and non-viral substrates suggests that no subsite has absolute specificity, and that a combination of moderate interactions may be sufficient to confer catalytic specificity [4"]. Heterogeneity in the composition of viral polyprotein cleavage sites probably plays a role in determining the rate, and consequently the order, of cleavage at different sites.

A proteinase-deficient HIV mutant produced noninfectious immature virions containing unprocessed polyprotein [5] , an observation that is crucial to the consideration of PR as a therapeutic target. HIV-1 PR is a C 2 symmetric homodimeric aspartyl proteinase that consists of two identical 99-amino-acid subunits. Their termini interdigitate at the dimer interface, but otherwise the topology of HIV-1 PR is similar to that of pepsin-like aspartyl proteinases [6"'].

Techniques for the large scale purification of recombinant HIV-1 PR, and for the routine assay of its proteolytic activity are fundamental prerequisites for the development of inhibitors, and various methods have been reported [7"']. The successful design of substratebased inhibitors of other aspartyl proteinases, such as pepsin and renin, suggested the strategy of designing analogous peptide-mimetic inhibitors of HIV-1 PR by incorporating non-hydrolyzable 'transition-state' mimics into substrate analogues [7"',8",9-]. This approach is based on a key step in aspartyl proteinase catalysis: generation of a tetrahedral diol by hydration of a trigonal amide (Fig. 2a) . The high (millimolar) K m values of peptide substrates indicate that potent (i.e. nanomolar) PR inhibitors must incorporate structural features that significantly increase their binding affinity. Peptide-based inhibitors have several disadvantages, including vulnerability to degradative enzymes, rapid clearance, and poor oral absorption. These problems are commonly addressed by minimizing size and peptide-like character of promising lead compounds.

The first reported PR inhibitor was pepstatin [10], a diagnostic inhibitor of aspartyl proteinases. This weak inhibitor contains two statine residues that embody transition state analogue I (Fig. 2b ). To identify more potent inhibitors, Dreyer et al. [11] compared the effectiveness as PR inhibitors of five different classes of dipeptide isosteres inserted into a consensus heptapeptide template. Heptapeptides (P4-P3' or P3-P4') are the shortest substrates that are cleaved efficiently by PR. Statine-based (I), reduced amide (II) and phosphinate (III) transition state analogues exhibited modest potency, but placement of Phe-Gly hydroxyethylene dipeptide isosteres (IV) into the consensus template yielded compounds that inhibited HIV PR at nanomolar concentrations in vitro and prevented polyprotein processing, virion maturation and viral spread at 25-100btM in cell culture. Truncation and extensive structure-activity analysis at the P1, PI" and P2' positions led to the identification of highly potent (subnanomolar) PR inhibitors based on dihydroxyethylene (V) [12-] and hydroxyethylene (IV) [13"] isostere transition state analogues. Potency was enhanced by incorporating residues that stabilize the extended inhibitor structure, presumably due to optimized hydrogenbonding in the substrate binding cleft. For example, the P2' and P3' residues (Leu-Phe) can effectively be replaced by various substituted aminobenzocycloalkanes [14" ]. The PI' position can accept side-chains unrelated to natural amino acids, allowing modifications to be made that enhance solubility and thus cell penetration [15", 16-] . Such substitutions may reduce binding affinity (K i) but the enhanced solubility may nevertheless result in a net increase in antiviral activity.

The ability to cleave the amino terminus to proline distinguishes HIV PR from non-viral aspartyl proteinases. Hydroxyethylamine (VI) structures that readily accomodate the prolyl imino acid have been incorporated into a number of potent inhibitors [17,18",19,20" ]. Modification of a protected tripeptide incorporating this structure by substituting the imino residue decahydroisoquinoline at the PI' position yielded highly potent inhibitors of PR in vitro and in cell culture, such as the compound Ro 31-8959 [19,21" ]. This inhibitor was expected to have considerable selectivity, and indeed it inhibited human aspartyl proteinases such as gastricsin, renin and pepsin by less than 50% at a concentration of 10pM. Typical IC90 values for Ro 31-8959 (5-30 nM) are 1000-fold below its cytotoxic concentration in uninfected host cells. Recent reports indicate that a 600 mg oral dose every 8 hours is sufficient to maintain the mean human plasma concentration at about 70nM. The potency of Ro 31-8959 is strongly dependent on tight binding by the P2 and particularly the P3 substituents, whereas binding of a second class of hydroxyethylamine inhibitors (which contain a noncyclic, secondary amine in place of the decahydroisoquinoline residue) [22" ] is less dependent on these interactions, and this second class binds more tightly in the PI"-P2' region.

Peptide substrates are inherently asymmetric and the PR dimer must therefore lose its perfect C 2 symmetry during catalysis. Symmetry is permissible for inhibitors, however, and might even improve binding affinity and selectivity over endogenous aspartyl proteinases. These considerations have led to the design of a serieS-of diaminoalcohol-and diaminodiolbased inhibitors with C 2 (VII) or pseudo-C 2 symmetry [23,24"-26" ]. These inhibitors are potent even at subnanomolar concentrations and highly selective in vitro, but most have suffered from poor solubility, leading to modest potency in cell culture. Strategies to circumvent this deficiency and thus enhance activity in cell culture have included modification of terminal residues and their linkage groups to increase solubility. These interactions include extensive Van der Waal's contacts with residues that define the hydrophobic 82-82' binding pockets, and a hydrogen bonding system that sandwiches the inhibitor strand between the catalytic cleft and the flaps. The hydroxyl groups of type V, VI and VII inhibitors form hydrogen bonds with both catalytic aspartates. Significantly, all complexes contain a tetrahedrally coordinated active-site water molecule, which bridges two flap residues and two inhibitor carbonyl groups, prompting suggestions that an improved inhibitor would contain a functional replacement for the water [24",27] . The similarities in the extended conformation of all inhibitors, as well as in the induced conformational changes that they cause in the enzyme, indicate that it is possible to model and improve peptidic inhibitors on the basis of these known structures. An alternative approach to discover novel templates for the design of non-peptide inhibitors is to search three-dimensional structure databases for molecules with a shape that is complementary to the active-site cleft. To date, this approach has led to identification of the antipsychotic agent haloperidol as a weak PR inhibitor [29] .

The Picornaviridae are a family of small icosahedral viruses that includes the etiological agents of several important human and animal diseases. It consists of five genera, including rhinovirus (the common cold virus) and enterovirus (e.g. poliovirus and hepatitis A virus).

Picornaviruses have a positive-sense monopartite RNA genome that encodes a single large polyprotein. It is processed by three different proteolytic activities which can each be regarded as serving a distinct function ( Fig. 3 and Table 2 ) [30" ]. The initial event in this cascade is cleavage by 2AP r° at its own amino terminus, separating the P~ structural protein precursor from the nascent polyprotein. Secondly, functional proteins are released from the P1 and P2-P3 (non-structural) protein precursors by 3CP r° or its precursors. Finally, maturation cleavage of the VP0 capsid protein occurs on encapsidation of viral RNA to yield infectious virus particles. In addition to their role in viral replication, the 2A and 3C proteinases of poliovirus (and by implication, of other picornaviruses) are responsible for aspects of the dramatic inhibition of host cell RNA and protein synthesis that occurs on infection. The 2A proteinase is involved in degradation of the eukaryotic initiation factor elF~Fy, which is correlated with shut-off of cap-dependent translation [31] , and 3CP ro inactivates transcription factor IIIC, inhibiting polymerase III transcription [32" ].

Sequence alignment and inhibitor studies suggested that both 2A and 3C proteinases are related structurally to trypsin-like serine proteinases, with the no- at all eight sites within the polyprotein (Fig. 3) . Sites in other picornaviruses are slightly more heterogenous. Poliovirus 2Apro cleaves Tyr-Gly dipeptides at the P1-2A junction and within the three-dimensional polymerase, but although all corresponding sites in other picornaviruses have a Gly residue at the PI' position, various residues occur at the P1 position. Aliphatic residues occur at the P4 positions of most 2AP r° and 3CP r° sites. Mutagenesis and peptide cleavage experiments indicate that cleavage site recognition depends on a minimum substrate length (six residues for 3CP r°) and the presence of specific residues at positions that differ according to both the virus and the proteinase [38,39",40",41"- 43" ]. There are additional conformational determinants of recognition of cleavage sites within polyproteins, so the large (millimolar) K m values of peptide substrates may reflect their greater conformational freedom. Potential peptidemimetic in- hibitors are likely to exhibit similar flexibility, and must therefore be conformationally constrained and incorporate structural features that increase their binding affinity. The lack of absolute specificity at most subsites, and the requirement for peptide substrates to extend to the P4 position indicates that the substrate binding clefts of 3CP r° and probably 2AP r° are capable of extensive hydrogen bond interactions with such inhibitors. However, only a few inhibitors of 3CP ro have been reported [44",45] .

Proteases are encoded by several DNA viruses and numerous RNA viruses in addition to the picornaviruses and retrovimses discussed above. Although they are all potential targets for chemotherapeutic intervention, significant progress in inhibitor development has only been reported for HIV-1 PR. In the few years since its identification, the structure of PR and numerous inhibitor complexes have been determined, and highly potent peptidemimetic inhibitors have been developed. Knowledge of the strategies used in enhancing the potency and specificity of PR inhibitors, and in overcoming the inherent limitations of peptide-based inhibitors is likely to prove invaluable in the development of peptidemimetic inhibitors of other viral proteinases. HEINRIKSON RL: The Complexities of AIDS: an Assessment of the HIV Protease as a Therapeutic Target. Chem Today 1991, 9:6-27. An account of the recent progress in substrate-based inhibitor design is complemented by a concise summary of the structure of HIV PR and a clear exposition of its application to structure-based inhibitor design.

NORBECK :1225-1228. The first of" series of disclosures [I3"-15",16"'] from a group at Merck illustrating strategies to selectively modify termini and side-chain residues to increase potency, reduce size and peptidic character, and to increase solubility (and hence antiviral activity) of hydroxyethylene isostere-based inhibitors. This paper describes the effect of modification of the PI', P2' and P3' residues.

Benzocycloalkyl Amines as Novel C-termini for HIV-1 Protease Inhibitors

HIV-1 Protease Inhibitors Based on Hydroxyethylene Dipeptide Isosteres: an Investigation into the Role of the PI' Side Chain on Structure-activity

Systematic modification of the P]' residue of the lead inhibitors described in [13",14"] established that side chains unrelated to natural amino acids are tolerated at this position, permitting substitutions that increase solubility and cell penetration

Synthesis and Antiviral Activity of a Series of HIV-1 Protease Inhibitors with Functionality Tethered to the P1 or PI' Substituents: X-ray Crystal Structure Assisted Design

Computer-assisted molecular modelling was used to design derivatives of the lead inhibitor L-685,434 [14"] with increased cell penetration and antiviral potency. An X-ray crystal structure of the

Hydroxyethylamine Analogues of the p17/p24 Substrate Cleavage Site are Tight-binding Inhibitors of the HIV Protease

Effect of Hydroxyl Group Configuration in Hydroxyethylamine Dipeptide Isosteres on HIV Protease Inhibition. Evidence for Multiple Binding Modes

In a series of hydroxyethylamine isostere inhibitors, the preferred diastereomeric configuration of an essential hydroxyl group depended on both the length and nature of the peptide framework. This hydroxyl group is hydrogen bonded to the two catalytic Asp residues

Rational Design of Peptide-based HIV Proteinase Inhibitors

Novel Binding Mode of Highly Potent HIV-proteinase Inhibitors Incorporating the (R)-Hydroxyethylamine Isostere

Binding of the hydroxyethylamine inhibitor Ro 31-8959 described in

NA: Effects of a Specific Inhibitor of HIV Proteinase (RO 31-8959) on Virus Maturation in a Chronically Infected Promonocytic Cell Line (U1)

Antiviral activity of the potent HIV inhibitor Ro 31-8959 is sufficient to inhibit acute and chronic infections. The low toxicity of this compound renders it a highly promising antiviral agent in AIDS chemotherapy

A Series of Potent HIV-1 Protease Inhibitors Containing a Hydroxyethyl Secondary Amine Transition State Isostere: Synthesis, Enzyme Inhibition, and Antiviral Activity

A novel subclass of potent hydroxyethylamine inhibitors containing a secondary amine isostere in place of the cyclic amine of Ro 31-8959 show differences in structure-activity relationships and in binding mode

Structure-based C a Symmetric Inhibitors of tIIV Protease

Design, Activity and 2.8A Struc-Lure of a C2 Symmetric Inhibitor Complexed to HlV-1

Design of C 2 symmetric inhibitors exploiting the perfect symmetry of HIV PR revealed by X-ray crystallography

X-ray Crystal Structure of the HIV Protease Complex with L-700,417, an Inhibitor with Pseudo-C 2 Symmetry

An illustration of the application of X-ray crystallography to rational drug design, which in this instance revealed unoptimized hydrogen bonding and several water-mediated PR-inhibitor interactions

Antiviral and Pharmokinetic Properties of C 2 Symmetric Inhibitors of the Human Immunodeficiency Virus Type 1 Protease

This paper illustrates the difficulties in reconciling the competing demands on PR inhibitors for tight hydrophobic interactions with PR subsites, and aqueous solubility required for bioavailability and in vivo efficacy

X-ray Crystallographic Structure of a Complex Between a Synthetic Protease of Human Immunodeficiency Virus 1 and a Substrate-based Hydroxyethyalmine Inhibitor

Structure at 2.5A Resolution of Chemically Synthesized Human Immunodeficiency Virus Type 1 Protease Complexed with a Hydroxyethylene-based Inhibitor

Structure-based Design of Nonpeptide Inhibitors Specific for the Htmmal Immunodeficiency Virus 1

Maturation of Poliovirus Capsid Proteins

A concise review of the role of three distinct proteolytic activities in the release of capsid proteins from the poliovims polyprotein and their subsequent assembly into virions

Poliovirus Protease 2A Induces Cleavage of EuLkaryotic Initiation Factor 4F Polypeptide p220

Poliovirus Proteinase 3C Converts an Active Form of Transcription Factor HIC to an Inactive form: a Mechanism for Inhibition of Host Cell Polymerase Ill Transcription by Poliovirus

The severe inhibition of RNA polymerase III-mediated transcription in polio-infected cells is a result of inactivation of transcription factor IIIC by the proteolytic activity of 3CP r°

Site-directed Mutagenesis of the Putative Catalytic Triad of Poliovirus 3C Proteinase

Site-directed mutagenesis experiments suggest that the catalytic triad of polio 3CP ro (His40, Glu71, Cys149) structurally resembles trypsinlike serine proteinases but differs significantly in its constituent residues

Analysis of Putative Active Site Residues of the Poliovirus 3C Protease

Experiments designed to evaluate two conflicting structural models of polio 3CP ro suggest that Glu71 is a constituent residue of the catalytic triad whereas Asp85 is involved in polyprotein substrate recognition

Characterization of Poliovirus 2A Proteinase by Mutational Analysis: Residues Required for Autocatalytic Activity are Essential for Induction of Cleavage of Eukaryotic Initiation Factor p220

Polio 2APrO containing a Cysl09Ser substitution within the putative His20, Asp38, Cysl09 catalytic triad retains significant autocatalytic activity

Expression and Characterization of Recombinant Hepatitis A Virus 3C Protease

A colorimetric assay was used to characterize cleavage by purified hepatitis A virus 3CP to

Putative Papainrelated Thiol Proteases of Positive-strand RNA Viruses. Identification of Rubi-and Aphthovirus Proteases and Delineation of a Novel Conserved Domain Associated with Proteases of Rubi, c~-and Coronaviruses

Computer-assisted analysis was used to identify the first viral proteirlases that are distantly related to papain-like thiol proteinases

Role for the P4 Amino Acid Residue in Substrate Utilization by the PoUovirus 3CD Proteinase

BA: A Rapid Method for Determination of Endoproteinase Substrate Specificity: Specificity of the 3C Proteinase from Hepatitis A Virus

A novel and rapid technique was used to demonstrate that hepatitis A vires 3CP ro has a strong preference for smalI residues (Ala, Ser, Gly) at the Pl' position, but has little specificity at P2'. 41. WEIDNER JR, DUNN BM: Development of Synthetic Peptide Substrates for the Poliovirus 3C Proteinase

A high performance liquid chromatography assay reveals that polio 3CP ro has a strong preference for a proline P2' residue, and a continuous fluorescence assay is reported

Determinants of Substrate Recognition by Poliovirus 2A Proteinase

positions are strict determinants of substrate recognition by polio 2APrO, but P2', PI' and P3 positions are broadly tolerant of substitution. Substrate requirements for cleavage in trans are more stringent than for cleavage in cis

Hepatitis A Virus 3C Proteinase Substrate Specificity

Hepatitis A virus 3CP ro has strong preferences for residues at P4 and Pa positions, and differs in specificity from enteroviral 3C proteinases

Structure and Stereochemistry of Thysananone: a Novel Human Rhinovirus 3C Protease Inhibitor from Thyanophora penicilloides. Telrahedron Lett

A novel non-peptide (naphthoquinone) inhibitor of rhinovirus 3C proteinase

Spiro lndoline Beta-lactams, Inhibitors of Poliovirus and Rhinovirus 3C-proteinases