Báo cáo Y học: Modulation of the oligomeric structures of HIV-1 retroviral enzymes by synthetic peptides and small molecules

Eur. J. Biochem. 269, 5103–5111 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03216.x REVIEW ARTICLE Modulation of the oligomeric structures of HIV-1 retroviral enzymes by synthetic peptides and small molecules Nicolas Sluis-Cremer1 and Gilda Tachedjian2 1Department of Medicine, Division of Infectious Diseases, University of Pittsburgh, PA, USA; 2AIDS Molecular Biology Unit, Macfarlane Burnet Institute for Medical Research and Public Health, Melbourne, Victoria, Australia The efficacy of antiretroviral agents approved for the treat-ment of HIV-1 infection is limited by the virus’s ability to develop resistance. As such there is an urgent need for new ways of thinking about anti-HIV drug development, and accordingly novel viral and cellular targets critical to HIV-1 replicationneedtobeexploredandexploited.Theretroviral RNA genome encodes for three enzymes essential for viral replication: HIV-1 protease (PR), HIV-1 reverse transcrip-tase (RT) and HIV-1 integrase (IN). The enzymatic func-tioning of each of these enzymes is entirely dependent on their oligomeric structures, suggesting that inhibition of In 1983 HIV was identified as the etiologic agent of AIDS [1,2]. During the past 18 years a tremendous effort has been placed in the identification and/or development of com-pounds that effectively attenuate HIV-1 infection. To date, 16 anti-HIV agents have been approved by the United States FDA for administration to HIV-1 infected individ-uals. These antiviral agents target the active sites of two retroviral enzymes, protease (PR) and reverse transcriptase (RT), and can be further divided into three different therapeuticclasses;PRactivesiteinhibitors,nucleoside(and nucleotide) reverse transcriptase inhibitors (NRTI) and non-nucleoside reverse transcriptase inhibitors (NNRTI). However, due to the long-lived nature of the HIV-1 infection as well as the genetic plasticity inherent to the virus, emergence of viral resistance to these antiretroviral agents is inevitable. Furthermore, as many of the com- Correspondence to N. Sluis-Cremer, Department of Medicine, Division of Infectious Diseases, University of Pittsburgh, S808 Scaife Hall, 3550 Terrace Street, Pittsburgh 15261, PA, USA., Fax: + 412 6489653, Tel.: + 412 3838525, E-mail: CremerN@msx.dept-med.pitt.edu Abbreviations: IN, integrase; NNRTI, non-nucleoside reverse transcriptase inhibitor; PR, protease; RNase H, ribonuclease H; RT, reverse transcriptase; TSAOe3T, 1-{spiro[4¢-amino-2¢,2¢-dioxo-1¢,2¢-oxathiole-5¢,3¢-[2¢,5¢-bis-O-(tert-butyldimethylsilyl)-b-D-ribofurano-syl]]}-3-ethylthymine. Enzymes: HIV-1 integrase (EC 2.7.7.49); HIV-1 protease (EC 3.4.23.16); HIV-1 reverse transcriptase (EC 2.7.7.49); HIV-1 ribonuclease H(EC 3.1.26.4, SWISS-PROT entry name: POL_HV1B1, Pol polyprotein of HIV-1 (BH10 isolate)). (Received 22 May 2002, revised 28 June 2002, accepted 29 August 2002) subunit-subunitassemblyormodulationoftheirquaternary structures provide alternative targets for HIV-1 inhibition. This review discusses the recent advances in the design and/or identification of synthetic peptides and small mole-culesthatspecificallytargetthesubunit–subunitinterfacesof theseretroviralenzymes,resultingintheinactivationoftheir enzymatic functioning. Keywords: protease; reverse transcriptase; integrase; oligo-meric structure; inhibiting protein–protein interactions. pounds from the same therapeutic class exhibit similar chemical structures and mechanisms of action, the emer-gence of viral resistance to one drug frequently results in cross-resistance to other compounds. Thus, the identifica-tion of additional viral targets and the development of new classes of antiviral compounds are essential in the fight against HIV/AIDS. In this regard, many promising com-pounds have been identified that target different steps in the HIV-1 viral life cycle including viral entry and fusion, proviral DNA integration as well as viral assembly (for reviews see [3,4]). Physical interactions between proteins play a critical role in many biological processes including signal trans-duction, cell cycle and gene regulation, and viral assembly and replication [5–7]. Furthermore, many protein–protein interactions provide therapeutically worthwhile targets. In this regard, inhibitors of protein– protein interactions have been successfully developed that target, amongst others, the interface of the large and small subunits of herpes simplex virus ribonucleotide reductase [8], cytokines (IL-2/IL-2Ra) [9], and growth hormone/receptor binding [10]. The three enzymes of HIV (PR, RT and integrase (IN)) are all oligomeric proteins (Fig. 1). The enzymatic functioning of each of these enzymes is entirely dependent on their quaternary structure [11–13]. Therefore, inhibition of retroviral enzyme protein-protein assembly, or drug-mediated modulation of retroviral enzyme oligomers, provide alternative targets for HIV-1 inhibition. The objective of this review is to describe the unique structuralfeaturesoftheHIV-1oligomericenzymesPR,RT and IN and the strategies that have been developed to inhibit enzyme function by modulation of the interfaces between thesubunits oftheenzymes.Eachviralenzymewill be dealt with individually. 5104 N. Sluis-Cremer and G. Tachedjian (Eur. J. Biochem. 269) Ó FEBS 2002 (structure-assisted) design target this region (discussed below). Fig. 1. Oligomeric structures of HIV-1 PR (1A3O.pdb), RT (1HMV.pdb)and IN (1EX4.pdb). The two subunits for each retroviral enzyme are depicted in magenta and cyan, respectively. Residues contributing to the protein–protein interface are illustrated using a surface representation. HIV-1 PROTEASE Structure and function of HIV-1 PR HIV-1 PR catalyzes the hydrolysis of specific peptide bonds withintheHIV-1GagandGag-Polpolyproteinstogenerate the various structural and functional proteins essential for viral replication. HIV-1 PR is a symmetrically arranged homodimeric protein composed of two chemically identical subunits of 99 amino acids (Fig. 1). The PR subunit fold consists of a compact structure of b strands with a short ahelix near the C-terminus [14]. The antiparallel b strands constituted by residues 44–57 from both subunits, form a flexible ÔflapÕ region that is thought to fold down over the active site during catalysis to both bind substrate and exclude water. Protein–protein interactions in the dimer include interactions between the catalytic triad residues (D25-G27), I50 and G51 at the tip of the flaps, and the antiparallel b sheet formed by the four termini in the dimer (residues 1–5 and 95–99). Additional interactions include a complex salt bridge between D29 and R87 of one subunit and R8 of the other subunit. Thermodynamic analyses of the dimeric PR molecule indicates a Gibbs energy of dimer stabilization of 10 kcal/mol at 25 °C (pH3.4), consistent with a dissociation constant of 5 · 10)8 M [15]. Interest-ingly,theGibbsfreeenergy ofdimerizationis notuniformly distributed along the protein–protein interface [15]. Instead, the interface is characterized by the presence of clusters of residues (Ôhot spotsÕ) that significantly contribute to subunit association, and other regions that contribute very little. In particular, the four-stranded b sheet formed by the amino-acid residues atthe N-and C-termini of PR contribute close to 75% of the total Gibbs energy [15]. The importance of this four-stranded b sheet is further emphasized by the fact that all PR dimerization inhibitors developed by ÔrationalÕ Peptide-based inhibitors of PR dimerization Short synthetic peptides corresponding to the amino-acid sequences of the N- and C-termini of HIV-1 PR have been shown to inhibit proteolytic activity by binding to the inactive PR subunits and preventing their association into active dimer [16–19]. Peptides corresponding to the C-terminal segment of the HIV-1 matrix protein have also been found to elicit the same effect [20]. However, the concentration of these peptides (both PR- and matrix-derived) required to effectively inhibit the PR monomer-dimer equilibrium by 50% (IC50) is relatively high (30–100 mM, Table 1). Schramm et al. demonstrated that it was possible to significantly improve their inhibitory properties ( 50–200 fold) through modification of their amino-acid composition and the addition of a hydrophobic moiety, such aminocaproyl or palmitoyl, to the N-terminus of the peptide [20]. The development of these Ômore potentÕ peptide inhibitors firmly established that PR dimerization was a rational target for the development of AIDS therapeutics,andthatsmall-sizepeptidemimeticsexhibiting goodbio-availabilitycouldbederivedforHIV-1therapy.In this regard, N-terminally palmitoyl-blocked peptides con-sisting of only three residues, one of which is a non-natural aminoacid,havebeendevelopedandshowntoexhibitgood potency against PR dimerization [21]. Cross-linking of the N- and C-terminal peptides to form a mimic of the HIV)1 PR dimerization interface has provided an alternative strategy for the development of more potent PR dimerization inhibitors. The principle of this strategy is illustrated in Fig. 2. This approach was initially adopted by Babe et al., who cross-linked the N- and C-terminal PR-derived peptides by a 3.5-A tether composed of three glycine residues, however, the resulting compounds were not potent inhibitors of PR dimerization [18]. In the crystal structures of HIV-1 PR, the polypep-tide termini are held at a distance of approximately 10 A (see Fig. 2). Accordingly, more potent cross-linked inter-facial peptide compounds have been developed using tethers that bridge this gap [22–24]. For example, Zutshi et al. used flexible alkyl-tethers to link the peptide strands [22], while Bouras et al. took advantage of a pyridinediol-or naphthalenediol-based scaffold [23]. The supposed advantage of the aromatic Ôconformationally constrainedÕ scaffold is that it may allow the two peptide strands to be initially more suitably oriented to permit formation of the antiparallel b sheet with one PR monomer [23]. However, very little difference in relative potency is observed between the different tethers (Table 1). Irreversible inhibi-tion of PR dimerization has also been achieved by designing a cross-linked interfacial peptide molecule that can form a disulfide bond with C95 in HIV-1 PR [25]. Other novel strategies involve tethering an active-site peptide inhibitor with the dimerization inhibiting C-terminus peptide, thereby generating a compound that exhibits synergistic inhibition of PR activity [26]. Thepeptideinhibitorsdescribedabovewerealldeveloped using peptide sequences corresponding to the N- and C-termini of PR, which themselves had been initially tested followingtheobservationoftheiressentialroleinlinkingthe Ó FEBS 2002 Modulation of the quaternary structure of HIV-1 enzymes (Eur. J. Biochem. 269) 5105 Table 1. Peptide and small molecule inhibitors of HIV-1 PR dimerization. Peptides Derived from the N- and C- Termini of HIV-1 PR and MA Ac-Thr-Leu-Asn-Phe-COOH45 N-Pro-Gln-Ile-Thr-Leu-Trp-OH>100 Ac-Gln-Val-Ser-Gln-Asn-Tyr-COOH100 Modified Peptides Derived from the C-terminus of HIV-1 PR Thr-Val-Ser-Tyr-Glu-Leu-OH12 Palmitoyl-Thr-Val-Ser-Tyr-Glu-Leu-OH0.5 Palmitoyl-Tyr-Glu-Leu-OH0.15 Palmitoyl-Tyr-Glu-(L-threonine)-OH0.05 IC50 (lM) Kinetic analysis Kinetic analysis Kinetic analysis Kinetic analysis Kinetic analysis Kinetic analysis Kinetic analysis Method of analysis of PR dimerization Ref a [16] [19] [20] [20] [20] [21] [21] Palmitoyl-Tyr-Glu-(p-biphenyl-alanine)-OH0.025 Kinetic analysis [21] Cross-linked Interfacial Peptides >50 Protein cross-linking [18] 2.0 Gel-filtration; protein cross-linking [22] PR fluorescence 4.2 Kinetic analysis [23] Other Peptides Ile-Val-Gln-Val-Asp-Ala-Glu-Gly-Gly Above peptide cross-linked using 1,6-hexane-bis-vinylsulfone Non-Peptide Based Inhibitors Ursolic acid Didemnaketal A penta-ester derivative 32 Genetic selection [27] Kinetic analysis; gel-filtration 0.78 Kinetic analysis [27] 3.4 Kinetic analysis [28] 2.1 Kinetic analysis [29] a Kinetic analyses were carried out according to the method described by Zhang et al. (1991) [16]. twoPRsubunitsthroughtheformationofthefour-stranded b sheet [14]. Recently, an elegant strategy for the genetic selection of dissociative peptide-based inhibitors of HIV-1 PR (and virtually all other designated protein–protein interactions) has been reported [27]. Briefly, this strategy takes advantage of k-bacteriophage repressor protein (cI) that binds to its operator (kPR) as a homodimer. The C-terminal dimerization domain of cI can be replaced by another protein that homodimerizes, in this instance an inactive variant of PR was used. When bacteria are transformed with a reporter plasmid (that contains the selection module kPR-lacZ-tet and directs the production of the cI-PR fusion protein), cI-PR represses the transcription of the reporter genes and the transformants show a LacZ(negative)-Tet(sensitive) phenotype [27]. Co-transfor-mation of the reporter plasmid with a peptide plasmid library allows for the selection of peptides that prevent NcI-PRdimerizationandgeneratetransformantsexhibiting a LacZ(positive)Tet(resistant) phenotype [27]. The power and utility of this technique was demonstrated by the selection of approximately 300 peptides from 3 · 108 cotransformants that exhibited a Ôpositive phenotypeÕ, rep-resenting a selection frequency of 1 in 106. Further analyses oftheselectedpeptidesidentifiedthepeptideIVQVDAEGG asaninhibitorofPRdimerization,whichwhentetheredina head-to-head or a tail-to-tail fashion generated a relatively potent inhibitor of PR dimerization (Table 1). Non-peptide based inhibitors of PR dimerization To date, two structurally unrelated classes of small-organic molecules have been identified which inhibit PR dimeriza-tion [28,29]. The first class of molecules, which exhibit a polycyclic triterpene structure, were identified following a search of the Cambridge Structural Database (www.ccdc.cam.ac.uk) for pharmacophores that could bridge the 10 A gap between the termini of a PR subunit [28]. Extensive kinetic analysis of one of these triterpenes, ursolic acid, demonstrated that these compounds inhibited PRdimerizationwithrelativelyhighpotency(Ki ¼ 3.4 lM). 5106 N. Sluis-Cremer and G. Tachedjian (Eur. J. Biochem. 269) Ó FEBS 2002 Fig. 2. Schematic representation of the strategy used to inhibit PR dimerization by cross-linked interfacial peptides. The N-terminus of PR is indicated (N). The four-stranded b sheet formed by amino-acid residues at the N- and C-termini of PR is a major binding determinant in the formation of dimeric PR. Cross-linked interfacial peptides containing a tether region of approximately 10 A inhibit PR dimeri-zation by permitting the formation of a pseudo antiparallel b sheet with one of the PR subunits. The second class of molecules that inhibited PR dimeriza-tion include pentaester derivatives of of didemnaketal A [29]. The identification of these classes of small molecules is significant, as in general many empirical searches for low molecular mass pharmacological inhibitors (<400) of protein–protein interactions have routinely failed. HIV-1 REVERSE TRANSCRIPTASE Structure and function of HIV-1 RT HIV-1 RT is required for conversion of the viral genomic RNA into a double-stranded proviral DNA precursor. This process is catalyzed by the RNA- and DNA-dependent polymerase and ribonuclease H(RNase H) activities of the enzyme in a reverse transcription complex in the cell cytoplasm [30]. HIV-1 RT is an asymmetric heterodimer composed of a 560-residue 66 kDa subunit (p66) compri-sing two domains termed DNA polymerase and RNase H, and a p66-derived 440-residue 51 kDa subunit (p51). The p51 subunit is produced during viral assembly and matur-ation via HIV-1 protease-mediated cleavage of the C-terminal (RNase H) domain of a p66 subunit [31]. A fascinating feature of the HIV-1 RT heterodimer is the structural asymmetry which exists between the p66 and p51 subunits despite the fact that they are products of the same gene and exhibit identical amino-acid sequences for the first 440 residues [32–40]. The overall shape of the p66 subunit has been likened to that of a Ôright-handÕ [35]. The major subdomains of the polymerase domain of p66 are termed fingers (residues 1–85, 118–155), palm (86–117, 156–237) and thumb (238– 318). The DNA polymerase catalytic aspartate residues (D110, D185, and D186) reside in the palm subdomain. A fourth subdomain, termed the ÔconnectionÕ subdomain (residues 319–426), acts as a tether between the DNA polymerase and C-terminal RNase H(427–565) domains. Thep51subunitcontainsthesamefingers,thumb,palmand connection subdomains, however, their spatial arrangement differs markedly to those of the p66 subunit [35]. UponformationoftheRTheterodimerfromthep66and p51 monomers, large surface areas of the individual subunits become inaccessible to water [33,41]. Approxi-mately 4800 A2 of protein surface is buried in the RT dimer complex of which 3050A2 corresponds to nonpolar atoms. Amino-acid residues in the p66 subunit that form part of the dimer interface are derived primarily from the palm, connection and RNase Hdomain, while in the p51 they arise from the fingers, thumb and connection domains. Dissection of the contributions of each individual residue to thetotalburied surface areaupon dimerization revealseight stretches of residues that make the largest contribution to total binding strength. These include residues D86-L92, Q373-G384, W406-W410 and P537-G546 in p66 subunit, and P52-P55, I135-P140, C280-T290 and P392-W401 in the p51 subunit [41,42]. Three clearly visible clusters are formed between these interfacial residues [33]. A single region in the palmdomainofp66(D86-L92)interactswithtworegionsof the fingers of p51 (P52-P55 and S135-P140). The RNase H residues 537–546 in the p66 subunit interact with the p51 thumb residues 280–290, and the p66 connection residues W406-W410 interact with residues in the p51 connection domain residues (P392-W401). Evident from these clusters, the two subunits are completely asymmetric with respect to one another in that the subunit interface on p51 involves different amino acids than the p66 [33]. Contacts between the connection subdomains form the only interactions between equivalent subdomains from each subunit. How-ever, even in this case, many equivalent residues make different protein–protein interactions in such a way that the contacts between the two connection subdomains are also intrinsicallyasymmetric.Thermodynamicevaluationsofthe association between the p66 and p51 subunits of RT have estimated a Gibbs free energy of dimer stabilization of approximately 10–12 kcalÆmol)1, corresponding to a disso-ciation constant of approximately 10 nM [43,44]. Peptide-based inhibitors of RT dimerization As described above, one of the three clusters of residues formed between the RT p66 and p51 dimer interface are formed through the interactions between the RT p66 connection residues W406-W410 and residues P392-W401 inthep51connectiondomain.Interestingly,a19amino-acid synthetic peptide corresponding to residues 389–407 of the connectiondomain(N-FKLPIQKETWETWWTEYWQ-C) ofRTwasdemonstratedtoberelativelyefficientinretarding the heterodimerization process of HIV-1 RT [45]. Further studies indicated that the same peptide was as efficient at inhibiting the heterodimerization process of HIV-2 RT as HIV-1 RT [46]. This result is not surprising given that this regionintheconnectiondomainisconservedinHIV-1,HIV-2 and the simian immunodeficiency virus RTs [46] (the corresponding amino-acid sequence in HIV-2 RT is N-FHLPVERDTWEQWWDNYWQ-C). More recently, thelengthofthepeptidewasoptimizedtogenerateashorter (10 residue) peptide (corresponding to residues 395–404 of RT)thatwassynthesizedwithanacetylatedN-terminusand a cysteamide group at the C-terminus to improve stability and cellular uptake [47]. The resulting peptide was a more Ó FEBS 2002 Modulation of the quaternary structure of HIV-1 enzymes (Eur. J. Biochem. 269) 5107 efficient inhibitor of RT dimerization in vitro and was also shown to inhibit HIV-1 replication in cell culture [47]. The antiviral activity of the peptide was further enhanced by conjugationtoapeptidylcarrierwithoutadversetoxiceffects to cells [47]. Remarkably, the concentration of peptide-carrier complex required to inhibit HIV-1 replication was significantlyless than the peptide concentration required for the inhibition of RT dimerization in vitro. For example, 0.1 nM of peptide-carrier completely suppressed HIV-1 replication for 15 days, whereas a peptide concentration of 240 mM was required to inhibit RT heterodimerization by 50% in vitro [47]. This may suggest that the mechanism of inhibitionofsubunitassociationinvitroisdifferentfromthe processin HIV-1 infectedcells. InHIV-infected cells the RT polypeptides are translated as part of the Gag–Pol polypro-tein which is subsequently cleaved by HIV-1 PR to release thevariousstructuralandfunctionalproteins.Recentstudies have shown that HIV-1 PR cleaves the Pol region of Gag– Polina sequentialmannerinwhichtheRTp66polypeptide isinitiallyreleasedfromthepolyproteinprecursor.Cleavage of the p66 subunit to generate RT p51 appears to require a p66/p66 homodimeric intermediate (D. Arion, N. Sluis-Cremer&M.A.Parniak,unpublishedresults).Thedissocia-tion constant for p66 homodimerization is 10)6 M, a value approximately 1000-fold weaker than the interaction be-tween RT p66 and p51 [43,44]. Thus, one could anticipate thatthe10-residuepeptideshouldbeamorepotentinhibitor of p66 homodimerization. Furthermore, it is interesting to considerthatmodulationofRTdimerizationmayalsoaffect the interaction between two Gag–Pol molecules that must dimerize to activate HIV-1 PR [48]. Any affects on this interaction may adversely affect PR activity [49]. Hence, the in vitro study of the effect of peptides on p66 and p51 dimerization does not necessarily reflect the process that is occurring in HIV-infected cells and may not accurately predicttheirimpactonHIV-1replicationincellbasedassays. Synthetic peptides that inhibit conformational changes during HIV-1 RT heterodimerization In vitro formation of active heterodimeric p66/p51 HIV-1 RT from the p66 and p51 monomeric subunits occurs in a two step process involving an initial bimolecular association followed by a slow conformational change [50]. The conformational change (or maturation step) appears to be essential for the complete enzymatic activation of RT [50]. As discussed previously, the RNase Hresidues 537–546 in the p66 subunit interact with the p51 thumb residues 280– 290. A synthetic peptide derived from a sequence within the thumb subdomain of HIV-1 RT (residues 284–300) was found to bind to heterodimeric HIV-1 RT with an apparent dissociation constant in the nanomolar range and interfere with the conformational change (or maturation step) required for activation of heterodimeric RT [51]. Based on this work it was suggested that the activation of RT might also represent an important target for the design of novel antiviral compounds. Destabilization of the HIV-1 RT dimer interface by small nonpeptidic molecules The complete dissociation of the p66 and p51 subunits of HIV-1 RT heterodimer may not be entirely necessary for there to be a negative impact on RT enzymatic function. Indeed, recent studies have shown that small molecule binding to the dimer interface of HIV-1 RT may induce conformational changes that impact on the overall stability of the heterodimeric complex without dissociating the heterodimer complex [44,52]. Two structurally unrelated classes of compounds have been found to elicit this effect. 2¢,5¢-Bis-O-(tert-butyldimethylsilyl)-b-D-ribofuranosyl]-3¢spiro-5¢¢-(4¢¢-amino-1¢,2¢-oxathiole-2¢,2¢-dioxide)thymine (TSAO-T) is the prototype of an unusual class of non-nucleoside reverse transcriptase inhibitors (NNRTI) which have structures (Fig. 3) and mechanism of actions quite distinct from conventional NNRTI [53,54]. The N3-ethyl derivative of TSAO-T, TSAO-e3T has been shown to destabilize both the p66/p51 and p66/p66 dimeric forms of HIV-1 RT [44]. The Gibbs free energy of RT dimer dissociation is decreased in the presence of increasing concentrations of TSAOe3T, resulting in loss of dimer stability of 4.0 and 3.2 kcalÆmol)1 for p66/p51 and p66/p66 forms of HIV-1 RT, respectively [44]. This loss of energy is not sufficient to induce subunit dissociation in the absence of denaturant [44]. High-level drug resistance to TSAO is mediated by the E138K mutation in the p51 subunit of HIV-1 RT [55]. The introduction of this mutation into RT significantly diminishes the ability of TSAO to bind to and inhibittheenzyme[55]andaccordinglyTSAO-e3Tisunable to destabilize the subunit interactions of the E138K mutant enzyme [44]. Modeling experiments have suggested that TSAOmaybindtoasiteinRTthatisoverlappingwith,but Fig. 3. Chemical structures of NNRTI that modulate the RT dimeri-zation process. The NNRTI nevirapine, efavirenz, and UC781 act as chemical enhancers of HIV-1 RT dimerization [57]. Unlike other NNRTI, delavirdine has no effect on RT dimerization [57]. TSAOe3T and BBNHbinding to HIV-1 RT destabilizes the quaternary structure of the enzyme [43,51]. ... - tailieumienphi.vn