Predicting cleavability of peptide sequences by HIV protease via correlation-angle approach

In designing HIV protease inhibitors as potential drugs for AIDS therapy, knowledge about what peptide sequences in polyproteins are cleavable by HIV proteases is very useful. In this article, based on the formulation that any octapeptide can be uniquely expressed as a 160-dimensional vector and the principle that the similarity of any two such vectors is associated with their correlation angle, a new method is proposed to predict the cleavability of a peptide sequence by HIV-1 and HIV-2 proteases. The average predicted accuracy the new method for the 105 peptide sequences whose cleavability by HIV-1 protease is known is 96/105=9.14%, which is about 8% higher than that by the existing method for the same set of data. A considerably high rate of correct prediction was also obtained when the new method was used to predict the HIV-2 protease-cleaved sites in some proteins.     

A vector projection approach to predicting HIV protease cleavage sites in proteins

A vector projection method is proposed to predict the cleavability of oligopeptides by extended-specificity site proteases. For an enzyme with eight specificity subsites the substrate octapeptide can be uniquely expressed as a vector in an 8-dimensional space, whose eight bases correspond to the amino acids at the eight subsites, P₁, P_1′, P_2′, P_3′ and P_4′, respectively. The component of such a characteristic vector on each of the eight bases is defined as the frequency of an amino acid occurring at a given site. These frequencies were derived from a set of octapeptides known to be cleaved by HIV protease. The cleavability of an octapeptide can then be estimated from the projection of its characteristic vector on an idealized, optimally cleavable vector. The high ratio of correct prediction vs. total prediction for the data in both the training and the testing sets indicates that the new method is self-consistent and efficient. It provides a rapid and accurate algorithm for analyzing the specificity of any multisubsite enzyme for which there is no coupling between subsites. In particular, it is useful for predicting the cleavability of an oligopeptide by either HIV-1 or HIV-2 protease, and hence offers a supplementary means for finding effective inhibitors of HIV protease as potential drugs against AIDS. © Wiley-Liss, Inc.     

Variable context Markov chains for HIV protease cleavage site prediction

Deciphering the knowledge of HIV protease specificity and developing computational tools for detecting its cleavage sites in protein polypeptide chain are very desirable for designing efficient and specific chemical inhibitors to prevent acquired immunodeficiency syndrome. In this study, we developed a generative model based on a generalization of variable order Markov chains (VOMC) for peptide sequences and adapted the model for prediction of their cleavability by certain proteases. The new method, called variable context Markov chains (VCMC), attempts to identify the context equivalence based on the evolutionary similarities between individual amino acids. It was applied for HIV-1 protease cleavage site prediction problem and shown to outperform existing methods in terms of prediction accuracy on a common dataset. In general, the method is a promising tool for prediction of cleavage sites of all proteases and encouraged to be used for any kind of peptide classification problem as well.     

Narrow substrate specificity and sensitivity toward ligand-binding site mutations of human T-cell Leukemia virus type 1 protease

Human T-cell leukemia virus type 1 (HTLV-1) is associated with a number of human diseases; therefore, its protease is a potential target for chemotherapy. To compare the specificity of HTLV-1 protease with that of human immunodeficiency virus type 1 (HIV-1) protease, oligopeptides representing naturally occurring cleavage sites in various retroviruses were tested. The number of hydrolyzed peptides as well as the specificity constants suggested a substantially broader specificity of the HIV protease. Amino acid residues of HTLV-1 protease substrate-binding sites were replaced by equivalent ones of HIV-1 protease. Most of the single and multiple mutants had altered specificity and a dramatically reduced folding and catalytic capability, suggesting that mutations are not well tolerated in HTLV-1 protease. The catalytically most efficient mutant was that with the flap residues of HIV-1 protease. The inhibition profile of the mutants was also determined for five inhibitors used in clinical practice and inhibitor analogs of HTLV-1 cleavage sites. Except for indinavir, the HIV-1 protease inhibitors did not inhibit wild type and most of the mutant HTLV-1 proteases. The wild type HTLV-1 protease was inhibited by the reduced peptide bond-containing substrate analogs, whereas the mutants showed various degrees of weakened binding capability. Most interesting, the enzyme with HIV-1-like residues in the flap region was the most sensitive to the HIV-1 protease inhibitors and least sensitive to the HTLV-1 protease inhibitors, indicating that the flap plays an important role in defining the specificity differences of retroviral proteases.     

Studies on the specificity of HIV protease: An application of Markov chain theory

A sequence-coupled (Markov chain) model is proposed to predict the cleavage sites in proteins by proteases with extended specificity subsites. In addition to the probability of an amino acid occurring at each of these subsites as observed from a training set of oligopeptides known cleavable by HIV protease, the conditional probabilities as reflected by the neighbor-coupled effect along the subsite sequence are also taken into account. These conditional probabilities are derived from an expanded training set consisting of sufficiently large peptide sequences generated by the Monte Carlo sampling process. Very high accuracy was obtained in predicting protein cleavage sites by both HIV-1 and HIV-2 proteases. The new method provides a rapid and accurate means for analyzing the specificity of HIV protease, and hence can be used to help find effective inhibitors of HIV protease as potential drugs against AIDS. The principle of this method can also be used to study the specificity of any multisubsite enzyme.     

Artificial neural network method for predicting HIV protease cleavage sites in protein

Knowledge of the polyprotein cleavage sites by HIV protease will refine our understanding of its specificity, and the information thus acquired will be useful for designing specific and efficient HIV protease inhibitors. The search for inhibitors of HIV protease will be greatly expedited if one can find and accurate, robust, and rapid method for predicting the cleavage sites in proteins by HIV protease. In this paper, Kohonen’s self-organization model, which uses typical artificial neural networks, is applied to predict the cleavability of oligopeptides by proteases with multiple and extended specificity subsites. We selected HIV-1 protease as the subject of study. We chose 299 oligopeptides for the training set, and another 63 oligopeptides for the test set. Because of its high rate of correct prediction (58/63=92.06%) and stronger fault-tolerant ability, the neural network method should be a useful technique for finding effective inhibitors of HIV protease, which is one of the targets in designing potential drugs against AIDS. The principle of the artificial neural network method can also be applied to analyzing the specificity of any multisubsite enzyme.     

Artificial neural network method for predicting HIV protease cleavage sites in protein

Knowledge of the polyprotein cleavage sites by HIV protease will refine our understanding of its specificity, and the information thus acquired will be useful for designing specific and efficient HIV protease inhibitors. The search for inhibitors of HIV protease will be greatly expedited if one can find and accurate, robust, and rapid method for predicting the cleavage sites in proteins by HIV protease. In this paper, Kohonen’s self-organization model, which uses typical artificial neural networks, is applied to predict the cleavability of oligopeptides by proteases with multiple and extended specificity subsites. We selected HIV-1 protease as the subject of study. We chose 299 oligopeptides for the training set, and another 63 oligopeptides for the test set. Because of its high rate of correct prediction (58/63=92.06%) and stronger fault-tolerant ability, the neural network method should be a useful technique for finding effective inhibitors of HIV protease, which is one of the targets in designing potential drugs against AIDS. The principle of the artificial neural network method can also be applied to analyzing the specificity of any multisubsite enzyme.     

Structural and kinetic analysis of pyrrolidine-based inhibitors of the drug-resistant Ile84Val mutant of HIV-1 protease

Human immunodeficiency virus (HIV) protease is a well-established drug target in HIV chemotherapy. However, continuously increasing resistance towards approved drugs inevitably requires the development of new inhibitors preferably showing no susceptibility against resistant HIV protease strains. Recently, symmetric pyrrolidine-3,4-bis-N-benzyl-sulfonamides have been developed as a new class of HIV-1 protease inhibitors. The most promising candidate exhibited a K_i of 74 nM towards a wild-type protease. Herein, we report the influence of the active-site mutations Ile50Val and Ile84Val on these inhibitors by structural and kinetic analysis. Although the Ile50Val mutation leads to a significant decrease in affinity for all compounds in this series, they retain or even show increased affinity towards the important Ile84Val mutation. By detailed analysis of the crystal structures of two representatives in complex with wild-type and mutant proteases, we were able to elucidate the structural basis of this phenomenon.     

Structural basis for coevolution of a human immunodeficiency virus type 1 nucleocapsid-p1 cleavage site with a V82A drug-resistant mutation in viral protease

Maturation of human immunodeficiency virus (HIV) depends on the processing of Gag and Pol polyproteins by the viral protease, making this enzyme a prime target for anti-HIV therapy. Among the protease substrates, the nucleocapsid-p1 (NC-p1) sequence is the least homologous, and its cleavage is the rate-determining step in viral maturation. In the other substrates of HIV-1 protease, P1 is usually either a hydrophobic or an aromatic residue, and P2 is usually a branched residue. NC-p1, however, contains Asn at P1 and Ala at P2. In response to the V82A drug-resistant protease mutation, the P2 alanine of NC-p1 mutates to valine (AP2V). To provide a structural rationale for HIV-1 protease binding to the NC-p1 cleavage site, we solved the crystal structures of inactive (D25N) WT and V82A HIV-1 proteases in complex with their respective WT and AP2V mutant NC-p1 substrates. Overall, the WT NC-p1 peptide binds HIV-1 protease less optimally than the AP2V mutant, as indicated by the presence of fewer hydrogen bonds and fewer van der Waals contacts. AlaP2 does not fill the P2 pocket completely; PheP1' makes van der Waals interactions with Val82 that are lost with the V82A protease mutation. This loss is compensated by the AP2V mutation, which reorients the peptide to a conformation more similar to that observed in other substrate-protease complexes. Thus, the mutant substrate not only binds the mutant protease more optimally but also reveals the interdependency between the P1' and P2 substrate sites. This structural interdependency results from coevolution of the substrate with the viral protease.     

Mechanism of aspartyl proteinase action. VII. Noncovalent complexes of HIV-1 aspartyl proteinase with substrate and substrate-like inhibitors

A computer model of a noncovalent complex of HIV-1 aspartyl protease with substrate-like inhibitor JG-365 was a priori constructed by using the approaches of theoretical conformational analysis and molecular mechanics. The root mean square deviation of the calculated conformation of the inhibitor from the X-ray diffraction analysis data was 0.87 A. These results enabled the a priori calculation of the structure of noncovalent complex of HIV-1 protease with a hexapeptide fragment of its native specific substrate Ser-Gln-Asn-Tyr-Pro-Ile-Val. The only possible orientation of the cleavable peptide bond in this and the nucleophilic water molecule relative to the catalytically active Asp residues of the enzyme (Asp25 and Asp125) was found that provides for the chemical transformation of the substrate to a tetrahedral intermediate. An action mechanism of enzymes of this class was proposed on the basis of the analysis of calculated distances. We showed that neither steric distortion of the cleavable bond nor the formation of unfavorable contacts in molecules of the enzymes and their substrates accompany the optimum orientation of substrate molecules at the active sites of HIV-1 aspartyl proteases and rhizopuspepsin.     

Synthetic HIV-2 protease cleaves the GAG precursor of HIV-1 with the same specificity as HIV-1 protease

A 99-amino acid protein having the deduced sequence of the protease from human immunodeficiency virus type 2 (HIV-2) was synthesized by the solid phase method and tested for specificity. The folded peptide catalyzes specific processing of a recombinant 43-kDa GAG precursor protein (F-16) of HIV-1. Although the protease of HIV-2 shares only 48% amino acid identity with that of HIV-1, the HIV-2 enzyme exhibits the same specificity toward the HIV-1 GAG precursor. Fragments of 34, 32, 24, 10, and 9 kDa were generated from F-16 GAG incubated with the protease. N-terminal amino acid sequence analysis of proteolytic fragments indicate that cleavage sites recognized by HIV-2 protease are identical to those of HIV-1 protease. The verified cleavage sites in F-16 GAG appear to be processed independently, as indicated by the formation of the intermediate fragments P32 and P34 in nearly equal ratios. The site nearest the amino terminus is quite conserved between the two viral GAG proteins (...VSQNY-PIVQN...in HIV-1,...KGGNY-PVQHV...in HIV-2). In contrast, the putative second site (...IPFAA-AQQKG...) of HIV-2 GAG shares minimal sequence identity with site 2 of HIV-1 GAG (...SATIM-MQRGN...). These sequence variations in the substrates suggest higher order structural features that may influence recognition by the proteases. Pepstatin A inhibits HIV-2 protease, whereas 1,10-phenanthroline and phenylmethylsulfonylfluoride do not; these results are in agreement with the finding that proteases of HIV and other retroviruses are aspartyl proteases.     

Substrate-Driven Mapping of the Degradome by Comparison of Sequence Logos

Sequence logos are frequently used to illustrate substrate preferences and specificity of proteases. Here, we employed the compiled substrates of the MEROPS database to introduce a novel metric for comparison of protease substrate preferences. The constructed similarity matrix of 62 proteases can be used to intuitively visualize similarities in protease substrate readout via principal component analysis and construction of protease specificity trees. Since our new metric is solely based on substrate data, we can engraft the protease tree including proteolytic enzymes of different evolutionary origin. Thereby, our analyses confirm pronounced overlaps in substrate recognition not only between proteases closely related on sequence basis but also between proteolytic enzymes of different evolutionary origin and catalytic type. To illustrate the applicability of our approach we analyze the distribution of targets of small molecules from the ChEMBL database in our substrate-based protease specificity trees. We observe a striking clustering of annotated targets in tree branches even though these grouped targets do not necessarily share similarity on protein sequence level. This highlights the value and applicability of knowledge acquired from peptide substrates in drug design of small molecules, e.g., for the prediction of off-target effects or drug repurposing. Consequently, our similarity metric allows to map the degradome and its associated drug target network via comparison of known substrate peptides. The substrate-driven view of protein-protein interfaces is not limited to the field of proteases but can be applied to any target class where a sufficient amount of known substrate data is available.     

A Computational Module Assembled from Different Protease Family Motifs Identifies PI PLC from Bacillus cereus as a Putative Prolyl Peptidase with a Serine Protease Scaffold

Proteolytic enzymes have evolved several mechanisms to cleave peptide bonds. These distinct types have been systematically categorized in the MEROPS database. While a BLAST search on these proteases identifies homologous proteins, sequence alignment methods often fail to identify relationships arising from convergent evolution, exon shuffling, and modular reuse of catalytic units. We have previously established a computational method to detect functions in proteins based on the spatial and electrostatic properties of the catalytic residues (CLASP). CLASP identified a promiscuous serine protease scaffold in alkaline phosphatases (AP) and a scaffold recognizing a β-lactam (imipenem) in a cold-active Vibrio AP. Subsequently, we defined a methodology to quantify promiscuous activities in a wide range of proteins. Here, we assemble a module which encapsulates the multifarious motifs used by protease families listed in the MEROPS database. Since APs and proteases are an integral component of outer membrane vesicles (OMV), we sought to query other OMV proteins, like phospholipase C (PLC), using this search module. Our analysis indicated that phosphoinositide-specific PLC from Bacillus cereus is a serine protease. This was validated by protease assays, mass spectrometry and by inhibition of the native phospholipase activity of PI-PLC by the well-known serine protease inhibitor AEBSF (IC50 = 0.018 mM). Edman degradation analysis linked the specificity of the protease activity to a proline in the amino terminal, suggesting that the PI-PLC is a prolyl peptidase. Thus, we propose a computational method of extending protein families based on the spatial and electrostatic congruence of active site residues.     

Crystal structure of a cross-reaction complex between an anti-HIV-1 protease antibody and an HIV-2 protease peptide

The monoclonal antibody 1696, elicited by HIV-1 protease, inhibits the activity of both HIV-1 and HIV-2 proteases with inhibition constants in the low nanomolar range. The antibody cross-reacts with peptides derived from the N-terminal region of both proteases. The crystal structure of the recombinant single-chain Fv fragment of 1696 complexed with an N-terminal peptide from the HIV-2 protease has been determined at 1.88A resolution. Interactions of the peptide with scFv1696 are compared with the previously reported structure of scFv1696 in complex with the corresponding peptide from HIV-1 protease. The origin of cross-reactivity of mAb1696 with HIV proteases is discussed.     

Accessory mutations balance the marginal stability of the HIV-1 protease in drug resistance

The HIV-1 protease is a major target of inhibitor drugs in AIDS therapies. The therapies are impaired by mutations of the HIV-1 protease that can lead to resistance to protease inhibitors. These mutations are classified into major mutations, which usually occur first and clearly reduce the susceptibility to protease inhibitors, and minor, accessory mutations that occur later and individually do not substantially affect the susceptibility to inhibitors. Major mutations are predominantly located in the active site of the HIV-1 protease and can directly interfere with inhibitor binding. Minor mutations, in contrast, are typically located distal to the active site. A central question is how these distal mutations contribute to resistance development. In this article, we present a systematic computational investigation of stability changes caused by major and minor mutations of the HIV-1 protease. As most small single-domain proteins, the HIV-1 protease is only marginally stable. Mutations that destabilize the folded, active state of the protease therefore can shift the conformational equilibrium towards the unfolded, inactive state. We find that the most frequent major mutations destabilize the HIV-1 protease, whereas roughly half of the frequent minor mutations are stabilizing. An analysis of protease sequences from patients in treatment indicates that the stabilizing minor mutations are frequently correlated with destabilizing major mutations, and that highly resistant HIV-1 proteases exhibit significant fractions of stabilizing mutations. Our results thus indicate a central role of minor mutations in balancing the marginal stability of the protease against the destabilization induced by the most frequent major mutations.     

Overexpression,Purification and Functional Characterisation of Wild-Type HIV-1 Subtype C Protease and Two Variants Using a Thioredoxin and His-Tag Protein Fusion System

In recent years, various strategies have been used to overexpress and purify HIV-1 protease because it is an essential drug target in anti-retroviral therapy. Obtaining sufficient quantities of the enzyme, however, remains challenging. Overexpression of large quantities is prevented due to the enzyme’s autolytic nature and its inherent cytotoxicity in Escherichia coli cells. Here, we describe a novel HIV-1 protease purification method using a thioredoxin–hexahistidine fusion system for the wild-type and two variant proteases. The fusion proteases were overexpressed in E. coli and recovered by immobilised metal ion affinity chromatography. The proteases were cleaved from the fusion constructs using thrombin. When compared to the standard overexpression and purification protocol in use in our laboratory, the expression of the fusion-derived wild-type protease was increased from 0.83 to 2.5 mg/l of culture medium. The expression levels of the two variant proteases ranged from 1.5 to 2 mg/l of culture medium. The fusion wild-type and variant proteases were inactive before the cleavage of the thioredoxin–hexahistidine fusion tag as no enzymatic activity was observed. The proteases were, however, active after cleavage of the tag. The novel thioredoxin–hexahistidine fusion system, therefore, enables the successful overexpression and purification of catalytically active HIV-1 proteases.     

Unraveling sub-site specificities of peptidic serine protease inhibitors by substitutional and structural analysis

The interaction of peptidic inhibitors with serine proteases is investigated using peptide spot synthesis on cellulose sheets. This method permits a highly parallel analysis of subsite specificities and an optimization of peptidic inhibitors towards higher affinity and specificity for a given target protease. Information from crystal structures of corresponding protease/inhibitor complexes may help to explain the observed binding pattern.     

Insight into the structural similarity between HIV protease and secreted aspartic protease-2 and binding mode analysis of HIV-Candida albicans inhibitors

The analysis of the structural similarity between Candida albicans Sap2 and HIV-1 aspartic proteases by molecular modeling gave insight into the common requirements for inhibition of both targets. Structure superimposition of Sap2 and HIV-1 protease confirmed the similarity between their active sites and flap regions. HIV-1 protease inhibitors herein investigated can fit the active site of Sap2, adopting very similar ligand-backbone conformations. In particular, key anchoring sites consisting of Gly85 in Sap2 and Ile50 in HIV-1 protease, both belonging to their corresponding flap regions, were found as elements of a similar binding-mode interaction. The knowledge of the molecular basis for binding to both Sap2 and HIV-1 proteases may ultimately lead to the development of single inhibitor acting on both targets.