PROSPER: An Integrated Feature-Based Tool for Predicting Protease Substrate Cleavage Sites

The ability to catalytically cleave protein substrates after synthesis is fundamental for all forms of life. Accordingly, site-specific proteolysis is one of the most important post-translational modifications. The key to understanding the physiological role of a protease is to identify its natural substrate(s). Knowledge of the substrate specificity of a protease can dramatically improve our ability to predict its target protein substrates, but this information must be utilized in an effective manner in order to efficiently identify protein substrates by in silico approaches. To address this problem, we present PROSPER, an integrated feature-based server for in silico identification of protease substrates and their cleavage sites for twenty-four different proteases. PROSPER utilizes established specificity information for these proteases (derived from the MEROPS database) with a machine learning approach to predict protease cleavage sites by using different, but complementary sequence and structure characteristics. Features used by PROSPER include local amino acid sequence profile, predicted secondary structure, solvent accessibility and predicted native disorder. Thus, for proteases with known amino acid specificity, PROSPER provides a convenient, pre-prepared tool for use in identifying protein substrates for the enzymes. Systematic prediction analysis for the twenty-four proteases thus far included in the database revealed that the features we have included in the tool strongly improve performance in terms of cleavage site prediction, as evidenced by their contribution to performance improvement in terms of identifying known cleavage sites in substrates for these enzymes. In comparison with two state-of-the-art prediction tools, PoPS and SitePrediction, PROSPER achieves greater accuracy and coverage. To our knowledge, PROSPER is the first comprehensive server capable of predicting cleavage sites of multiple proteases within a single substrate sequence using machine learning techniques. It is freely available at http://lightning.med.monash.edu.au/PROSPER/.     

Bioinformatic approaches for predicting substrates of proteases

Proteases have central roles in "life and death" processes due to their important ability to catalytically hydrolyze protein substrates, usually altering the function and/or activity of the target in the process. Knowledge of the substrate specificity of a protease should, in theory, dramatically improve the ability to predict target protein substrates. However, experimental identification and characterization of protease substrates is often difficult and time-consuming. Thus solving the "substrate identification" problem is fundamental to both understanding protease biology and the development of therapeutics that target specific protease-regulated pathways. In this context, bioinformatic prediction of protease substrates may provide useful and experimentally testable information about novel potential cleavage sites in candidate substrates. In this article, we provide an overview of recent advances in developing bioinformatic approaches for predicting protease substrate cleavage sites and identifying novel putative substrates. We discuss the advantages and drawbacks of the current methods and detail how more accurate models can be built by deriving multiple sequence and structural features of substrates. We also provide some suggestions about how future studies might further improve the accuracy of protease substrate specificity prediction.     

Global sequencing of proteolytic cleavage sites in apoptosis by specific labeling of protein N termini

The nearly 600 proteases in the human genome regulate a diversity of biological processes, including programmed cell death. Comprehensive characterization of protease signaling in complex biological samples is limited by available proteomic methods. We have developed a general approach for global identification of proteolytic cleavage sites using an engineered enzyme to selectively biotinylate free protein N termini for positive enrichment of corresponding N-terminal peptides. Using this method to study apoptosis, we have sequenced 333 caspase-like cleavage sites distributed among 292 protein substrates. These sites are generally not predicted by in vitro caspase substrate specificity but can be used to predict other physiological caspase cleavage sites. Structural bioinformatic studies show that caspase cleavage sites often appear in surface-accessible loops and even occasionally in helical regions. Strikingly, we also find that a disproportionate number of caspase substrates physically interact, suggesting that these dimeric proteases target protein complexes and networks to elicit apoptosis.     

Proteolysis to Identify Protease Substrates: Cleave to Decipher

Proteolysis is an irreversible post‐translational modification process, characterized by highly precise yet stable cleavage of proteins. Downstream events in signaling processes are reliant on proteolysis triggered by the protease activity. Studies indicate that abnormal proteolytic activity may lead to the manifestation of diseased conditions. Therefore, characterization of proteases may provide clues to understand their role in fundamental cellular processes like cellular growth, differentiation, apoptosis, and survival. The relevance of proteases and their substrates as clinical targets are being studied. Understanding the mechanism of proteolytic activity, the identity, and the role of repertoire of its substrates in a physiological pathway has opened avenues for novel drug designing. However, only a limited knowledge of protease substrates is currently available. In this review, the authors recapitulate the library screening, proteomics, and bioinformatics based approaches that have been employed for the identification of protease substrates.     

Determination of protease cleavage site motifs using mixture-based oriented peptide libraries.

The number of known proteases is increasing at a tremendous rate as a consequence of genome sequencing projects. Although one can guess at the functions of these novel enzymes by considering sequence homology to known proteases, there is a need for new tools to rapidly provide functional information on large numbers of proteins. We describe a method for determining the cleavage site specificity of proteolytic enzymes that involves pooled sequencing of peptide library mixtures. The method was used to determine cleavage site motifs for six enzymes in the matrix metalloprotease (MMP) family. The results were validated by comparison with previous literature and by analyzing the cleavage of individually synthesized peptide substrates. The library data led us to identify the proteoglycan neurocan as a novel MMP-2 substrate. Our results indicate that a small set of libraries can be used to quickly profile an expanding protease family, providing information applicable to the design of inhibitors and to the identification of protein substrates.     

Prediction of substrate sites for protein phosphatases 1B,SHP-1, and SHP-2 based on sequence features

Tyrosine phosphorylation plays crucial roles in numerous physiological processes. The level of phosphorylation state depends on the combined action of protein tyrosine kinases and protein tyrosine phosphatases. Detection of possible phosphorylation and dephosphorylation sites can provide useful information to the functional studies of relevant proteins. Several studies have focused on the identification of protein tyrosine kinase substrates. However, compared with protein tyrosine kinases, the prediction of protein tyrosine phosphatase substrates involved in the balance of protein phosphorylation level falls behind. This paper described a method that utilized the k-nearest neighbor algorithm to identity the substrate sites of three protein tyrosine phosphatases based on the sequence features of manually collected dephosphorylation sites. In the performance evaluation, both sensitivities and specificities could reach above 75 % for all three protein tyrosine phosphatases. Finally, the method was applied on a set of known tyrosine phosphorylation sites to search for candidate substrates.     

Understanding the substrate specificity of conventional calpains

Abstract: Calpains are intracellular Ca2+-dependent Cys proteases that play important roles in a wide range of biological phenomena via the limited proteolysis of their substrates. Genetic defects in calpain genes cause lethality and/or functional deficits in many organisms, including humans. Despite their biological importance, the mechanisms underlying the action of calpains, particularly of their substrate specificities, remain largely unknown. Studies show that certain sequence preferences influence calpain substrate recognition, and some properties of amino acids have been related successfully to substrate specificity and to the calpains' 3D structure. The full spectrum of this substrate specificity, however, has not been clarified using standard sequence analysis algorithms, e.g., the position-specific scoring-matrix method. More advanced bioinformatics techniques were used recently to identify the substrate specificities of calpains and to develop a predictor for calpain cleavage sites, demonstrating the potential of combining empirical data acquisition and machine learning. This review discusses the calpains' substrate specificities, introducing the benefits of bioinformatics applications. In conclusion, machine learning has led to the development of useful predictors for calpain cleavage sites, although the accuracy of the predictions still needs improvement. Machine learning has also elucidated information about the properties of calpains' substrate specificities, including a preference for sequences over secondary structures and the existence of a substrate specificity difference between two similar conventional calpains, which has never been indicated biochemically.     

The papain-like protease of severe acute respiratory syndrome coronavirus has deubiquitinating activity

Replication of the genomic RNA of severe acute respiratory syndrome coronavirus (SARS-CoV) is mediated by replicase polyproteins that are processed by two viral proteases, papain-like protease (PLpro) and 3C-like protease (3CLpro). Previously, we showed that SARS-CoV PLpro processes the replicase polyprotein at three conserved cleavage sites. Here, we report the identification and characterization of a 316-amino-acid catalytic core domain of PLpro that can efficiently cleave replicase substrates in trans-cleavage assays and peptide substrates in fluorescent resonance energy transfer-based protease assays. We performed bioinformatics analysis on 16 papain-like protease domains from nine different coronaviruses and identified a putative catalytic triad (Cys1651-His1812-Asp1826) and zinc-binding site. Mutagenesis studies revealed that Asp1826 and the four cysteine residues involved in zinc binding are essential for SARS-CoV PLpro activity. Molecular modeling of SARS-CoV PLpro suggested that this catalytic core may also have deubiquitinating activity. We tested this hypothesis by measuring the deubiquitinating activity of PLpro by two independent assays. SARS CoV-PLpro hydrolyzed both diubiquitin and ubiquitin-7-amino-4-methylcoumarin (AMC) substrates, and hydrolysis of ubiquitin-AMC is approximately 180-fold more efficient than hydrolysis of a peptide substrate that mimics the PLpro replicase recognition sequence. To investigate the critical determinants recognized by PLpro, we performed site-directed mutagenesis on the P6 to P2' residues at each of the three PLpro cleavage sites. We found that PLpro recognizes the consensus cleavage sequence LXGG, which is also the consensus sequence recognized by cellular deubiquitinating enzymes. This similarity in the substrate recognition sites should be considered during the development of SARS-CoV PLpro inhibitors.     

Binding of nonphysiological protein and peptide substrates to proteases: differences between urokinase-type plasminogen activator and trypsin and contributions to the evolution of regulated proteolysis

Understanding the regulation of physiological processes requires detailed knowledge of the recognition of substrates by enzymes. One of the most productive model systems for the study of enzyme-substrate interactions is the serine protease family; however, most studies of protease action have used small substrates that contain an activated, non-natural scissile bond. Because few kinetic or structural studies have used protein substrates, the physiologically relevant target of most proteases, it seems likely that important mechanisms of substrate recognition and processing by proteases have not yet been fully elucidated. Consistent with this hypothesis, we have observed that K(m) values for protein substrates are reduced as much as 200-15000-fold relative to those of analogous peptide substrates. Here we examine the thermodynamic consequences of interactions between proteases and their substrates using staphylococcal nuclease (SNase) and SNase variants as model protein substrates. We have obtained values for enthalpy, entropy, and K(d) for binding of proteins and peptides by the nonspecific protease trypsin and the highly specific protease urokinase-type plasminogen activator (u-PA). To avoid cleavage of substrates during these measurements, we used inactive variants of trypsin and u-PA whose catalytic serine S195 had been replaced by alanine. Differences in the K(d) values for binding of protein and peptide substrates closely approximate the large differences observed in the corresponding K(m) values. Improved binding of protein substrates is due to decreased enthalpy, and this effect is pronounced for the selective protease u-PA. Fundamental differences in recognition of analogous protein and peptide substrates may have influenced the evolution of protease specificity.     

Proteomic analyses reveal an acidic prime side specificity for the astacin metalloprotease family reflected by physiological substrates

Astacins are secreted and membrane-bound metalloproteases with clear associations to many important pathological and physiological processes. Yet with only a few substrates described their biological roles are enigmatic. Moreover, the lack of knowledge of astacin cleavage site specificities hampers assay and drug development. Using PICS (proteomic identification of protease cleavage site specificity) and TAILS (terminal amine isotopic labeling of substrates) degradomics approaches >3000 cleavage sites were proteomically identified for five different astacins. Such broad coverage enables family-wide determination of specificities N- and C-terminal to the scissile peptide bond. Remarkably, meprin α, meprin β, and LAST_MAM proteases exhibit a strong preference for aspartate in the peptide (P)1' position because of a conserved positively charged residue in the active cleft subsite (S)1'. This unparalleled specificity has not been found for other families of extracellular proteases. Interestingly, cleavage specificity is also strongly influenced by proline in P2' or P3' leading to a rare example of subsite cooperativity. This specificity characterizes the astacins as unique contributors to extracellular proteolysis that is corroborated by known cleavage sites in procollagen I+III, VEGF (vascular endothelial growth factor)-A, IL (interleukin)-1β, and pro-kallikrein 7. Indeed, cleavage sites in VEGF-A and pro-kallikrein 7 identified by terminal amine isotopic labeling of substrates matched those reported by Edman degradation. Moreover, the novel substrate FGF-19 was validated biochemically and shown to exhibit altered biological activity after meprin processing.     

PACMANS: A bioinformatically informed algorithm to predict,design, and disrupt protease‐on‐protease hydrolysis

Multiple proteases in a system hydrolyze target substrates, but recent evidence indicates that some proteases will degrade other proteases as well. Cathepsin S hydrolysis of cathepsin K is one such example. These interactions may be uni‐ or bi‐directional and change the expected kinetics. To explore potential protease‐on‐protease interactions in silico, a program was developed for users to input two proteases: (1) the protease‐ase that hydrolyzes (2) the substrate, protease. This program identifies putative sites on the substrate protease highly susceptible to cleavage by the protease‐ase, using a sliding‐window approach that scores amino acid sequences by their preference in the protease‐ase active site, culled from MEROPS database. We call this PACMANS, Protease‐Ase Cleavage from MEROPS ANalyzed Specificities, and test and validate this algorithm with cathepsins S and K. PACMANS cumulative likelihood scoring identified L253 and V171 as sites on cathepsin K subject to cathepsin S hydrolysis. Mutations made at these locations were tested to block hydrolysis and validate PACMANS predictions. L253A and L253V cathepsin K mutants significantly reduced cathepsin S hydrolysis, validating PACMANS unbiased identification of these sites. Interfamilial protease interactions between cathepsin S and MMP‐2 or MMP‐9 were tested after predictions by PACMANS, confirming its utility for these systems as well. PACMANS is unique compared to other putative site cleavage programs by allowing users to define the proteases of interest and target, and can also be employed for non‐protease substrate proteins, as well as short peptide sequences.     

Substrate specificity of rhomboid intramembrane proteases is governed by helix-breaking residues in the substrate transmembrane domain

Rhomboid intramembrane proteases initiate cell signaling during Drosophila development and Providencia bacterial growth by cleaving transmembrane ligand precursors. We have determined how specificity is achieved: Drosophila Rhomboid-1 is a site-specific protease that recognizes its substrate Spitz by a small region of the Spitz transmembrane domain (TMD). This substrate motif is necessary and sufficient for cleavage and is composed of residues known to disrupt helices. Rhomboids from diverse organisms including bacteria and vertebrates recognize the same substrate motif, suggesting that they use a universal targeting strategy. We used this information to search for other rhomboid substrates and identified a family of adhesion proteins from the human parasite Toxoplasma gondii, the TMDs of which were efficient substrates for rhomboid proteases. Intramembrane cleavage of these proteins is required for host cell invasion. These results provide an explanation of how rhomboid proteases achieve specificity, and allow some rhomboid substrates to be predicted from sequence information.     

N- and C-terminal degradomics: new approaches to reveal biological roles for plant proteases from substrate identification

Proteolysis is an irreversible post-translational modification that regulates many intra- and intercellular processes, including essential go/no-go decisions during cell proliferation, development and cell death. Hundreds of protease-coding genes have been identified in plants, but few have been linked to specific substrates. Conversely, proteolytic processes are frequently observed in plant biology but rarely have they been ascribed to specific proteases. In mammalian systems, unbiased system-wide proteomics analyses of protease activities have recently been tremendously successful in the identification of protease substrate repertoires, also known as substrate degradomes. Knowledge of the substrate degradome is key to understand the role of proteases in vivo. Quantitative shotgun proteomic studies have been successful in identifying protease substrates, but while simple to perform they are biased toward abundant proteins and do not reveal precise cleavage sites. Current degradomics techniques overcome these limitations by focusing on the information-rich amino- and carboxy-terminal peptides of the original mature proteins and the protease-generated neo-termini. Targeted quantitative analysis of protein termini identifies precise cleavage sites in protease substrates with exquisite sensitivity and dynamic range in in vitro and in vivo systems. This review provides an overview of state-of-the-art methods for enrichment of protein terminal peptides, and their application to protease research. These emerging degradomics techniques promise to clarify the elusive biological roles of proteases and proteolysis in plants.     

Cleavage Specificity Analysis of Six Type II Transmembrane Serine Proteases (TTSPs) Using PICS with Proteome-Derived Peptide Libraries

Background

Type II transmembrane serine proteases (TTSPs) are a family of cell membrane tethered serine proteases with unclear roles as their cleavage site specificities and substrate degradomes have not been fully elucidated. Indeed just 52 cleavage sites are annotated in MEROPS, the database of proteases, their substrates and inhibitors.

Methodology/Principal Finding

To profile the active site specificities of the TTSPs, we applied Proteomic Identification of protease Cleavage Sites (PICS). Human proteome-derived database searchable peptide libraries were assayed with six human TTSPs (matriptase, matriptase-2, matriptase-3, HAT, DESC and hepsin) to simultaneously determine sequence preferences on the N-terminal non-prime (P) and C-terminal prime (P’) sides of the scissile bond. Prime-side cleavage products were isolated following biotinylation and identified by tandem mass spectrometry. The corresponding non-prime side sequences were derived from human proteome databases using bioinformatics. Sequencing of 2,405 individual cleaved peptides allowed for the development of the family consensus protease cleavage site specificity revealing a strong specificity for arginine in the P1 position and surprisingly a lysine in P1′ position. TTSP cleavage between R↓K was confirmed using synthetic peptides. By parsing through known substrates and known structures of TTSP catalytic domains, and by modeling the remainder, structural explanations for this strong specificity were derived.

Conclusions

Degradomics analysis of 2,405 cleavage sites revealed a similar and characteristic TTSP family specificity at the P1 and P1′ positions for arginine and lysine in unfolded peptides. The prime side is important for cleavage specificity, thus making these proteases unusual within the tryptic-enzyme class that generally has overriding non-prime side specificity.     

Characterizing Protease Specificity: How Many Substrates Do We Need?

Calculation of cleavage entropies allows to quantify, map and compare protease substrate specificity by an information entropy based approach. The metric intrinsically depends on the number of experimentally determined substrates (data points). Thus a statistical analysis of its numerical stability is crucial to estimate the systematic error made by estimating specificity based on a limited number of substrates. In this contribution, we show the mathematical basis for estimating the uncertainty in cleavage entropies. Sets of cleavage entropies are calculated using experimental cleavage data and modeled extreme cases. By analyzing the underlying mathematics and applying statistical tools, a linear dependence of the metric in respect to 1/n was found. This allows us to extrapolate the values to an infinite number of samples and to estimate the errors. Analyzing the errors, a minimum number of 30 substrates was found to be necessary to characterize substrate specificity, in terms of amino acid variability, for a protease (S4-S4’) with an uncertainty of 5 percent. Therefore, we encourage experimental researchers in the protease field to record specificity profiles of novel proteases aiming to identify at least 30 peptide substrates of maximum sequence diversity. We expect a full characterization of protease specificity helpful to rationalize biological functions of proteases and to assist rational drug design.     

Mechanisms of Gasdermin Recognition by Proteases

Members of the gasdermin family contain positively charged N-terminal domains (NTDs) capable of binding phospholipids and assembling membrane pores, and C-terminal domains (CTDs) that bind the NTDs to prevent pore formation in the resting states. The flexible NTD-CTD linker regions of gasdermins are highly variable in length and sequences, which may be attributable to gasdermin recognition by diverse proteases. In addition, protease cleavage within the NTDs is known to inactivate several gasdermin family members. Recognition and cleavage of the gasdermin family members by different proteases share common and distinct features at the protease active sites, as well as exosites recently identified for the inflammatory caspases. Utilization of exosites may strengthen enzyme-substrate interaction, improve efficiency of proteolysis, and enhance substrate selectivity. It remains to be determined if the dual site recognition of gasdermin D (GSDMD) by the inflammatory caspases is employed by other GSDMD-targeting proteases, or is involved in proteolytic processing of other gasdermins. Biochemical and structural approaches will be instrumental in revealing how potential exosites in diverse proteases engage different gasdermin substrates. Different features of gasdermin sequence, structure, expression characteristics, and post-translational modifications may dictate distinct mechanisms of protease-dependent activation or inactivation. Such diverse mechanisms may underlie the divergent physiological and pathological functions of gasdermins, and furnish opportunities for therapeutic targeting of gasdermins in infectious diseases and inflammatory disorders.     

Direct measurement of acylenzyme hydrolysis demonstrates rate-limiting deacylation in cleavage of physiological sequences by the processing protease Kex2

Saccharomyces cerevisiae Kex2 protease is the prototype for the family of eukaryotic proprotein convertases that includes furin, PC1/3, and PC2. These enzymes belong to the subtilase superfamily of serine proteases and are distinguished from degradative subtilisins by structural features and by their much more stringent substrate specificity. Pre-steady-state studies have shown that both Kex2 and furin exhibit an initial burst of 7-amino-4-methylcoumarin release in cleavage of peptidyl methylcoumarinamide substrates that are based on physiological cleavage sites. Thus, in cleavage of such substrates, formation of the acylenzyme intermediate is fast relative to some later step (deacylation or N-terminal product release). This behavior is significant, because Kex2 also exhibits burst kinetics in cleavage of peptide bonds. k(cat) for cleavage of a tetrapeptidyl methylcoumarinamide substrate based on the physiological yeast substrate pro-alpha-factor exhibits a weak solvent isotope effect, but neither this isotope effect nor temperature dependence studies with this substrate conclusively identify the rate-limiting step for Kex2 cleavage of this substrate. We therefore developed an assay to measure deacylation directly by pulse-chase incorporation of H(2)(18)O in a rapid-quenched-flow mixer followed by mass spectrometric quantitation. The results given by this assay rule out rate-limiting product release for cleavage of this substrate by Kex2. These experiments demonstrate that cleavage of the acylenzyme ester bond, as opposed to either the initial attack on the amide bond or product release, is rate-limiting for the action of Kex2 at physiological sequences. This work demonstrates a fundamental difference in the catalytic strategy of proprotein processing enzymes and degradative subtilisins.     

CASVM: web server for SVM-based prediction of caspase substrates cleavage sites

Caspases belong to a unique class of cysteine proteases which function as critical effectors of apoptosis, inflammation and other important cellular processes. Caspases cleave substrates at specific tetrapeptide sites after a highly conserved aspartic acid residue. Prediction of such cleavage sites will complement structural and functional studies on substrates cleavage as well as discovery of new substrates. We have recently developed a support vector machines (SVM) method to address this issue. Our algorithm achieved an accuracy ranging from 81.25 to 97.92%, making it one of the best methods currently available. CASVM is the web server implementation of our SVM algorithms, written in Perl and hosted on a Linux platform. The server can be used for predicting non-canonical caspase substrate cleavage sites. We have also included a relational database containing experimentally verified caspase substrates retrievable using accession IDs, keywords or sequence similarity. AVAILABILITY: http://www.casbase.org/casvm/index.html     

Pripper: prediction of caspase cleavage sites from whole proteomes

Background  

Caspases are a family of proteases that have central functions in programmed cell death (apoptosis) and inflammation. Caspases mediate their effects through aspartate-specific cleavage of their target proteins, and at present almost 400 caspase substrates are known. There are several methods developed to predict caspase cleavage sites from individual proteins, but currently none of them can be used to predict caspase cleavage sites from multiple proteins or entire proteomes, or to use several classifiers in combination. The possibility to create a database from predicted caspase cleavage products for the whole genome could significantly aid in identifying novel caspase targets from tandem mass spectrometry based proteomic experiments.