Procleave: Predicting Protease-specific Substrate Cleavage Sites by Combining Sequence and Structural Information

Abstract

Proteases are enzymes that cleave and hydrolyse the peptide bonds between two specific amino acid residues of target substrate proteins. Protease-controlled proteolysis plays a key role in the degradation and recycling of proteins, which is essential for various physiological processes. Thus, solving the substrate identification problem will have important implications for the precise understanding of functions and physiological roles of proteases, as well as for therapeutic target identification and pharmaceutical applicability. Consequently, there is a great demand for bioinformatics methods that can predict novel substrate cleavage events with high accuracy by utilizing both sequence and structural information. In this study, we present Procleave, a novel bioinformatics approach for predicting protease-specific substrates and specific cleavage sites by taking into account both their sequence and 3D structural information. Structural features of known cleavage sites were represented by discrete values using a LOWESS data-smoothing optimization method, which turned out to be critical for the performance of Procleave. The optimal approximations of all structural parameter values were encoded in a conditional random field (CRF) computational framework, alongside sequence and chemical group-based features. Here, we demonstrate the outstanding performance of Procleave through extensive benchmarking and independent tests. Procleave is capable of correctly identifying most cleavage sites in the case study. Importantly, when applied to the human structural proteome encompassing 17,628 protein structures, Procleave suggests a number of potential novel target substrates and their corresponding cleavage sites of different proteases. Procleave is implemented as a webserver and is freely accessible at http://procleave.erc.monash.edu/.

Keywords: Cleavage site prediction; Conditional random field; Machine learning; Protease; Structural determinants.

Figure 1

The overall framework of Procleave There are five major steps in the framework of Procleave, including data pre-processing, feature extraction, model training and optimization, model testing and evaluation, as well as web server development.

Figure 2

Structural determinants of the substrate specificity of nine proteases across the P4–P4′ cleavage sites A. Cathepsin D. B. Cathepsin E. C. HIV-1 retropepsin. D. Cathepsin B. E. Caspase-3. F. MMP-2. G. MMP-9. H. Granzyme B (human). I. Cathepsin G. MMP, matrix metallopeptidase. The secondary structure information was extracted from DSSP results. H, helix; E, strand; L, loop.

Figure 3

Performance comparison of CRF models trained using different feature combinations in terms of AUC values A. Cathepsin D. B. Cathepsin E. C. HIV-1 retropepsin. D. Cathepsin B. E. Caspase-3. F. MMP-2. G MMP-9. H. Granzyme B (human). I. Cathepsin G. The evaluation was based on 10 times of 5-fold cross-validation tests on training datasets.

Figure 4

Comparison of cleavage site prediction performance of Procleave and other methods in terms of AUC values for 5 different proteases A. Cathepsin E. B. Caspase-3. C. Caspase-6. D. MMP-2. E. Granzyme B. PoPS, PROSPER, and iProt-Sub cannot predict cleavage sites of cathepsin E; SitePrediction and PROSPER cannot predict cleavage sites of granzyme B. SVM and RF were included to test whether the conditional random field model employed in Procleave provides better performance.

Figure 5

Predicted cleavage sites of four substrate protein structures A. Human αB crystalline (PDB ID: 3L1G, chain: A) cleaved by MMP-9. B. Human Interferon β (PDB ID: 1AU1, chain: A) cleaved by MMP-9. C. ATPase p97 mutant (PDB ID: 3HU2, chain A) cleaved by caspase-6. D. Human enolase 1 (PDB ID: 3B97, chain A) cleaved by meprin β.