Fig 1.
Basis for quantitative specificity profiling of the MMP catalytic cleft across S3 –S1՛.
A. Combinatorics of MMP–substrate interactions define the limits of substrate specificity and substrate fitness in experiments using hexapeptide probes. S3 and S1՛ are the most selective binding sites in the catalytic cleft of the MMPs (bold blue lettering). Together with the S2 and S1, they interact with P3-P1՛ tetramers in substrates (red lettering). To interrogate specificity of MMP-2 and 9, we used a library of randomized hexapeptides displayed on PIII gene product of M13 phage. The theoretical maximum of hexamer combinations is 64,000,000. The theoretical maximum for the number of hexapeptides containing identical tetramers (tetramer cluster) is 1200. There are 160,000 combinations of natural amino acid residues in random tetramers. B. Results of phage display analysis can be interpreted to quantify MMP specificity as well as the fitness of individual P3-P1՛ substrates. The number of tetramer clusters defines the amount of specificity of proteases recognizing P3-P1՛ positions in substrates, which ranges from absolutely specific (1 tetramer cluster) to absolutely non-specific (160,000 tetramer clusters). The number of hexamers per tetramer cluster is a measure of substrate fitness of all hexamers comprising it up to the kcat/KM threshold defined by experimental conditions.
Fig 2.
Probability of finding a tetramer cluster in substrate selections relative to the naïve library (RP) correlates with substrate fitness.
K(obs) values for 1369 individual phage substrates or kcat/KM values of 100 peptides derived from substrate phage sequences were experimentally determined as described in the text. The substrates were binned into evenly distributed groups based on the RP values of the tetramer clusters corresponding to their P3-P1՛ positions in substrates. The average K(obs) (A) or kcat/KM (B) values for each bin were plotted as a function of the corresponding average RP values and the data were subjected to linear regression analysis. The equations and goodness of fit parameters (R2) of the linear regression analyses of the MMP-2 and 9 data are shown at the top and bottom of the graph, respectively. These results can be compared to raw, unbinned data presented in the last graphs of the S6 Table, in the corresponding MMP specific tabs. The corresponding Pearson correlation coefficient R values for the raw (unbinned) data are 0.66 and 0.76 for MMP-2 and MMP-9, respectively, indicating already well correlated data.
Fig 3.
Divergence between probability distributions of tetramer clusters in substrate selections and the naïve library is a measure of substrate specificity.
A. Distributions of relative abundances of tetramer clusters in the MMP substrate selections are significantly different from that in the naïve phage display library. Tetramer clusters within evenly spaced ranges of probabilities were binned together and their relative abundances were plotted as a function of log10 of average probabilities in the respective sets. B. A subset of tetramer clusters in substrate selections with positive cumulative contribution to K-L divergence relative to the naïve library constitutes the selectome of a protease. Cumulative contribution of individual tetramer clusters with RP values between of 0 and 4.5 for MMP-2 and 0 and 4.7 for MMP-9 to the K-L divergence relative to the naïve library is equal to 0. The K-L divergence between the probability distributions in the substrate sets of a protease with no definable specificity and the naïve library is always equal to 0. Therefore, the substrate sets with overall positive cumulative individual contributions to K-L divergences constitute the selectomes of proteases with definable specificities (MMP-2 and 9, red arrows).
Fig 4.
Composition of the selectomes of MMP-2 and 9 reflect differences in their specificities.
A. Selectome-based specificity profiles of MMP-2 and 9 reflect changes in substrate composition as a function of substrate fitness. Peptide hexamers in the selectomes of MMP-2 and 9 were aligned along the P5-P3՛ positions based on P3-P1՛ matches in the corresponding tetramer clusters and divided into 10 groups according to their RP values relative to the maximum (RP/RPMax). Relative abundances of amino acid residues at each position were calculated and presented in the form of a logo plot for each of the groups. B. Aggregate specificity profiles of MMP-2 and 9 reveal the major selectivity features of MMP-2 and 9. Hexamers belonging to the tetramer clusters constituting the selectomes of MMP-2 and 9 were aligned along the P5-P3՛ positions based on P3-P1՛ matches in the corresponding tetramer clusters, and relative abundances of residues at each position were plotted in the form of a logo. Zoom-in ovals show the relative contributions of residues to the P2 specificity profile of each enzyme.
Fig 5.
Comparative analysis of selectomes reveals divergent and conserved features in substrate recognition between MMP-2 and 9.
A. Venn diagram of substrate specificity overlap and distinction between the selectomes of MMP-2 and 9 shows significant selectivity between the two enzymes. Tetramer clusters comprising the selectomes of MMP-2 and 9 were grouped based on their occurrence in the individual and overlapping substrate sets and the corresponding numbers are shown in the Venn diagram. For making Venn diagram we used on-line service at: http://bioinformatics.psb.ugent.be/webtools/Venn/. B. Aggregate specificity profiles based on the unique and overlapping tetramer clusters of MMP-2 and 9 reveal the distribution of selectivity across the catalytic cleft. Hexamer peptides belonging to the tetramer clusters constituting the selective and common substrate sets of MMP-2 and 9 were aligned along P5-P3՛ positions based on the P3-P1՛ matches and the relative abundances of residues at each position were plotted as logos. MMP-2&9/2 denotes residue frequencies based on MMP-2 RP values, while MMP-2&9/9 denotes residue frequencies based on MMP-9 RP values.
Fig 6.
Changes in composition of selective substrates correlate with changes in selectivity determinants between MMP-2 and 9.
A. Distribution of the selectivity determinants across the subsites of the catalytic clefts of MMP-2 and 9. Sequences of the catalytic domains of MMP-2 and 9 were aligned based on the crystal structures of the catalytic domains of the respective enzymes (PDB IDs: 1QID for MMP2, 1GKC for MMP9). Residue numberings from the native N-termini are shown above (MMP-2) or below (MMP-9) the sequence. Selectivity Determining Positions (SDPs) are shown in color and marked in larger font. The catalytic cleft subsites they contribute to are shown directly above each SDP. Residues marked in bold differ between the two proteases (see text for details). B. Structural features of the SDPs in the catalytic clefts of MMP-2 and 9 provide basis for experimentally determined subsite selectivity. Residues contributing to the SDPs at S3, S2, S1 and S1՛ binding pockets are shown on the surface representations of the three-dimensional structures of MMP-2 and 9 in colors matching the sequence alignments in A. See text for more details. PyMOL molecular visualization system was used for display and analysis of 3D structures. C. Single substitutions in substrate tetramers illustrate distinctions in substrate recognition by MMP-2 and 9. Substrates selective for MMP-2 (KELAN↓Q), MMP-9 (KEPFN↓Q) and in common between the two enzymes (KEPAN↓Q) were docked into the catalytic clefts of MMP-2 and MMP-9 (See the Methods section for details). The P3-P1՛ sequences are shown above the structures. The docked residues below corresponding to the P3 and P2 residues in tetramers are shown as spheres, while the rest of the sequence is shown as sticks. Positions of residues relative to the scissile bond in the docked peptides are marked by white lettering. The RP values for the corresponding P3-P1՛ tetramer clusters are shown directly below the structures of each complex.
Fig 7.
Enrichment in novel N-termini following hydrolysis by MMP-2 or 9 reflects the proportion of protein substrates in respective selectomes.
Following hydrolysis with MMP-2 or 9, the secretome of HEK293 cells was labeled with TMT isobaric tags. Isotopic enrichment (IE) of the novel N-terminally labeled peptides in the MMP-treated samples relative to the untreated controls, was determined by LC/MS analysis of tryptic digests of the labeled secretomes (See text and Methods section for details). The graphs show 10-90th percentile box and whiskers plots of the RP values of the P3-P1՛ sequences corresponding to cleavages in human proteins by MMP-2 (A) and 9 (B) deduced from sequences of N-terminally labeled peptides, as a function of IE expressed as multiples of standard deviation relative to the population mean. The median and mean RP values are marked in the boxes by horizontal lines and crosses, respectively. The red dotted lines mark the selectome thresholds for MMP-2 (RP = 4.5) and MMP-9 (RP = 4.7).