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

Abstract

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.

Figure 1

A subtiligase-based method for positive selection of peptides corresponding to N-termini of proteins from complex mixtures. (A) Workflow for the biotinylation of protein N-termini in complex mixtures using subtiligase and a biotinylated peptide ester that contains a TEV protease cleavage site, trypsinization of labeled proteins, capture of biotinylated N-terminal peptides with immobilized avidin, recovery of captured peptides using TEV protease, and analysis of N-terminal peptides by 1D or 2D LC/MS/MS for identification of corresponding proteins and cleavage sites. The representative MS/MS spectrum corresponds to semi-tryptic peptide GSAVNGTSSAETNLEALQK from MEK1 (MP2K1_HUMAN) and identifies a previously unknown caspase-like cleavage site at Asp 16. The a₂ and b₂ ions at m/z 223 and 251 are characteristic hallmarks of a ligated, SY-bearing, N-terminal peptide. (B) Structure of the biotinylated peptide glycolate ester used in the proteomic workflow.

Figure 2

N-termini derived from caspase-like proteolytic processing are a hallmark of apoptotic cells. (A) Frequencies of P1 and P1′ amino acid residues corresponding to non-homologous N-termini identified in small-scale 1D LC/MS/MS experiments with untreated and apoptotic Jurkat cells. Data are represented as mean ± SD (n=2 for untreated and n=4 for apoptotic). (B) Frequencies of P1 and P1′ amino acid residues corresponding to non-homologous N-termini identified in large-scale 2D LC/MS/MS experiments with untreated and apoptotic Jurkat cells. Data are represented as mean ± SD (n=2 for untreated and n=3 for apoptotic). “ – “ indicates lack of a putative P1 residue in cases where the P1′ residue is an initiator methionine.

Figure 3

Substrate specificity of the caspase-like proteolytic activity in etoposide-stimulated Jurkat cells. (A) Sequence logo representation (Crooks et al., 2004) of the frequency of amino acid residues in the identified caspase cleavage sites. (B) Sequence logo representation of the in vitro substrate specificity of caspase-3 (Stennicke et al., 2000; Thornberry et al., 1997). (C) Sequence logo representation of the frequency of amino acid residues in known human caspase cleavage sites (Lüthi and Martin, 2007). (D) Frequency of P4-P1 motifs in the identified caspase cleavage sites. (E) Receiver operator characteristic curves showing the discrimination ability of HMMs constructed from three different cleavage site training sets (Jurkat, literature, and merged). Three representative HMM score threshold values for the merged dataset are indicates (TPR = true positive rate, FPR = false positive rate).

Figure 4

Structural determinants of caspase substrate specificity. (A) Solvent accessibility (> 30% surface area exposed) at each position of all known P4-P4′ caspase cleavage sites and each position of all eight residue sequences containing aspartate in the fourth position found in PDB protein structures (control peptides). Differences between cleavage sites and control peptides have associated p-values < 0.0001 by chi-square at all positions. (B) Sequence logo representations of secondary structure at each position of all known P4-P4′ caspase cleavage sites and each position of the control peptides described above, where L = loop, A = α-helix, and B = β-sheet. Differences between cleavage sites and control peptides have associated p-values < 0.0001 by chi-square at all positions, except for P2 with 0.0014 and P1 with 0.0007. (C) Distribution of secondary structure motifs for P4-P4′ caspase cleavage sites and for the control peptides described above. (D) Localization of caspase cleavage sites in substrates relative to functional domain boundaries annotated in Pfam compared to localization of all eight residue sequences containing aspartate in the fourth position found in the human Swiss-Prot database (control peptides). Differences between cleavage sites and control peptides have associated p-values < 0.0001 by chi-square in all cases.

Figure 5

Network analysis of protein interactions between caspase substrates, including those identified in this work and previously reported human caspase substrates (Lüthi and Martin, 2007). (A) Enrichment in protein-protein interactions between the 602 total caspase substrates relative to an equally sized reference control set of protein interactors randomly selected from protein interaction databases. Data for the control set are represented as mean ± SD (n=10). (B–K) Caspase substrate protein interaction subnetworks encompassing substrates annotated to overrepresentated GO biological process terms relative to the entire human GO annotation. Substrates are labeled using gene symbols. Corrected p-values, number of nodes, and number of edges are indicated in each case.

Figure 6

Analysis of proteolysis of N-CoR/SMRT corepressor complex components during apoptosis in Jurkat cells following treatment with etoposide (50 μM). (A) Caspase substrate protein interaction subnetwork encompassing components of the N-CoR/SMRT corepressor complex (N-CoR, SPEN, TBLR1, RBBP7, and HDAC7), and transcription factors such as retinoic acid receptor, androgen receptor, and SP1 (green = from this work, red = from literature, blue = in both datasets). (B) Schematic representation of N-CoR/SMRT corepressor complex resident components and visiting interactors (red label = resident component, white label = visiting interactor, black fill = target of proteolysis in apoptosis). (C) Time courses for the proteolysis of procaspase-3 and DFF35/45, and for oligonucleosomal DNA fragmentation. (D) Full cleavage of N-CoR, HDAC7, SHARP, and TBLR1, and partial cleavage of RBBP7. (E) Full cleavage of SMRT and of HDAC-3, a previously identified caspase substrate. Black arrows indicate full-length proteins. Red arrows indicate expected cleavage products for cleavage at the sites identified in our studies (cleavage products were not detected in all cases).

Figure 7

Caspases cleave resident N-CoR/SMRT complex components and visiting interactors at regions leading to separation of functional domains. Functional domain organization and candidate caspase cleavage sites in SMRT and TBL1 are similar to those indicated for the respective homologs, N-CoR and TBL1R.