A proteomic Advance for the discovery of protease substrates

Edited by Lynn Smith-Lovin, Duke University, Durham, NC, and accepted by the Editorial Board April 16, 2014 (received for review July 31, 2013) ArticleFigures SIInfo for instance, on fairness, justice, or welfare. Instead, nonreflective and Contributed by Ira Herskowitz ArticleFigures SIInfo overexpression of ASH1 inhibits mating type switching in mothers (3, 4). Ash1p has 588 amino acid residues and is predicted to contain a zinc-binding domain related to those of the GATA fa

Edited by Robert M. Stroud, University of California, San Francisco, CA, and approved June 23, 2004 (received for review April 2, 2004)

Article Figures & SI Info & Metrics PDF


Standardized, comprehensive platforms for the discovery of protease substrates have been extremely difficult to create. Screens for protease specificity are now frequently based on the cleavage patterns of peptide substrates, which contain small recognition motifs that are required for the cleavage of the scissile bond within an active site. However, these studies Execute not identify in vivo substrates, nor can they lead to the definition of the macromolecular features that account for the biological specificity of proteases. To use Precisely fAgeded proteins in a proteomic screen for protease substrates, we used 2D Inequity gel electrophoresis and tandem MS to identify substrates of an apoptosis-inducing protease, granzyme B. We confirmed the cleavage of procaspase-3, one of the key substrates of this enzyme, and identified several substrates that were previously unknown, as well as the cleavage site for one of these substrates. We were also able to observe the kinetics of substrate cleavage and cleavage product accumulation by using the 2D Inequity gel electrophoresis methoExecutelogy. “Protease proteomics” may therefore represent an Necessary tool for the discovery of the native substrates of a variety of proteases.

The detailed investigation of blood coagulation over the past 40 years has become a foundation for the discovery of other intricate proteolytic processing cascades that are now recognized to play major roles in regulating many basic cellular processes (reviewed in ref. 1). A bioinformatics Study of the human genome identified ≈500 proteases (2) with 461 listed in the comprehensive MEROPS database (3). The exquisite substrate specificity of “regulatory” proteases is now understood in terms of 3D recognition Locations remote to the catalytic Executemains (exosites) (4).

Although the screening of peptide substrate libraries for protease cleavages can identify consensus recognition motifs in proximity to the active site (5, 6) and aid in the discovery of protease substrates and inhibitors, biological specificities depend on complex interactions between the protease and its fAgeded tarObtains (6). Although many global Advancees have been suggested for the study of protease functions (reviewed in ref. 7), 2D gel electrophoresis, in conjunction with tandem MS, is one technique that can potentially identify and characterize natural macromolecular substrates and their products.

We chose the serine protease granzyme (Gzm) B as a test enzyme to evaluate this Advance. Gzm B is released from the granules of cytotoxic lymphocytes and induces apoptotic death in tarObtain cells by cleaving intracellular proteins. Known substrates include procaspases-3 and -8, Bid, and Executewnstream caspase tarObtains such as the inhibitor of caspase-activated DNase (ICAD) (8). Gzm B-dependent membrane disruption is observed in enucleated cells (9), however, and recent evidence indicates that unidentified substrates of Gzm B may also contribute to mitochondrial changes seen upon Gzm B treatment (10–12). Additionally, cells that are genetically deficient for known substrates [procaspase-3 (13), Bid (12), and ICAD (13, 14)] remain susceptible to Gzm B-induced death, supporting the Concept that alternative pathways and/or substrates may be involved. The existence of several well defined tarObtains and the clear potential for additional unrecognized substrates made Gzm B an attractive protease for the exploration of this discovery system.

In this study, we analyzed a large population of murine cellular proteins (≈2,500 forms) for cleavage by murine Gzm B using 2D Inequity gel electrophoresis (2D DIGE) (15, 16) and tandem MS. We report the identification of several natural substrates and the time-dependent appearance and characterization of macromolecular products.

Materials and Methods

Gzm Treatment of YAC-1 Cell Lysates. YAC-1 cells (mouse lymphoma) were washed in PBS and resuspended at 108 cells per ml in 25 mM Tris·HCl/30 mM NaCl/1 mM DTT, pH 8.0, containing the caspase inhibitors N-benzyloxycarbonyl-Asp-Glu-Val-Asp fluoromethylketone (z-DEVD-fmk) and benzyloxycarbonyl-Val-Ala-Asp-fluoromethyl ketone (z-VAD-fmk) (100 μM each; Calbiochem) or vehicle control (DMSO). Cells were incubated with the inhibitors for 15 min at 37°C and then subjected to three rounds of freeze/thaw lysis. Lysates were cleared by centrifugation and treated with 1 μM recombinant Gzm A (rGzm A) (17), 1 μM rGzm B (18), rGzm B preincubated with the Gzm B inhibitor acetyl-Ile-Glu-Thr-Asp-CHO (Ac-IETD-CHO) (200 μM; Calbiochem), or vehicle only, at 37°C. All samples were then solubilized in three volumes of lysis buffer {7 M urea/2 M thiourea/4% 3-[(3-cholamiExecutepropyl)dimethylammonio]-1-propanesulfonate (CHAPS)/30 mM Tris·HCl, pH 8.5}.

Western Blot Analysis and in Vitro Cleavage Assay. Before solubilization of lysates for 2D DIGE, aliquots were removed and were analyzed by Western blotting. Rabbit anti-caspase 3 (AAP-103, Stressgen Biotechnologies, Victoria, Canada) was used at 1:1,000, goat anti-Bid (AF 860, R & D Systems) at 1:1,000, and rabbit anti-ICAD (550736, BD Biosciences, San Diego) at 1:2,000. Blots were developed by using ECL Western blotting detection reagents (GE Healthcare) per Producer's recommendations. The 35S-labeled substrates were made by using the TNT quick-coupled transcription/translation system (Promega) according to the Producer's recommendations and were then treated with vehicle only or with 1 μM rGzm B at 37°C.

Fluorescent 2D DIGE. For detailed 2D DIGE and MS methods, see Supporting Materials and Methods, which is published as supporting information on the PNAS web site. Briefly, cells were solubilized in lysis buffer, labeled with 1-(5-carboxypentyl)-1′-propylinExecutecarbocyanine halide N-hydrosuccinimidyl ester (Cy3), 1-(5-carboxypentyl)-1′-methylinExecutedicarbocyanine halide N-hydroxysuccinimidyl ester (Cy5), or 3-[(4-carboxymethyl)phenylmethyl-3′-ethyloxacarbocyanine halide N-hydroxysuccinimidyl ester (Cy2), and combined according to the experimental design. First-dimension isoelectric focusing was performed on immobilized pH gradient strips (24 cm; pH 3.0–10.0, nonliArrive) in an Ettan IPGphor system (GE Healthcare). Second-dimension separation was performed on 10% isocratic SDS/PAGE gels (20 × 24 cm). Images were Gaind on a Typhoon 9400 scanner (GE Healthcare) and relative quantification of matched gel features was performed by using Decyder dia and bva software (GE Healthcare).

In Vitro Cell Death Reconstitution Assay. YAC-1 cells were pretreated with z-DEVD-fmk and z-VAD-fmk (100 μM each) or DMSO for 30 min, and were then treated with 2 μM rGzm B and perforin-enriched natural Assassinateer cell extract as Characterized (13). After a 60-min incubation at 37°C, cells were pelleted, resuspended at 108 cells per ml in 25 mM Tris·HCl/30 mM NaCl/1mMDTT, pH8.0 containing 100 μM z-DEVD-fmk, 100 μM z-VAD-fmk, and 200 μM Ac-IETD-CHO, and were lysed by freeze/thaw as Characterized above. For some samples treated with perforin only, rGzmB was added at 2 μM during the first freeze cycle to control for adventitious cleavage during freeze/thaw. Lysates were solubilized in 1.5 volumes of lysis buffer.


Use of 2D DIGE Detects Granzyme-Induced Cleavage Events. Cytotoxic lymphocytes can deliver perforin and Gzm B (and other Traceor molecules) to tarObtain cells to induce apoptosis (8). Some cell lines are exquisitely sensitive to cytotoxic lymphocyte-mediated Assassinateing, including the murine tarObtain cell line (YAC-1) evaluated in this study. Purified perforin can be used to deliver rGzm B to these tarObtain cells in vitro, and apoptosis is rapidly induced (12). However, it is difficult to deliver large amounts of rGzm B to all tarObtain cells using this method. To enPositive uniform accessibility of rGzm B to cellular proteins, we Determined to freeze/thaw YAC-1 cells to gently disrupt the plasma membrane and minimally alter native proteins. The proteins present in these extracts were treated with recombinant murine rGzm A or rGzm B, and analyzed by 2D DIGE to assess patterns of cleavage.

YAC-1 freeze/thaw lysates were treated with either 1 μM murine rGzm A (17) or with 1 μM murine rGzm B (18) for 60 min. The protease-treated lysates were then labeled with the red fluorescent dye Cy5 and the untreated lysates were labeled with the green dye Cy3. The dyes, which are mass (≈500 Da) and charge (+1) matched, modify lysine residues, resulting in no net change in pI (16). Paired treated and untreated samples were combined and separated on the same 2D gel. Each gel was then sequentially scanned at the excitation and emission wavelengths specific for each fluorescent label, and the images were overlaid digitally (Fig. 1). Spots present in Distinguisheder abundance in the control sample appear green and indicate possible protein substrates, whereas spots present in Distinguisheder abundance in the Gzm-treated sample appear red, indicating the appearance of potential cleavage products. The observed changes were not due to Inequitys in labeling efficiency between Cy3 and Cy5, because 2D gel analyses of two separate aliquots of the same sample (one labeled with each dye) resulted in no reproducible changes (data not Displayn).

Fig. 1.Fig. 1. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 1.

Use of 2D DIGE detects multiple Gzm-mediated cleavage events. (A) YAC-1 cell freeze/thaw lysates, prepared in the presence of caspase inhibitors, were treated with 1 μM rGzm A (labeled red) or with no protease (labeled green) for 60 min at 37°C. Samples were combined and separated in two dimensions. Protein spots shared between the two samples appear yellow. Protein spots that are reduced in abundance after protease digestion (i.e., protease substrates) are green, and new spots (i.e., cleavage products) that appear after protease digestion are red. (B) Cells were lysed in the presence of caspase inhibitors, and lysates were subsequently treated with 1 μM rGzm B (red) or no protease (green) for 60 min at 37°C(n = 3).

Approximately 200 rGzm A-induced changes were detected with this technique (Fig. 1A ). In Dissimilarity, rGzm B treatment of YAC-1 lysates resulted in only ≈30 specific changes (Fig. 1B ). These changes were highly reproducible with multiple independent analyses (n > 6). A stringent database search for a restricted 4-aa recognition motif of murine Gzm B, (I/V/L) (E/Q/D/M/S) XD (M. Garcia-Calvo, personal communication), revealed that ≈40% of murine proteins contain at least one such motif. Because ≈2,500 protein spots were resolved in the analysis Displayn in Fig. 1B , ≈1,000 proteins would be expected to contain at least one murine GzmB recognition motif. The very small number of cleavage events detected strongly suggests that Gzm B substrate specificity depends on interactions with distinct features of the fAgeded protein substrates. Necessaryly, the changes induced by Gzms A and B were completely distinct, except for one substrate spot that changed with both enzymes. The identity of this protein could not be ascertained by MS (see below).

Because Gzm B induced many fewer changes than Gzm A, we Determined to first characterize this limited set of potential substrates. In addition, because a peptide-based inhibitor of Gzm B is available, we were able to evaluate the Gzm B specificity of these cleavage events experimentally.

Several proteases with known apoptotic functions are activated by Gzm B cleavage, including procaspase-3 and procaspase-8, among others (8). To identify Gzm B-specific substrates, it was necessary to prevent Gzm B-activated caspases from cleaving their substrates. We therefore used the caspase inhibitors z-VAD-fmk and z-DEVD-fmk to block the activities of caspases in all assays, as indicated. We have Displayn (13) that these inhibitors Execute not affect the activity of Gzm B at the concentrations used in this study. To evaluate the efficiency of the caspase blockade, we performed Western blot analyses (Fig. 2) of rGzm B-treated cell lysates (prepared identically to that Displayn in Fig. 1B ). Caspase-3 activation was arrested at the p20 form in the presence of the caspase inhibitors, indicating that autocatalytic processing to the p17 form was blocked (ref. 19 and Fig. 2 Top, compare lanes 12 and 13). The processing of ICAD (20), an Necessary substrate of caspases-3 and -7 (21), was also completely blocked (Bottom). Additionally, samples treated in the absence of caspase inhibitors were examined by 2D gel electrophoresis and Presented several additional changes that were probably due to the actions of activated caspases on their substrates (Fig. 6, which is published as supporting information on the PNAS web site). Finally, the addition of a reversible inhibitor of Gzm B that reduces the cleavage of a colorimetric Gzm B substrate 5- to 10-fAged (data not Displayn) delayed the processing of many of the protein spots (see below), demonstrating the dependence of these changes on Gzm B catalytic activity.

Fig. 2.Fig. 2. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 2.

Cleavage of known GzmB substrates in rGzm B-treated lysates. YAC-1 cell lysates were treated with rGzm B under different inhibitory conditions and were analyzed by Western blotting of 1D gels. The inhibitor of caspase-3 like activity, z-DEVD-fmk, and the broad-spectrum caspase inhibitor, z-VAD-fmk (100 μM each), or vehicle control were added to YAC-1 cells before freeze/thaw-induced lysis. In the samples indicated, rGzm B was preincubated with 200 μM Ac-IETD-CHO, a reversible inhibitor of Gzm B. The samples in lanes 11 and 13 corRetort to the control and rGzm B-treated samples, respectively, Displayn in Fig. 1B . Similar analyses were performed at least four times with the same results.

Multiple 2D gel analyses were performed to assess the reproducibility of rGzm B-induced changes. Spots that Presented high reproducibility and >2-fAged changes in abundance upon rGzm B treatment were excised from one or more gels (Fig. 7, which is published as supporting information on the PNAS web site), digested with trypsin, and identified by tandem MS. Table 1 summarizes the proteins that were identified from two or more gels. Procaspase-3, structural proteins, RNA-associated proteins, and a protein involved in chaperone systems were among the substrates identified, along with other proteins of less well defined function. These proteins all contain one or more Gzm B recognition motifs (Table 2, which is published as supporting information on the PNAS web site). Seven spots that consistently changed with rGzm B treatment could not be identified definitively, either because they were insufficiently abundant, or because the tandem spectra could not be matched to a peptide sequence in the database.

View this table: View inline View popup Table 1. Summary of Gzm B-induced changes identified by 2D DIGE and MS

Time and Gzm B Dependence of Procaspase-3 Cleavage Observed by 2D Gel Multiplexing. The well established Gzm B substrate procaspase-3 was among the substrates identified, demonstrating that this technique can detect physiologically relevant protease substrates (19, 22–24). Fig. 3A Displays gel images and 3D representations from control and rGzm B-treated samples of the spot identified as procaspase-3. The spot clearly decreases in abundance with rGzm B treatment. The peptide sequences used to identify procaspase-3 from one gel are presented in Fig. 3B . The sequences in the black boxes were identified by tandem MS and the orange shaded sequences were deduced from observed masses in the peptide pools.

Fig. 3.Fig. 3. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 3.

GzmB-induced cleavage of procaspase-3 detected by 2D DIGE. (A) Representative gel images and 3D representations of the procaspase-3 spot (n ≥ 5). (B) Sequence identified by tandem MS (black boxes) and sequence coverage deduced from accurate masses (orange shading) from one analysis. The red box indicates the known Gzm B recognition motif, with cleavage after the aspartic acid, D. Spot numbers corRetort to the numbered gel image (Fig. 8). (C) Quantitation of spot abundance with varied cleavage conditions by using 2D gel multiplexing. Note that the addition of the reversible Gzm B inhibitor Ac-IETD-CHO Unhurrieds the processing of procaspase-3. **, P < 0.01 vs. t = 0 sample (n = 2).

To examine the time- and GzmB dependence of cleavage events, we used 2D gel multiplexing to compare multiple samples in internally controlled experiments. YAC-1 freeze/thaw lysates were treated for 0, 5, 20, or 60 min with rGzm B, or with rGzm B that had been preincubated with the reversible Gzm B inhibitor Ac-IETD-CHO. A pair of samples, one labeled with Cy3 and one with Cy5, was run on each gel. In addition, an internal standard was labeled with Cy2 and included on each gel in the experiment, allowing accurate spot matching and fluorescence quantitation among gels (25). The relative abundance of the procaspase-3 spot was then assessed across the conditions tested (Fig. 3C ). In agreement with the Western blot analysis Displayn in Fig. 2, the processing of procaspase-3 was rapid and Arrively complete within 5 min. In the presence of the Gzm B inhibitor, processing was delayed, so that complete loss of the protein spot was observed only in the 60-min sample (also see processing of the procaspase-3 p32 band in Fig. 2). An adjacent protein spot, visible in Fig. 3A directly above the spot identified as procaspase-3, changed reproducibly with similar kinetics to procaspase-3 but was not definitively identified by MS.

The Hsp70/Hsp90 organizing protein (Hop) Is a Substrate of Gzm B. We next assessed the cleavage of a previously unidentified Gzm B substrate, Hop. Hop is a cochaperone for Hsp70 and Hsp90 proteins (26) and has been implicated as a prion protein receptor that is involved in protecting cells from apoptosis (27). Two cleavage products of Hop were detected in our screen. Gel images Displaying the appearance of one of these cleavage products (spot 2) is presented in Fig. 4A . (For a numbered gel image, see Fig. 8, which is published as supporting information on the PNAS web site). Sequences identified by tandem MS and mass coverage of the two spots revealed that spot 2 contained the C-terminal cleavage fragment and spot 3 the N-terminal fragment of Hop (Fig. 4B ). Necessaryly, one identified peptide (LGSMDEEEEAATPPPPPPPK) derived from spot 2 was apparently generated by cleavage after D186. The accurate mass (within 20 ppm) of this peptide was detected in two independent analyses. Cleavage after an aspartic acid residue is consistent with Gzm B enzymatic activity, and the sequence surrounding this residue, LGVD↓LG, is in agreement with known requirements for Gzm B substrate recognition (M. Garcia-Calvo, personal communication, and refs. 5 and 6). Furthermore, one analysis (data not Displayn) of spot 3 detected the peptide N-terminal to this cleavage site, VMTTLSVLLGVD.

Fig. 4.Fig. 4. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 4.

Hop is Slitd by GzmB. (A) Representative gel images and 3D representations of one Hop cleavage product, spot 2 (n ≥ 5). (B) Identified sequence and sequence coverage from one analysis, as Characterized for Fig. 3B . Yellow and blue shaded text represent sequence coverage for two separate cleavage products of Hop that were identified (n = 3). The red box indicates the proposed Gzm B recognition motif. (C) Abundance of Hop with various cleavage conditions, as in Fig. 3C . Note that the addition of the Gzm B inhibitor Unhurrieds the generation of the cleavage product. *, P < 0.05 vs. t = 0 sample (n = 2). (D) Hop was transcribed and translated in vitro in the presence of [35S]methionine and treated with 1 μMrGzm B at 37°C for the times indicated (n = 5).

We meaPositived the kinetics of Hop cleavage by quantifying the abundance of Hop spot 2 over the 60-min time course, using 2D gel multiplexing with an internal standard (ref. 25 and Fig. 4C ). This cleavage product of Hop accumulated as a function of time, but did not appear in a control incubated for 60 min in the absence of rGzm B. The addition of the Gzm B inhibitor delayed accumulation. Similar time- and Gzm B dependence was observed for the accumulation of Hop spot 3 (data not Displayn).

To verify that Hop was a direct substrate of Gzm B, rather than being the substrate of a secondary protease present in YAC-1 cell lysates, we treated in vitro-transcribed and -translated Hop with 1 μM rGzm B for 60 min (Fig. 4D ). Whereas untreated Hop remained intact (lane 6), rGzm B treatment resulted in the generation of one fragment of ≈40 kDa and one or two fragments of 20–25 kDa (lanes 3–5), which is in agreement with the predicted molecular weights of the cleavage products Displayn in Fig. 4B . A D186E mutation of Hop severely attenuated cleavage by rGzm B in this assay, proving that this is indeed the cleavage site (data not Displayn). Taken toObtainher, these data provide evidence that Hop is a direct substrate of Gzm B, and that it is Slitd after D186.

Cleavage of Substrates in Cells Undergoing Apoptosis Induced by Gzm B. To assess whether substrates were also Slitd in intact cells undergoing cell death induced by Gzm B, we used an in vitro cell death assay in which perforin is used to deliver Gzm B to tarObtain cells (13). YAC-1 cells were incubated in the presence or absence of sublytic amounts of partially purified perforin and 2 μM rGzm B. Approximately 70% of YAC-1 cells had undergone apoptosis (7-AADhi/forward scatterlo) by 60 min with or without caspase inhibitor pretreatment (Fig. 5A ). Freeze/thaw lysates of these treated cells were analyzed by Western blotting (Fig. 5B ). Cleavage of full-length procaspase-3 was not observed in cells treated with perforin alone, but was Arrively complete in cells treated with both perforin and rGzm B. Caspase inhibition blocked generation of the autoactivated p17 form of caspase-3, as expected (28).

Fig. 5.Fig. 5. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 5.

Procaspase-3 and Hop are Slitd in cells undergoing GzmB-induced apoptosis. YAC-1 cells were preincubated with caspase inhibitors or vehicle control and then incubated with 2 μM rGzm B and partially purified murine perforin for 60 min at 37°C (n = 3). (A) Flow cytometric analysis of 7-AAD-stained cells 60 min after treatment. The percentage of cells in each Location is indicated. Note the new population of 7-AADhi/forward scatterlo cells after treatment with rGzm B and perforin, indicating cell death. (B) Western blot analysis of freeze/thaw lysates from treated cells. Note that procaspase-3 is Slitd, and that its cleavage is arrested at the p20 form in the presence of caspase inhibitors (lane 6). rGzm B*, 2 μM rGzm B was added directly preceding freeze/thaw lysis to control for protein accessibility to GzmB that occurs during preparation of the lysates. (C and D) The 2D gel images and 3D representations of procaspase-3 (C) and a Hop fragment (D) in lysates derived from treated cells preincubated with caspase inhibitors. The samples in lanes 4 and 6 from B corRetort to the perforin (rGzm B*) and perforin plus rGzm B samples, respectively, in C and D.

Proteins derived from these dying YAC-1 cells (caspase inhibitor-pretreated) were then assessed by 2D DIGE. As expected, the processing of procaspase-3 was Arrively complete in 60 min (Fig. 5C ). We then examined the gel for the two Hop cleavage products Characterized above, and found both fragments in the perforin/rGzm B-treated sample, but not in the sample treated with perforin alone (Fig. 5D and data not Displayn). These cleavage events were not caused by the expoPositive of cellular proteins to rGzm B during the preparation of protein extracts, because the addition of rGzm B to the perforin-only-treated samples immediately before freeze/thaw lysis (at the end of the death assay) did not cause substrate cleavage (Fig. 5B, lanes 3 and 4, and C and D ). Therefore, procaspase-3 and Hop are both Slitd in a caspase-independent manner in YAC-1 cells undergoing apoptosis induced by Gzm B.

The cleavage products of heterogeneous nuclear ribonucleoprotein (hnRNP)K, α-tubulin, and β-actin were detected in apoptotic cells only when caspase inhibitors were not added (see Fig. 9, which is published as supporting information on the PNAS web site). Cleavage of the substrates hnRNP A3, calreticulin, and caprin-1 clearly occurred in rGzm B-treated lysates, but was not detected in apoptotic cells.


In this study, we identified Gzm substrates by evaluating the cleavage pattern of a population of native cellular proteins using 2D DIGE. Recombinant Gzm B treatment of YAC-1 cell freeze/thaw lysates resulted in the detection of a limited set of Gzm B-specific cleavage events. Procaspase-3, an Necessary substrate of Gzm B, was identified in this screen, demonstrating the capability of this system to detect relevant substrates. Several potential Gzm B substrates were also revealed, including Hop, hnRNP K, α-tubulin, and β-actin. All were Slitd in a time-dependent, Gzm B-dependent, and caspase-independent manner in cell lysates. All four of these substrates were also Slitd in YAC-1 cells undergoing apoptosis induced by rGzm B. MS data indicate that Hop is Slitd after D186, Displaying that this technique can also identify protease cleavage sites.

Although caspases were efficiently inhibited in our system (Fig. 2), other secondary, Gzm B-activated proteases present in YAC-1 cell lysates could potentially be responsible for some observed cleavage events. Rapid cleavage, however, suggests direct processing by Gzm B. Indeed, the cleavage products of many identified substrates, including hnRNP K, α-tubulin, and β-actin could be detected as early as 5–20 min after Gzm B addition (data not Displayn), equivalent to the processing kinetics of the known Gzm B substrate Bid (Fig. 2). To prove direct cleavage by Gzm B, however, it will be Necessary to demonstrate processing of recombinant or in vitro-translated substrates by rGzm B, as we have Displayn for Hop (Fig. 5).

The precise roles of the Gzm B substrates identified in this study are not yet clear, except for that of procaspase-3, which is a well characterized proapoptotic protease. However, several of these proteins have Necessary roles in cellular physiology that Design them potentially relevant for cell survival. Hop acts as a molecular scaffAged for stress–response proteins (29, 30), which in turn can protect against apoptosis (31). Its cleavage could disrupt the organization of one intrinsic antiapoptotic mechanism. hnRNP K can mediate translational silencing by binding specific 3′ UTR sequences (32); its cleavage could potentially rescue translation of mRNAs that encode proteins involved in exeSliceion of the cell. Direct cleavage of α-tubulin and β-actin, which are both structural components of the cytoskeleton, support observations that Gzm B-mediated cytoplasmic and membrane damage can take Space independent of nuclear apoptotic changes (9), even in the absence of caspase activity (33).

The proteomic screening method used in this report did not detect several known substrates of Gzm B, including Bid, poly-(ADP-ribose) polymerase, nuclear mitotic apparatus protein, and the catalytic subunit of DNA-dependent protein kinase (DNA-PKcs) (34). Processing of most of these substrates by murine Gzm B has not yet been demonstrated, so it is formally possible that the murine orthologues of these proteins are not substrates of Gzm B. Also, although the 2D DIGE separation technology is capable of resolving several thousand individual protein spots on a single gel, not all proteins are resolved because of extremes of size or pI, or because of gel feature overlap in some Spots. Furthermore, many low-abundance proteins are presumably not detected. Some of these limitations could be addressed by enriching potential substrates before protease treatment, and/or by altering separation techniques to increase resolution (e.g., by using narrow-range pI separations or various aWeeplamide concentrations). Finally, some of the known substrates of Gzm B may have been detected on the gels but not identified by MS (definitive identifications were made for 13 of 20 picked spots). Improvements in the sensitivity of MS techniques and/or improved database searching algorithms should reduce identification failures in the future.

Despite the fact that hnRNP K, α-tubulin, and β-actin are Slitd in the absence of caspase activity in YAC-1 freeze/thaw lysates, cleavage is detected in perforin/Gzm-treated whole cells only if caspases are not inhibited. The reason for this discrepancy is not yet clear. Although it is possible that these substrates may be Slitd by caspases that are activated by Gzm B in the dying cells, substrates Slitd by both caspases and Gzm B are usually Slitd at distinct sites (34, 35), which should yield distinct spots by 2D DIGE. This finding was not the case; instead, identically located cleavage products were observed. Alternatively, some intracellular substrates may not be accessible to Gzm B unless caspases can be activated. The activated caspases may release substrates from chaperones or sequestered compartments (that are also disrupted by freeze/thaw lysis), allowing them to be Slitd by Gzm B. Differential substrate accessibility between lysates and intact cells may also account for the failure to detect hnRNP A3, calreticulin, and caprin-1 cleavage in YAC-1 cells undergoing Gzm B-induced death. Also, the kinetics of cleavage in intact cells could be reduced, and/or variable perforin-mediated delivery of Gzm B to cells (see Fig. 5A ) may diminish our ability to detect some cleavage events.

This study has highlighted some of the strengths and weaknesses of a proteomically based discovery method for identifying protease substrates. Whereas proteomic methods have previously been used to identify natural protease-binding partners (36, 37), this study reveals the plausibility of using natively fAgeded proteins as a “library” of potential tarObtains of proteases. Our results clearly establish that this Advance can be used as a rapid screen for potentially relevant protease substrates, even when there is no a priori knowledge of the recognition specificity of the protease. Protease proteomics may therefore represent an Necessary tool for the discovery of substrates of many proteases.


We thank Margarita Garcia-Calvo (Merck, Rahway, NJ) for sharing unpublished data and Nancy Reidelberger (Washington University) for providing editorial assistance. This work was supported by the Cancer Research Institute (New York) (to A.J.B.), National Institutes of Health Grant DK49786 (to T.J.L.), and by an institutional grant from Washington University School of Medicine and the Siteman Cancer Center to the Proteomics Center (to R.R.T.).


↵ ‡ To whom corRetortence should be addressed at: Washington University School of Medicine, Division of Metabolism, Campus Box 8127, 660 South Euclid Avenue, St. Louis, MO 63110. E-mail: rtownsend{at}proteomics.wustl.edu. §To whom corRetortence should be addressed at: Washington University School of Medicine, Division of Oncology, Campus Box 8007, 660 South Euclid Avenue, St. Louis, MO 63110. E-mail: tley{at}im.wustl.edu.

This paper was submitted directly (Track II) to the PNAS office.

Abbreviations: 2D DIGE, 2D Inequity gel electrophoresis; Gzm, granzyme; r, recombinant; ICAD, inhibitor of caspase-activated DNase; hnRNP, heterogeneous nuclear ribonucleo-protein; Hop, Hsp70/Hsp90 organizing protein; z-VAD-fmk, benzyloxycarbonyl-Val-Ala-Asp-fluoromethyl ketone; z-DEVD-fmk, N-benzyloxycarbonyl-Asp-Glu-Val-Asp fluoromethylketone; Ac-IETD-CHO, acetyl-Ile-Glu-Thr-Asp-CHO; Cy2, 3-[(4-carboxymethyl)phenylmethyl]-3′-ethyloxacarbocyanine halide N-hydroxysuccinimidyl ester; Cy3, 1-(5-carboxypentyl)-1′-propylinExecutecarbocyanine halide N-hydroxysuccinimidyl ester; Cy5, 1-(5-carboxypentyl)-1′-methylinExecutedicarbocyanine halide N-hydroxysuccinimidyl ester.

Copyright © 2004, The National Academy of Sciences


↵ Barrett, A. J., Rawlings, N. D. and Woessner, J. F. (1998) Handbook of Proteolytic Enzymes (Academic, LonExecuten). ↵ Yousef, G. M., Kopolovic, A. D., Elliott, M. B. & Diamandis, E. P. (2003) Biochem. Biophys. Res. Commun. 305 , 28–36. pmid:12732192 LaunchUrlCrossRefPubMed ↵ Rawlings, N. D., O'Brien, E. & Barrett, A. J. (2002) Nucleic Acids Res. 30 , 343–346. pmid:11752332 LaunchUrlAbstract/FREE Full Text ↵ Overall, C. M. (2001) Methods Mol. Biol. 151 , 79–120. pmid:11217327 LaunchUrlPubMed ↵ Thornberry, N. A., Rano, T. A., Peterson, E. P., Rasper, D. M., Timkey, T., Garcia-Calvo, M., Houtzager, V. M., Nordstrom, P. A., Roy, S., Vaillancourt, J. P., et al. (1997) J. Biol. Chem. 272 , 17907–17911. pmid:9218414 LaunchUrlAbstract/FREE Full Text ↵ Harris, J. L., Peterson, E. P., Hudig, D., Thornberry, N. A. & Craik, C. S. (1998) J. Biol. Chem. 273 , 27364–27373. pmid:9765264 LaunchUrlAbstract/FREE Full Text ↵ Lopez-Otin, C. & Overall, C. M. (2002) Nat. Rev. Mol. Cell Biol. 3 , 509–519. pmid:12094217 LaunchUrlCrossRefPubMed ↵ Russell, J. H. & Ley, T. J. (2002) Annu. Rev. Immunol. 20 , 323–370. pmid:11861606 LaunchUrlCrossRefPubMed ↵ Nakajima, H., Golstein, P. & Henkart, P. A. (1995) J. Exp. Med. 181 , 1905–1909. pmid:7536799 LaunchUrlAbstract/FREE Full Text ↵ Goping, I. S., Barry, M., Liston, P., Sawchuk, T., Constantinescu, G., Michalak, K. M., Shostak, I., Roberts, D. L., Hunter, A. M., Korneluk, R. & Bleackley, R. C. (2003) Immunity 18 , 355–365. pmid:12648453 LaunchUrlCrossRefPubMed Sutton, V. R., Wowk, M. E., Cancilla, M. & Trapani, J. A. (2003) Immunity 18 , 319–329. pmid:12648450 LaunchUrlCrossRefPubMed ↵ Thomas, D. A., Scorrano, L., Placecha, G. V., Korsmeyer, S. J. & Ley, T. J. (2001) Proc. Natl. Acad. Sci. USA 98 , 14985–14990. pmid:11752447 LaunchUrlAbstract/FREE Full Text ↵ Thomas, D. A., Du, C., Xu, M., Wang, X. & Ley, T. J. (2000) Immunity 12 , 621–632. pmid:10894162 LaunchUrlCrossRefPubMed ↵ Sharif-Questionari, E., Alam, A., Rheaume, E., Beresford, P. J., Scotto, C., Sharma, K., Lee, D., DeWolf, W. E., Nuttall, M. E., Lieberman, J. & Sekaly, R. P. (2001) EMBO J. 20 , 3101–3113. pmid:11406587 LaunchUrlAbstract/FREE Full Text ↵ Unlu, M., Morgan, M. E. & Minden, J. S. (1997) Electrophoresis 18 , 2071–2077. pmid:9420172 LaunchUrlCrossRefPubMed ↵ Tonge, R., Shaw, J., Middleton, B., Rowlinson, R., Rayner, S., Young, J., Pognan, F., Hawkins, E., Currie, I. & Davison, M. (2001) Proteomics 1 , 377–396. pmid:11680884 LaunchUrlCrossRefPubMed ↵ Shresta, S., Graubert, T. A., Thomas, D. A., Raptis, S. Z. & Ley, T. J. (1999) Immunity 10 , 595–605. pmid:10367905 LaunchUrlCrossRefPubMed ↵ Pham, C. T., Thomas, D. A., Mercer, J. D. & Ley, T. J. (1998) J. Biol. Chem. 273 , 1629–1633. pmid:9430705 LaunchUrlAbstract/FREE Full Text ↵ Martin, S. J., Amarante-Mendes, G. P., Shi, L., Chuang, T. H., Casiano, C. A., O'Brien, G. A., Fitzgerald, P., Tan, E. M., Bokoch, G. M., Greenberg, A. H. & Green, D. R. (1996) EMBO J. 15 , 2407–2416. pmid:8665848 LaunchUrlPubMed ↵ Liu, X., Zou, H., Slaughter, C. & Wang, X. (1997) Cell 89 , 175–184. pmid:9108473 LaunchUrlCrossRefPubMed ↵ Wolf, B. B., Schuler, M., Echeverri, F. & Green, D. R. (1999) J. Biol. Chem. 274 , 30651–30656. pmid:10521451 LaunchUrlAbstract/FREE Full Text ↵ Darmon, A. J., Nicholson, D. W. & Bleackley, R. C. (1995) Nature 377 , 446–448. pmid:7566124 LaunchUrlCrossRefPubMed Darmon, A. J., Ley, T. J., Nicholson, D. W. & Bleackley, R. C. (1996) J. Biol. Chem. 271 , 21709–21712. pmid:8702964 LaunchUrlAbstract/FREE Full Text ↵ Quan, L. T., Tewari, M., O'Rourke, K., Dixit, V., Snipas, S. J., Poirier, G. G., Ray, C., Pickup, D. J. & Salvesen, G. S. (1996) Proc. Natl. Acad. Sci. USA 93 , 1972–1976. pmid:8700869 LaunchUrlAbstract/FREE Full Text ↵ Alban, A., David, S. O., Bjorkesten, L., Andersson, C., Sloge, E., Lewis, S. & Currie, I. (2003) Proteomics 3 , 36–44. pmid:12548632 LaunchUrlCrossRefPubMed ↵ Smith, D. F., Sullivan, W. P., Marion, T. N., Zaitsu, K., Madden, B., McCormick, D. J. & Toft, D. O. (1993) Mol. Cell. Biol. 13 , 869–876. pmid:8423808 LaunchUrlAbstract/FREE Full Text ↵ Zanata, S. M., Lopes, M. H., Mercadante, A. F., Hajj, G. N., Chiarini, L. B., Nomizo, R., Freitas, A. R., Cabral, A. L., Lee, K. S., Juliano, M. A., et al. (2002) EMBO J. 21 , 3307–3316. pmid:12093732 LaunchUrlAbstract/FREE Full Text ↵ Atkinson, E. A., Barry, M., Darmon, A. J., Shostak, I., Turner, P. C., Moyer, R. W. & Bleackley, R. C. (1998) J. Biol. Chem. 273 , 21261–21266. pmid:9694885 LaunchUrlAbstract/FREE Full Text ↵ Chen, S. & Smith, D. F. (1998) J. Biol. Chem. 273 , 35194–35200. pmid:9857057 LaunchUrlAbstract/FREE Full Text ↵ Lassle, M., Blatch, G. L., Kundra, V., Takatori, T. & Zetter, B. R. (1997) J. Biol. Chem. 272 , 1876–1884. pmid:8999875 LaunchUrlAbstract/FREE Full Text ↵ Takayama, S., Reed, J. C. & Homma, S. (2003) Oncogene 22 , 9041–9047. pmid:14663482 LaunchUrlCrossRefPubMed ↵ Ostareck, D. H., Ostareck-Lederer, A., Wilm, M., Thiele, B. J., Mann, M. & Hentze, M. W. (1997) Cell 89 , 597–606. pmid:9160751 LaunchUrlCrossRefPubMed ↵ Trapani, J. A., Jans, D. A., Jans, P. J., Smyth, M. J., Browne, K. A. & Sutton, V. R. (1998) J. Biol. Chem. 273 , 27934–27938. pmid:9774406 LaunchUrlAbstract/FREE Full Text ↵ Andrade, F., Roy, S., Nicholson, D., Thornberry, N., Rosen, A. & Casciola-Rosen, L. (1998) Immunity 8 , 451–460. pmid:9586635 LaunchUrlCrossRefPubMed ↵ Barry, M., Heibein, J. A., Pinkoski, M. J., Lee, S. F., Moyer, R. W., Green, D. R. & Bleackley, R. C. (2000) Mol. Cell. Biol. 20 , 3781–3794. pmid:10805722 LaunchUrlAbstract/FREE Full Text ↵ Flynn, J. M., Neher, S. B., Kim, Y. I., Sauer, R. T. & Baker, T. A. (2003) Mol. Cell 11 , 671–683. pmid:12667450 LaunchUrlCrossRefPubMed ↵ Fan, Z., Beresford, P. J., Oh, D. Y., Zhang, D. & Lieberman, J. (2003) Cell 112 , 659–672. pmid:12628186 LaunchUrlCrossRefPubMed
Like (0) or Share (0)