A Competitive Chemical-Proteomic Platform To Identify Zinc-Binding

Oct 10, 2013 - Department of Chemistry, Boston College, Chestnut Hill, Massachusetts 02467, United States. ACS Chem. Biol. , 2014, 9 (1), pp 258–265...
0 downloads 0 Views 2MB Size
Download PDF
Articles pubs.acs.org/acschemicalbiology

A Competitive Chemical-Proteomic Platform To Identify Zinc-Binding Cysteines Nicholas J. Pace and Eranthie Weerapana* Department of Chemistry, Boston College, Chestnut Hill, Massachusetts 02467, United States S Supporting Information *

ABSTRACT: Zinc ions (Zn2+) play vital catalytic, structural, and regulatory roles in protein function and are commonly chelated to cysteine residues within the protein framework. Current methods to identify Zn2+-binding cysteines rely on computational studies based on known Zn2+-chelating motifs, as well as high-resolution structural data. These available approaches preclude the global identification of putative Zn2+chelating cysteines, particularly on poorly characterized proteins in the proteome. Herein, we describe an experimental platform that identifies metal-binding cysteines on the basis of their reduced nucleophilicity upon treatment with metal ions. As validation of our platform, we utilize a peptide-based cysteine-reactive probe to show that the known Zn2+-chelating cysteine in sorbitol dehydrogenase (SORD) demonstrates an expected loss in nucleophilicity in the presence of Zn2+ ions and a gain in nucleophilicity upon treatment with a Zn2+ chelator. We also identified the active-site cysteine in glutathione S-transferase omega-1 (GSTO1) as a potential Zn2+-chelation site, albeit with lower metal affinity relative to SORD. Treatment of recombinant GSTO1 with Zn2+ ions results in a dose-dependent decrease in GSTO1 activity. Furthermore, we apply a promiscuous cysteine-reactive probe to globally identify putative Zn2+-binding cysteines across ∼900 cysteines in the human proteome. This proteomic study identified several well-characterized Zn2+-binding proteins, as well as numerous uncharacterized proteins from functionally distinct classes. This platform is highly versatile and provides an experimental tool that complements existing computational and structural methods to identify metal-binding cysteine residues.

A

enzymes has been well-demonstrated and shown to play important roles in regulating protein activities such as caspases, protein tyrosine phosphatases, and aldehyde dehydrogenases.10,11 Furthermore, differentiation between proteinbound Zn2+ and other metal ions remains challenging, thereby complicating the identification of the physiologically relevant metal species using these methods. Due to the limitations of existing structural and computational methods, new experimental platforms are needed to identify putative Zn2+-binding cysteines in the proteome. Herein, we describe a chemical proteomic strategy for identifying stable and transient Zn2+-binding cysteines within the context of a human cellular proteome. Our strategy is based on the hypothesis that Zn 2+ -binding will reduce the nucleophilicity of a cysteine and this loss in reactivity can be monitored using cysteine-reactive electrophilic probes. Utilizing cysteine probes specific to a subset of reactive cysteines or a more promiscuous probe that more globally interrogates cysteine reactivities, we can identify those residues that are particularly sensitive to either Zn2+ ions or a metal chelator. These reactivity changes can be visualized by gel-based studies and quantified using high-resolution mass spectrometry so as to identify putative Zn2+-binding cysteines in the proteome.

lthough cysteine is one of the least abundant amino acids incorporated into proteins, it is often found to concentrate at functional loci within diverse protein classes.1,2 Cysteine residues are known to facilitate a broad array of functions, acting as catalytic nucleophiles, redox-active sites, structural disulfides, and regulatory switches.3 Moreover, cysteine is one of the most commonly observed metal-binding residues, forming complexes with iron (Fe2+/3+), zinc (Zn2+), cadmium (Cd2+), and copper (Cu+). Zn2+ ions are essential for the function of many proteins, with recent studies estimating that Zn2+-binding proteins account for 10% of the entire human proteome.4 Zn2+-cysteine complexes participate in a variety of functional roles that are essential for catalysis, structural stability, and regulation across diverse protein classes.5,6 Current methods to identify Zn2+-binding proteins comprise a combination of experimental approaches, including structural genomics, and theoretical approaches, including homology searches of sequence databases.7,8 Despite advances in both of these fields, they frequently encounter a number of limitations. The structure-based methods are reliant on a high resolution structure being available for the protein of interest or a close homologue and are generally unreliable for identifying residues that transiently or weakly bind Zn2+ ions.9 Similarly, computational approaches are well-suited for identifying Zn2+-finger type motifs with defined sequence conservation; however, these approaches are ineffective at predicting transient, regulatory Zn2+-binding sites for which defining structural features are unknown.10 Such transient binding of Zn2+ ions to a variety of © XXXX American Chemical Society

Received: August 21, 2013 Accepted: October 10, 2013

A

dx.doi.org/10.1021/cb400622q | ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology

■

Articles

RESULTS AND DISCUSSION Cysteine-Reactive Probes Can Identify Zn2+-Binding Residues. Functionally important cysteine residues, including metal-binding cysteines, typically exhibit hyper-reactivity12 and are therefore susceptible to covalent modification by small molecules containing electrophiles such as haloacetamides, Michael acceptors, and sulfonate esters.13 Since direct chelation to a metal ion will quench the nucleophilicity of cysteines, we hypothesized that metal-binding cysteines would show a decreased reactivity with a thiol-specific chemical probe upon pretreatment with metal ions (Figure 1A). We have utilized a

human proteome as identified by LC/LC−MS/MS studies (Supplementary Table S2). The alkyne moiety in both IAalkyne and NJP14 serves as a click-chemistry handle to append either a fluorophore for visualization by in-gel fluorescence or a biotin for enrichment and identification by mass spectrometry.16 Initially, we evaluated both IA-alkyne and NJP14 using an in-gel fluorescence-based screen to detect cysteine reactivities mitigated in the presence of metal ions. Soluble HeLa cell lysates were treated with a panel of biologically relevant divalent metals: Zn2+, Ca2+, Mg2+, and Mn2+. After pretreatment with the metals, lysates were subjected to labeling by either IAalkyne or NJP14, fractionated on a size exclusion column to eliminate any unbound metal or probe, and visualized by in-gel fluorescence after incorporation of rhodamine-azide (Rh−N3) using click chemistry (Supplementary Figure S1, and Figure 1C).17 Although gel-based analysis of the highly reactive IAalkyne revealed observable decreases in the intensity of specific labeled bands upon Zn2+ treatment (Supplementary Figure S1), these changes were significantly more pronounced in the more selective NJP14 gel (Figure 1C). Two bands in particular (bands A and B in Figure 1C) showed a dramatic loss of labeling in the presence of Zn2+ ions. This observed loss in reactivity was specific to Zn2+ ions, with no apparent change in the presence of other metal ions. To eliminate the possibility that the loss in labeling could be due to Zn2+-induced precipitation of proteins,18 we demonstrated that the loss of probe labeling observed for both bands upon Zn2+ treatment is recovered by subsequent treatment with a Zn2+-chelating agent, ethylenediaminetetraacetic acid (EDTA) (Figure 2A). This supports the notion that the observed decrease in cysteine reactivity can be attributed to reversible binding of Zn2+, as opposed to an irreversible Zn 2+ -induced precipitation. Furthermore, treatment of lysates with only EDTA (no Zn2+) resulted in an observable increase in cysteine reactivity (especially for band A, Figure 2A), suggesting that EDTAmediated removal of prebound metal ions likely results in increased cysteine exposure and enhanced reactivity. This provides further support of our hypothesis that this competitive platform is selecting for endogenous Zn2+-chelating cysteines. To quantitatively compare the relative affinities of the proteins signified by bands A and B for Zn2+ binding, we treated HeLa cell lysates with increasing concentrations of Zn2+ (0−10 μM) for in-gel fluorescence analysis. The fluorescence intensity of the gel-bands at ∼38 kD (band A) and ∼27 kD (band B) were integrated from three trials (Supplementary Figure S2) and plotted to calculate relative EC50 values (for inhibition of NJP14 labeling) as 365 nM for band A and 2.1 μM for band B (Figure 2B). To identify the proteins signified by bands A and B, we utilized streptavidin enrichment followed by mass spectrometry. To achieve this, HeLa lysates were preincubated with Zn2+, Mg2+, or vehicle and subsequently labeled with NJP14. Probelabeled proteins were tagged with biotin-azide using click chemistry, enriched on streptavidin beads, and subjected to onbead tryptic digestion and LC/LC−MS/MS analysis.19 Proteins identified with high spectral counts in the control and Mg2+treated samples with significantly decreased spectral counts in the Zn2+-treated samples (Table 1; Supplementary Table S3) were designated as putative Zn2+-chelating proteins. Spectral counts refer to the total fragmentation spectra for peptides identified from a particular protein and provide a semiquantitative measure of protein abundance across numerous proteomic samples. Our mass spectrometry analysis identified

Figure 1. A competitive proteomic platform to assess cysteine reactivity upon metal-ion treatment. (A) Schematic representation of the in-gel fluorescence platform to visualize Zn2+-binding proteins. (B) The structure of the cysteine-reactive peptide-based probe, NJP14. This probe contains a cysteine-reactive chloroacetamide electrophile for covalent modification and an alkyne group for click chemistry. (C) In-gel fluorescence analysis of HeLa lysates treated with Zn2+, Ca2+, Mg2+, or Mn2+, followed by NJP14 labeling, reveals several proteins that exhibit a loss in fluorescent signal upon incubation with Zn2+. Two bands (A and B) appeared particularly susceptible to Zn2+ treatment.

variety of thiol-reactive chemical probes to covalently modify reactive cysteines, ranging from a highly reactive iodoacetamide-alkyne (IA-alkyne; Supplementary Figure S1), to milder small-molecule and peptide-based probes that are selective for a subset of reactive cysteines in the proteome.12,14,15 To validate the effectiveness of our competitive strategy to identify Zn2+binding cysteines, we utilized the IA-alkyne and a peptide-based probe, NJP14, consisting of a moderately cysteine-reactive chloroacetamide moiety coupled to a Ser-Pro-Pra-Phe-Phe peptide-sequence (Pra: propargyl glycine) (Figure 1B). This pentapeptide sequence was selected from a proteome-wide screen of a peptide-based library,15 since it selectively labeled a distinct subset of reactive cysteines from diverse proteins in the B

dx.doi.org/10.1021/cb400622q | ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology

Articles

Figure 2. Characterization of Zn2+-sensitive cysteines. (A) The effects of Zn2+ (10 μM) and EDTA (1 mM) on bands A and B were evaluated. Both bands showed an increase in labeling upon EDTA treatment and a decrease in labeling upon Zn2+ treatment. The loss in labeling upon Zn2+ treatment was recovered by EDTA addition, thereby demonstrating the reversibility of the Zn2+-induced loss of cysteine reactivity. (B) Proteomes were treated with increasing concentrations of Zn2+, and the fluorescence intensities of bands A and B were integrated to calculate the relative affinity of each enzyme for Zn2+. EC50 values were determined for band A and band B as 365 nM and 2.1 μM, respectively. (C) Mass spectrometry analysis and spectral counting identified SORD and GSTO1 as highly sensitive to Zn2+ treatment and BLMH as insensitive. This data is corroborated by subsequent in-gel fluorescence studies of overexpressed lysates for these proteins that reflect the same pattern as the spectral counts from the mass spectrometry experiment.

Table 1. Cysteines That Demonstrate a Loss in Reactivity upon Treatment with Zn2+ Ionsa spectral counts 2+

protein

molecular weight

cntrl

Zn (10 μM)

Zn2+ (20 μM)

Mg2+ (20 μM)

% change Zn2+ (20 μM)

SORD similar to sorbitol dehydrogenase GSTO1 glutathione S-transferase omega-1 TXNRD1 thioredoxin reductase 1 isoform 3 RRM1 ribonucleoside-diphosphate reductase large subunit RPS3 40S ribosomal protein S3 EEF1A2 elongation factor 1-α 2 TUBB4 tubulin β-4 chain TGM2 isoform 1 of protein-glutamine γ-glutamyltransferase 2

38687 27566 71153 90070 26688 50470 49586 77329

250 324 535 136 40 68 351 53

12 269 86 114 21 45 246 27

3 24 42 26 13 25 171 26

262 267 464 124 38 15 357 63

−98.8 −92.6 −92.1 −80.9 −67.5 −63.2 −51.3 −50.9

The data were sorted by % change in spectral counts when comparing the control sample to the Zn2+ (20 μM)-treated sample. The top hits (largest decrease in spectral counts upon Zn2+ treatment) from our analysis are SORD and GSTO1, which were subjected to further characterization. a

In order to corroborate our mass spectrometry data, we performed gel-based studies to confirm the observed Zn2+sensitive labeling of SORD and GSTO1, as well the Zn2+insensitivity of BLMH labeling. To achieve this, we overexpressed SORD, GSTO1, and BLMH with C-terminal myc/His tags using transient transfection in HEK293T cells. Probe labeling and Zn2+-sensitivity of these three proteins was assessed using the competitive gel-based analysis platform described previously. As predicted by our mass spectrometry data, Zn2+ treatment of SORD and GSTO1 resulted in a significant decrease in probe labeling, whereas Zn2+ treatment

the two proteins visualized by gel as sorbitol dehydrogenase (SORD; band A) and glutathione S-transferase omega 1 (GSTO1; band B). Other proteins such as thioredoxin reductase also displayed diminished NJP14 labeling in the presence of Zn2+ ions, while the majority of proteins, such as bleomycin hydrolase (BLMH), showed no reduction in spectral counts (Figure 2C; Supplementary Table S3). These data indicate that cysteine residues in the proteome demonstrate varied affinities toward Zn2+-binding, with the majority of NJP14 targets showing no change in cysteine reactivity in the presence of Zn2+ ions (Supplementary Figure S3). C

dx.doi.org/10.1021/cb400622q | ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology

Articles

Figure 3. Effect of Zn2+ treatment on SORD and GSTO1. (A) Overexpressed SORD-wildtype (WT), C44A, C179A, and mock lysates were treated with NJP14. The C44A mutant demonstrates no fluorescent signal, indicating that NJP14 covalently modifies C44 in SORD. (B) SORD activity assays of mock and SORD WT, C44A, and C179A mutant overexpressing lysates reveals complete loss of reductive activity for C44A mutant. (C) Overexpressed SORD WT lysates were treated with EDTA (1 mM), Zn2+ (1 mM), and/or NJP14 (100 μM) to study the effects on activity. EDTA completely inhibits SORD, and this inhibition is reversed by Zn2+ addition. NJP14 treatment results in a 50% decrease in SORD activity, and this inhibition is impeded by the addition of Zn2+ and enhanced by EDTA. (D) Recombinant GSTO1 WT showed a dose-dependent decrease in labeling by NJP14 in the presence of Zn2+ as illustrated by in-gel fluorescence analysis. At high Zn2+ concentrations, labeling of the WT protein equals that of the C32A mutant. (E) GSTO1 activity assay showing a dose-dependent decrease in activity of recombinant GSTO1 in the presence of Zn2+.

To fully assess the effects of Zn2+ ions, EDTA, and NJP14 on SORD function, we established SORD activity assays using SORD-overexpressing HEK293T lysates. First we evaluated WT, C44A, and C179A SORD-overexpressing lysates using previously reported oxidative and reductive SORD activity assays.24,25 As expected, both WT and C179A mutant lysates exhibited significant enzymatic activity relative to the mocktransfected lysates, whereas the C44A mutant was inactive (Figure 3B; Supplementary Figure S4). This served to confirm the functional importance of the Zn2+-cysteine complex for both oxidative and reductive catalysis. This was further corroborated by treatment of WT SORD lysates with EDTA, which resulted in complete inactivation of the enzyme (Figure 3C; Supplementary Figure S6). This inactivation was reversible and could be restored by subsequent treatment with Zn2+ ions. This recovery was partial (∼60%) likely due to residual EDTA remaining in the buffer. Unlike EDTA treatment, NJP14 only partially inhibited SORD activity (∼50%) (Figure 3C; Supplementary Figure S5), suggesting that NJP14 can only bind to the Zn2+-free population of SORD and is unable to displace bound Zn2+ ions from the active site. Pretreatment with Zn2+ ions completely abolished the inhibitory effect of NJP14, confirming that saturation of the SORD active site with Zn2+ ions prevents NJP14 binding. Lastly, we pretreated WT SORD-overexpressing lysates with EDTA to remove the catalytic Zn2+ ion, followed by treatment with NJP14 to covalently modify the now accessible C44. These labeled lysates were then assayed for activity upon treatment with Zn2+ ions (Figure 3C; Supplementary Figure S6), revealing complete inhibition of SORD activity. These activity assays serve to verify that NJP14 covalently modifies Cys44 of Zn2+-free SORD and thereby serves to inhibit SORD activity. This inhibition is mitigated by the addition of Zn2+ ions and enhanced by EDTA.

of BLMH showed no effect on cysteine reactivity (Figure 2C). Furthermore, Mg2+ ions had no effect on the overexpressed proteins, confirming that the observed decrease in cysteine reactivity was specific to Zn2+ ions. NJP14 Modifies the Catalytic Zn2+-Chelating Cysteine in SORD. SORD is a member of the polyol pathway that catalyzes the reversible conversion of sorbitol to fructose to reduce aberrantly high glucose levels. Consequently, this metabolic pathway produces significant oxidative stress resulting in characteristic polyol-induced oxidative damage, thus implicating SORD as a potential therapeutic target for diabetes mellitus.20 Notably, SORD has long been known to utilize a Zn2+ ion to promote catalysis.21,22 Recent crystallographic evidence determined that each subunit of this tetrameric protein contains a Zn2+ ion coordinated with Cys44, His69, Glu70, and an activated water molecule to form a distorted tetrahedral structure.23 We hypothesized that our probe, NJP14, covalently bound to this predicted catalytic Zn2+-chelating cysteine (Cys44); however, another cysteine (Cys179) was previously identified as hyper-reactive in a proteomic study.12 To determine which of these cysteines are directly modified by NJP14, SORD wild-type (WT), C44A, and C179A mutants were generated and overexpressed with Cterminal myc/His tags in HEK293T cells by transient transfection. After NJP14 treatment, click chemistry to install Rh-N3, and in-gel fluorescence analysis, the WT and C179A mutant were shown to be labeled by NJP14, whereas no labeling was detected in the C44A mutant sample (Figure 3A). These data confirm that NJP14 modifies the known Zn2+binding cysteine of SORD and serves to validate our competitive cysteine-labeling strategy to identify putative Zn2+-binding cysteines in the human proteome. D

dx.doi.org/10.1021/cb400622q | ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology

Articles

Figure 4. A global quantitative mass spectrometry platform to identify putative Zn2+-binding cysteines. (A) HeLa cell lysates were treated with either Zn2+ or EDTA, followed by labeling with IA-alkyne. The Zn2+/EDTA-treated proteomes were tagged with the Azo-L linker, and the untreated proteomes were tagged with the Azo-H linker. The samples were mixed together and subjected to avidin enrichment, on-bead trypsin digestion, sodium dithionite treatment, and LC/LC−MS/MS analysis to obtain light:heavy ratios (R) for each labeled cysteine. A cysteine with R > 1 indicates an increase in reactivity upon Zn2+/EDTA treatment, and R < 1 signifies a decrease in reactivity. (B) To identify cysteines that show decreased reactivity in the presence of Zn2+ and increased reactivity upon EDTA treatment, we prioritized cysteines with R < 0.66 in the Zn2+ sample and R > 1.50 in the EDTA sample. This analysis afforded 48 putative Zn2+-binding cysteines. (C) Structure of alcohol dehydrogenase class-3 (ADH5) highlighting the identified Cys174 (red) that binds to the catalytic Zn2+ cofactor (purple) (PDB 1MC5).

previously described activity assay that couples the thioltransferase activity of GSTO1 to the oxidation of NADPH by glutathione reductase.27 Human GSTO1-WT and the C32A mutant were recombinantly expressed in E. coli and purified using a Ni-NTA column (Supplementary Figure S7). We confirmed the dose-dependent decrease in labeling of the recombinant GSTO1 in the presence of Zn2+ ions, using a fluorescence gel-based readout as described previously (Figure 3D). The purified protein was then treated with increasing concentrations of Zn2+ ions prior to measuring enzymatic activity (Figure 3E). While low concentrations (1 μM) had little effect on GSTO1 activity, high concentrations (>10 μM) resulted in complete inhibition of enzyme function. Although the physiological relevance of GSTO1 inhibition by Zn2+ is still in question, our studies demonstrate that, at least in vitro, Zn2+ ions can regulate the activity of proteins with reactive cysteines such as GSTO1. Quantitative Mass Spectrometry Can Globally Identify Zn2+-Sensitive Cysteines. To globally and quantitatively characterize other reactive and functional cysteines that can chelate to Zn2+ and to modify our platform to prioritize those cysteines that are likely to be endogenously Zn2+-bound, we revisited our highly promiscuous IA-alkyne probe that is known to react with hundreds of functional cysteines in the proteome.12 The gel-based studies on this probe indicated several proteins that demonstrated Zn2+-sensitive binding effects (Supplementary Figure S1). Since the high reactivity of the IA-alkyne can result in labeling of numerous cysteines within a single protein, we required a mass spectrometry

Interestingly, the fact that we observe labeling of SORD in nonoverexpressing HeLa cell lysates (Figure 1C) implies that a certain population of SORD endogenously exists in a Zn2+-free state and is susceptible to inhibition by cysteine-modifying agents. Therefore, depending on the cellular concentrations of Zn2+-bound versus unbound protein, cysteine-reactive small molecules could be utilized as potential inhibitors of SORD as well as other enzymes that rely on Zn2+ ions for catalysis. Glutathione S-Transferase Omega 1 Activity Is Sensitive to Zn2+. Identification of the known Zn2+-binding cysteine in SORD serves as validation of our competitive profiling method to identify putative metal-binding sites. In contrast, the cysteine in GSTO1 has not been previously reported to chelate Zn2+ ions. Given the significant Zn2+mediated decrease in NJP14-labeling that we observed for this protein (Figures 1C and 2C), we chose to investigate the effect of Zn2+ on GSTO1 activity. GSTO1 is a member of the GST superfamily of phase II detoxifying enzymes that conjugate glutathione to a wide spectrum of endogenous and exogenous electrophiles as a cellular defense mechanism against oxidative damage, therapeutic drugs, and carcinogens.26 Notably, GSTO1 is a unique member of this family in that its catalytic mechanism relies on an active-site cysteine (Cys32) as opposed to the canonical serine or tyrosine present in the other GST isoforms.27 Although Cys32 has long been functionally characterized as a catalytic nucleophile in GSTO1 activity, its ability to bind Zn2+ ions has never been defined. To explore the functional role of Zn2+-binding, we evaluated the effect of Zn2+ ions on GSTO1 activity using recombinant enzyme and a E

dx.doi.org/10.1021/cb400622q | ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology

Articles

Table 2. Cysteine Residues Identified As Endogenously Bound to Zn2+ Ions in HeLa Cell Lysatesa ratio (light:heavy) protein

peptide sequence

Zn2+ 20 μM

EDTA 1 μM

cellular role

ACO2 Aconitase 2, mitochondrial RPS27 40S ribosomal protein S27 RPS3 40S ribosomal protein S3 TUBB2c, TUBB, TUBB4, TUBB2b, TUBB8, TUBB6, Tubulin β chain TUBA4a, TUBA1b, TUBA1a, TUBA1c, Tubulin α/β chain ADH5 Alcohol dehydrogenase class-3 RPS11 40S ribosomal protein S11 RPL23 60S ribosomal protein L23 USP22 Ubiquitin carboxyl-terminal hydrolase 22

R.VGLIGSC*TNSSYEDMGR.S R.LTEGC*SFR.R K.GC*EVVVSGK.L K.NMMAAC*DPR.H

0.12 0.39 0.44 0.44

1.58 4.55 8.83 1.95

metabolic ribosomal protein ribosomal protein tubulin protein

K.AYHEQLSVAEITNAC*FEPANQMVK.C K.VCLLGC*GISTGYGAAVNTAK.L R.DVQIGDIVTVGEC*RPLSK.T R.ISLGLPVGAVINC*ADNTGAK.N K.C*DDAIITK.A

0.46 0.57 0.57 0.61 0.64

1.98 3.92 4.98 1.72 1.52

MRPS12 28S Ribosomal protein S12, mitochondrial

K.GVVLC*TFTR.K

0.66

1.50

tubulin protein metabolic ribosomal protein ribosomal protein protein degradation ribosomal protein

a

The data were sorted to identify cysteines that demonstrated a 1.5-fold decrease in reactivity upon treatment with Zn2+, and a 1.5-fold increase in reactivity upon EDTA-treatment. A subset of these proteins is shown below, with the complete list in Supporting Information Table S6.

We therefore hypothesized that refining the data from these two mass spectrometry analyses by focusing our attention on those cysteines that showed a ratio R > 1.5 upon EDTA treatment and R < 0.66 upon Zn2+ treatment would likely indicate a cysteine that is endogenously chelated to Zn2+ in these HeLa cell lysates (Figure 4B). These are cysteines that show rescued reactivity upon EDTA-mediated metal removal and decreased reactivity upon addition of Zn2+ ions. The proteins that fulfilled these criteria represent diverse functional classes including oxidoreductases, metabolic enzymes, ribosomal proteins, and microtubule assembly proteins (Table 2 and Supplementary Table S6). Notably, many of these proteins have been previously annotated to bind Zn2+, which helps validate our workflow. In particular, the Cys174 identified by our platform within alcohol dehydrogenase class-3 (ADH5) binds a catalytic Zn2+ cofactor according to a previously solved structure (Figure 4C).29 Additionally, bacterial ribosomal proteins are known to possess the capacity to bind up to 11 equiv of Zn2+ within a single ribosome, and these interactions can account for up to 65% of the total cellular Zn2+.30 After stringent filtering of our data, 10 unique ribosomal cysteine residues were detected by our platform, confirming that this Zn2+ affinity extends to human ribosomes as well. Tubulin-Zn2+ interactions have been well documented, and these interactions have been proven essential for the formation of protofilament sheets and subsequent polymerization.31,32 We identified numerous Zn2+-binding cysteines in both alpha and beta isoforms of tubulin. In addition to these previously characterized Zn2+-binding cysteines, we identified several other putative Zn2+-binding cysteines with unknown function for future exploration. Moreover, these data suggest that diverse cellular pathways could potentially be modulated by fluxes in localized Zn2+ concentrations resulting from cellular damage or dysregulation of Zn2+ homeostasis. In summary, we report a competitive platform to identify cysteine residues that are susceptible to Zn2+ binding within a complex proteome. This platform relies on monitoring a loss in cysteine nucleophilicity induced by direct chelation of Zn2+ to the thiol group, as well as an increase in reactivity resulting from removal of prebound Zn2+ from endogenous chelation sites in proteins. Using a mildly reactive cysteine probe (NJP14), we identified the known Zn2+-binding cysteine in SORD, thereby validating the reliability of this platform. Furthermore, we also identified and characterized a cysteine

platform that specifically identifies each IA-targeted cysteine. Furthermore, we envisioned a platform that provides a more quantitative read-out relative to the semiquantitative method of spectral counting used previously. To achieve this, we utilized a quantitative mass spectrometry platform that facilitates the identification and accurate quantitation of sites of IA-alkyne labeling from two or more proteomes.28 IA-alkyne labeled proteins are tagged with a linker that contains (a) an azide for click-chemistry-based attachment to labeled proteins, (b) a biotin for enrichment on streptavidin beads, (c) an azobenzenebased cleavable unit for release of tagged peptides upon sodium dithionite treatment, and (d) an isotopically tagged valine (light: Azo-L; heavy: Azo-H) for quantification of IA-alkynelabeled peptides across two proteomes.28 To identify Zn2+binding cysteines and furthermore elucidate those that are endogenously Zn2+-bound in cell lysates, we performed two quantitative mass spectrometry analyses with the IA-alkyne probe (Figure 4A). First, an untreated control sample conjugated to Azo-H was compared to either Zn2+- or Mg2+treated lysates tagged with Azo-L. Upon avidin enrichment, onbead trypsin digestion, and selective probe-labeled peptide release with sodium dithionite treatment, the resulting peptide mixtures were analyzed by high resolution LC/LC−MS/MS. A light:heavy ratio (R) was calculated for each IA-labeled peptide and provides an accurate measure of the extent of IA-labeling in the untreated versus metal-treated samples. A cysteine with a R value of 1 is unaffected by metal ion treatment, whereas a cysteine with R < 0.66 signifies a residue that demonstrated at least a 1.5-fold decrease in reactivity in the presence of the metal ions. Similar to our NJP14 studies, most cysteines in the Mg2+-treated sample showed ratios of ∼1, whereas several cysteines in the Zn2+-treated samples showed significantly reduced R values (Supplementary Table S4; Supplementary Figure S8). To narrow the large number of putative Zn2+binding cysteines that we identified, we did a second comparison, whereby we compared a proteome preadministered EDTA and conjugated to the Azo-L linker to an untreated control sample conjugated to the Azo-H linker. IA-labeled peptides identified with R > 1.5 suggest a 1.5-fold increase in cysteine reactivity upon EDTA treatment, which likely implies that the innate reactivity of the cysteine is quenched by a bound-metal ion (Supplementary Table S5). The data from this comparison identify cysteines that are endogenously complexed to any metal ion and are not specific to Zn2+-binding residues. F

dx.doi.org/10.1021/cb400622q | ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology

Articles

samples containing high spectral counts in the water/Mg2+-treated samples with decreased spectral counts in Zn2+-treated samples. SORD Activity Assays. For the oxidative activity assay, the oxidative assay buffer (OAB) stock consisted of glycine (50 mM) at pH 9.9. OAB, sorbitol (10 mM), NAD+ (18.0 μM), and HEK293T SORD-overexpressed protein lysates (15 μg) were combined and aliquoted into a clear 96-well plate, 100 μL per well. The increase of absorbance at 340 nm was monitored to determine enzymatic activity. For the reductive activity assay, the reductive assay buffer (RAB) stock consisted of sodium phosphate (10 mM) at pH 7.0. RAB, fructose (100 mM), NADH (46 μM), and HEK293T SORD-overexpressed protein lysates (15 μg) were combined and aliquoted into a clear 96well plate, 100 μL per well. The decrease of absorbance at 340 nm was monitored to determine enzymatic activity. All activities were calculated as an average of three trials. GSTO1 Activity Assays. Tris (100 mM) pH 8.0 buffer, bovine serum albumin (0.1 mg mL−1), reduced glutathione (1 mM), glutathione reductase (∼2 units), and NADPH (0.3 mM) were combined for each sample. Hydroxyethyl disulfide (0.75 mM) was added to each sample, and the samples were incubated at RT for about 2 min in order for the disulfides to equilibrate. Recombinant purified GSTO1 protein was added to each sample, and the samples were subsequently aliquoted into a clear 96-well plate, 100 μL per well. The plate was monitored for absorbance at 340 nm to determine enzymatic activity. All activities were calculated as an average of three trials. Global Quantitative Mass Spectrometry Analyses. A detailed protocol is described within the Supporting Information for IAlabeling, click chemistry to append the azo-tags, and quantitative mass spectrometry analyses.

disposed to Zn2+-binding in GSTO1 and demonstrated the effect of Zn2+ in inhibiting GSTO1 activity. We then expanded our platform to apply a highly promiscuous cysteine-reactive probe (IA-alkyne), which enabled identification of potential Zn2+-chelating cysteines among ∼900 reactive cysteines in the human proteome. In addition to the many known Zn2+-binding proteins that we identified, there were numerous putative Zn2+binding proteins derived from a diverse array of cellular pathways. These novel Zn2+-binding proteins allude to the potential regulation of a variety of cellular functions through transient fluctuations in intracellular Zn2+ concentrations. Since the concentrations of Zn2+ used in our study are significantly higher than endogenous concentrations, the physiological significance of the putative Zn2+-binding sites that we identified is unclear. However, we hypothesize that the combined identification of Zn2+-sensitive cysteines that also show an EDTA-mediated increase in reactivity suggests that (1) the cysteine has high affinity to chelate to Zn2+, and (2) a certain population of the protein is found to be endogenously bound to a metal ion. Therefore, combining these two analyses gives added confidence in the assignment of these cysteines as physiologically relevant Zn2+ binders. It is of interest to note that a small number of cysteines demonstrated an increase in reactivity upon treatment with Zn2+ ions. These cysteines were found on proteins such as a ribonuclease inhibitor (RNH1) and inorganic pyrophosphatase (PPA1) (Supplementary Table S4). Although the relevance of these increases in cysteine reactivity is still unclear, we hypothesize that these are likely cysteines that do not directly chelate to the Zn2+ but instead are located allosteric to a Zn2+-binding site, such that the metal-binding event perturbs the local environment of the cysteine and thereby affects reactivity. Future studies will delve into the functional relevance of such cases of metal-induced increases in cysteine reactivity. Importantly, from a technological standpoint, our competitive platform can be easily expanded to examine other biologically relevant metals across diverse proteomes, thereby providing an experimental method to complement available structural and computational studies to identify both stable and transient metal-binding sites in proteomes.

■

■

ASSOCIATED CONTENT

S Supporting Information *

Detailed experimental procedures and supplementary figures and tables. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work is supported by the Damon Runyon Cancer Research Foundation (DRR-18-12), the Smith Family Foundation, and Boston College. We thank C. Wang for help with the CIMAGE data analysis algorithm and members of the Weerapana Lab for comments and critical reading of the manuscript.

METHODS

Competitive In-Gel Fluorescence Analysis. HeLa protein lysates (50 μL, 2 mg mL−1) were treated with either Zn2+, Ca2+, Mg2+, or Mn2+ (10 and 20 μM) or water as a control. The samples were incubated at RT for 1 h. NJP14 (50 μM) was added, and the samples were incubated at RT for 1 h. The protein samples were fractionated by size exclusion chromatography and subjected to click chemistry and in-gel fluorescence analysis as described in detail in the Supplemental Methods. Mass Spectrometry Platform To Identify Zn2+-Sensitive NJP14 Targets. HeLa protein lysates (500 μL, 4 mg mL−1) were aliquoted, and Zn2+ (10 and 20 μM), Mg2+ (20 μM), or water was added to the appropriate samples at the designated concentration. The samples were incubated at RT for 1 h. NJP14 (100 μM) was added to all the samples, and they were incubated at RT for an additional 1 h. Unbound metal was removed by size exclusion chromatography, and the proteomes were subjected to click chemistry, streptavidin enrichment, on-bead trypsin digestion, and LC/LC−MS/MS analysis according to previous studies and as described in the Supplemental Methods.19 The generated tandem MS data were searched using the SEQUEST algorithm against the human IPI database. A static modification of +57 on Cys was specified to account for iodoacetamide alkylation. The SEQUEST output files generated from the digests were filtered using DTASelect 2.0. The data were sorted to identify those

■

REFERENCES

(1) Pe’er, I., Felder, C. E., Man, O., Silman, I., Sussman, J., and Beckmann, J. S. (2004) Proteomic signatures: amino acid and oligopeptide compositions differentiate among phyla. Proteins 54, 20−40. (2) Marino, S. M., and Gladyshev, V. N. (2010) Cysteine function governs its conservation and degeneration and restricts its utilization on protein surfaces. J. Mol. Biol. 404, 902−916. (3) Pace, N. J., and Weerapana, E. (2012) Diverse functional roles of reactive cysteines. ACS Chem. Biol. 8, 283−296. (4) Andreini, C., Banci, L., Bertini, I., and Rosato, A. (2006) Counting the zinc-proteins encoded in the human genome. J. Proteome Res. 5, 196−201. (5) Giles, N. M., Watts, A. B., Giles, G. I., Fry, F. H., Littlechild, J. A., and Jacob, C. (2003) Metal and redox modulation of cysteine protein function. Chem. Biol. 10, 677−693.

G

dx.doi.org/10.1021/cb400622q | ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology

Articles

(6) Tainer, J. A., Roberts, V. A., and Getzoff, E. D. (1991) Metalbinding sites in proteins. Curr. Opin. Biotechnol. 2, 582−591. (7) Maret, W. (2010) Metalloproteomics, metalloproteomes, and the annotation of metalloproteins. Metallomics 2, 117−125. (8) Bertini, I., Decaria, L., and Rosato, A. (2010) The annotation of full zinc proteomes. J. Biol. Inorg. Chem. 15, 1071−1078. (9) Maret, W. (2008) Zinc proteomics and the annotation of the human zinc proteome. Pure Appl. Chem. 80, 2679−2687. (10) Maret, W. (2013) Inhibitory zinc sites in enzymes. Biometals 26, 197−204. (11) Maret, W., Jacob, C., Vallee, B. L., and Fischer, E. H. (1999) Inhibitory sites in enzymes: zinc removal and reactivation by thionein. Proc. Natl. Acad. Sci. U.S.A. 96, 1936−1940. (12) Weerapana, E., Wang, C., Simon, G. M., Richter, F., Khare, S., Dillon, M. B. D., Bachovchin, D. A., Mowen, K., Baker, D., and Cravatt, B. F. (2010) Quantitative reactivity profiling predicts functional cysteines in proteomes. Nature 468, 790−795. (13) Weerapana, E., Simon, G. M., and Cravatt, B. F. (2008) Disparate proteome reactivity profiles carbon electrophiles. Nat. Chem. Biol. 4, 405−407. (14) Banerjee, R., Pace, N. J., Brown, D. R., and Weerapana, E. (2013) 1,3,5-Triazine as a modular scaffold for covalent inhibitors with streamlined target identification. J. Am. Chem. Soc. 135, 2497−2500. (15) Pace, N. J., Pimental, D. R., and Weerapana, E. (2012) An inhibitor of glutathione S-transferase omega 1 that selectively target apoptotic cells. Angew. Chem., Int. Ed. 51, 8365−8368. (16) Speers, A. E., and Cravatt, B. F. (2004) Profiling enzyme activities in vivo using click chemistry methods. Chem. Biol. 11, 535− 546. (17) Speers, A. E., Adam, G. C., and Cravatt, B. F. (2003) Activitybased protein profiling in vivo using copper(I)-catalyzed azide-alkyne [3 + 2] cycloaddition. J. Am. Chem. Soc. 125, 4686−4687. (18) Zaworski, P. G., and Gill, G. S. (1988) Precipitation and recovery of proteins from culture supernatants using zinc. Anal. Biochem. 173, 440−444. (19) Weerapana, E., Speers, A. E., and Cravatt, B. F. (2007) Tandem orthogonal proteolysis-activity-based protein profiling (TOP-ABPP)− a general method for mapping sites of probe modifications in proteomes. Nat. Protoc. 2, 1414−1425. (20) El-Kabbani, O., Darmanin, C., and Chung, R. P.-T. (2004) Sorbitol dehydrogenase: structure, function, and ligand design. Curr. Med. Chem. 11, 465−476. (21) Jeffrey, J., Chesters, J., Mills, C., Sadler, P. J., and Jornvall, H. (1984) Sorbitol dehydrogenase is a zinc enzyme. EMBO J. 3, 357−360. (22) Maret, W. (1996) Human Sorbitol Dehydrogenase - a secondary alcohol dehydrogenase with distinct pathophysiological roles: pH-dependent kinetic studies. Adv. Exp. Med. Biol. 414, 383− 393. (23) Pauly, T. A., Ekstrom, J. L., Beebe, D. A., Chrunyk, B., Cunningham, D., Griffor, M., Kamath, A., Lee, S. E., Madura, R., Mcguire, D., Subashi, T., Wasilko, D., Watts, P., Mylari, B. L., Oates, P. J., Adams, P. D., and Rath, V. L. (2003) X-Ray crystallographic and kinetic studies of human sorbitol dehydrogenase. Structure 11, 1072− 1085. (24) Rose, C. I., and Henderson, A. R. (1975) Reaction-rate assay of serum sorbitol dehydrogenase activity at 37 °C. Clin. Chem. 21, 1619− 1626. (25) Lindstad, R. I., and McKinley-McKee, J. S. (1997) Reversible inhibition of sheep liver sorbitol dehydrogenase by the antidiabetogenic drug 2-hydroxymethyl-4-(4-N,N-dimethylaminosulfonyl-1-piperazino) pyrimidine. FEBS Lett. 408, 57−61. (26) Hayes, J. D., and Pulford, D. J. (1995) The glutathione stransferase supergene family: regulation of GST and the contribution of the isoenzymes to cancer chemoprotection and drug resistance. Crit. Rev. Biochem. Mol. Biol. 30, 445−600. (27) Board, P. G., Coggan, M., Chelvanayagam, G., Easteal, S., Jermiin, L. S., Schulte, G. K., Danley, D. E., Hoth, L. R., Griffor, M. C., Kamath, A. V., Rosner, M. H., Chrunyk, B. A., Perregaux, D. E., Gabel, C. A., Geoghegan, K. R., and Pandit, J. (2000) Identification,

characterization, and crystal structure of the omega class of glutathione transferases. J. Biol. Chem. 275, 24798−24806. (28) Qian, Y., Martell, J., Pace, N. J., Ballard, T. E., Johnson, D. S., and Weerapana, E. (2013) An isotopically tagged azobenzene-based cleavable linker for quantitative proteomics. ChemBioChem 14, 1410− 1414. (29) Sanghani, P. C., Bosron, W. F., and Hurley, T. D. (2002) Human glutathione-dependent formaldehyde dehydrogenase. Structural changes associated with ternary complex formation. Biochemistry 41, 15189−15194. (30) Hensley, M. P., Tierney, D. L., and Crowder, M. W. (2011) Zn(II) binding to Escherichia coli 70S ribosomes. Biochemistry 50, 9937−9939. (31) Gaskin, F., and Kress, Y. (1977) Zinc ion-induced assembly of tubulin. J. Biol. Chem. 252, 6918−6924. (32) Eagle, G. R., Zombola, R. R., and Himes, R. H. (1983) Tubulinzinc interactions: binding and polymerization studies. Biochemistry 22, 221−228.

H

dx.doi.org/10.1021/cb400622q | ACS Chem. Biol. XXXX, XXX, XXX−XXX