Characterization of surface-exposed reactive cysteine residues in Saccharomyces cerevisiae

Abstract

Numerous cellular processes are subject to redox regulation, and thiol-dependent redox control, acting through reactive cysteine (Cys) residues, is among the major mechanisms of redox regulation. However, information on the sets of proteins that provide thiol-based redox regulation or are affected by it is limited. Here, we describe proteomic approaches to characterize proteins that contain reactive thiols and methods to identify redox Cys in these proteins. Using Saccharomyces cerevisiae as a eukaryotic model organism, we identified 284 proteins with exposed reactive Cys and determined the identities of 185 of these residues. We then characterized subsets of these proteins as in vitro targets of major cellular thiol oxidoreductases, thioredoxin and glutaredoxin, and found that these enzymes can control the redox state of a significant number of thiols in target proteins. We further examined common features of exposed reactive Cys and compared them with an unbiased control set of Cys using computational approaches. This analysis (i) validated the efficacy of targeting exposed Cys in proteins in their native, folded state, (ii) quantified the proportion of targets that can be redox regulated via thiol oxidoreductase systems, and (iii) revealed the theoretical range of the experimental approach with regard to protein abundance and physicochemical properties of reactive Cys. From these analyses, we estimate that approximately one-fourth of exposed Cys in the yeast proteome can be regarded as functional sites, either subject to regulation by thiol oxidoreductases or involved in structural disulfides and metal binding.

Fig. 1. Proteomic approaches employed in the current study

Key steps in the overall procedure are shown for the detection of yeast proteins containing surface-exposed reactive Cys residues. (A) An in-gel approach. (B) An LC-MS/MS approach. (C) An alternative representation of the LCMS/MS approach. In (B) and (C), green number labels refer to different steps in the LC-MS/MS method wherein a number in (B) corresponds to the number in (C).

Fig. 2. BIAM-reactive and TCEP-reducible proteins

In the TCEP-reducible sample, proteins were reduced with TCEP and then labeled with BIAM. In the sample without addition of reducing agents, cells were broken in the presence of BIAM, which reacted with exposed Cys. After isolation on an avidin column, proteins were analyzed by SDS-PAGE. Coomassie Blue stained bands were cut and identified by in-gel tryptic digestion and MS/MS. Lane 1, markers; lane 2, initial cell lysate; lane 3, unbound fraction; (not biotinylated proteins); lane 4, enriched BIAM-labeled proteins after elution from the avidin column; lane 5, cell lysate treated with TCEP; lane 6, flow-through from the avidin column of the BIAM-labeled sample; lane 7, BIAM-labeled proteins from the avidin column.

Fig. 3. TRX1-reducible proteins

Cys residues in the lysate prepared from diamide-treated yeast cells were blocked by IAM (cells were lysed in the presence of this compound). Proteins were reduced by TRX1 and labeled with BIAM, followed by isolation on an avidin column and SDS-PAGE to visualize the target proteins. Coomassie Blue stained protein bands were labeled, cut and identified by in-gel digestion and mass spectrometry. Lane 1, protein standards (molecular masses in kDa are indicated on the left); lane 2, cell lysate; lane 3, flow-through fraction from the affinity column of the TRX1-reduced sample; lane 4, BIAM-labeled proteins isolated on the affinity column (each number represents one protein band extracted from the gel and analyzed by MS/MS); lane 5, flow-through fraction of the control; lane 6, BIAM-labeled proteins isolated from the control group (i.e., diamide-treated and not subsequently reduced).

Fig. 4. TTR1-reducible proteins

Cys residues in the lysate prepared from diamide-treated yeast cells were blocked by IAM (cells were lysed in the presence of this compound). Proteins were reduced by TTR1 and labeled with BIAM, followed by isolation on an avidin column and SDS-PAGE to visualize the target proteins. Protein bands were cut and identified by in-gel tryptic digestion and MS/MS. Lane 1, protein standards (molecular masses in kDa are indicated on the left); lane 2, cell lysate; lane 3, flow through fraction of the TTR1-reduced sample; lane 4, BIAM-labeled proteins isolated from the control sample that lacked GSH (each number represents one band extracted and analyzed by MS/MS); lane 5, flow-through fraction of the TTR1-reduced sample; lane 6, BIAM-labeled TTR1 target proteins (each number represents one protein band extracted and analyzed by MS/MS).

Fig. 5. Functional categories of redox target proteins

Proteins identified in all experiments, after removal of redundancy, are shown. They are subdivided according to the reducing agent capable of their reduction; ready-to-react thiols (i.e., targets detected by BIAM in cell lysate without addition of any reducing agent) and TCEP-reducible are labeled as “SH + TCEP targets” in the figure ; TRX1-reducible targets are labeled as “TRX1 targets”, and TTR1-reducible proteins are shown as “TTR1 targets”. The complete lists of proteins targeted exclusively by TRX1 or TTR1 are provided (blue contoured and red contoured tables respectively). Additionally, among the set of “SH + TCEP targets”, the list of proteins containing “ready to react” thiols (labeled as “SH”, i.e. BIAM reactive Cys in absence of reducing agent) is provided.

Fig. 6. Common features of detected Cys-containing peptides

(A) Percentage of buried residues for each amino acid type in a set of 505 yeast protein structures. Cys proved to be a mostly buried residue (~ 60% of Cys had a whole residue exposure < 0.1 Ų). In contrast, 80% of Cys detected in our analysis were accessible to solvent. (B) Average Kyte-Doolittle (KD) score per residue (red bars, with standard error of mean) and average net charge per residue (blue bars, with standard error of mean) are shown for detected reactive Cys (left red bar and left blue bar, respectively) and control Cys (blue bars). (C) Amino acid composition around reactive Cys (6 Å) (blue bars) compared to control (red contoured bars). Average frequency (from 0 to 1) of each amino acid is shown in the Y-axis. (D) Protein abundance for proteins with reactive Cys residues (blue line) and control proteins (red line). Abundance (X-axis) is in the logarithmic scale (log₂); frequency (Y-axis) is normalized to the highest. The average abundance for yeast proteins containing BIAM-reactive Cys was 70,736 molecules/cell, while the overall abundance for all yeast ORFs was 12,064 molecules/cell, representing an approximately 6-fold increase.

Fig. 7. Proportion of reactive Cys in the yeast proteome

(A) Theoretical approach that was used to screen an unbiased set of yeast protein structures (depicted in ribbons) to detect reactive thiols, defined as having features that characterize the experimentally detected peptides (see Fig. 6 and text). In our structural test case, some proteins had Cys detected by the BIAM-based proteomic approach (marked with an asterisk in the figure), and some were not. By screening the test case for the Cys exposure, hydrophobicity, net charge balance, and abundance, we detected 382 Cys among 1,429 Cys residues analyzed. (B) A pie graph showing the proportion of detected BIAM-reactive Cys (reactive with probe, red shaded slices) among all theoretically-derived Cys. BIAM-reactive Cys found to be regulated via thioredoxin and/or glutaredoxin systems are shown in dark shaded red. Among other BIAM-reactive Cys (light shaded red slice), the majority (92%) were regulated via TCEP; thus, the fraction of non-redox regulated BIAM-reactive Cys would be a narrow sub-slice of ~ 1% (not shown), within the ‘reactive with probe’ section. Exposed and clustered Cys (yellow shaded slices) are further divided into two groups, where the dark shaded slice represents a theoretical estimation of interacting proximal Cys (disulfides or metal binding Cys; details are provided in the text). The remaining Cys (labeled as ‘Non reactive’) were not identified in our proteomic analysis.