Proteome-wide analysis of cysteine oxidation reveals metabolic sensitivity to redox stress


Sample preparation for proteomics

For cellular proteomics, Fh1fl/fl cells were treated with 500 µM hydrogen peroxide 15 min prior to extraction for acute oxidation experiments. For chronic oxidation experiments, Fh1fl/fl and Fh1−/− cells were extracted without prior treatment. Plates were washed three times with phosphate-buffered saline (PBS) and extraction solution was added at 80 µl per well. Extraction solution consisted of 55 mM heavy (13C2D2H2INO) or light (12C2H4INO) iodoacetamide in 100 mM Tris-HCl buffer (pH 8.5) with 4% SDS to alkylate free cysteine thiols upon lysis. Cells were scraped immediately and lysates were collected in Eppendorf tubes and sonicated (speed 2, amp 10) for 4 × 5 s. For tissue proteomics, mice were sacrificed by cervical dislocation and kidney tissues were harvested and snap frozen immediately using liquid nitrogen. Representative samples were excised on dry ice, added to extraction solution, and homogenised under dry ice vapour using a Precellys24 bead-based homogeniser (Bertin Instruments) for 3 × 20 s at 5000 rpm. All samples were centrifuged at 16,000g for 5 min at 4 °C. Supernatant was transferred to new tubes and incubated in a table top shaker in the dark at 1400 rpm for 1 h at room temperature. Protein concentration was determined by BCA assay and samples were stored at −80 °C until further processing.

For each independent experiment, 150 µg of differentially labelled protein extracts were mixed using a label-swap replication strategy (i.e. untreated-IAM-heavy mixed with hydrogen peroxide-treated-IAM-light generates forward replicate 1; untreated-IAM-light with hydrogen peroxide-treated-IAM-heavy generates reverse replicate 1, etc.), finally obtaining four replicates for each experimental model. Reversibly oxidised thiols were reduced with 70 mM DTT, incubated at room temperature (24 °C) for 45 min and diluted 1:2 using 50 mM ammonium bicarbonate solution (pH 7.0). Newly generated free thiols were subsequently alkylated using 80 mM NEM. Proteins were then precipitated in two steps using 24 and 10% solutions of trichloroacetic acid (TCA). In both steps, pellets were incubated at 4 °C for 10 min and centrifuged at 18,000 g for 5 min. Supernatants were carefully aspirated and pellets were finally washed with water until the supernatant reached neutral pH. Pellets were reconstituted in 50 µl of 8 M urea solution and submitted to a two-step digestion. First using Endoproteinase Lys-C (ratio 1:33 enzyme:lysate, Alpha Laboratories) for 1 h at room temperature (24 °C), after which partial digests were further diluted to 500 µl with 50 mM ammonium bicarbonate (pH 7.0). This reduced the urea concentration below 1 M to allow for the second digestion with trypsin (ratio 1:33 enzyme:lysate, Promega) overnight at room temperature (24 °C).

Dimethyl labelling

Dimethyl labelling was carried out on alkylated protein digests using a label-swap replication strategy following the on-column protocol described by Boersema et al.26.

Off-line HPLC fractionation

For SICyLIA proteomic experiments, both IAM and dimethyl-modified protein digests were fractionated using reverse phase chromatography. A C18 column (150 × 2.1 mm i.d. – Kinetex EVO (5 µm, 100 Å)) was used with a Dionex HPLC system (Ultimate LPG-3000 binary pump and UVD170U Ultraviolet detector). Modules were controlled by Chromeleon version 6.7. Solvent A (98% water, 2% Acetonitrile) and solvent B (90% Acetonitrile and 10% water) were adjusted to pH 10 using ammonium hydroxide. Samples were injected manually through a Rheodyne valve onto the RP-HPLC column equilibrated with 4% solvent B and kept at this percentage for 6 min. A two-step gradient was applied at a flow-rate of 200 µl/min (from 4–27% B in 36 min, then from 27–48% B in 8 min) followed by a 5 min washing step at 80% solvent B and a 10 min re-equilibration step, for a total run time of 65 min. Column eluate was monitored at 220 and 280 nm, and collected using a Foxy Jr. FC144 fraction collector (Dionex). Collection was allowed from 9 to 54 min for 90 s per vial (300 µl) for a total of 30 fractions. No fraction concatenation strategy was used; only the first 4 and the last 5 fractions were pooled resulting in 21 fractions in total.

UHPLC-MS/MS analysis

For SICyLIA proteomic experiments, fractionated tryptic digests were separated by nanoscale C18 reverse-phase liquid chromatography using an EASY-nLC II 1200 (Thermo Scientific) coupled to a Q-Exactive HF mass spectrometer (Thermo Scientific). Elution was carried out using a binary gradient with buffer A (2% acetonitrile) and B (80% acetonitrile), both containing 0.1% formic acid. Samples were loaded with 8 µl of buffer A into a 20 cm fused silica emitter (New Objective) packed in-house with ReproSil-Pur C18-AQ, 1.9 μm resin (Dr Maisch GmbH). Packed emitter was kept at 35 °C by means of a column oven (Sonation) integrated into the nanoelectrospray ion source (Thermo Scientific). Peptides were eluted at a flow rate of 300 nl/min using different gradients optimised for three sets of fractions: 1–7, 8–15, and 16–21. Two-step gradients were used, all with 20 min for step one and 7 min for step two. Percentages of buffer B (%B) were changed as follows. For F1-7, %B was 2 at the start, 20 at step one, and 39 at step two. For F8-14, %B was 4 at the start, 23 at step one, and 43 at step two. For F15-21, %B was 6 at the start, 28 at step one, and 48 at step two.

All gradients were followed by a washing step (100% B) for 10 min followed by a 5 min re-equilibration step (5%), for a total run time of 40 min. Eluting peptides were electrosprayed into the mass spectrometer using a nanoelectrospray ion source (Thermo Scientific). An Active Background Ion Reduction Device was used to decrease air contaminants signal level.

For label-free quantitation analysis, proteins were digested as described above and resulting tryptic peptides were desalted with a C-18 Stage-Tip52 and subsequently injected using an EASY-nLC 1200 (Thermo Fisher Scientific) coupled online to an Orbitrap Q-Exactive HF mass spectrometer. Peptides were eluted at 300 nl/min flow into a 50 cm fused silica emitter (New Objective) packed in-house with ReproSil-Pur C18-AQ, 1.9 μm resin (Dr Maisch GmbH). The gradient used started at 2% of buffer B and was increased to 16% over 185 min, and then to 28% over 30 min. Finally, a washing step at 95% of B was carried out over 10 min followed by a 13 min re-equilibration at 5% B for a total duration of 243 min.

Ionisation conditions used include: spray voltage 2.1 kV, ion transfer tube temperature 250 °C. Data were acquired using Xcalibur software (Thermo Scientific) and acquisition was carried out in positive ion mode using data dependent acquisition. A full scan (FT-MS) over mass range of 375–1400 m/z was acquired at 60,000 resolution at 200 m/z, with a target value of 3,000,000 ions for a maximum injection time of 20 ms. Higher energy collisional dissociation fragmentation was performed on the 15 most intense ions, for a maximum injection time of 50 ms, or a target value of 50,000 ions. Multiply charged ions having intensity greater than 12,000 counts were selected through a 1.5 m/z window and fragmented using normalised collision energy of 27. Former target ions selected for MS/MS were dynamically excluded for 25 s.

Proteomics data analysis

The MS Raw data were processed with MaxQuant software20 version and searched with Andromeda search engine53, querying UniProt28Mus musculus (20/06/2016; 57,258 entries). First and main searches were performed with precursor mass tolerances of 20 ppm and 4.5 ppm, respectively, and MS/MS tolerance of 20 ppm. The minimum peptide length was set to six amino acids and specificity for trypsin cleavage was required, allowing up to two missed cleavage sites. Methionine oxidation and N-terminal acetylation were specified as variable modifications, no fixed modifications were specified. The peptide, protein, and site false discovery rate (FDR) was set to 1 %. Modification by light and heavy iodoacetamide on cysteine residues (carbamidomethylation) was set as label type modification in Andromeda configuration. Compositions set in the software were: HNOCx(2)Hx(2) for heavy and H(3)NOC(2) for light label. As such, cysteine-containing peptide pairs were treated in the same way as SILAC pairs, and highly accurate median peptide ratios were obtained. These ratios were then further normalised to the median of all ratios in each replicate. As the MaxQuant algorithm searches for and identifies only peptides that contain either light or heavy carbamidomethylation on cysteine residues, peptides with NEM labelling were not considered. Therefore, peptides with multiple cysteine residues that carried both IAM and NEM modifications were excluded from the analysis. In our data sets, 15–16% of peptides contain multiple cysteine residues and less than 3% (H2O2 model), 2% (Fh1 cell model), and 4% (Fh1 tissue model) carried mixed labelling and were excluded.

For Fh1 cell and kidney tissue samples that required protein expression normalisation, dimethylated samples were processed using: DimethLys0/Nter0 and DimethLys8/Nter8 as light and heavy labels, respectively. Protein abundance was then determined by MaxQuant, which calculates the median of the ratios between light and heavy dimethyl modifications measured for all unique peptides from each protein. Both data sets (iodoacetamide heavy/light and dimethyl heavy/light) were processed at the same time in MaxQuant using different parameters, which were defined with the Parameter Groups option. Quantitation of cysteine oxidation reported in the MaxQuant output peptide.txt file was used for the analysis. For LFQ of proteins in hydrogen peroxide-treated and untreated samples, proteins were quantified according to the LFQ algorithm in MaxQuant54.

MaxQuant output was further processed and analysed using Perseus software version Peptides with Cys count lower than one were excluded, together with Reverse and Potential Contaminant flagged peptides. Protein level quantitation was done using the ProteinGroups.txt file. From the ProteinGroups.txt file, Reverse and Potential Contaminant flagged proteins were removed, and at least one uniquely assigned peptide and a minimum ratio count of 2 were required for a protein to be quantified. Only cysteine-containing peptides uniquely assigned to one protein group within each replicate experiment were normalised and included in the analysis. Only cysteine-containing peptides and protein groups that were robustly quantified in three out of four replicate experiments (cells) or replicate samples (tissues) were used for the analysis. QC was imposed by detecting and excluding outliers by calculating upper fences (Q3 + 1.5IQR, where Q3 is the third quartile and IQR is the interquartile range) of the distribution of CV% for all median peptide oxidation ratios in each dataset and using these as cut-off. This led to the exclusion of 578 (H2O2 model), 499 (Fh1 cell model), and 315 (Fh1 tissue model) peptides. Throughout all further analyses, median peptide and protein ratios were used to further minimise the effect of outliers.

Cell culture

Fh1fl/fl and Fh1−/− mouse primary kidney epithelial cells were isolated, immortalised, and authenticated in our laboratory as described previously22. Briefly, kidney epithelial cells from mice with a homozygous conditionally targeted Fh1 allele were immortalised (referred to as Fh1fl/fl cells; these express Fh1 protein at a normal level). Next, Fh1−/− cells were obtained by infecting Fh1fl/fl cells with adenovirus expressing Cre recombinase. Cells were maintained in DMEM (Invitrogen 21969-035, Thermo Fisher Scientific) supplemented with 2 mM glutamine and 10% FBS. All experiments were carried out in medium containing physiological concentrations of nutrients based on a formulation previously reported by our lab55, containing 5.56 mM glucose, 0.65 mM glutamine, 0.1 mM sodium pyruvate, and 2.5% FBS (referred to as experimental medium). Cells were plated in six-well plates (Fh1fl/fl at 2.5 × 105 cells per well, Fh1−/− at 3.75 × 105 cells per well) using 2 ml medium per well and experiments were started after incubating for 20 h at 37 °C and 5% CO2 for all experiments, except Seahorse analysis (specified below). Cell lines were negative for mycoplasma as determined by MycoAlert™ Mycoplasma Detection Kit (Lonza).

Mouse studies

For tissue proteomics, previously described male Fh1fl/fl Ksp1.3/Cre mice were used that express Cre recombinase under the kidney-specific cadherin (Ksp-cadherin) promoter21. One Cre-positive mouse (referred to as Fh1−/−, 232 days old) was compared to a Cre-negative control mouse (referred to as Fh1fl/fl, 239 days old). Mice were sacrificed by cervical dislocation and kidneys were harvested and snap frozen immediately. Four representative kidney tissue slices were used per mice, two from each kidney. No statistical method was used to predetermine sample size. The experiments were not randomised. The investigators were not blinded to allocation during experiments and outcome assessment. Animal work was carried out with ethical approval from the University of Glasgow under the Animal (Scientific Procedures) Act 1986 and the EU Directive 2010 (PPL 70/8645).

Hydrogen peroxide quantification

Phenylboronate (C6H7BO2, Sigma 78181) was used as a probe to detect and quantify hydrogen peroxide, developed based on the methodology described in56. In alkaline conditions, the nucleophilic peroxide attacks the boronic acid and forms a negatively charged tetrahedral boronate intermediate. The weak C–B bond undergoes a 1,2-insertion, which makes the C-bond migrate to the peroxide oxygen atoms. The newly formed borate ester is quickly hydrolysed by water to phenol. Briefly, reaction solution was prepared by diluting phenylboronate (stock solution 60 mM in EtOH) to 1 mM using TBS-T (pH 11 for experimental samples and positive control, pH 7.5 for negative control). To detect hydrogen peroxide in media, 30 µl of medium was added to 200 µl of reaction solution. As a positive control, 30 µl of 500 µM hydrogen peroxide in water was added to 200 µl of reaction solution. Additionally a standard curve was generated for quantification purposes using various concentrations of hydrogen peroxide. As a negative control, 30 µl of 500 µM hydrogen peroxide was added to 20 µl of 0.2 mg/ml catalase (Sigma) in PBS and incubated for 10 min at room temperature, before adding to reaction solution. All samples were vortexed for 2 s and incubated at 80 °C for 7 min, before 400 µl of LC-MS extraction solution (5:3:2 methanol:acetonitrile:H2O) was added. Samples were incubated in a shaker at max speed for 10 min at room temperature before centrifuging at 16,000 g for 10 min. Supernatants were transferred to glass vials and stored at −80 °C until LC-MS analysis. Phenol (C6H6O) generated by the reaction between phenylboronate and hydrogen peroxide was detected using LC-MS as described below.

Cell proliferation assay

Hydrogen peroxide was spiked into the medium at indicated concentrations. After 15 min, residual hydrogen peroxide was removed by aspirating and replenishing the medium. Cell proliferation was monitored by counting live cells at indicated time points.

Colony formation assay

Hydrogen peroxide was spiked into the medium at indicated concentrations. After 15 min, residual hydrogen peroxide was removed by aspirating the medium and cells were trypsinised. Cells were diluted 1:40 in PBS and 200 µl of cell solution was plated in six-well plates with 2 ml of medium per well. Medium was replaced every 3 days and experiments were stopped after 8 days by fixing colonies with 10% TCA overnight at 4 °C. Next day, colonies were stained by addition of 0.04% SRB in 1% acetic acid for 30 min. After four wash cycles with 1% acetic acid, plates were air dried and scanned using an Odyssey scanner (LI-COR, Inc.). Images were exported using Image Studio software (Lite Ver 5.2, LI-COR, Inc.) in grayscale (signal range 60–1800, K:1) and quantified using an ImageJ plugin designed in-house.

Isotope tracing experiments

Uniformly labelled glucose (13C6-glucose) was used at equimolar amounts as found in experimental medium (5.56 mM) and both 13C6-glucose and 500 µM hydrogen peroxide were spiked into the medium at time point 0 and metabolites were extracted after 15 min (Fig. 3d and Fig. 5a, b). For recovery experiments (Fig. 6a–c), residual hydrogen peroxide was removed after 15 min by aspirating and replenishing the medium and metabolites were extracted at indicated time points.

Extraction of metabolites and HPLC-MS analysis

Metabolites were extracted and analysed as described previously57. Briefly, cells were washed twice with ice-cold PBS before LC-MS extraction solution (5:3:2 methanol:acetonitrile:H2O) was added at 1 ml per well of a six-well plate. Plates were incubated for 5 min at 4 °C. Next, extraction solution was transferred to microcentrifuge tubes and centrifuged at 16,000g for 10 min at 4 °C. Supernatants were transferred to glass HPLC vials and stored at −80 °C until analysis LC-MS analysis. For HPLC-MS analysis a ZIC-pHILIC column (SeQuant, VWR) was used and Exactive and Q-Exactive mass spectrometers (Thermo Fisher Scientific) were operated with electrospray (ESI) ionisation and polarity switching mode at scan range (m/z) 75–1000 at a resolution of 25,000 at 200 m/z. Data were acquired using Xcalibur software (Thermo Fisher Scientific) and peak areas of metabolites were determined using TraceFinder software (Thermo Fisher Scientific) by identifying metabolites using mass and known retention time following in-house analysis of commercial standards on our systems. Metabolites were normalised to cellular protein content by Lowry assay performed on extracted cell monolayers.

OCR measurements

Fh1fl/fl cells were plated at 3 × 104 cells per well using 200 µl medium per well onto Seahorse XFe96 plates (Agilent) and incubated overnight at 37 °C and 5% CO2. Then, the medium was replaced with 150 µl unbuffered Seahorse XF Base Medium (Agilent) supplemented with 5.56 mM glucose, 0.65 mM glutamine, 0.1 mM sodium pyruvate (equal to experimental medium used in other experiments) and 1% FBS (pH 7.4) and cells were placed in a CO2-free incubator at 37 °C for 30 min. OCR was recorded using a Seahorse XFe96 Extracellular Flux Analyzer (Agilent) at baseline and after injection of medium (untreated condition) or indicated concentrations of hydrogen peroxide. OCR was normalised to basal OCR of untreated cells.


Mouse kidney tissues were fixed in 10% buffered formalin and paraffin embedded. Sections (4 µm) were stained with H&E and digital images were acquired using a Leica SCN400f slide scanner.


Feature Key annotations for cysteine residues were downloaded from UniProt (Mus musculus proteome downloaded on 03/11/2017; 51,950 sequences)28. Annotations from subcategories Function (Active site, Binding site, Metal binding and Site) and PTM/Processing (Disulphide bond, Modified residues and Post-translational modifications) were included and linked to Leading razor protein and Cysteine position of all identified cysteine-containing peptides using Perseus software version The 1789 enzyme-encoding genes included in Recon 241 were converted to mouse associated genes using Ensembl BioMart tool (Ensembl version 85, ref. 58), resulting in a list of 1740 mouse metabolic genes that was used for enrichment analysis. Enrichment analyses of Recon 2 and GO categories30,31 were performed using Perseus. To further analyse the individual proteins contained in enriched GO categories, they were queried using the AmiGO web application59. The GOBP category of cell redox homoeostasis (GO:0045454) for Mus musculus was downloaded on 07/10/2017. Global protein cysteine content analysis of the metabolic and mitochondrial proteome was done based on annotation of GOBP category of metabolic process (GO:0008152) and GOCC category of mitochondrion (GO:0005739), respectively, which were downloaded for Mus musculus on 22/02/2017.

Statistical methods

The numbers of independent experiments and replicates are indicated in figure legends. Error bars represent standard deviations. For hydrogen peroxide quantification experiments, linear regression analyses were performed using GraphPad Prism 7.02 software (GraphPad Software Inc, San Diego, USA). For LFQ proteomics experiments, to determine whether protein abundance differed significantly between samples in hydrogen peroxide experiments, a two-sided t-test was applied using the recommended settings in Perseus software version The eight individual LFQ measurements for untreated and hydrogen peroxide-treated samples were grouped according to treatment, no grouping was preserved in randomisations, and the number of randomisations was 250 (FDR 0.05, S0 = 0.1). To define significantly oxidised or reduced cysteine peptides for all SICyLIA proteomic experiments, the two-sided Significance B algorithm20 was applied to the median oxidation ratio of the replicate samples for each unique peptide using Perseus. This algorithm calculates the probability that a log-ratio of at least the magnitude observed is obtained (with the Benjamini-Hochberg correction for multiple hypothesis testing). By creating intensity bins of equal occupancy, the effect of peptide intensity on their statistical spread is taken into account60. Cysteine peptides were considered significantly oxidised or reduced if they passed the threshold value of 0.05 (Benjamini–Hochberg FDR used for truncation). For a more detailed description of this algorithm and a substantiation of the statistical approach used, see Supplementary Note 1. Kernel density plots of the p-values were constructed in R version 3.4.3 using RStudio version 1.1.423 (bw = 0.4). In order to define significantly enriched GO categories within the significantly modified peptides, the Fisher exact test was applied using Perseus. Categories were considered significantly enriched if they passed the threshold value of 0.05 (Benjamini-Hochberg FDR used for truncation, with relative enrichment on leading razor protein). For Seahorse experiments, repeated measures two-way ANOVA with Dunnet’s test for multiple comparisons correction was performed using GraphPad Prism 7.02 software. Significance threshold for the multiplicity adjusted p value was 0.05 (95% confidence interval). No statistical method was used to predetermine sample size for the experiments.


13C6-glucose (CLM-1396-5) was obtained from Cambridge Isotopes Laboratories (Tewksbury, MA, USA). Labelled (13C2D2H2INO, 721328) and unlabelled (C2H4INO, I6125) iodoacetamide and all other remaining reagents were obtained from Sigma-Aldrich (Merck KGaA, Darmstadt, Germany). For labelled iodoacetamide, the purity (GC), 13C enrichment, and D enrichment were all 99% as determined by Sigma-Aldrich.

Data availability

The raw MS files and search/identification files obtained with MaxQuant have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository61 with the dataset identifiers PXD006363 for the H2O2 model, PXD006372 for the Fh1 cell model, and PXD006373 for the Fh1 tissue model. All other data from this study are available from the authors upon reasonable request.

Source link


Please enter your comment!
Please enter your name here