Proteomic detection of hydrogen peroxide-sensitive thiol proteins in Jurkat cells

Materials

Cell-culture materials were from Gibco BRL supplied by Invitrogen New Zealand Ltd. (Auckland, New Zealand). Complete™ protease cocktail tablets, sequencing-grade trypsin, Tris and CHAPS were from Roche Diagnostics (Mannheim, Germany). The caspase substrate DEVD-AMC (Asp-Glu-Val-Asp-7-amino-4-methylcoumarin) was purchased from Peptide Institute Inc. (Osaka, Japan). A phospho-MAP kinase (mitogen-activated protein kinase) antibody kit was obtained from Cell Signaling Technology (Beverly, MA, U.S.A.), and peroxiredoxin antibodies were from LabFrontier (Seoul, Korea). IAF was from Molecular Probes Inc. (Eugene, OR, U.S.A.). Micro Bio-Spin® 6 chromatography columns, acrylamide/bisacrylamide solution, Bio-Rad D_c protein assay, ReadyStrip™ (17 cm, pH 3–10) IPG (immobilized pH gradient) strips, Bio-Lyte® 3/10 ampholytes and PDQuest™ 2D gel analysis software were from Bio-Rad Laboratories (Hercules, CA, U.S.A.). Other reagents were from Sigma Chemical Co. (St Louis, MO, U.S.A.). ZipTip® C18 pipette tips were from Millipore (Billerica, MA, U.S.A.).

Cell culture

The Jurkat T-lymphocyte cell line was obtained from the A.T.C.C. [American Type Culture Collection, Rockville, MD, U.S.A. (now at Manassas, VA, U.S.A)]. Cells were maintained in RPMI-1640 supplemented with 10% (v/v) heat-inactivated fetal-bovine serum, 2 mM glutamine, 100 units/ml penicillin and 100 μg/ml streptomycin at 37 °C in a humidified atmosphere with 5% CO₂. Before use, cells were suspended at 1×10⁶ cells/ml in fresh media.

Proliferation assay

A colorimetric BrdU (5-bromo-2′-deoxyuridine) ELISA kit available from Roche Diagnostics (Mannheim, Germany) was used to measure cell proliferation. Jurkat cells (1×10⁶ cells/ml) were treated for 1 h with a range of H₂O₂ concentrations, then diluted 1:1 with fresh medium and incubated at 37 °C for another 1 h to optimize the rate of cell proliferation. BrdU solution was added to each sample and the DNA labelling reaction carried out for 2 h at 37 °C. Cells were then harvested, washed in PBS and transferred to 96-well plates (5×10⁴ cells/well). The cells were sedimented by 10 min centrifugation at 500 g and BrdU-labelled DNA was detected as described in the manufacturer's instructions.

Caspase activity assay

Jurkat cells were harvested by centrifugation after 6 h treatment with H₂O₂. The activity of caspase 3-like enzymes was measured by cleavage of the caspase substrate DEVD-AMC [3].

MAP kinase activity and peroxiredoxin Western blotting

After H₂O₂ treatment for various times, Jurkat cells were washed and extracted with 1 ml of 40 mM Hepes buffer, pH 7.4, containing 1% (w/v) CHAPS, 50 mM NaCl, 1 mM EDTA, 1 mM EGTA, and Complete™ protease inhibitors per 5×10⁶ cells. Cell extracts (50 μg of protein) were separated by SDS/PAGE and Western-blotted with an antibody against either the dually phosphorylated form of ERK (extracellular-signal-regulated kinase) 1/2, peroxiredoxin 2 or overoxidized peroxiredoxin (anti-Prx-SO₃). Peroxidase-conjugated secondary antibodies and chemiluminescence were used for detection.

Assesment of cell viability

After treatment, cells (approx. 2×10⁵) were harvested and incubated for 5 min with 2 μg of propidium iodide. Cell fluorescence was analysed with a bivariate flow cytometer (Becton Dickinson, Mountain View, CA, U.S.A.). Cells that did not stain with propidium iodide were classified as viable, and the viability was expressed as a percentage of the total cells analysed.

Thioredoxin reductase assay

Thioredoxin reductase activity was estimated indirectly as NADPH-dependent reduction of DTNB [5,5′-dithiobis-(2-nitrobenzoic acid)] on cell extracts containing approx. 200 μg of protein [20]. Enzyme activity was estimated as the difference in rate of absorbance increase at 412 nm before and after the addition of NADPH.

Glutathione assay

The intracellular concentration of GSH was measured by HPLC with fluorescence detection following derivatization with monobromobimane [21].

Fluorescent labelling of oxidized thiol proteins and 2D electrophoresis

Reversibly oxidized thiol proteins were monitored as described previously [19], and summarized in Scheme 1. Briefly, reduced protein thiols were blocked by resuspending Jurkat cells in buffer containing NEM (N-ethylmaleimide) and incubating them at room temperature for 15 min [22]. After lysing the cells, excess NEM was removed and the oxidized thiols were reduced with DTT (dithiothreitol). The unblocked protein thiols were then labelled with IAF. Samples were desalted and excess IAF was removed before 2D electrophoresis, which was performed as described in [19], except that both SDS/10%- and 15%-(w/v)-PAGE gels were used for the second dimension.

MALDI-TOF MS

Fluorescence-labelled thiol proteins were isolated from 2D SDS/PAGE gels for identification prior to staining the gels for total protein. The spots that decreased in fluorescence intensity were excised from untreated gels and those that increased in intensity were excised from H₂O₂-treated gels. Gels were immediately scanned for fluorescence at the conclusion of electrophoresis. An actual size image of the gel was printed and the gel was placed on a glass plate and carefully oriented over the gel image to visualize the fluorescent protein spots. The protein spots of choice were excised and the gels re-scanned to confirm that the desired protein spots had been accurately obtained.

The gel pieces were washed in 100 mM ammonium bicarbonate before being dehydrated in 100% acetonitrile and dried in a vacuum concentrator at 45 °C. The gel pieces were rehydrated in 50 mM ammonium bicarbonate solution containing 0.25 μg of trypsin for 45 min on ice. Unabsorbed trypsin solution was replaced with 50 mM ammonium bicarbonate solution and in-gel enzymatic proteolysis was performed overnight at 37 °C. Peptides were eluted from the gel pieces by repeated incubations in 1:1 (v/v) solution of acetonitrile and 0.5% (v/v) trifluoroacetic acid. The eluted peptides were dried in a vacuum concentrator at 60 °C and then resuspended in 0.5% trifluoroacetic acid [23]. C18 reversed-phase ZipTip® pipette tips were used to concentrate and desalt the samples according the manufacturer's instructions. The peptides were eluted from the ZipTips® with 0.5% (w/v) α-cyano-4-hydroxycinnamic acid in 50% (v/v) acetonitrile and 0.25% (v/v) trifluoroacetic acid. Spectra were acquired in positive-ion reflector mode with accelerating voltage 20000 V, grid voltage 75%, guide-wire voltage 0.002% and extraction delay time 180 ns. Masses were calibrated with external standards [angiotensin, 1 m/z 1296.6853; ACTH (adrenocorticotropic hormone) fragment 1–17, m/z 2093.0876; and ACTH fragment 18–39, m/z 2465.1989]. The peptide mass fingerprints were used to search online protein databases found at: http://129.85.19.192/profound_bin/WebProFound.exe [24].

Cellular responses to H₂O₂

H₂O₂ triggers a range of cellular responses, including proliferation, growth arrest, induction of apoptosis and MAP kinase activation [3,15,25]. The effect of H₂O₂ treatment on these processes in Jurkat cells was investigated in order to establish the concentration range over which signalling pathways were affected. To detect activation of ERK 1/2, the dually phosphorylated form was visualized by Western blotting (Figure 1A). Slight activation was observed after 20 min with 100 μM H₂O₂, and maximal activation at 200–500 μM H₂O₂. Cell proliferation, measured at 2 h, was almost completely inhibited by 100 μM H₂O₂ (Figure 1B). These cells subsequently underwent apoptosis, as determined by an increase in activity of caspase 3-like enzymes 6 h after H₂O₂ treatment (Figure 1C). Some apoptosis was also detected with 50 μM H₂O₂.

We selected two concentrations of H₂O₂ for investigating thiol-protein oxidation. A low concentration of 20 μM was used to detect proteins that undergo oxidation without long-term changes in cell physiology. A higher dose of 200 μM H₂O₂ was used to detect thiol protein changes that occur with more severe oxidative stress that eventually leads to growth arrest and cell death. The cells were treated for 10 min. As previously [3], approx. 50% of the peroxide was consumed in this time. There was no loss of viability, measured as propidium iodide uptake (Figure 1D). Only after 6 h was there a slight decrease with 200 μM H₂O₂. DNCB, which was used to disrupt thiol homoeostasis (see below), had no effect at 10 min, but subsequently caused a progressive loss in viability that was not enhanced by H₂O₂ (Figure 1D).

Detection of H₂O₂-sensitive thiol proteins

Using a protocol that labels reversibly oxidized thiol proteins with IAF (Scheme 1), we have previously shown that only a minor fraction of the cysteine residues in proteins of resting Jurkat cells are oxidized [19]. Nevertheless, approx. 400 proteins with oxidized cysteine residues can be detected by 2D electrophoresis. These represent proteins normally present as disulphides, those that undergo redox cycling and exist in a partially oxidized form, and any that were incompletely blocked with NEM. In the present study, cells were treated with H₂O₂ for 10 min so as to focus on redox changes rather than changes in protein expression. Each treatment was compared with control cells, and spots that doubled in intensity or decreased by at least half in at least three out of the four experiments were deemed to reflect genuine changes.

When Jurkat cells were treated with 200 μM H₂O₂, the IAF labelling intensity of the vast majority of the spots was unchanged (Figures 2A and 2B). However, there were selective changes to specific proteins. Some of these, including those highlighted in the 37, 29 and 26 kDa regions, stood out when comparing the whole gels (Figures 2A and 2B). Others, such as those in Figures 2C–2F, became evident in zoomed images. Treatment with H₂O₂ increased labelling of 28 spots. Over half of the increases well exceeded the 2-fold threshold. Unexpectedly, H₂O₂ also caused decreased labelling of 24 spots, including the 26 and 29 kDa clusters (as shown in Figures 2A and 2B). In most of these cases, the spots became scarcely detectable.

The molecular masses and pI values of the peroxide-sensitive spots are given in Figure 3. The identities of those that could be excised from the gels and characterized by MALDI-TOF mass fingerprinting are in Table 1. With a few exceptions, there is good agreement between the pI values and molecular masses of the identified proteins predicted from their sequences and estimates from their positions on the 2D gels. The identified proteins have a range of structural and metabolic functions. The nine 37 kDa spots that increased in intensity are isoforms of GAPDH. One cluster of 26 kDa proteins (pI≈5) that showed a marked decrease includes several spots identified as peroxiredoxin 2. Proteins in the peroxiredoxin 1 region (pI≈7) also decreased, but not as strikingly. The 29 kDa spots that almost disappeared could not be identified. Other proteins that showed marked increases included ubiquitin thiolesterase, proteosome activator complex subunit 1, HSP90β (heat-shock protein 90β) and enolase. A substantial decrease in labelling was seen for PKA (cAMP-dependent kinase) type II-α regulatory chain. One spot, identified as the actin-binding protein moesin, decreased, whereas another moesin spot at a higher molecular mass increased, suggesting possible modification in the H₂O₂-treated cells.

At 20 μM, H₂O₂ still caused a dramatic increase in labelling of the GAPDH cluster (results not shown), but there were many fewer changes than at the higher concentration. These were evident only on zoomed images, where eight consistent but modest increases were detected along with six decreases, including three (two identified) in the peroxiredoxin 2 region and PKA type IIα regulatory subunit (Figure 3).

Thiol-protein oxidation in cells treated with DNCB

GSH and thioredoxin can regulate the redox state of thiol proteins either by reducing oxidized cysteine residues or by consuming oxidants such as H₂O₂ via the glutathione peroxidase/glutathione reductase and peroxiredoxin/thioredoxin reductase pathways [5,6]. Their influence on the pattern of protein oxidation induced by H₂O₂ was probed by treating Jurkat cells with DNCB [26] to inhibit thioredoxin reductase and deplete GSH. Treatment for 10 min inhibited thioredoxin reductase with an IC₅₀ of 8 μM (Figure 4, upper panel). Almost complete inhibition was observed with 30 μM, which was used for all following experiments. Intracellular GSH content decreased after 1 min incubation with DNCB to 57%, and after 11 min to 31%, of that measured in untreated cells. When cells were incubated with 200 μM H₂O₂ for 10 min, there was no decrease in GSH, and the decrease produced by a combination of H₂O₂ and DNCB was comparable with that produced by DNCB alone (Figure 4, lower panel).

Even in the absence of H₂O₂, incubation of Jurkat cells with DNCB caused changes in IAF-labelling pattern of oxidized proteins (Figures 5A and 5B). A total of 29 spots increased in intensity and 13 decreased (Figure 3). The most obvious changes were the increase in labelling of three proteins in the 14 kDa region (box 3 in Figure 5), one of which was identified as thioredoxin 1, and increases in clusters of spots around 25 kDa were shown to include members of the peroxiredoxin 1 and 2 families (box 2). The accumulation of oxidized thioredoxin and peroxiredoxins is consistent with inhibition of thioredoxin reductase. Three spots with mass fingerprints corresponding to glutathione S-transferase also increased in intensity.

As illustrated in the cluster diagram (Figure 3), the set of spots that changed with DNCB was largely independent of the set that changed with H₂O₂. In particular, labelling of the GAPDH series (box 1, Figure 5) increased only with H₂O₂, and oxidized thioredoxin increased only with DNCB. There were also several spots, including the peroxiredoxins, GST P1-1 and the PKA regulatory subunit, which were reversibly oxidized with DNCB, but decreased in intensity with H₂O₂.

When the cells were treated with a combination of DNCB and 200 μM H₂O₂ (Figure 5C), many of the changes seen with either treatment alone (Figures 2B and 5B) were evident. However, the peroxiredoxin and GST spots that increased in intensity with DNCB and decreased with H₂O₂ were unchanged compared with controls. The overall pattern (summarized in the cluster diagram in Figure 3) was largely a summation of the changes seen with either treatment alone. There was little evidence of DNCB sensitizing the cells to H₂O₂, as relatively few spots changed only with the combination. Most obvious of these were the increases in intensity seen for deoxycytidine kinase and proteosome activator complex subunit 1 and the decrease in actin (marked with arrows in Figure 5C). These proteins may have cysteine residues that are sensitive to oxidation and rapidly reduced by the glutathione and thioredoxin systems or become oxidant-sensitive only when these pathways are compromised. Although not assessed quantitatively, the effect of H₂O₂ on spots such as GAPDH did not appear to be greater in DNCB-treated cells.

Detection of irreversible oxidation of peroxiredoxins

An unexpected finding was that IAF labelling of a number of proteins including peroxiredoxins was less after treatment with H₂O₂. This implies a decrease in reversibly oxidized thiols. To test whether this represented irreversible oxidation, we investigated the fate of the peroxiredoxins, which have active-site cysteine residues that can be oxidized to sulphinic acids by excess H₂O₂ [9–11]. Western blotting was performed using an antibody that recognizes the overoxidized proteins [9]. Control Jurkat cells gave little staining, but cells treated with H₂O₂ in the presence or absence of DNCB gave a strongly positive band at about 25 kDa (Figure 6A), which co-localized with peroxiredoxin 2. These treatments gave either a decrease (Figure 2B) or no change (Figure 5C) in IAF labelling of the peroxiredoxins. DNCB alone, which increased IAF labelling (Figure 5B), did not cause irreversible oxidation (Figure 6A). Blotting for overoxidized peroxiredoxins on 2D gels of proteins from H₂O₂-treated cells showed labelling of several spots in the peroxiredoxin region (Figure 6B). These findings are consistent with irreversible oxidation being one cause of decreased IAF labelling.