Endogenous, regulatory cysteine sulfenylation of ERK kinases in response to proliferative signals

Abstract

ERK-dependent signaling is key to many pathways through which extracellular signals are transduced into cell-fate decisions. One conundrum is the way in which disparate signals induce specific responses through a common, ERK-dependent kinase cascade. While studies have revealed intricate ways of controlling ERK signaling through spatiotemporal localization and phosphorylation dynamics, additional modes of ERK regulation undoubtedly remain to be discovered. We hypothesized that fine-tuning of ERK signaling could occur by cysteine oxidation. We report that ERK is actively and directly oxidized by signal-generated H₂O₂ during proliferative signaling, and that ERK oxidation occurs downstream of a variety of receptor classes tested in four cell lines. Furthermore, within the tested cell lines and proliferative signals, we observed that both activation loop-phosphorylated and non-phosphorylated ERK undergo sulfenylation in cells and that dynamics of ERK sulfenylation is dependent on the cell growth conditions prior to stimulation. We also tested the effect of endogenous ERK oxidation on kinase activity and report that phosphotransfer reactions are reversibly inhibited by oxidation by as much as 80-90%, underscoring the importance of considering this additional modification when assessing ERK activation in response to extracellular signals.

Keywords: Growth factor signaling; Kinase; Reactive cysteine, ERK, MAPK; Redox regulation; Sulfenic acid.

Figure 1

Trapping of sulfenic acids within cellular proteins during lysis. A) Labeling of oxidized proteins with DCP-Bio1. The reactive carbon of DCP-Bio1 selectively reacts with sulfenic acids to form a covalent adduct, thereby biotinylating proteins containing sulfenic acids at the time of lysis. B) Overall experimental approach for labeling of proteins undergoing active oxidation to sulfenic acid. Cells are cultured to ~80% confluency, treated with an exogenous signal of interest such as platelet-derived growth factor (PDGF), and harvested in lysis buffer containing DCP-Bio1 to trap sulfenic acids, N-ethylmaleimide (NEM) and iodoacetamide (IAM) to alkylate free, reduced thiols (to minimize thiol oxidation during lysis), and catalase to remove any H₂O₂ formed during lysis. After incubation for 30 min on ice, lysates are clarified by centrifugation, excess DCP-Bio1 is removed with a molecular sieve (BioGel P6) spin column, and eluates are subjected to avidin-based affinity capture. Beads are then stringently washed to remove any unlabeled proteins carried over during capture. Captured proteins are eluted by incubation in 2% sodium dodecyl sulfate (SDS) at 100°C for 10 min, followed by electrophoresis and immunoblotting for the protein of interest. To control for differences in capture efficiency and gel loading, a pre-biotinylated bacterial protein, AhpC, is added to lysates at a ratio of 500 μg lysate to 0.5 μg AhpC before affinity capture. Photo courtesy of ATCC.

Figure 2

ERK1/2 cysteines are oxidized to sulfenic acid in response to PDGF in NIH 3T3 cells. ERK oxidation was monitored during proliferative signaling as described in Materials and Methods. Briefly, cells were treated with 20 ng/mL PDGF after 18 h in serum-free media and lysed in the presence of the sulfenic acid trap DCP-Bio1, thereby biotinylating proteins undergoing cysteine oxidation. After addition of prebiotinylated AhpC as an internal control (see Figure 1 legend and Methods), DCP-Bio1 labeled proteins were captured via streptavidin-agarose beads, resolved by SDS-PAGE, and subjected to Western blot for total and dual-(TEY) phosphorylated ERK (ppERK). Upper panels of (A) and (B) show representative blots of total (A) and phosphorylated (B) ERK oxidation, as well as corresponding blots of whole cell lysates, detected in NIH 3T3 cells treated with PDGF. Lower panels summarize ERK band intensity from multiple experiments (n=4) of captured proteins following normalization to AhpC and then to the band intensity at the 10 min time point (set to 1) for each replicate. *, p<0.05; ***, p<0.001; ****, p<2x10⁻¹⁹.

Figure 3

H₂O₂ generated in response to PDGF in NIH 3T3 cells is responsible for observed sulfenylation of ERK1/2. NIH 3T3 cells were grown as in Fig. 2 and treated or not with PDGF for 10 min with or without preincubation with PEG-Catalase as described in Materials and Methods. PEG-Catalase significantly reduced the amount of sulfenic acid formation on total ERK (A) and phosphorylated ERK (B) in response to PDGF, but had no effect on total or phosphorylated ERK content in whole lysates, indicating that sulfenic acid formation on ERK is a result of cysteine reaction with H₂O₂ (or a derivative of H₂O₂). Lower panels summarize ERK band intensity from multiple experiments (n=4) normalized as described in Figure 2. Images are representative of four biological replicates. *, p<0.05; **, p< 0.01; ***, p<0.001; ****, p<1x10⁻⁸

Figure 4

ERK1/2 cysteines are oxidized to sulfenic acid in response to PDGF in WI-38 fibroblasts. Cells were serum-starved for 18 h, treated with PDGF, and lysed in the presence of DCP-Bio1 as described in Figures 1 and 2. Affinity-captured proteins were immunoblotted for total (A) or phospho-ERK1/2 (B). Lower panels summarize ERK band intensity from multiple experiments (n=4) normalized to AhpC, then to the sample with highest band intensity. *, p<0.05; **, p< 0.01; ****, p<5x10⁻¹⁰

Figure 5

ERK1/2 oxidation in ovarian cancer-derived SK-OV-3 cells treated with lysophosphatidic acid (LPA). SK-OV-3 cells depleted of serum for 18 h were treated with 100 nM LPA and harvested in the presence of DCP-Bio1 as described in Figures 1 and 2. The upper panels show representative immunoblots of DCP-Bio1 labeled proteins for total ERK (A) and TEY-phosphorylated ERK1/2 (B). The lower panels depict averaged relative intensity for each sample after LPA treatment, normalized as in Figure 4 (n=6 for total ERK, n=3 for phospho-ERK). *** p<0.001

Figure 6

Endogenous ERK oxidation inhibits kinase activity towards Elk1. Dual (TEY) phosphorylated ERK1/2 was immunoprecipitated from NIH 3T3 cells treated with PDGF for various time points as in Fig. 2 and as described in Materials and Methods. Each sample of immunoprecipitated ERK1/2 was then split into two: one sample was diluted into dithiothreitol (DTT)-containing buffer while the other was left in its native redox state (diluted into buffer lacking DTT). Recombinant Elk1 and ATP were added and samples were incubated at 30°C for 30 min. An immunoblot of S383-phosphorylated Elk1 was conducted to measure relative activities of immunoprecipitated ERK1/2. (A) Representative immunoblot of phosphorylated Elk1. (B) Normalized Western blot data from 5 independent replicates. DTT-treated samples are at least 5-fold more active toward Elk1 than non-reduced immunoprecipitates at all time points tested after PDGF addition. ***p< 0.001, ****p< 0.000005. (C) Representative immunoblot for dually (TEY) phosphorylated (top) and total (bottom) ERK1/2 from cell lysates.

Figure 7

Model for ERK phosphorylation and oxidation controlling activity in response to extracellular signals. Receptor (R) activation due to stimulus (S) binding (1) leads to MEK1/2 activation (2p) as well as concurrent NADPH oxidase (Nox) activation to produce H₂O₂ (2o). Activated MEK1/2 phosphorylates inactive ERK1/2 (gray or red) on Thr and Tyr residues (3p), and ERK1/2 is reversibly oxidized by H₂O₂ or a derivative thereof (3o) (our data suggest that neither modification precludes the other). Our data support a strong inhibitory effect of this oxidation on phospho-ERK. Note that dually phosphorylated AND oxidized ERK (yellow) is “primed” to be active once reduced (e.g., by thioredoxin or other dithiol). Note that this diagram does not emphasize the high degree of localization that is expected to be important in the posttranslational modification of ERK due to signal-mediated receptor activation.