Comprehensive analyses of the cysteine thiol oxidation of PKM2 reveals the effects of multiple oxidation on cellular oxidative stress response

Redox regulation of proteins via cysteine residue oxidation is known to be involved in the control of various cellular signal pathways. Pyruvate kinase M2 (PKM2), a rate-limiting enzyme in glycolysis, is critical for the metabolic shift from glycolysis to the pentose phosphate pathway under oxidative stress in cancer cell growth. The PKM2 tetramer acts as pyruvate kinase (PK), whereas the PKM2 dimer, which is induced by Cys358 oxidation, has reduced PK activity. Here, we identified four oxidation-sensitive cysteine residues (Cys152, Cys358, Cys423, and Cys424) responsible for three different oxidation forms. Possibly due to obstruction of the dimer-dimer interface, sulfenylation (-SOH) at Cys424 inhibited tetramer formation and PK activity. Cys423 is responsible for intermolecular disulfide bonds with heterologous proteins. In addition, intramolecular polysulfide linkage (–Sn–, n≧3) possible between Cys152 and Cys358 also is induced. We found that cells expressing the oxidation-resistant, constitutive-tetramer PKM2 (PKM2C358,424A) show a higher intracellular reactive oxygen species (ROS) and greater sensitivity to ROS-generating reagents and ROS-inducible anti-cancer drugs. These results highlight the possibility that PKM2 inhibition via Cys358 and Cys424 oxidation contributes to the elimination of excess ROS and oxidative stress.


Introduction
Oxidative modifications of protein cysteine residues represent post-translational modifications (PTMs). Notably, cysteine (Cys) residue can react with electrophilic-structured compounds such as reactive oxygen species (ROS), resulting in various chemical modifications. Hydrogen peroxide (H2O2), which is the most abundant ROS in aerobic organisms , reacts to Cys thiol (Cys-SH) and forms various types of oxidized cysteine residues such as disulfide, sulfenic acid, and sulfinic acid (1). Such oxidations alter the function of target proteins and, in many cases, are involved in a signaling cascade for defense mechanisms against oxidative stress (2)(3)(4)(5)(6)(7).
Activation of glutaredoxins and peroxiredoxins, which are major antioxidant systems, requires NADPH as a proton donor (8,9). The major pathway for NADPH production is the pentose phosphate pathway (PPP), which is a branch pathway of glycolysis. Therefore, under oxidative stress conditions, increases of the metabolic flux to PPP by glycolysis inhibition serves an important role in the detoxification of H2O2 (10)(11)(12). This metabolic alteration is known to be induced by the ROS-induced inactivation of glyceraldehyde-3phosphate dehydrogenase (GAPDH) and pyruvate kinase (PK) (13). GAPDH has an oxidationsensitive (redox-active) cysteine residue in the enzymatic active site. The cysteine residues are evolutionarily conserved and oxidized by various SH-oxidizing reagents and intracellular levels of H2O2, resulting in many modifications, such as disulfide bond formation (14)(15)(16)(17)(18)(19)(20). However, much of the mechanism of PK redox regulation remains unknown. In the present study, we identified multiple redox-active cysteine residues and oxidation forms that affect PK activity.
The PK of a rate-limiting enzyme for glycolysis catalyzes the conversion of phosphoenolpyruvate (PEP) and ADP to pyruvate and ATP (PK activity). This reaction is irreversible and represents the final step in glycolysis. In the yeast model system, inhibition of PK activity induces metabolic change under oxidative stress (21,22). In mammalians, there are four isoforms and two genes, with each isoform expressing in different tissues. The expression of PKL, PKR, and PKM1 is tissuespecific (23), whereas that of PKM2 is ubiquitous. PKM1 and PKM2, encoded by the PKM gene, depend on alternative splicing using two mutually exclusive exons (24): PKM1 that includes exon 9 is expressed in limited tissues (such as heart, muscle, brain), whereas PKM2 that includes exon 10 is ubiquitously expressed in many tissues and cells. Interestingly, the regulation of PKM2 enzymatic activity plays a critical role in cancer cell metabolism (25,26). As such, the appropriate regulation of PKM2 might confer metabolic advantages for tumor growth and progression since cancer cells tend to prefer metabolism via glycolysis rather than the oxidative phosphorylation pathway even in aerobic conditions. This is referred to as the Warburg effect (27,28). Tetramer PKM2 and dimer/monomer PKM2 exhibit higher and lower PK activity, respectively. Allosteric activators of PKM2 such as fructose-1,6-bisphosphate (FBP) and serine increase PK activity by promoting the tetramer form (29,30), while the PTMs of PKM2 (e.g., oxidation, phosphorylation, acetylation, and glycosylation) decrease PK activity by blocking the formation of the tetramer. The decrease of PKM2 activity via PTMs provides a metabolic advantage for the Warburg effect and facilitates cancer cell growth (31)(32)(33)(34). Thus, the appropriate inhibition of PK activity might induce the accumulation of intermediate metabolite of glycolysis and promote the synthesis of several amino acids and nucleotides including NADPH. As a result, this change fulfills the metabolic requirement for cancer cells (35). Moreover, dimeric PKM2 that is induced by several specific PTMs is responsible for regulation of gene transcription via interacting transcription factors in the nucleus (36,37). Several studies have reported that the moonlighting (noncanonical) functions of PKM2 serve a critical role in tumorigenesis (38). Taken together, in cancer cell growth, the tetramer-to-dimer transition of PKM2 provides two benefits. These are the downregulation of PK activity and activation of moonlighting functions.
Cysteine modifications (oxidations) are important PTMs in PKM2. An initial report indicated that Cys358 oxidation was crucial for oxidative stressinduced downregulation and the intermolecular disulfide bond formation of PKM2 (39). However, the molecular basis of PKM2 oxidation has not been elucidated. Whether another Cys might be responsible for the disulfide bond as a partner has not been resolved. Recently, two reports revealed that cysteine residues at 152, 326, 358, 423, and 424 were involved in modification (40,41). Thus, comprehensive observation of Cys oxidation of PKM2 gives insight into the role of PKM2 in the cellular oxidative-stress response.

Identification of novel oxidation-sensitive cysteine residues of PKM2
A previous report indicated that a specific cysteine residue (Cys358) of PKM2 was being involved in its redox regulation using diamide (39). Diamide is a widely utilized thiol oxidant and ROS generator. Recent global analyses on the oxidative modification of protein cysteine residues indicated that another cysteine residue (Cys424) in PKM2 was modified under oxidative stress (42,43). Moreover, as shown in Fig. S1, the PK activity of H1299 cells expressing PKM2 with a loss of Cys358 mutation (PKM2 C358A ) was still downregulated in the presence of H2O2 (See "Redox regulation of PKM2 contributes to oxidative stress response" section).
Thus, multiple cysteine residues might be responsible for the redox downregulation of PK activity. To identify cysteine residues that might be oxidation-sensitive, we performed a PEG maleimide-based gel shift assay. Sulfhydryl residues of whole-cell proteins were first blocked with N-ethyl maleimide (NEM). Then a reduction of preoxidized Cys via DTT treatment was performed, followed by the addition of PEGlyated maleimide (PEGM, MW 2,333) to NEM-free Cys (NEM-DTT-PEGM assay) (see Fig. 1A). As shown in Fig. 1B, the mobility of PKM2 (63 kDa) in SDS-PAGE shifted to a higher molecular size (100 kDa) when all 10 PKM2 Cys were assumed to be modified by PEGM under the condition without NEM. In contrast, the 63 kDa band was shifted to one band (66 kDa) and two bands (66 kDa and 70 kDa) when HEK293T and HepG2 cells were treated with H2O2 and diamide, respectively. However, tert-butyl hydroperoxide (tBHP) treatment did not affect the mobility of PKM2 (Fig.  1B). Although this oxidized PKM2 may be specific to the oxidant used, it seems likely that all three oxidation forms were efficiently induced by diamide. We found that H2O2 level in the diamidetreated cell culture was enhanced (Fig. 1C). Thus, possibly due to the oxidation of cellular reduction systems such as glutathione, diamide induces an intracellular H2O2 level that might result in more efficient Cys oxidation than the addition of H2O2 in the culture medium.
Next, to identify cysteine residues in PKM2 responsible for the shift (oxidation), we created 10 PKM2 mutants, in which each of 10 cysteine residues was replaced with alanine (Ala). H1299 shPKM2 cells expressing each of the PKM2 mutants were treated with oxidants. As shown in Fig. 1D, the appearance of the two molecularweight shifts by diamide were almost completely abolished by Cys424 mutation (PKM2 C424A ), while only the upper-shift band (70 kDa) was absent when Cys423 was mutated (PKM2 C423A , indicated as '**'; Fig. 1D). By contrast, a mutation on Cys358 (PKM2 C358A and PKM2 C358S ) did not affect PKM2 oxidation (Fig. 1D). The H2O2-induced 66 kDa shift was decreased by the mutation of either Cys424 (PKM2 C424A ) or Cys423 (PKM2 C423A ) (Fig.  1E). These results suggest that Cys424 and Cys423-but not Cys358-are found by NEM-DTT-PEGM assay to be sensitive to oxidation induced by H2O2 and diamide. Notably, Cys424 is a unique Cys mutant present in PKM2 but not in PKM1. Notably, we found that PKM1 was not oxidized by H2O2, and it was only partially oxidized by diamide (Fig. 1F). Taken together, we identified Cys423 and Cys424 as novel oxidationsensitive cysteine residues.

The characterization of oxidative modification of all cysteine residues
The formation of an intramolecular disulfide bond in particular proteins enhances the protein migration in SDS-PAGE under non-reducing conditions ( Fig. 2A) (3,44,45). A previous study indicated that faster-migrating PKM2 bands appearing in response to diamide were completely abolished by the mutation of Cys358 to serine (Ser) (39). We observed that the faster-migrating PKM2 appeared in response to 0.4 mM diamide, but not in response to H2O2 (Fig. 2B). Treatment with DTT ( Fig. 2B) and the substitution of Cys358 to Alabut not Cys424 to Ala- (Fig. 2C) abolished the formation. Moreover, the in vitro oxidation assay of recombinant PKM2 proteins (rPKM2) demonstrated that Cys358-not Cys424-is involved in the formation of the faster-migrating band, which is induced by both diamide and H2O2 ( Fig. 2D). Notably, we observed that this fastermigrating band was absent with NEM but not iodoacetamide (IAA) treatment (Fig. 2E). In general, a disulfide bond is not replaced by NEM. Therefore, the DTT-sensitive faster-migrating band of PKM2 may be a NEM-sensitive linkage between Cys358 and other Cys. Recent findings have indicated that polysulfide can link between two cysteine residues (-Sn-, n≧3), and that the linkage is cleaved by NEM but not by IAA (46). Therefore, we investigated the possibility of polysulfide formation using a generator of Cys polysulfur, Na2S4. The faster-migrating band of PKM2 was strongly induced by Na2S4, but it disappeared with the addition of DTT and NEM (Fig. 2F). Furthermore, the faster-migrating band induced by Na2S4 was not detected in the recombinant PKM2 C358A mutant (Fig. S2). Based on these results, we concluded that the diamideinduced faster-migrating PKM2 could be due to an intramolecular linkage via polysulfide bond, but not via a disulfide bond between Cys358 and other Cys within the same PKM2 molecule.
To identify a putative partner Cys for the Cys358polysulfide linkage, we examined the fastermigrating band formation by using cells expressing each of the Cys mutants of PKM2. As shown in Fig.  2G, the faster-migrating band almost completely disappeared when Cys152 was mutated (PKM2 C152A in Fig. 2G). The distance in 3D structure between Cys358 and Cys152 is approximately 36 Å (PDB ID: 4B2D). Therefore, the formation of polysulfide may be required to form the linkage. In addition, we noticed multiple bands with high molecular weight (MW130-180 proteins, as indicated by # in Figs. 2B and 2G) that were sensitive to DTT treatment ( Fig. 2B) were induced in response to oxidative stress by both diamide and H2O2 (Fig. 2B). Since PKM2 C423A failed to form a multiple band with high molecular weight (Fig. 2G), Cys423 could be responsible for the intermolecular disulfide bond between PKM2 molecules and/or between a PKM2 molecule and unknown protein(s).
Next, we investigated the effects of Cys424 oxidation on the multimer formation of PKM2. The PK activity of PKM2 is activated by tetramer formation, whereas dimers and monomers of PKM2 are less active (51,52). We treated cell lysate with GA to crosslink multimer proteins, which are separated by SDS-PAGE under reducing conditions. As previously indicated (53), the tetramer formation of PKM2 was abrogated by S437Y mutation (Fig. S4). We observed that the tetramer was markedly decreased by treatment with diamide ( Fig. S4), while the remaining levels of the tetramer under diamide treatment were higher in PKM2 with the simultaneous mutation of C358A and C424A (PKM2 C358,424A ), or even only a C424A mutation (PKM2 C424A ). Thus, the inhibition of tetramer formation may be partly due to the oxidation of Cys424. To clarify the effect of sulfenylation Cys424 on multimer formation, recombinant PKM2s (rPKM2s) were used for further investigation. The tetramer formation of rPKM2 WT was significantly decreased by treatment with H2O2, while its formation of rPKM2 C424A was not affected (Fig. 3D). Interestingly, Cys424 is located on the surface of dimer-dimer interaction, and it is considered a crucial residue for tetramer formation (54). Furthermore, a previous study has demonstrated that the mutation of Cys424 to a hydrophobic residue (e.g., leucine, C424L) increases PK activity, whereas, in the case of hydrophilic residues (e.g., serine, C424S) this mutation decreases PK activity (40). As shown in Figs. 3E and 3F, the tetramer ratio of rPKM2 C424A is not affected by its mutation (tetramer ratio: WT = 0.6, C424A = 0.63), while the tetramer ratio of the hydrophilic mutations rPKM2 C424D (aspartate, C424D) and rPKM2 C424S , which might mimic Cys424 sulfenylation, were lower than that of rPKM2 WT (tetramer ratio: C424D = 0.37, C424S = 0.32, Fig. 3F). Furthermore, the PK activity of these mutants (rPKM2 C424D and rPKM2 C424S ) was lower than that of rPKM2 WT (Fig. 3G). Taken together, it is possible that sulfenylation of Cys424-which increases the hydrophilicity of Cys-SH-disrupts tetramer formation by increasing hydrophilicity on the dimer-dimer interface and contributes to the inhibition of PK activity.

Redox regulation of PKM2 contributes to oxidative stress response
To assess the functional significance of the redox regulation of PKM2, we examined the importance of the redox-sensitive cysteine residues on the PK activity and oxidative stress sensitivity of cells. PKM2 WT -Flag and its Cys mutants were introduced in H1299 cells, of which endogenous PKM2 was stably knocked down using a lentivirus-based shRNA expression vector (H1299 shPKM2 cells; Figs. S5A and S5B). First, we investigated the effect of cysteine oxidation on the PK activity. The aforementioned results indicate that mutations on Cys358 and Cys424 do not affect oxidative stressinduced suppression of PK activity (Fig. S1). Thus, we created a PKM2 mutant with a simultaneous mutation on Cys358 and Cys424 (PKM2 C358,424A ).
The oxidation of PKM2 treated with H2O2 or diamide was almost completely abrogated in cultured cells expressing PKM2 C358,424A (Fig. S6). In addition, the induction of the faster-migrating (the intramolecular polysulfide) rPKM2 by H2O2 and diamide (Fig. 4A) and the sulfenylated rPKM2 by H2O2 (Fig. 3B) were not detected in rPKM2 C358,424A in vitro. Although the PK activity in the lysate of H1299 cells expressing PKM2 WT was significantly increased by DTT treatment, the levels of PKM2 C358,424A were not increased as mach (Fig. 4B). Similarly, the decreased level of PK activity in the cells expressing PKM2 WT in response to H2O2 and diamide was lower in the cells expressing PKM2 C358,424A (Fig. 4B). Thus, these results indicated that Cys358 and Cys424 are responsible for oxidation-induced suppression of PK activity of PKM2.
The inhibition of PKM2 enzymatic activity alters the flow of metabolites into PPP, and it contributes to NADPH generation and oxidative stress response (39,55). Therefore, we investigated the intracellular level of NADPH and the accumulation of intracellular ROS under oxidative stress conditions.
The NADPH ratios (NADPH/(NADPH+NADP)) in cells were decreased in response to the diamide treatment (Fig.  4C). These decreases were marked in cells expressing PKM2 C358,424A (Fig. 4C). In addition, the intracellular ROS levels were enhanced in response to oxidative stress (H2O2 and diamide). Again, ROS levels in response to H2O2 and diamide in cells expressing PKM2 C358,424A were higher than in those expressing PKM2 WT (Fig. 4D). Next, to examine the effect of continuous oxidative stress on cell viability, we added glucose oxidase (GO) to the culture medium. This generates H2O2 by glucose catalysis (56). Cells expressing PKM2 C358,424A were more sensitive than those expressing PKM2 WT (Fig.  4E). These results highlight the possibility that PKM2 inhibition via Cys358 and Cys424 oxidation contributes to the elimination of excess ROS and oxidative stress.
Some anti-cancer drugs induce intracellular ROS, which gives rise to cytotoxic cancer cells (57,58). Therefore, the ability to eliminate ROS is an important factor for chemosensitivity. We tested whether the oxidation of PKM2 is involved against the chemosensitivity of cisplatin (CDDP) and doxorubicin (DOX). As shown in Fig. 4F, the cells expressing PKM2 C358,424A were more sensitive than those expressing PKM2 WT . These results suggest that the oxidative stress response via the Cys358 and Cys424 oxidation in PKM2 might be involved in the sensitivity of CDDP and DOX. Collectively, our findings indicate a mechanism for oxidative stress resistance via the downregulation of PKM2 by redox-based modification of Cys358 and Cys424 to enhance NADPH levels through the potential activation of PPP under oxidative stress conditions.

Discussion
In the present study, we demonstrated the existence of three different oxidation statuses of PKM2 (sulfenylation, disulfide, polysulfide). First, we identified Cys423 and Cys424 as novel redoxactive cysteines using NEM-DTT-PEGM assays (Figs. 1A, 1B, and 1D). This method facilitates the detection of NEM-non-reactive yet DTT-sensitive cysteine oxidation, which includes Cyssulfenylation (SOH) and Cys-Cys disulfide. Our results suggest that oxidation induces Cys424 sulfenylation (Figs. 3A and 3B). Cys424 is located at the interface of the dimer-dimer interaction of the tetramer PKM2. Interestingly, Cys424 is a unique cysteine residue in PKM2, but not in PKM1. The corresponding residue of PKM1 is leucine. It has been shown that increasing hydrophobicity at the 424 residues (e.g., from C424A to C424L) enhanced PKM2 in an active state (tetramer) because of increased dimer-dimer surface interaction (54). Conversely, Asp and Ser substitution of PKM2 Cys424, which might decrease hydrophobicity, suppresses the formation of the active form (Figs. 3E and 3F). Thus, Cys424 sulfenylation might increase hydrophilicity on the dimer-dimer interface, thereby leading to decreased tetramer formation (Figs. 3D-3F), suppression of PK activity (Fig. 3G), and oxidative stress resistance (Fig. 4E). Second, we also demonstrated that PKM2 Cys423 is involved in intermolecular disulfide bonding with other proteins (Fig. 2G). However, since the PKM2 C423A mutant was enzymatically inactive (Fig. S7), we could not pursue further analysis to understand the effect of Cys423 oxidation on PK activity. Since the corresponding cysteine residue in PKM1 was conserved, the possible partner protein linking Cys423 may be crucial for PK activity and important for identification in future studies. Third, we revealed that Cys358 and Cys152 are essential for the formation of a faster-migrating band under non-reduced SDS-PAGE ( Fig. 2C and 2G). The band completely disappeared following treatment with an SH-alkylation reagent, NEM (Figs. 2E and  2F). An intramolecular disulfide bond, which is induced by H2O2 in many cases, is not reactive to NEM. Recently, Ida et al. have reported protein Spolythiolation, which is more unstable and highly reactive (59). We observed that an intracellular generator of reactive-sulfur species (RSS) induced intramolecular Cys-Cys linkage (Fig. 2F). This suggests that an intramolecular polysulfide linkage between Cys152 and Cys358 enhances the fastermigrating band. Although diamide induced possible Cys S-polythiolation of PKM2, it was unknown whether intracellular RSS was induced by diamide, and its oxidation mechanism remains unclear.
The inhibition of PKM2 by oxidants contributes to the metabolic change required to enhance metabolites in PPP (39). Our findings indicated that cells expressing the oxidation-resistant form of PKM2 (PKM2 C358,424A ) increased intracellular ROS, and they were more sensitive to ROS-generating reagents, such as glucose oxidase, and ROSinducible anti-cancer drugs, such as cisplatin and doxorubicin (Figs. 4D-4F). Thus, the inhibition of PKM2 via Cys358 and Cys424 oxidation is important to the oxidative stress response.
In the tumor progression process, cancer cells are exposed to higher levels of oxidative stress (60). As such, compounds that increase intracellular ROS are considered anti-cancer agents that enhance cell death (61,62). PK activity may be repressed due to the oxidation of PKM2 in the oxidative tumor environment. Our findings suggest that the specific inhibition of Cys358 and Cys424 oxidation to reduce PK suppression may represent ideal targets for anti-cancer drugs.

Experimental procedures
Cell lines, cell culture, and the establishment of PKM2 knockdown A human non-small cell lung carcinoma cell line H1299 and HepG2 were obtained from American Type Culture Collection (ATCC). H1299 is a human non-small cell lung carcinoma cell line. HepG2 is a hepatocellular carcinoma HepG2 cell line. HEK293T (Human Embryonic Kidney cells 293) cells were obtained from Thermo Fisher Scientific, MA, USA. These cells were cultured in DMEM (Nissui, Tokyo, Japan or Nacalai Tesque, Kyoto, Japan) supplemented with 10% fetal bovine serum (Biosera, Kansas, MO, USA), 3% glutamine (Nacalai Tesque), penicillin/streptomycin (FUJIFILM Wako, Osaka, Japan), and 7.5% NaHCO3 (Nacalai Tesque) at 37℃ and 5% CO2 concentration. Stable knockdown of endogenous PKM2 was performed with a lentivirus expressing a high pyruvate kinase knockdown efficiency shRNA (63). The shPKM2 sequence was cloned into AgeI-EcoRI sites of a pLKO.1 puro vector (pLKO-shPKM). The lentivirus vectors were produced in HEK293FT cells by the co-transfection of plasmids pMD2.G and pCMVR8.74 with the pLKO.

Expression of PKM2 in H1299 shPKM2 cells
Human PKM2 cDNA (NM_002654.6) was fused with a Flag-tag to the corresponding C-terminus of the PKM2 sequences. PKM2-Flag was cloned into a pEB multi-hygro vector (FUJIFILM Wako) between KpnⅠ and NotⅠ. The PKM2 coding sequence corresponding codon 106-112 was modified as 5'-GCcGTcGCcCTg-3' (the lowercased letters indicate substitutions) to perform shRNA-resistant expression. To constract PKM2 cysteine mutants, corresponding cysteine codon was mutated to Ala (GCC) or Asp (GAC) or Ser (AGC) using PCR. These plasmids were then transfected in H1299 shPKM2 cells using FuGENE HD transfection reagent (Promega, Madison, WI, USA) and cultured in DMEM for 48 h (Fig. S5A).

Maleimide-based gel shift assay (NEM-DTT-PEGM assay)
After cells were treated with each oxidant at the indicated time, they were incubated for 5 min in 100 mM NEM in PBS (Nacalai Tesque) to block free thiols. Cells were then lysed in PK lysis buffer (50 mM Tris-HCl, pH 7.5, 1 mM EDTA, 150 mM NaCl, 1% Igepal-630 (Merck)) supplemented with 50 mM DTT and protease inhibitors (cOmplete TM protease inhibitor cocktail EDTA-free, Roche, Basel, Switzerland)). Lysates were centrifuged (13,800 g, 5 min, 4℃). Supernatants were collected in new tubes and then incubated for 10 min at room temperature to reduce the oxidized thiols. To remove DTT, protein in these lysates were precipitated in 5% trichloroacetic acid (TCA)-75% acetone for 10 min on ice. The proteins were precipitated by centrifugation (13,800 g, 2 min, 4℃). The pellets were then washed with 500 µL acetone and collected by centrifugation. This step was repeated three times to neutralize the pellets. After neutralization, the pellets were suspended again in urea PEGM buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 8 M urea, 1% SDS, 5 mM PEGM). These samples were then mixed in SDS sample buffer (50 mM Tris-HCl, pH 6.8, 10% glycerol, 2% SDS, 0.01% bromophenol blue) supplemented with 50 mM DTT. SDS-PAGE was then performed within 30 min of sample preparation.

Non-reduced SDS-PAGE
After cells were treated with each oxidant for the indicated time, they were incubated for 5 min in 100 mM IAA in PBS to block free thiols. Cells were then lysed in PK lysis buffer supplemented with 50 mM IAA and protease inhibitors. Lysates were centrifuged (13,800 g, 5 min, 4℃), and the supernatants were collected in new tubes. These samples were mixed with SDS sample buffer (without DTT), and SDS-PAGE was performed.

Detection of PKM2 sulfenylation in cultured cells
H1299 shPKM2 cells transfected with pEB PKM2-Flag were treated with oxidants. To block free thiols, cells were incubated with 100 mM NEM in PBS at 37℃. Then, cells were washed with PBS and lysed in a dimedone-containing lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 5 mM dimedone (Nacalai Tesque), 1% Igepal, protease inhibitors). After removal of the unresolved fraction by centrifugation, the supernatants as whole-cell lysates (WCLs) were quantified using a DC Protein Assay kit (Bio-Rad). Then, 600 µg WCL was mixed with anti-Flag antibody beads (Merck) at 4℃ for 3 hr with rotation. The beads were then washed with the lysis buffer three times. Bound proteins were eluted from the beads with 90 µL of sample buffer containing 50 mM DTT at 95℃ for 5 min. The immunoprecipitates were analyzed by SDS-PAGE and western blotting using the anti-PKM2 antibody and rabbit anti-SOH (antidimedone) antibody, which specifically reacts with dimedone-bound cysteine residues.
Detection of rPKM2 sulfenylation rPKM2 WT or its mutants (0.5 µg) were diluted in PBS. These proteins were treated with H2O2 for 10 min. Then, dimedone was added to the protein sample for a final concentration of 0.25 mM and incubated at room temperature for 30 min. The reaction mixtures were combined with a sample buffer supplemented with 2-mercaptoethanol (Nacalai Tesque) and analyzed by SDS-PAGE and western blotting, as previously described.
Detection of rPKM2 multimer rPKM2 WT or its mutants (0.5 µg) were diluted in PBS and pre-incubated with 1 mM FBP and 5 mM DTT at 37℃ for 30 min. Glutaraldehyde (GA) (Nacalai Tesque) was added to the protein sample for a final concentration of 0.025% and incubated at 37℃ for 3 min. The reaction mixtures were then mixed with a sample buffer supplemented with 50 mM DTT and analyzed by SDS-PAGE and western blotting using the anti-PKM2 antibody.

Detection of PKM2 multimer in cultured cell
Cells were lysed in PBS with the addition of 1% Igepal, and protease inhibitors. GA was added to the cell lysate for a final concentration of 0.025% and incubated at 37℃ for 3 min. The reaction mixtures were then mixed with a sample buffer supplemented with 50 mM DTT and analyzed by SDS-PAGE and western blotting using the anti-Flag antibody. Measurement of NADPH ratio and intracellular ROS Measurement of NADPH ratio was performed using an NADP/NADPH-Glo TM assay kit (Promega), according to the manufacturer's protocol. We used CellRox Orange Reagent (Thermo Fisher Scientific) and ROS-Glo (Promega) to measure total H2O2 level in the cell culture (medium and cells    The non-GA-treated sample is indicated as "GA (-)". D. Inhibition of tetramer formation of rPKM2 by H2O2 treatment. The indicated concentration of H2O2 was treated for 120 min before GA treatment. E. rPKM2 WT and Cys424 mutants were pre-incubated with DTT and FBP and treated with GA (+) or untreated (-). F. The relative intensity of the specific bands of PKM2 tetramer, dimer, and monomer was calculated using Image Lab software (Bio-Rad). The tetramer ratios are expressed as the mean +/-the standard error of the mean (N = 3). Statistical comparisons were performed using Dunnett's post-hoc test.

Measurement of pyruvate kinase activity
Differences were considered to be significant at *p < 0.05, vs. WT. G. rPKM2 WT and rPKM2 Cys424 mutants that were pre-incubated with DTT were diluted with PBS, and the measurement of pyruvate kinase activity was performed. Results were normalized to WT and expressed as the mean +/-the standard error of the mean (N = 3). Statistical comparisons were performed using Dunnett's post-hoc test.
Differences were considered to be significant at *p < 0.05, vs. WT.