Introduction

Altered metabolism is a key feature of tumor cells [1,2,3,4]. In early 1920’s, Otto Warburg proposed a unique metabolic model for tumor cells [5]: they generate energy even under aerobic conditions by glycolysis to produce enough ATP to satisfy the requirements of proliferative cells [6,7,8]. For these cells to proliferate, a glycolytic flux needs to be constrained to allow a higher percentage of intermediates to be diverted to biosynthesis, e.g., availing G6P (glucose-6-phosphate) to the pentose phosphate pathway (PPP), thus an exquisite balance between fluxes of glycolysis and PPP that benefits tumor development. The pentose phosphate pathway is the major path to produce NADPH via two key enzymes G6PD (Glucose-6-phosphate Dehydrogenase) and PGD (6-phosphogluconate dehydrogenase), which provides the reducing power to sustain redox balance. In addition, ribose-5-phosphate as a key metabolic intermediate of the pentose phosphate pathway is the precursor for DNA/RNA synthesis, vital for the growth of cancer cells [9,10,11].

Due to aerobic glycolysis and deformed angiogenesis in solid tumors, nutrients such as glucose and oxygen are very scarce, especially in the inner part. Thus, glucose deprivation and high levels of lactic acid eventually form the tumor metabolic microenvironment. Reportedly, glucose deprivation boosts up the reactive oxygen species and causes the death of cancer cells through inhibiting glycolysis and the pentose phosphate pathway [12], and lactic acid is known to promote NADPH production and maintains redox balance [13]. However, an involvement of the pentose phosphate pathway and mechanistic description of lactic acid-mediated up-regulation of the major reducing power production path remain unaddressed.

Lactic acid is the end product of aerobic glycolysis [14, 15]. This metabolite is quite often dismissed as a waste product, but recent evidence points to its roles in tumorigenesis beyond serving a fuel molecule [16, 17]. For example, lactic acid helps cancer cells to acquire growth advantage in metabolic stress via a feedback mechanism modulating the activities or expression of key glycolytic enzymes such as hexokinase and PKM2 [18, 50]. To investigate a clinical relevance of the lactic acid-responsive G6PDY249/322 Tyr phosphorylation and G6PD-SRC-GSTP1 complex, and to establish their potential correlation with breast cancer types, we obtained 13 human breast cancer (T) samples along with adjacent tissues (N) characterized to be normal, and made lysate for detailed IP analyses (Fig. S6A). The G6PD Y249/Y322 phosphorylation was drastically decreased or even absent in most tested human breast cancer samples (Fig. 6A, B). Moreover, the levels of immuno-enriched SRC and GSTP1 were lower in human breast cancer samples than in normal tissues when normalized to endogenous G6PD protein controls (Fig. 6A,C), which is consistent with in vitro data implying that the dynamic G6PD Tyr phosphorylation was related to a changing tripartite complex in response to lactic acid.

Fig. 6: The phosphorylation status of G6PD was positively related to the immunoprecipitated SRC/GSTP1 in the ER/PR positive type or Her (-) type of breast cancer.
figure 6

A Immunoprecipitated SRC/GSTP1 and Y249/Y322 phosphorylation status of G6PD that was immuno-enriched from clinical human breast cancer samples, as determined by Western-blots against the tumor (T) and adjacent normal tissues (N). B The G6PD Tyr phosphorylation status at 2 sites in (A) were quantified and statistically analyzed. Data are normalized to G6PD (IP) protein, presented as mean ± SD and analyzed by unpaired Students’ test, n = 13 pairs. P(Y249) = 0.0507, P(Y322) = 0.0009. C The immunoprecipitated GSTP1 and SRC expression levels in (A) were quantified and statistically analyzed. Data are normalized to G6PD (IP) protein, presented as mean ± SD and analyzed by unpaired Students’ test, n = 13 pairs. P(SRC) = 0.0009, P(GSTP1) = 0.0162. D A possible correlation between relative SRC/GSTP1 (IP) levels and relative p-Tyr levels of G6PD was determined in various types of breast cancer samples. Data are analyzed and shown by Pearson Correlation Coefficient. ER + /PR + : 5pairs; ER-/PR-: 4 pairs; ER-PR + : 3 pairs; HER-: 10 pairs.

Full size image

We next examined the relationship between the IP-enriched SRC/GSTP1 protein and the overall G6PDY249/322 phosphorylation levels in these clinical samples upon classifying them into types based on ER/PR/HER2 expression patterns. The progesterone receptor is the downstream target of the estrogen receptor, so we take progesterone receptor and estrogen receptor together into consideration. SRC/GSTP1 expression and G6PDY249/322 phosphorylation were positively correlated in breast cancer tissues characterized to be ER (+) and PR (+) or HER (-); however, this correlation was not evident in ER(-)/PR(-) or ER(-)/PR(+) or HER (+) tumor tissues (Fig. 6D, Fig. S6B). Further, the IP-enriched SRC/GSTP1 protein and G6PDY249/322 Tyr phosphorylation status were strongly and positively correlated with the breast cancer that exhibit ER expression and PR expression, to which our experimental cell line MCF-7 belongs. Hence, the G6PDY249/322 Tyr phosphorylation sites and LAC-G6PD-SRC-GSTP1 axis are potentially a novel target for the ER/PR positive type or Her (-) type of breast cancer.

Discussion

Many a signaling pathway necessitates multiplex protein-protein associations via dynamic complex that receive and forward diverse proliferative or other signals, which in essence function as “signalosomes”. We, for the first time, show that the tumor metabolic microenvironment lactic acid activates a “signalosome” comprising GSTP1, SRC and G6PD by attenuating the complex, which facilitates a reprogrammed PPP to promote cancer cell growth. Such a dynamic pattern of a complex is also manifested in certain breast cancer subtypes of human clinical specimens. Hence, these findings unravel crucial roles of lactic acid during malignancies and imply strategies for cancer intervention through metabolic means by targeting individual components involved in, or by perturbing the formation of, the complex as a druggable “signalosome”.

Roles of lactic acid during malignancies

Wealth of evidence points to essential roles of lactic acid during malignancies. Aside from serving a fuel molecule [51] or a carbon source for other metabolites, and aside from an involvement in gene regulation transcriptionally or post-translationally [52], lactic acid can function without net chemical consumption thus in essence a signaling molecule. Indeed, we found that the proton reduced the glycolysis flux by suppressing the catalytic functions of multiple glycolytic enzymes, whereas a high level of lactate facilitated the glycolysis towards reaching an equilibrium [19]. Such a combinatorial form of metabolic regulation employing both elements of lactic acidosis would in theory limit a glycolytic flux preventing further glucose fermentation in the direction of lactate production thus intercepting glycolytic intermediates. Given ample glucose (a likely scenario with outer-sphere cells of solid tumors), cells can afford the Warburg Effect to generate sufficient intermediates for biomass building; however, the glycolytic flux must be kept in check in case of glucose deprivation (a scenario with inner-sphere cells of solid tumors). In the latter scenario, the lactic acidosis-based interception mechanism would allow more glycolytic intermediates to be diverted to biomass building, which, as our current work unravels, also includes metabolizing G6P towards the PPP for nucleotides synthesis (R5P) and reducing power NADPH (anti-oxidation and reductive biosynthesis). Thus, lactic acid constitutes a crucial part of a signaling cascade that allows optimal partitioning of glucose metabolism intermediates for energetics and reductive biosynthetic and anti-oxidation capabilities, and in so doing plays essential roles during malignancy of solid tumors.

This work mechanistically addresses the question regarding how the lactic acid signaling functions. G6PD, the first rate-limiting enzyme of the PPP oxidative branch, is regulated by phosphorylation at Y249/322 by SRC (Fig. 2). GSTP1 up-regulates the G6PD phosphorylation by recruiting SRC as part of the tripartite complex, which might completely impede PPP; however, lactate binds to GSTP1 (Fig. 1) that in turn attenuates the tripartite GSTP1-SRC-G6PD interaction (Fig. 3), altering the G6PD phosphorylation status and augmenting NADPH levels by up-regulating the G6PD catalytic activity via GSTP1 (Figs. 4, 5). We emphasize that both elements of lactic acidosis contribute: a high level of lactate facilitates an equilibrium of glycolysis and proton squelches a number of glycolytic enzymes, which in a combinatorial way intercepts more intermediates for biosynthesis especially diverting G6P to PPP for more R5P and NADPH. Hence, lactic acid can function as a tumor metabolic micro-environment that reverses the above-mentioned PPP impediment, which in turn robustly supports proliferation of (severely) glucose-deprived cells, a situation that inner-sphere solid tumor cells quite often encounter.

Function of the GSTP1-G6PD-SRC complex as a metabolic “signalosome”

Our work implies that GSTP1 plays a dominant role linking G6PD and SRC to form a “signalosome” that provides a basis, under metabolic stresses, for robust cancer cell proliferation upon lactic acid signaling, which could regulate many aspects of cellular metabolism. Given roles of lactate in promoting redox (re)balance and PPP pathway via GSTP1 (Fig. 5), the defined GSTP1 regions involved in binding with lactate (Fig. 1) are likely potential targets for blocking the lactic acid signaling, paving the way of searching for antagonist(s). Of note, defined GSTP1 lactate binding regions are also regions that mediate interactions with G6PD/SRC.

GSTP1

Of the trio involved in the above-mentioned “signalosome”, high GSTP1 expression levels have in fact been reported in diverse human malignancies including breast, colon and ovarian cancer, and an involvement of GSTP1 in tumorigenesis and drug resistance in general has long been recognized [53, 54]. In other studies, NBDHEX allosterically binds to GSTP1 leading to activation of JNK via the interaction of the JNK-GSTP1 axis, and the conformation of GSTP1 is modulated upon binding to certain anti-cancer drugs [55]. But the interaction between GSTP1 and JNK didn’t change when treated with or without lactic acid (Fig. S7A). In addition, GSTP1 can mediate the S-glutathionylation of some proteins and interact with them [56] such as Keap1 [57] and PKM2 [58]. We also checked the S-glutathionylation of G6PD and SRC under lactic acidosis. G6PD could be S-glutathionylated (Fig. S7B) and the modification level decreased under lactic acid (Fig. S7C). Previous paper reported that GSTP1 promoted the S-glutathionylation of SRC under TNF-α stimulation [59], but is not found under lactic acidosis in our paper (Fig. S7D). Hence, especially given a crucial role of GSTP1 in receiving and forwarding the lactic acid signaling as we unravel, intervention targeting the GSTP1 protein could impede multiple metabolic pathways in diverse cancers. Furthermore, small molecules such as LAS17 [60] that suppresses the GSTP1 enzymatic activity or protein expression are predicted to have therapeutic potentials. We observed a tumor suppressing activity by a GSTP1 enzyme inhibitor NBDHEX in animal models, presumably by blocking the function of GSTP1 as a lactic acid sensor that impacts the GSTP1-G6PD-SRC complex (Fig. 5).

SRC

The oncoprotein c-SRC once activated is capable of potently promoting cell proliferation, adhesion and migration [61, 62]. SRC is known to take part in regulating the activities of glycolytic enzymes thus overall flux. In scenarios with activated c-SRC, however, the ROS levels are dramatically elevated [63, 64]. A balanced ROS is vital for cellular physiology and is regulated via redox pairs, particularly by fluctuating the NADPH/NADP+ ratio and/or NADPH/NADP+ recycling in conjunction with the GSH/GSSH recycling [65, 66]. In extreme, cancer cells acquire anti-ROS capacities by establishing new redox balances that can counteract even more than 100-fold-elevated ROS levels. This might be how augmented PPP comes into play: generating more NADPH or facilitating an NADPH/NADP+ recycling while accumulating the building block R5P for DNA/RNA synthesis. It is especially crucial if cancer cells encounter scarce glucose but necessitates a signaling from lactic acid to facilitate the PPP. SRC inhibited the G6PD catalytic activity via Tyr-phosphorylation (Fig. 4A). Manifesting opposing effect(s) of a presumed onco-protein as it may seem, it allows lactic acid to be sensed by GSTP1 to activate a cascade leading to augmented PPP that overcomes ROS levels elevated by activated c-SRC. That lactic acid attenuated the G6PD Y249/Y322 phosphorylation and boosted its enzyme activity (Fig. 4B) is consistent with a scenario in which the elevated ROS levels by c-SRC activation are squelched by augmented PPP under glucose deprivation and with a lactic acid signaling.

G6PD

G6PD is a rate-limiting enzyme of the PPP, and its expression is elevated in diverse cancers. The TCGA database positively correlated elevated abundance of G6PD with poor prognosis of cancer patients [67, 68]. We identified two new (Y249 and Y322) phosphorylation sites (Fig. 2) influenced by the tumor metabolic microenvironment lactic acid sensing by GSTP1 (Fig. 1), which frees up G6PD from a tripartite complex (Fig. 3) to facilitate the PPP oxidative branch (Fig. 4). Thus, a GSTP1-G6PD-SRC complex might allow this metabolic “signalosome” to orchestrate responses of cancer cells to diverse forms of metabolic stresses typified by glucose deprivation and, with a lactic acid-triggered mechanism, facilitate biomass building and redox (re)balance to support robust cell proliferation (Graphical Abstracts). Our newly-defined G6PD Y249/Y322 phosphorylation sites and a regulatory mechanism centered upon the GSTP1-G6PD-SRC complex, especially given observations on clinical samples, could be more physiologically relevant during malignancy. However, G6PD reportedly undergoes modifications such as phosphorylation at Y112 and Y428 [69, 70] in HCT116 cells, acetylation at K403 [71] in leukaemia cells and glycosylation at S84 in lung cancer cells [72]. The above modifications reported were detected under different growth factors stimulation, and might not reflect the genuine situation of cancer cells that quite often encounter metabolic stresses in diverse forms. Therefore, Such difference may be caused by different cell types or other factors, which needs to be investigated later.

Perspectives

A more detailed description of molecular impacts exerted by GSTP1 as a lactic acid sensor remains to be elucidated, along with the normal physiological function of the LAC-G6PD-GSTP1 axis. Our findings imply that GSTP1 plays key roles in regulating metabolism subject to modulation upon sensing the tumor metabolic microenvironment lactic acidosis. In a hindsight, all components of the tripartite complex are metabolic enzymes that participate in maintaining the redox balance. Targeting cellular redox systems, in particular those that involve the NADPH/NADP+ pair, might pave the way for more efficacious metabolic intervention than targeting cellular energetics. Our findings would broaden strategic options for tumor intervention by metabolic means.

Methods and Materials

Human breast cancer samples

Human breast cancer samples were obtained from the Breast Surgery Department of the Second Affiliated Hospital of Zhejiang University School of Medicine with approval of the ethics committee of the hospital and informed consents from all patients. In total, 13 cases of human breast cancer tissues were obtained, all from women, and all individuals were over 40 years old. Each breast cancer sample was paired with an adjacent normal sample from the same patient. Prior to analyses, all samples were stored in liquid nitrogen to keep the integrity of cellular proteins.

Cell lines

Human breast cancer cell lines MCF-7 and Bcap37 were gifts from, respectively, Laboratories of DONG Chenfang and HU Xun of Zhejiang University. Other cell lines (HEK293T, HepG2 and HeLa) were purchased from Cell Bank of Type Culture Collection of the Chinese Academy of Sciences, Shanghai, China. All cell lines were proven to be free of mycoplasma contamination.

Cell culture conditions

Unless otherwise noted, cells were cultured in Dulbecco Minimal Essential Medium (DMEM) supplemented with 1% penicillin/streptomycin sulfate and 10% fetal bovine serum (FBS) in a 37 °C incubator with 5% CO2 atmosphere. For plasmid and siRNA transfection assays, media were devoid of antibiotics. For glucose starvation, media with normal glucose level were removed, cells washed twice with 1X PBS, and glucose-deprived media (final glucose level at 0.5 mM) with or without lactic acid added for 12 h. For kinase inhibitor treatments, cells were administered by the following: Dasatinib, 5 µM, 12 h, Selleck Chemicals; Ponasatinib, 5 µM, 12 h, Selleck Chemicals; Ruxolitinib 5 µM,12 h, Selleck Chemicals; IPA-3, 10 µM, 12 h, Selleck Chemicals; Crizatinib, 10 µM,12 h, Selleck Chemicals; Sorafenib Tosylate, 20 µM,12 h, Selleck Chemicals; Getifinib, 2 µM,12 h, Selleck Chemicals; AZD4547, 20 µM,12 h, Selleck Chemicals,NBHDEX, 5 µM or 10 µM, 12 h, MCE. All inhibitors were used at dosages suggested by the suppliers.

Antibodies

The following antibodies were purchased from Cell Signaling Technology: GSTP1 (3F2) (3369; 1: 2,000), SRC (36D10) (2109; 1:2,000), Rabbit IgG (Light-Chain Specific) (D4W3E) (93702; 1:2,000), Mouse IgG (Light-Chain Specific) (D3V2A) (58802;1:2,000), and (Ser/Thr) Phe (9631;1:1,000). The following antibodies were purchased from Santa Cruz Biotechnology: p-Tyr (PY99) (sc-7020; 1:1,000), and HA-tag (F-7) (sc-7392; 1:5,000). G6PD (SAB5300459; 1:8,000) was purchased from Sigma-Aldrich; another GSTP1 (ab153949; 1:2,000) was purchased from Abcam; Flag (30502ES20; 1:10,000) was obtained from YEASEN; GFP (R1312-2; 1:4,000) and Actin (EM21002; 1:10,000) was purchased from HuaAn Biotechnology. To create site-specific antibodies for Tyr 249 and 322 of G6PD (Anti-pG6PD (Y249) and Anti-pG6PD (Y322)), synthesized peptides CEPFGTEGRGGY(pi)FDEF and CGEATKGY(pi)LDDPTVP (HuaAn Biotechnology) were coupled to a protein carrier prior to immunization (2 rabbits/antigen). High titer anti-sera that specifically recognized the pTyr-containing epitopes were obtained after one initial immunization and three boosts, and were stored at −70 °C.

Plasmid constructs

Plasmid encoding SRC was purchased from Sino Biological, and SRC mutant (KM) was kindly provided by Dr. ZHAO Bin (Zhejiang University). Human full-length G6PD cDNA, its mutants, full-length GSTP1 and its truncated GSTP1 versions were cloned into the PXJ40-based Flag-tag, HA-tag and GFP-tag vectors for cell transfection; pGEX-4T1 vector was used for GST-G6PD proteins (WT, 2YF or 2YA) expression in bacteria. GST tagged SRC protein was purchased from Sino Biological. His tagged GSTP1 and His tagged G6PD protein was also obtained from MCE.

The primers (forward and reverse) for constructing plasmids are listed below:

human G6PD (5’-ATATAAGCTTATGGCAGAGCAGGTGGCCCTGA-3’ and 5’-ATATATACCCGGGTCAGAGCTTGTGGGGGTTCACCC-3’);

human G6PD Y249F (5’-GGGCTTTTTCGATGAATTTGGGATCATCCGGG-3’ and 5’-ATTCATCGAAAAAGCCCCCGCGACCCTCAGTG-3’);

human G6PD Y322F (5’- AAAGGGTTCCTGGACGACCCCACGGTGCCCCG

-3’ and 5’- TCGTCCAGAGGCCCTTTGGTGGCCTCGCCCTC -3’);

human G6PD Y249A (5’-GGGCGCTTTCGATGAATTTGGGATCATCCGGG-3’ and 5’-ATTCATCGAAAGCGCCCCCGCGACCCTCAGTG-3’);

human G6PD Y322A (5’-AAAGGGGCCCTGGACGACCCCACGGTGCCCCG-3’ and 5’- TCGTCCAGGGCCCCTTTGGTGGCCTCGCCCTC -3’);

human GST-G6PD (5’-ATATGAATTCATGGCAGAGCAGGTGGCCCTGAGC-3’ and 5’-ATATCTCGAGTCAGAGCTTGTGGGGGTTCACCCACTTG-3’);

human GSTP1 (5’-ATATAAGGATCCATGCCGCCCTACACCGTGGTC-3’ and 5’-AGTGCCTCGAGTCACTGTTTCCCGTTGCCATTG-3’);

Human GSTP1-D1 (5’-ATATAAGGATCCATGCCGCCCTACACCGTGGTC-3’ and 5’-AAGCGGCCGCTCACAGGATGGTATTGGACTGGT-3’);

Human GSTP1-D2 (5’-TTAAGCTTCGTCACCTGGGCCGCACCCT-3’ and 5’-TTTGCGGCCGCTCAGCCTCCCTGGTTCTGGGACA-3’)

Human GSTP1-D3 (5’-GCCCAAGCTTAAGACCTTCATTGTGGGAGA-3’ and 5’-TTTGCGGCCGCTCACTGTTTCCCGTTGCCAT-3’)

Crosslinking assay

Half gram Epoxy-activated sepharose-6B (GE Healthcare Life Sciences) was swollen in 50 mL distilled water for 1 hour, and 50 ml coupling buffer (0.1 M Na2CO3, pH 13.0 or pH 9.5) was applied for washing beads through a sintered glass filter. Lactic acid (Sigma-Aldrich) (0.5 g in 15 ml coupling buffer) was mixed with Epoxy-activated sepharose-6B, the pH adjusted to either 9.5 or 13.0, and the coupling allowed for 16 h at 37 °C with gentle shaking. The beads were washed to get rid of free ligand with coupling buffer, and remaining active chemical groups blocked by 1 M ethanolamine at pH 8.0. The amounts of cross-linked ligand were much higher in the 6B-LAC-2 (pH 13.0) than the 6B-LAC-1 (pH 9.5) resins (text). The beads were preserved in a solution containing 20% ethanol until use. Before use, ethanol was washed off, and the beads were incubated with cell lysates for 4 h, washed 4 times with HEPES-lysis buffer (50 mM HEPES-KOH, pH7.4, 180 mM NaCl, 1.5 mM MgCl2, 5 mM EDTA, 10% glycerol, 1% NP-40, 100X NaVO3), and re-suspended with HEPES-lysis buffer for Western-Blot and mass spectrometry analyses.

Immunoprecipitation (IP)

For immunoprecipitation, cells were lysed in HEPES-lysis buffer (50 mM HEPES, 150 mM NaCl, 1.5 mM MgCl2, 5 mM EDTA, 10% glycerol and 1% Triton X-100, pH7.4) supplemented with a protease inhibitor cocktail (Selleck Chemicals), and then centrifuged at 14,000 g for 15 min. For IP reactions with ectopically expressed and tagged proteins, Flag- (Selleck Chemicals) or HA-beads (MCE) were added to the lysates, stirred at 4 °C for 2 h and washed with HEPES-lysis buffer 4 times. Immuno-precipitated proteins were eluted with Flag or HA peptide (MCE) dissolved in 100 ul HEPES-lysis buffer for 3 h, and the supernatant collected for analyses. For IP reactions with endogenous G6PD, the supernatants were pre-cleared with protein A/G-coupled agarose (Thermo Fisher Scientific), anti-G6PD antibody was added to the cell lysates and stirred at 4 °C overnight, and then protein A/G-coupled agarose were added and incubated for 2 h, followed by washing 4 times and boiling in 5X loading buffer. Immuno-enriched samples were resolved by SDS-PAGE and analyzed by Western-Blot.

siRNA-mediated RNA Interference

For RNA interference, cells in 6-well plates were transfected with designed siRNAs (an irrelevant sequence [control] siRNA), GSTP1 siRNA-1 and GSTP1 siRNA-2, each at 5 nmol/L, using Lipofectamine RNAiMAX Transfection Reagent (Invitrogen) for 48 h according to the manufacturer’s instructions.

The sequences are:

control siRNA (5’- TTCTCCGAACGTGTCACGT-3’);

siGSTP1-1 (5’-CATCAATGGCAACGGGAAA-3’);

siGSTP1-2 (5’-ACACCGTGGTCTATTTCCCAGTTCG-3’);

siSRC-1 (5’-GCCTCAACGTGAAGCACTA-3’);

siSRC-2 (5’-GGTGGCCTACTACTCCAAA-3’).

Cell proliferation assay

Three thousand of MCF-7 cells were seeded with normal media in 12-well plates and switched to glucose-deprived media with varying doses of lactic acid next day. Post specified incubation days, cells were trypsinized, collected, stained with Trypan Blue, and numbers of viable cells counted.

Colony formation assay

Five hundred cells were seeded in 6 cm plates in triplicates and incubated in specified medium for a total of 14 days with the medium changed once on day 7. Viable colonies were stained with crystal violet.

NADPH/NADP+ ratio measurement

NADPH/NADP+ ratio was determined by enzymatic cycling methods as previously described [71], the values were normalized to the concentration of protein.

Cellular ROS detection assay

After treatment with 0.5 mM glucose with or without 20 mM lactic acid, MCF-7 cells were washed with PBS twice, cultured with the medium with 25 μΜ Η2DCFDA probe for 30 min, and the fluorescence intensity detected by excitation at 485 nm and emission at 535 nm.

Recombinant protein purification and in vitro kinase assay

Full-length human G6PD (wild-type, 2YA or 2YF mutant) sequences were cloned into the pGEX-4T1 vector and expressed as the GST-tagged fusion proteins in E. coli BL21 by induction with 0.15 mM isopropyl β-D-thiogalactopyranoside (Sigma Aldrich) for 20 h at 20 °C. The proteins were purified using glutathione-sepharose 4B beads (GE Healthcare Life Sciences). To assess phosphorylation in vitro, substrates GST-G6PD WT, GST-2YF or GST-2YA (1.5 µg) was incubated with recombinant SRC (0.5 µg) for 30 min at 37 °C in a 50 μl kinase reaction system (50 mM Tris-HCl pH 8.0, 150 mM NaCl, 100 mM DTT, 10 mM MgCl2) with 0.3 mM ATP. The reactions were terminated by loading buffer, samples resolved by SDS-PAGE and specific signals probed through Immuno-Blot using G6PD p-Tyr and site-specific antibodies.

G6PD enzyme activity measurement

G6PD enzyme activity was determined following the protocol [72], enzyme activities were normalized to the concentration of protein.

Lactic acid-responsive tripartite complex assay in vitro

GST-SRC (1.1 μg) protein expressed in bacteria was incubated with GST Sepharose beads for 2 h, washed and resuspended with 100 μl reaction buffer. His-G6PD (1.1 μg) and His-GSTP1 protein (1.1 μg) were incubated with GST-SRC beads in the reaction buffer (20 mM HEPES, pH 7.5, 100 mM NaCl) overnight with or without 20 mM lactic acid, washed for four times and analyzed by Western Blotting, which were stained by Coomassie Brilliant Blue.

GSTP1 activity measurement

Cells treated with or without 20 mM lactic acid were sonicated on ice and centrifuged. The supernatant was mixed with 160 μl solution II and 20ul solution III provided in the kit (Solarbio) and the absorbance at 340 nm measured at the interval of 30 s for 10 min. The activity of GSTP1 was calculated.

Lactate measurement

Briefly, three kinds of beads were collected for lactate measurement. The master reaction mix contained 5 ml reaction buffer (0.1 M Tris-HCl, pH8.6), 1 ml cosolvent (3% Triton X-100), 2 ml chromogenic reagent (120 mg Nitrotetrazolium blue chloride, 4 mg Phenazine methosulfate and 250 mg NAD+ dissolved in 50 ml sterilized water) and 130U LDH were prepared. For each well of 96-well plate, 175 µl reaction buffer and 25 μl beads were mixed and incubated at 37 °C for 30 min. The absorbance was measured at 570 nm (A570) on SpectraMax M5. The concentrations of the lactate used to generate the standard curve were 0, 0.0983, 0.1875, 0.375, 0.75 and 1.5 mM.

Dot blotting

A serial dilution with 0.5 pg, 0.005 ng, 0.05 ng, 0.5 ng and 1 ng of Y249(pi) or 322(pi) peptides along with unmodified peptides (HuaAn Biotechnology) were spotted onto nitrocellulose membranes. After complete absorption of the samples, the membranes were blocked with 5% skim milk for 1 h and incubated with Y249(pi) or Y322(pi) antibodies for 2 h at room temperature, followed by 1-hour incubation with a secondary antibody that was conjugated with horseradish peroxidase (HRP). After washing, HRP substrate was added to develop the images captured and recorded by exposure to X-ray films.

Simulation of GSTP1-Lactate and GSTP1-G6PD-SRC complex

The structure of the GSTP1 protein (PDB ID: 3GUS)was downloaded from the PDB website and small molecules removed through Pymol. The molecular structure of lactic acid was extracted from the structure of the protein (PDB ID: 3KB6). Autodock tools were used to simulate the interaction between GSTP1 and lactate and calculate the most possible binding energy. In addition, GSTP1, G6PD (PDB ID: 2bhl) and SRC (PDB ID: 2H8H) structures were each extracted as above and were modeled by GRAMM-X Protein-Protein Docking Web.

Quantification of metabolites by LC-MS/MS

To quantify polar metabolites concentrations, we used TSQ Vantage LC-MS interfaced with Ultimate 3000 Liquid Chromatography system (Thermo Scientific) and Triple quadruple mass spectrometry to detect metabolites from tissues or cells. Polar metabolites were extracted according to a protocol. Briefly, we added 80% HPLC-grade methanol (cooled to −80 °C) to fresh or frozen tissues or cells that were grinded for 1–2 min with tissue grinder in the tube on dry ice. After vortexing for 1 min at 4–8 °C and incubated for 4 h at −80 °C, the samples were then centrifuged at 14,000 g for 10 min at 4 °C. We transferred the supernatant to a new 2 ml tube and lyophilize the samples in a speedVac prior to LC-MS analysis. For quantification of metabolites, U-13C-glutamine was added to the extraction buffer in 80% methanol and used as internal standard. Samples and standards were measured using a TSQ Vantage equipped with a HILIC column (Amide 4.6 x 100 mm ID 3.5 μm; Part No: 186004868, Waters). The mobile phases and gradients were the same as those used for the sugar phosphate on the UHPLC-QTOF system (AB 6600 TripleTOF, SCIEX, Canada). The ion transfer tube temperature was set to 350 °C and vaporizer temperature was 270 °C. The instrument was run in the negative mode with a spray voltage of 3000, sheath gas 40 and Aux gas 5.0. The 7–8 concentrations (from low to high) of the different standard mix were measured using multiple reactions monitoring mode (MRM) with optimal collision energies to produce a standard curve.

Mice

All animal procedures were conducted with the approval of the Animal Research Ethics Committee at Zhejiang University. All animals were taken care of in a humane way following standard guidance information of the Care and Use of Laboratory Animals written by NIH. Female BALB/c nude mice (3 weeks old) in good conditions were purchased from GemPharmatech Company in Jiangsu, China. All the mice were housed in Laboratory Animal Center of Zhejiang University with stable temperature (23 ± 2 °C) and light/dark at the interval of 12 h.

Xenograft mouse models

Mice were injected subcutaneously with 1 × 107 stable cells (n = 6 for each group). At day 15, when tumors became visible, tumor growth was measured by caliper every 2 or 3 days, tumor volume was calculated as the formula of 1/2× major diameter × minor diameter2. When the tumor reached 100–150 mm2, mice were divided into three groups following the principle of randomness and respectively received subcutaneous Intraperitoneal injection every two days with Tween80 (vehicle control, 100 μl), GSPT1 inhibitor (HY135318; MCE) (5 µM and 10 µM in 100 μl of Tween80, PEG300 and saline solution). Mice were sacrificed when growth retardation of tumors was evident or when tumors reached the maximally permitted condition. After dissection, the weights of tumors were measured and recorded.

Statistical analysis

Statistical analyses were conducted on GraphPad Prism 7 or with appropriate computational tools. The data were analyzed with unpaired students’ t-test, multiple t test or two-way analysis of variance (ANOVA) as indicated in the figure legends. And the variation was estimated in every test. Error bars show sampling bias from independent samples or experiments. The correlation analyses were conducted by Pearson’s test. All data including the Western-blots were from at least two independent experiments with SEM (mean ± SD). ns, not significant, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. For animal models, we underwent a randomization process in mice within the same groups. Original western blots can be obtained in the “Supplementary Material” file.