Cellular signals converge at the NOX2-SHP-2 axis to induce reductive carboxylation in cancer cells

SUMMARY

Environmental stresses including hypoxia or detachment for anchorage independence, or attenuation of mitochondrial respiration through inhibition of electron transport chain induce reductive carboxylation in cells with an enhanced fraction of citrate arising through reductive metabolism of glutamine. This metabolic process contributes to redox homeostasis and sustains biosynthesis of lipids. Reductive carboxylation is often dependent on cytosolic isocitrate dehydrogenase 1 (IDH1). However, whether diverse cellular signals induce reductive carboxylation differentially, or through a common signaling converging node remains unclear. We found that induction of reductive carboxylation commonly requires enhanced tyrosine phosphorylation and activation of IDH1, which, surprisingly, is achieved by attenuation of a cytosolic protein-tyrosine phosphatase, Src homology region 2 domain-containing phosphatase-2 (SHP-2). Mechanistically, diverse signals induce reductive carboxylation by converging at upregulation of NADPH oxidase 2, leading to elevated cytosolic reactive oxygen species that consequently inhibit SHP-2. Together, our work elucidates the signaling basis underlying reductive carboxylation in cancer cells.

Graphical Abstract

eTOC Blurb

The signaling basis underlying induction of reductive carboxylation remains unclear. Zhang et. al. report that diverse signals converge at the NOX2-SHP-2 axis, leading to enhanced tyrosine phosphorylation and activation of IDH1 that is commonly required for induction of efficient reductive carboxylation.

INTRODUCTION

During the tricarboxylic acid (TCA) cycle, citrate can be generated from acetyl-CoA and oxaloacetate, which is at the center of cellular metabolism by breaking down glucose initiated in glycolysis and fueling ATP production (Williams and O'Neill, 2018). However, in cells where mitochondrial oxidation is suppressed due to defects in electron transport chain (ETC), or treatment with ETC inhibitors such as Complex I inhibitor rotenone, reductive carboxylation of glutamine-derived α-ketoglutarate (αKG)accounts for an increased fraction of cellular isocitrate and citrate (Mullen et al., 2012). Moreover, reductive carboxylation is also induced not only in cells experiencing hypoxia, where mitochondrial reactive oxygen species (ROS) accumulate due to inefficient electron transfer through the ETC (Metallo et al., 2012; Wise et al., 2011), but in cells during anchorage-independent growth, where detachment of cells from a monolayer also increases mitochondrial ROS, leading to aberrant oxidative stress that promotes reductive metabolism of glutamine (Jiang et al., 2016).

Reductive carboxylation is highly dependent on cytosolic isocitrate dehydrogenase 1 (IDH1) to generate isocitrate from αKG via a non-canonical reverse reaction by IDH1. Subsequently, isocitrate is important for producing citrate and acetyl-CoA that are essential for lipid synthesis in cells experiencing hypoxia, as well as for reducing mitochondrial ROS to sustain redox homeostasis during hypoxia, or to support anchorage-independent growth and survival of cells during extracellular matrix detachment and metastatic spread (Jiang et al., 2016; Metallo et al., 2012; Wise et al., 2011). Thus, sustained enzyme activity of cytosolic IDH1 must play a crucial role during reductive carboxylation.

We recently reported that tyrosine phosphorylation-enhanced IDH1 activation is important for glutamine-dependent reductive carboxylation and lipogenesis in cancer cells under hypoxia, which provides a metabolic advantage to cancer cell proliferation and tumor growth (Chen et al., 2019a). We found that Y42 and Y391 phosphorylation enhances IDH1 enzyme activity for both canonical (isocitrate→αKG) and non-canonical (αKG→isocitrate; reductive carboxylation) directions in cells. Diverse receptor tyrosine kinases including EGFR and FGFR1 activate Src to achieve Y42 and Y391 phosphorylation of IDH1, respectively, which contributes to reductive carboxylation and consequent cell proliferation and tumor growth. In addition, we demonstrated that IDH1 phosphorylation levels correlate with EGFR signal intensity and the contribution of reductive carboxylation to citrate metabolism in diverse lung cancer cell lines(Chen et al., 2019b). However, the signaling link between IDH1 activation and induction of reductive carboxylation remains unclear.

RESULTS AND DISCUSSION

Diverse signals commonly require enhanced tyrosine phosphorylation and activation of IDH1 to induce reductive carboxylation in cancer cells

We found that phosphorylation levels of IDH1 at Y42 and Y391 were induced in diverse human cells treated with different reductive carboxylation-inducing stresses, including lung cancer A549 and H1299 cells experiencing hypoxia, anchorage-independent detachment or treatment with Complex I inhibitor rotenone (Figure 1A, left, middle and right, respectively), as well as leukemia K562 and MOLM-14 cells treated with rotenone (Figure S1A). Moreover, we examined “rescue” A549 cells with siRNA-mediated knockdown of endogenous IDH1, followed by rescue expression of siRNA-resistant FLAG-IDH1 WT or the catalytically less active, phosphor-deficient FLAG-IDH1 Y42F/Y391F double mutant (Figure 1B). We found that IDH1 knockdown abolished the enhanced levels of reductive carboxylation induced by hypoxia, detachment, or rotenone treatment (Figure 1B, leftfour bars in left, middle and right panels, respectively), which were completely rescued by expression of FLAG-IDH1 WT, but not by expression of the phosphor-deficient Y42F/Y391F (Figure 1B, right four bars in left, middle and right panels, respectively). These data together suggest that tyrosine phosphorylation and subsequent activation of IDH1 are commonly enhanced and required for effective reductive carboxylation in cancer cells induced by diverse signals.

Figure 1. Diverse signals commonly require enhanced tyrosine phosphorylation and activation of IDH1 to induce reductive carboxylation.

(A) Effect of hypoxia treatment for 48 hours (left), detachment treatment for 24 hours (middle), or 10 μM rotenone treatment for 16 hours (Right)on tyrosine phosphorylation of IDH1 assessed by Western blot.

(B) Effect of “rescue” expression of siRNA-resistant FLAG-IDH1 WT or the catalytically less active FLAG-IDH1 Y42F/Y391F on reductive carboxylation indicated by ¹⁴C-lipid biosynthesis rate inIDH1 transient knockdown A549 cell with or without hypoxia (left), detachment (middle), or rotenone treatment (Right).

The fold changes of the intensity ratios between phosphor-Y42 and total IDH1 or phosphor-Y391 and total IDH1 were indicated. The results were presented as mean ± s.d. of triplicate experiments. p values were obtained by a two-tailed Student’s t-test (*0.01<p<0.05; **0.001<p<0.01;***p<0.001; ns, not significant).

Induction of reductive carboxylation attenuates SHP-2 activity to enhance phosphorylation and activation of IDH1

We next sought to examine whether diverse reductive carboxylation-inducing stresses promote tyrosine phosphorylation of IDH1 by activating upstream tyrosine kinases. We previously demonstrated that the EGFR-Src cascade in A549 cells, the FGFR1-Src cascade in H1299 cells, and the BCR-ABL-Src cascade in K562 cells represent distinct upstream IDH1 tyrosine kinase cascades in different cells (Chen et al., 2019a). Surprisingly, we found that treatments with hypoxia, detachment or rotenone resulted in rather decreased EGFR activity assessed by reduced tyrosine phosphorylation while Src activity was unaltered in A549 cells (Figure S1B). In H1299 cells, hypoxia resulted in reduced tyrosine phosphorylation of both FGFR1 and Src, while Rotenone and detachment treatments reduced Src phosphorylation levels but did not alter phosphorylation levels of FGFR1 (Figure S1C). Moreover, rotenone treatment resulted in slightly reduced tyrosine phosphorylation levels of both BCR-ABL and Src. Note that the leukemogenic fusion tyrosine kinase BCR-ABL is constitutively active and cannot be further activated(Gilliland, 2001) (Figure S1D). These data together suggest that all three different stresses may commonly attenuate an upstream protein tyrosine phosphatase (PTP) of IDH1, leading to enhanced IDH1 phosphorylation and activation that are required for effective reductive carboxylation.

We next examine three well known PTPs including Src homology region 2 domain-containing phosphatase-2 (SHP-2, a.k.a. tyrosine-protein phosphatase non-receptor type 11 (PTPN11)), PTP localized to the Mitochondrion 1 (PTPMT1), and PTP1B (a.k.a. tyrosine-protein phosphatase non-receptor type 1 (PTPN1)).We found that treatment with SHP099 (SHP-2 inhibitor), but not alexidine dihydrochloride (PTPMT1 inhibitor) or TCS401 (PTP1B inhibitor), resulted in elevated Y42 and Y391 phosphorylation levels of IDH1 and abolished rotenone-induced IDH1 phosphorylation in A549 cells (Figure 2A). Similar results were obtained in A549 cells treated with detachment or hypoxia (Figure S2A, upper and lower, respectively). In addition, treatment with SHP099, but not alexidine dihydrochloride or TCS401, resulted in elevated reductive carboxylation rate, which could not be further enhanced by hypoxia in A549 cells (Figure S2B). Consistent with these findings, knockdown of SHP-2, but not PTPMT1 or PTP1B, by specific siRNAs led to elevated Y42 and Y391 phosphorylation levels of IDH1 and abolished rotenone-induced IDH1 phosphorylation (Figure 2B). Moreover, rotenone treatment resulted in increased tyrosine phosphorylation of Y42 and Y391 of IDH1(Figure S2C) as well as increased reductive carboxylation in mouse embryonic fibroblasts (MEFs) assessed by elevated lipogenesis rate(Figure 2C), which was not observed in Shp-2 knockout MEFs (Shp-2^Δ/Δ) generated from Ptpn11 conditional knockout embryos (Ptpn11 ^fl/fl/ER-Cre) (Yu et al., 2003), further supporting our overall hypothesis.

Figure 2. Induction of reductive carboxylation attenuates SHP-2 activity to enhance phosphorylation and activation of IDH1.

(A) Effect of treatment with SHP099 (left), Alexidine dihydrochloride (middle), or TCS401 (Right) on tyrosine phosphorylation of IDH1 in A549 cells treated with or without rotenone assessed by Western blot.

(B) Effect of transient knockdown of endogenous SHP-2 (left), PTPMT1 (middle), or PTP1B (Right) on tyrosine phosphorylation of IDH1 in A549 cells treated with or without rotenone assessed by Western blot.

(C) Effect of rotenone treatment on reductive carboxylation in Shp-2 knockout MEFs (Shp-2^−/−).

(D) Tyrosine kinase rEGFR and/or rSrc pretreatedrIDH1 was incubated with purified SHP-2 (rSHP-2) in an in vitro dephosphorylation assay. Tyrosine phosphorylation of IDH1 were determined by Western blot.

(E) Effect of hypoxia, detachment, or rotenone treatment onSHP-2 phosphatase activity in A549 cells.

(F) Effect ofSHP099 and/or rotenone treatment on reductive carboxylation in IDH1 transientknockdown A549 cells with or without “rescue” expression of FLAG-IDH1 WT orFLAG-IDH1 Y42F/Y391F.

The results were presented as mean ± s.d. of triplicate experiments. p values were obtained by a two-tailed Student’s t-test (***p<0.001; ns, not significant). See also Figure S1 and S2.

We next confirmed SHP-2 as an IDH1 upstream phosphatase by performing an in vitro dephosphorylation assay using purified SHP-2 (rSHP-2) incubated with phosphorylated rIDH1as a substrate; rIDH1protein was pre-treated with rEGFR and/or rSrc to achieve phosphorylation at Y42 and/or Y391, respectively (Figure 2D). The results demonstrate that SHP-2 directly dephosphorylates IDH1.In addition, we found that SHP-2 phosphatase activity was commonly reduced in A549 (Figure 2E) and H1299 (Figure S2D) cells treated with rotenone, hypoxia, or detachment, which was assessed in an in vitro dephosphorylation assay using immunoprecipitated SHP-2 incubated with DiFMUP as a substrate. Finally, we found that treatment with SHP099resulted in elevated reductive carboxylation rate, which could not be further enhanced by rotenone in A549 cells (Figure 2F, left four bars), whileIDH1 knockdown abolished the enhanced reductive carboxylation induced by SHP99 and/or rotenone(Figure 2F, middle two bars), which was completely rescued by expression of FLAG-IDH1 WT, but not by expression of the phosphor-deficient Y42F/Y391F (Figure 2F, right four bars).Similar results were obtained using A549 cells treated with detachment (Figure S2E). These results together suggest that induction of reductive carboxylation is mediated through inhibition of SHP-2, leading to enhanced tyrosine phosphorylation and activation of IDH1.

Reductive carboxylation elevates cytosolic ROS to inhibits SHP-2

It was reported that reductive carboxylation-inducing stresses including hypoxia, detachment and rotenone suppress mitochondrial oxidation and cause mitochondrial ROS accumulation (Jiang et al., 2016; Metallo et al., 2012; Mullen et al., 2012; Wise et al., 2011); and that PTPs including SHP-2 are important targets of ROS, which inactivates PTPs by oxidizing catalytic Cys residues to the sulfenic acid state (Tanner et al., 2011). Since SHP-2 primarily localizes in cytosol (Qu, 2000), we were curious about whether cytosolic ROS levels are also elevated during reductive carboxylation and how this might influence cytosolic SHP-2 phosphatase activity.

We found that rotenone elevated both cytosolic ROS and mitochondrial ROS (Figure 3A, left and right, respectively). The cytosolic and mitochondrial ROS levels were assessed by confocal microscopic imaging to detect cytosolic Hyper-cyto or mitochondrial Hyper-mito, respectively, which are red fluorescent genetically encoded indicators (Belousov et al., 2006; Ermakova et al., 2014) being expressed in A549 cells with specific subcellular localizations (Figure S3A-S3B). Moreover, we found that treatment with antioxidant agent N-acetyl-L-cysteine (NAC; 1mM) effectively reversed elevated cytosolic ROS but not mitochondrial ROS in A549 cells treated with rotenone, whereas treatment with mitochondria-specific antioxidant agent mitoTEMPO (10μM) effectively reversed rotenone-elevated mitochondrial ROS but not cytosolic ROS in A549 cells (Figures 3A and S3A-S3B). We found that NAC but not MitoTEMPO treatment effectively reversed the decreased SHP-2 phosphatase activity by rotenone in A549 cells (Figure 3B), suggesting that cytosolic rather than mitochondrial ROS inhibit SHP-2 during the induction of reductive carboxylation. Consistent with this finding, NAC but not MitoTEMPO treatment effectively reversed the increased tyrosine phosphorylation of Y42 and Y391 of IDH1 (Figure 3C) as well as the increased reductive carboxylation assessed by increased lipogenesis rate in A549 cells treated with rotenone (Figure 3D).Similar results were obtained using H1299 cells treated with rotenone(Figures S3C) andA549 cells experiencing detachment (Figures S3D and 3E), or hypoxia (Figures S3E and 3F). Furthermore, treatment with another antioxidant agent, Ebselen, shows similar effects as NAC treatment in A549 cells treated with rotenone(Figures S3F).We also found that treatment with NAC abolished elevated reductive carboxylation rate by rotenone treatment (Figure 3G, first three bars on left), whileIDH1 knockdown abolished the effects on reductive carboxylation by rotenone or rotenone combined with NAC (Figure 3G, second three bars on left), which was completely rescued by expression of FLAG-IDH1 WT, but not by expression of the phosphor-deficient Y42F/Y391F (Figure 3G, right six bars).Moreover, we added indicated concentrations of hydrogen peroxide in an in vitro dephosphorylation assay using purified SHP-2 (rSHP-2) incubated with phosphorylated rIDH1 as a substrate and found that in vitro dephosphorylation of IDH1 by SHP-2 can be inhibited by hydrogen peroxide(Figure S3G).These data together suggest that different reductive carboxylation-inducing stresses commonly elevate cytosolic ROS for SHP-2 inhibition, leading to elevated tyrosine phosphorylation of IDH1in cancer cells.

Figure 3. Reductive carboxylation elevates cytosolic ROS to inhibits SHP-2.

(A) Effect of NAC or MitoTEMPO treatment on cytosolic ROS(left) and mitochondrial ROS (right) in A549 cells with rotenone treatment.

(B-D) Effect of NAC or MitoTEMPO treatment on SHP-2 phosphatase activity (B), tyrosine phosphorylation of IDH1 (C), and reductive carboxylation (D) in A549 cells with rotenone treatment.

(E-F) Effect of NAC or MitoTEMPO treatment on SHP-2 phosphatase activity (left), tyrosine phosphorylation of Y42 and Y391 of IDH1 (middle), and reductive carboxylation (right) in A549 cells with detachment (E)or hypoxia (F) treatment.

(G) Effect of NAC treatment on reductive carboxylation in IDH1transient knockdown A549 cells with or without “rescue” expression of FLAG-IDH1 WT orFLAG-IDH1 Y42F/Y391Fwith rotenone treatment.

The results were presented as mean ± s.d. of triplicate experiments. p values were obtained by a two-tailed Student’s t-test (*0.01<p<0.05; **0.001<p<0.01;***p<0.001; ns, not significant).See also Figure S3.

Induction of reductive carboxylation commonly upregulates NOX2 to achieve SHP-2 inhibition

SHP-2 oxidation was suggested to require NADPH oxidases (NOXs) (Tsutsumi et al., 2017). There are five NOX family members (NOX1-5)(Bedard and Krause, 2007). We found that detachment (Figure S4A), hypoxia (Figure S4B), and rotenone treatment (Figure S4C) resulted in upregulated mRNA levels of multiple NOX family members, with common upregulation of NOX2 and NOX3 under all three conditions. In order to identify which NOX member is responsible for SHP-2 inhibition during reductive carboxylation, we tested a group of NOX inhibitors. We found that treatment with a pan-NOX inhibitor diphenyleneiodonium (DPI) reversed the elevated reductive carboxylation assessed by increased lipogenesis rate in A549 cells experiencing hypoxia in a dose-dependent manner; NAC treatment was included as a positive control (Figure 4A). In contrast, among several selective NOX inhibitors including GSK2795039 (NOX2), GKT137831 (NOX1/4), GLX351322 (NOX4) and ML090 (NOX5), only the NOX2 inhibitor GSK2795039 (GSK) effectively reversed elevated reductive carboxylation in A549 cells treated with rotenone (Figure 4B). Moreover, GSK treatment similarly reversed increased reductive carboxylation in a dose-dependent manner in A549 cells induced by hypoxia or detachment (Figure S4D, left and right, respectively).

Figure 4. Induction of reductive carboxylation commonly upregulates NOX2 to achieve SHP-2 inhibition.

(A) Effect of treatment with diphenyleneiodonium (DPI) with indicated concentrations on reductive carboxylation in A549 cells under hypoxia.

(B) Effects of GSK2795039, GKT137831,GLX351322, or ML090treatments on reductive carboxylation in A549 cells with rotenone treatment.

(C) Effect of detachment (left), hypoxia (middle), or rotenone (right) treatment on protein level of NOX2 in A549 cells assessed by Western blot.

(D-E) Effect of rotenone treatment on NOX2 complex partner mRNA levels assessed by RT-qPCR

(D), and protein level assessed by Western blot (E) in A549 cells.

(F) Representative confocal microscopy images of NOX2 or p67phoxImmunocytochemistry(ICC)staining in A549 cells with or without rotenone treatment.

(G-H) Effect of transient knockdown ofNOX2 on reductive carboxylation (G; left) and cytosolic ROS (G; right), tyrosine phosphorylation of IDH1 (H; left) and SHP-2 phosphatase activity (H; right)in A549 cells with or without rotenone treatment.

(I) Effect of transient knockdown of NOX2 on reductive carboxylation in A549 cells with detachment treatment (left) or under hypoxia (right).

(J) Effect of GSK2795039 treatment on reductive carboxylation in transient IDH1-knockdown A549 cells with or without “rescue” expression of FLAG-IDH1 WT or FLAG-IDH1 Y42F/Y391F with rotenone treatment.

(K) Working model. The results were presented as mean ± s.d. of triplicate experiments. p values were obtained by a two-tailed Student’s t-test (*0.01<p<0.05; **0.001<p<0.01;***p<0.001; ns, not significant). See also Figure S4.

These results are consistent with findings that detachment, hypoxia or rotenone treatment resulted in increased protein expression levels of NOX2 in A549 cells (Figure 4C).Note that NOX2 requires the membrane subunit (p22phox), cytosolic subunits (p67phox and p47phox), and the Rac GTPase to form a complex for activation (Altenhofer et al., 2012; Brandes et al., 2014; Panday et al., 2015).Interestingly, we found that mRNA (Figure 4D) and protein (Figure 4E) levels of p67phox, p47phox, but not p22phox were elevated by rotenone treatment. Since p67phox and p47phox are NOX2-specific cytosolic subunits, while p22phox is a common binding partner for NOX1-4 in the membrane, these data not only further support our findings that reductive carboxylation activates NOX2 but also provide mechanistic insights into the selective activation of NOX2 that is ensured by upregulation of p67phox and p47phox induced by rotenone.

We next checked NOX2 and its partners subcellular localization by immunocytochemistry (ICC) staining using confocal microscopy(Figure 4F). We observed co-localization of NOX2 and p67phox in both cell membrane and cytosol, which was enriched to cell membrane with increased protein intensity upon rotenone treatment in A549 cells(Figure 4F).Furthermore, knockdown of NOX2 by siRNA in A549 cells (Figure 4G,left), but not siRNAs targeting NOX1 or NOX3 (Figure S4E), reversed the rotenone-induced reductive carboxylation and elevated cytosolic ROS levels (Figure 4G, left and right, respectively), as well as increased tyrosine phosphorylation levels of IDH1 and reduced SHP-2 activity (Figure 4H, left and right, respectively). Similar results were obtained in K562 and H1299 cells treated with rotenone(Figure S4F-S4I). NOX2 knockdown in A549 cells treated with detachment or hypoxia also led to induced reductive carboxylation rates (Figure 4I), elevated cytosolic ROS levels (Figure S4J), increased tyrosine phosphorylation levels of IDH1 (Figure S4K) and reduced SHP-2 activity (Figure S4L). Lastly, we found that treatment with GSK abolished elevated reductive carboxylation rate by rotenone treatment (Figure 4J, first three bars on left), whileIDH1 knockdown abolished the effects on reductive carboxylation by rotenone or rotenone combined with GSK(Figure 4J, second three bars on left), which was completely rescued by expression of FLAG-IDH1 WT, but not by expression of the phosphor-deficient Y42F/Y391F (Figure 4J, right six bars).Similar results were obtained using A549 cells treated with detachment (Figure S4M). Thus, these data together strongly support our hypothesis thatNOX2 is commonly required by reductive carboxylation induced by different stresses in cancer cells.

DISCUSSION

Our studies elucidate the underlying signaling mechanism by which diverse cellular responses to hypoxia, detachment and attenuation of ETC commonly converge at upregulation of NOX2, leading to consequent inhibition of cytosolic protein tyrosine phosphatase SHP-2, which is upstream of IDH1 and plays a crucial role in inducing effective reductive carboxylation in cancer cells through enhanced tyrosine phosphorylation and activation of cytosolic IDH1 (Figure 4K). In contrast, reductive carboxylation enhances tyrosine phosphorylation of IDH1 independent of its upstream Group I and II tyrosine kinases (Chen et al., 2019a) (Figure 4K).Our results revealed that diverse reductive carboxylation-inducing stresses simultaneously promote productions of both mitochondrial and cytosolic ROS, while although mitochondrial ROS accumulation suppresses mitochondrial oxidation, it is the cytosolic ROS that are responsible for the inhibition of SHP-2 and consequent phosphorylation and activation of IDH1, which is crucial for effective reductive carboxylation in cancer cells (Figure 4K).

Limitations of the study:

It was reported that reductive carboxylation is enhanced by hypoxia in a HIF-1α-dependent manner, whereas detachment associated reductive metabolism is independent of hypoxia or HIF-1α (Jiang et al., 2016). This may suggest that hypoxia signal through HIF-1α to upregulate NOX2, whereas detachment might activate NOX2 in a different way. Rotenonedependent activation of NOX2 in human lung cancer cells were reported but the underlying mechanism remains unknown (Hu et al., 2016). It remains also unknown about the connection between detachment and NOX2. A previous report suggested that inhibition of NOX1/4 with GKT137831 attenuates retinal detachment-induced photoreceptor apoptosis (Deliyanti and Wilkinson-Berka, 2015), whereas GKT137831 did not affect reductive carboxylation induced by detachment in A549 cells (Figure 4B), suggesting that NOX1/4 might not be involved. Future studies are warranted to decipher the underlying mechanisms by which NOX2 is upregulated by different signals.

STAR★Methods

RESOURCE AVAILABILITY

Lead contact

Further information requests should be directed to the lead contact, Jing Chen (jingchen@medicine.bsd.uchicago.edu)

Materials availability

Materials will be shared by the lead contact upon request.

Data and code availability

Data will be shared by the lead contact upon request. This paper does not report original code. Additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Primary Cell culture

MEF (from CK Qu’s lab; authentication and Mycoplasma not test) were cultured in DMEM medium with 10% FBS and 2 ng/ml recombinant human GM-CSF (R&D Systems).

Cell lines

HEK293T cells were cultured in Dulbecco Modified Eagle Medium (DMEM) with10% fetal bovine serum (FBS) (Sigma, F2442) and penicillin/streptomycin. NCI-H1299 (#CRL-5803, ATCC; purchased 2015; authentication and Mycoplasma tested 2018), K562, MOLM-14, and A549 cells were cultured in RPMI 1640 medium with 10% FBS. Please also refer to KEY RESOURCES TABLE for detailed information of each cell line. Cell lines experiments were conducted and designed according to protocols approved by Institutional Biosafety Committee (IBC) of the University of Chicago.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Mouse monoclonal Anti-β-actin antibody	Sigma-Aldrich	Cat# A1978; Clone# AC-15; RRID: AB_476692
Mouse monoclonal ANTI-FLAG M2 antibody	Sigma-Aldrich	Cat# F3165; Clone# M2; RRID: AB_259529
NOX2 antibody	Abcam	Cat# Ab129068; Clone# N/A; RRID:AB_11144496
SHP-2 antibody	Proteintech	Cat# 20145-1-AP; Clone# N/A; RRID: AB_10699877
p-IDH1 Y42	SHANGHAI GENOMICS, INC.(Chen et al., 2019a)	Cat# N/A; Clone# N/A; RRID: N/A
p-IDH1 Y391	SHANGHAI GENOMICS, INC.(Chen et al., 2019a)	Cat# N/A; Clone# N/A; RRID: N/A
Human Isocitrate Dehydrogenase1/IDH1 Antibody	R&D SYSTEMS	Cat# MAB7049; Clone# 843219; RRID: AB_2811299
Goat anti-Mouse IgG (H+L) Secondary Antibody, HRP	Thermo Fisher Scientific	Cat# 31430; Clone# N/A; RRID: AB_228307
Goat anti-Rabbit IgG (H+L) Secondary Antibody, HRP	Thermo Fisher Scientific	Cat# 31460; Clone# N/A; RRID: AB_ 228341
DYKDDDDK Tag Antibody	Cell Signaling Technology	Cat#2368S; Clone# N/A; RRID: AB_2217020
Anti-rabbit IgG (H+L), F(ab')2 Fragment (Alexa Fluor® 488 Conjugate)	Cell Signaling Technology	Cat#4412 Clone# N/A; RRID:AB_1904025
Anti-mouse IgG (H+L), F(ab')2 Fragment (Alexa Fluor® 555 Conjugate)	Cell Signaling Technology	Cat#4409 Clone# N/A; RRID:AB_1904022
NOX2/gp91phox Polyclonal Antibody	Bioss	Cat#BS-3889R Clone# N/A; RRID:AB_10855911
p67phox antibody	Santa Cruz	Cat#sc-374510 Clone# N/A; RRID:AB_10988074
Bacterial and Virus Strains
BL21(DE3) Chemically Competent Cells	Sigma-Aldrich	Cat# CMC0014
Chemicals, Peptides, and Recombinant Proteins
Rotenone	Sigma-Aldrich	Cat# R8875 CAS: 83-79-4
NAC	Sigma-Aldrich	Cat# A7250 CAS: 616-91-1
alexidine dihydrochloride	Sigma-Aldrich	Cat# A8986 CAS: 1715-30-6
SHP099	Selleckchem	Cat# S6388 CAS: 1801747-42-1
TCS401	CAYMAN	Cat# 20393 CAS: 243966-09-8
EGFR Recombinant Human Protein	Thermo Fisher	Cat# PR7295B
Src Recombinant Human Protein	Thermo Fisher	Cat# P3044
HEPES	Sigma-Aldrich	Cat# H4034 CAS: 7365-45-9
NaCl	Sigma-Aldrich	Cat# S9888 CAS: 7647-14-5
GSH	Sigma-Aldrich	Cat# G4251 CAS: 70-18-8
Glutamine L-[5-14C]	ARC	Cat# ARC 3562-50 uCi
MitoTEMPO	Thermo Fisher	Cat# NC1037796
DPI	Selleckchem	Cat# S8639 CAS: 4673-26-1
GSK2795039	Sigma-Aldrich	Cat# SML2770-5MG CAS: 1415925-18-6
GKT137831	Selleckchem	Cat# S7171 CAS: 1218942-37-0
GLX351322	MCE	Cat# HY-100111 CAS: 835598-94-2
ML090	CAYMAN	Cat# 15172-10 CAS: 531-46-4
DiFMUP	Thermo Fisher	Cat# D6567
si-NOX1	Sigma-Aldrich	Cat#EHU076051
si-NOX3	Sigma-Aldrich	Cat#EHU051651
si-SHP-2	Sigma-Aldrich	Cat#VPDSIRNA2D
si-PTP1B	Qiagen	Cat#SI00043827
Si-PTPMT1	Qiagen	Cat# SI4379921
iScript cDNA synthesize kit	BIO-RAD	Cat# 1708890
iTaq™ Universal SYBR® Green Supermix	BIO-RAD	Cat# 1725122
ANTI-FLAG® M2 Affinity Gel	Sigma-Aldrich	Cat# A2220
Protein G Sepharose® 4 Fast Flow	Sigma-Aldrich	Cat# GE17-0618-01
TransIT®-LT1 Transfection Reagent	Mirus	Cat# MIR 2305
TRIzol reagent	Thermo Fisher	Cat# 15596026
RNAi MAX Transfection Reagent	Thermo Fisher	Cat# 13778500
Ebselen	Millipore Sigma	Cat# E3520-25MG
Hydrogen peroxide	Thermo Fisher	Cat# BP2633-500
Triton X-100	Millipore Sigma	Cat# 64-846-410ML
4% Paraformaldehyde solution	Thermo Scientific	Cat# AAJ61899AK
Experimental Models: Cell Lines
Human: HEK293T cells	ATCC	Cat# CRL-3216; RRID: CVCL_0063
Human: A549	ATCC	Cat# CCL-185; RRID: CVCL_0023
Human: NCI-H1299	ATCC	Cat# CRL-5803; RRID: CVCL_0060
Human: MOLM-14	(Fan et al., 2016)	Cat# N/A; RRID: N/A
Human: K562	(Chen et al., 2019a)	Cat# N/A; RRID: N/A
Oligonucleotides
si-IDH1 sequence: Sense: 5’-CGAAUCAUUUGGGAAUUGAUU -3’ Antisense: 5’-UCAAUUCCCAAAUGAUUCGUU-3’	Sigma-Aldrich	N/A
si-NOX2 sequence: Sense: 5’- CCCUGAGUAAACAAAGCA-3’ Antisense: 5’- UUGGAGAUGCUUUGUUUA-3’	IDT	N/A
Primers: β-actin Forward: 5’- CACCATTGGCAATGAGCGGTTC-3’ Reverse: 5’- AGGTCTTTGCGGATGTCCACGT-3’	IDT	N/A
Primers: NOX1 Forward: 5’- TTGTTTGGTTAGGGCTGAATGT-3’ Reverse: 5’- GCCAATGTTGACCCAAGGATTTT-3’	IDT	N/A
Primers: NOX2 Forward: 5’- CCCTTTGGTACAGCCAGTGAAGAT-3’ Reverse: 5’- CAATCCCGGCTCCCACTAACATCA-3’	IDT	N/A
Primers: NOX3 Forward: 5’- ACCGTGGAGGAGGCAATTAGA-3’ Reverse: 5’- TGGTTGCATTAACAGCTATCCC-3’	IDT	N/A
Primers: NOX4 Forward: 5’- CAGATGTTGGGGCTAGGATTG-3’ Reverse: 5’- GAGTGTTCGGCACATGGGTA-3’	IDT	N/A
Primers: NOX5 Forward: 5’-CCACCATTGCTCGCTATGAGTG -3’ Reverse: 5’- GCCTTGAAGGACTCATACAGCC-3’	IDT	N/A
Primers: DUOX1 Forward: 5’- TCTCTGGCTGACAAGGATGGCA-3’ Reverse: 5’- AGGCGAGACTTTTCCTCAGGAG-3’	IDT	N/A
Primers: DUOX2 Forward: 5’- CAATGGCTACCTGTCCTTCCGA-3’ Reverse: 5’- GTCCTTGGAGAGGAAGCCATTC-3’	IDT	N/A
Primers: p67phox Forward: 5’-CCCACTCCCGGATTTGCTTC -3’ Reverse: 5’-GTCTCGGTTAATGCTTCTGGTAA -3’	IDT	N/A
Primers: p47phox Forward: 5’-GGGGCGATCAATCCAGAGAAC -3’ Reverse: 5’-GTACTCGGTAAGTGTGCCCTG -3’	IDT	N/A
Primers: p22phox Forward: 5’-CCCAGTGGTACTTTGGTGCC -3’ Reverse: 5’-GCGGTCATGTACTTCTGTCCC -3’	IDT	N/A
Recombinant DNA
pCS2+HyPer7-NES	Addgene	Cat# 136467
pCS2+MLS-HyPer7	Addgene	Cat# 136470
pGEX-4T1 SHP-2 WT	Addgene	Cat# 8322
PET53-His-IDH1-Flag	(Chen et al., 2019a)	N/A
PLHCX-IDH1-Flag	(Chen et al., 2019a)	N/A
PLHCX-IDH1-Flag Y42/391F	(Chen et al., 2019a)	N/A
PLHCX empty vector	(Chen et al., 2019a)	N/A
Software and Algorithms
GraphPad Prism 7 software	GraphPad Software	https://www.graphpad.com/
Image J	NIH	https://imagej.nih.gov/ij/index.html

METHOD DETAILS

Antibodies

Antibodies against DYKDDDDK (FLAG) tag, p-Tyrosine (p-Tyr-100) were from Cell Signaling Technology (CST). Antibodies against β-actin and were from Sigma-Aldrich. Antibody against IDH1 was from R&D SYSTEMS. Antibodies against SHP2, PTPMT1 and PTP1B was from PROTEINTECH. Goat anti-Mouse IgG (H+L) secondary antibody and goat anti-rabbit IgG (H+L) secondary antibody were from Thermo Fisher Scientific. Antibodies against p-IDH1 Y42 and p-IDH1 Y391 were custom-made by SHANGHAI GENOMICS, INC. Anti-NOX2 antibody was purchased from Abcam and Bioss. p67phox was purchased from Santa Cruz. Anti-rabbit IgG (H+L) F(ab')2 Fragment (Alexa Fluor 488 conjugate) and Anti-mouse IgG (H+L) F(ab')2 Fragment (Alexa Fluor 555 conjugate) was purchase from Cell Signaling Technology.

Reagents

Rotenone, NAC, GSK2795039,Ebselen, DPI, GSK2795039 and alexidine dihydrochloride were purchased from Sigma-Aldrich. Glutamine L-[5-14C] was from ARC. SHP099 and GKT137831 was purchased from Selleckchem. TCS401, ML090 was purchased from CAYMAN.GLX351322 was purchased from MCE. Mito-TEMPO, DiFMUP and hydrogen peroxide were purchase from Fisher Scientific. Si-IDH1, Si-NOX1, si-NOX3, si-SHP-2 were purchase from Sigma. si-PTP1B and si-PTPMT1 were purchased from Qiagen. si-NOX2 was synthesized by IDT. Primers was synthesized by IDT (The Sequence of siRNA are listed in Key Resources Table).

Retrovirus production and stable cell lines construction

Stable overexpression of IDH1 WT and mutants in A549 was conducted using retroviral vectors harboring RNAi-resistant FLAG-tagged IDH1 WT, and RNAi-resistant FLAG-tagged IDH1 Y42F/Y391F mutant. Briefly, to produce retrovirus, each construct was co-transfected with 0.1 μg VSVG, 0.9 μg EcoPak packaging plasmid, and 1 μg envelope plasmid (Addgene) into HEK293T cells seeded in 6 well plate using TransIT®-LT1 Transfection Reagent (Mirus) according to the manufacturer’s instructions. Retrovirus-containing supernatant medium was collected 48 hours after transfection and filtered by 0.45 μm filter before addition to the indicated host cell lines with 3 μL 10 mg/ml polybrene(1 mol/L HEPES was used to adjust the pH to 7.4 in culture medium). Twenty-four hours after infection, target cells were subjected to hygromycin selection (Invitrogen). The overexpression of proteins was confirmed by Western blotting using antibodies against IDH1.

Purification of Prokaryotic Recombinant IDH1 and SHP-2 Proteins

6 × His-FLAG-IDH1were purified by sonication of high expression BL21(DE3) cells obtained from a 250 mL culture subjected to IPTG induction for 16 hours at 30°C. Bacteria cell lysates were obtained by centrifugations and loaded onto a Ni-NTA column within 20 mmol/L imidazole. The bound proteins were eluted with 250 mmol/L imidazole, followed by desalting using a PD-10 column. SHP-2 expression plasmid was purchased from Addgene (Addgene8322). GST-SHP-2 protein was purified by sonication of high expression BL21(DE3) cells obtained from a 250 mL culture subjected to IPTG induction for 16 hours at 16 °C. Bacteria cell lysates were obtained by centrifugations and loaded onto a GSTrap™ High Performance (Cytiva 17-5281-01). The bound proteins were eluted with25 mM HEPES, pH 7.2, 50 mM NaCl, 5mM GSH, followed by desalting using a PD-10 column. The recombinant IDH1 and SHP-2 protein were stored at −80 °C fridge with adding 5 % glycerol (v/v).

In Vitro SHP-2 dephosphorylation assay

First, Flag-IDH1 protein was treated with EGFR or Src kinase as previous described (Chen et al., 2019a). 1 μg recombinant Flag-IDH1 protein was incubated with diverse recombinant active form of 800 ng EGFR (Thermo Fisher) or 100 ng SRC (Thermo Fisher) in the thermomixer in the presence of 800 μM ATP (Sigma) at 30°C, 300 rpm for 90 minutes in the following assay buffers, respectively. SRC buffer, 50 mM HEPES (pH 7.5), 10 mM MgCl₂, 10% glycerol, 2.5 mM DTT, and 0.01% Triton X-100 were used; for EGFR buffer, 20 mM Tris (pH 7.5), 10 mM MgCl2, 1 mM EGTA, 1 mM Na3VO4, 5 mM β-glycerophosphate, 2 mM DTT, and 0.02% Triton X-100 were used. After kinase reaction, the mixture was incubated with 30 μL of ANTI-FLAG M2 Affinity Gel (Sigma-Aldrich) for 4 hours at 4°C, followed by washing with phosphate-buffered saline (PBS) 3 times to remove unbound materials. Centrifuge the mixture at 3000 rpm for 3 min and discard the supernatants. Then the beads were resuspended using SHP-2 reaction buffer (25 mM HEPES, pH 7.2,50 mM NaCl, 2.5 mM EDTA, 5 mM DTT) including 10 μM SHP-2with or without hydrogen peroxide. The dephosphorylation assay was performed at room temperature for 120 min. Then the IDH1 phosphorylation was examined by Western Blotting.

Small interfering RNA-mediated knockdown

The transfection of small interfering RNA (siRNA) into A549 cells was carried out using Lipofectamine RNAiMAX Transfection Reagent (Thermo fisher), according to the manufacturer’s instructions. Briefly, 5 μL siRNA (10 μM) for each well and RNAiMAX reagent were mixed in Opti-MEM medium (Thermo fisher) and incubated for 5 min at room temperature to allow the complex formation. Then the cells seeded in 6 well plates were washed with Opti-MEM medium (Thermo fisher), and the mixtures were added. Twelve hours after transfection, the culture medium was replaced by fresh complete medium. The cells were harvested 72 hours after transfection, followed by further analysis (Chen et al., 2019a). The following siRNA sequences were used for knockdown: negative control siRNA (non-silencing; QIAGEN; SI03650325); PTP1B or PTPMT1 siRNA was purchased from QIAGEN; SHP2, IDH1 siRNA was synthesized from Sigma. NOX1, NOX2, NOX3 siRNA were synthesized from IDT.

Cell culture treatment

Treatment with Rotenone was performed by incubating 1 × 10⁷ cells with 10 μM Rotenone for 16 hours (Li et al., 2003). Treatment with detachment was performed by placing 1 × 10⁷ cells in Corning® 100 mm Ultra-Low Attachment Culture Dish for 24 hour (Lam et al., 2007). Treatment with hypoxia was performed by incubating 3 × 10⁶ cells under hypoxia (5% CO₂, 1% O₂ and 94% N₂) for 48 hours (Chen et al., 2019a). For PTP inhibitors treatment, SHP099 (10 μM), TCS401 (1 μM) and Alexidine dihydrochloride (1 μM) were added to medium synchronized with Rotenone, detachment or hypoxia (Du et al., 2015; Kenny et al., 2015; Niogret et al., 2019). For NAC, MitoTEMPO or Ebselen treatment, NAC (1 mM), MitoTEMPO (10 μM) or Ebselen (5 μM or 10 μM) was added to medium in synchronized with Rotenone, detachment or hypoxia (Sourbier et al., 2019; Yang et al., 2018). For MEF cells, 3 × 10⁶ MEF cells or MEF SHP-2-ff/ER-cre cells were seeded in 6 well plate, after 16 hours cells treated with 4-OHT (4-hydroxytamoxifen) (0.5 μM) to induce Cre expression and excision of floxed Ptpn11 gene fragment (Wu et al., 2019). For NOXs inhibitors treatment, GSK2795039 (20 μM), GKT137831 (100 nM), GLX351322 (5 μM), ML090 (10 nM) was added to medium synchronized with Rotenone, detachment or hypoxia treatment (Anvari et al., 2015; Casas et al., 2019; Hirano et al., 2015; Moon et al., 2016).

Lipid biosynthesis assay

Lipid biosynthesis assay was performed following the protocol as previously described (Chen et al., 2019a). For ¹⁴C-Lipid biosynthesis assay, cells were incubated with 4 μM Glutamine L-[5-¹⁴C] (ARC) for 48 hours under hypoxia (5% CO₂, 1% O₂ and 94% N₂). For rotenone treatment lipid biosynthesis measurement, cells were incubated with 4 μM Glutamine L-[5-¹⁴C] and rotenone. For detachment treatment lipid biosynthesis measurement, cells were incubated with 4 μM Glutamine L-[5-¹⁴C] in low attachment surface dish (Corning). Lipids were then extracted by the addition of 500 μL of hexane: isopropanol (3:2 v/v), air dried, resuspended in 50 μL of chloroform and transferred to 7 ml glass tubes followed by adding 5ml liquid scintillation cocktail to each tube. Then the bottles were gently inverted 5 times, and subjected to scintillation counting.

SHP2 activity assay

1 × 10⁷ cells that were treated by rotenone, detachment or hypoxia were collected and lysed by 800 μL NP-40 lysis buffer with protease inhibitor cocktail for 20 min under 4 °C. Then the mixture was centrifuged at 13000 rpm for 15 min at 4 °C and the supernatants were collected to 1.5 mL tubes. Then each tube of the cell lysates was incubated with 5 μL anti-SHP2 antibody (PROTEINTECH) at 4°C overnight, then wash with TBS for 3 times. Then 30 μL protein G Sepharose 4 Fast Flow beads were added to each tube following 3-hour incubation at 4 °C. The beads were centrifuged at 3000 rpm for 3min and discard the supernatants and washed by PBS for 3 times. Then the beads were resuspended by 150 μL SHP-2 reaction buffer (25 mM HEPES, pH 7.2,50 mM NaCl, 2.5 mM EDTA,5 mM DTT). 100 μM substrate DiFMUP (Thermo Fisher) were added to mixture to begin the reaction. Then the mixture was transferred to black 96-well plate (Corning). Gently shake the plate for 30 second and measure fluorescence at 455 nm every 1 min for 10 min under 358 nm excitation light using SpectralMax Plus spectrophotometer (Molecular Devices). The remaining beads were used for Western Blot.

Subcellular ROS detection

Cytosolic and mitochondrial ROS levels were measured with the organelle-specific HyPer system (Jiang et al., 2016). For rotenone or hypoxia treatment, 1 × 10⁵A549 cells were planted in 35 mm dishes (Ibidi 81158) and transfected with 1 μg Hyper-cyto (Addgene 136467) or 1 μg Hyper-mito vectors (Addgene 136470). After 48 hours, transfected cells were treated with treated with 10 μM rotenone for or 16 hours or treated with hypoxia for 48 hours. For detachment treatment, 1 × 10⁵cells were seeded in 6-well plate for vectors transfection. After 48 hours, transfected cells were trypsinized and transferred to low attachment 6-well plate (Corning 3471). After culturing 24 hours, cells were transferred to 35 mm dishes for imaging. Images were acquired using LeicaSP5 2photon microscope. ROS levels were calculated as the fluorescence intensity at 488 nm using Fiji software.

Quantitative RT-PCR

RNA was extracted by using TRIzol Reagent (Invitrogen). Quantitative RT-PCR was performed with PrimeScript 1st strand cDNA Synthesis Kit (Takara) and iTaq Universal SYBR Green Supermix (Bio-Rad). Real-time primers targeting NOX1, NOX2, NOX3, NOX4, NOX5, DUOX1, DUOX2, p22phox, p47phox, p67phox and actin were synthesized from Integrated DNA Technologies (IDT). The Sequence of primers are listed in Key Resources Table.

Immunocytochemistry(ICC)staining

For ICC staining, cells were seeded in 35 mm dishes for 24 hours and then treated with rotenone for 16 hours. Then the cells were washed by PBS for 3 times and fixed by 4% paraformaldehyde in PBS for 15 min at room temperature. Then discard the paraformaldehyde and washed the cells with PBS for 3 times and incubated the cells with 0.25% TritonX-100 in PBS for 10 min. After that, wash the cells with PBS 3 times for 5 min. Cells were blocked by 1% BSA in PBST for 30 min at room temperature. Then the NOX2 antibody and p67phox antibody were added to dish and incubated overnight at 4 °C. Discard the mixture solution and wash the cells three times in PBS, 5 mins each wash. Incubate cells with the mixture of anti-rabbit and anti-mouse secondary antibodies in 1% BSA for 1 hour at room temperature in dark. Discard the mixture of the secondary antibody solution and wash three times with PBS for 5 min each in dark. Incubate cells on 1 μg/ml DAPI for 10 min and then wash the cells three times in PBS. Cells were visualized and pictured using the LeicaSP5 2photon microscope.

Quantification and Statistical Analysis

Student’s t-test was used in studies in which statistical analyses were performed to generate p values. p values less than or equal to 0.05 were considered significant. Data with error bars represent mean ± s.d. There is no estimate of variation in each group of data, and the variance is similar between the groups. No statistical method was used to predetermine sample size. The investigators were not blinded to allocation during experiments and outcome assessment. All data are expected to have normal distribution. Statistical analysis and graphical presentation were performed using Prism 5.0 (Graph- Pad) and Microsoft Office Excel 2013.

DATA AND SOFTWARE AVAILABILITY

All software used in this study is listed in the Key Resource Table. Original imaging data have been deposited to Mendeley Data.

Highlights.

• Tyrosine phosphorylation of IDH1 is commonly required for reductive carboxylation

• Reductive carboxylation attenuates SHP-2 to activate IDH1

• Reductive carboxylation elevates cytosolic ROS to inhibit SHP-2

• Diverse signals converge at the NOX2-SHP-2 axis to induce reductive carboxylation

SIGNIFICANCE.

Reductive carboxylation describes an enhanced reductive formation of citrate from glutamine during diverse cellular stress processes including suppressed mitochondrial oxidation due to defects in ETC such as treatment with Complex I inhibitor rotenone, or hypoxia or detachment for anchorage-independent growth. Although lung cancer cells including A549 and H1299 have been extensively used for reductive carboxylation research, this phenomenon is not cancer-specific. Instead, reductive carboxylation is also important to support redox homeostasis as well as biosynthesis in quiescent fibroblasts, retinal pigment epithelium cells and many other types of “non-cancer” cells from different organs including liver, heart, and brown adipocytes. However, the signaling basis underlying induction of reductive carboxylation remains unclear. Here, we report that diverse signals converge at the NOX2-SHP-2 axis, leading to enhanced tyrosine phosphorylation and activation of IDH1 that is commonly required for induction of efficient reductive carboxylation.