Antibodies

Rabbit anti-phospho-p53-Ser58 (1:1000 dilution for immunoblotting (IB) and 1:1000 for immunohistochemistry (IHC); for detection of p-Ser58 in mouse tissues) was generated as described previously,63 except that the peptide GSH-QDVEEFFEGPS(P) was used. Rabbit anti-AXIN (cat# 2074, RRID: AB_2062419; 1:1000 for IB and 1:1000 for immunoprecipitation (IP)), anti-phospho-AMPKα-Thr172 (cat# 2535, RRID: AB_331250; 1:1000 for IB), anti-AMPKα (cat# 2532, RRID: AB_330331; 1:1000 for IB), anti-phospho-ACC-Ser79 (cat# 3661, RRID: AB_330337; 1:1000 for IB), anti-ACC (cat# 3662, RRID: AB_2219400; 1:1000 for IB), anti-phospho-p53-Ser46 (cat# 2521, RRID: AB_10828689; 1:500 for IB and 1:100 for IHC; for detection of p-Ser46 in human cells and tissues), anti-phospho-p53-Ser15 (cat# 9284, RRID: AB_331464; 1:1000 for IB), anti-PARP (cat# 9532, RRID: AB_659884; 1:1000 for IB), anti-BAX (cat# 2772, RRID: AB_10695870; 1:1000 for IB), anti-PUMA (cat#  98672; 1:1000 for IB), anti-PCNA (cat# 13110, RRID: AB_2636979; 1:1000 for IHC), anti-cleaved caspase-3 (cl-casp3; cat# 9661, RRID: AB_2341188; 1:1000 for IB and 1:400 for IHC), anti-caspase-3 (casp3; cat# 14220, RRID: AB_2798429; 1:1000 for IB), anti-HK1 (cat# 2024, RRID: AB_2116996; 1:1000 for IB), anti-HK2 (cat# 2867, RRID: AB_2232946; 1:1000 for IB), anti-PGK1 (cat# 63536; 1:1000 for IB), anti-G6PD (cat# 12263, RRID: AB_2797861; 1:1000 for IB), anti-GAPDH (cat#  5174, RRID: AB_10622025; 1:1000 for IB), anti-HA-tag (cat# 3724, RRID: AB_1549585; 1:1000 for IB), and mouse anti-Myc-tag (cat# 2276, RRID: AB_331783; 1:1000 for IB and 1:100 for IP), and HRP-conjugated mouse anti-rabbit IgG (conformation-specific, cat# 5127, RRID: AB_10892860; 1:2000 for IB) and rabbit anti-mouse IgG (light chain-specific, cat# 58802, RRID: AB_2799549; 1:2000 for IB) antibodies were purchased from Cell Signaling Technology. Rabbit anti-PHGDH (cat# HPA021241, RRID: AB_1855299; 1:1000 for IB) and rabbit anti-FLAG (cat# F7425, RRID: AB_439687; 1:4000 for IB) antibodies were purchased from Sigma. Rabbit anti-p53 (cat# sc-6243, RRID: AB_653753; 1:1000 for IB), mouse anti-p53 (cat# sc-126, RRID: AB_628082; 1:50 for IP, 1:1000 for IB and 1:100 for IHC), anti-PHGDH (cat# sc-100317, RRID: AB_2165393; 1:50 for IP), anti-HA (cat# sc-7392, RRID: AB_2894930; 1:100 for IP), and normal mouse control IgG (cat# sc-2025, RRID: AB_737182; 1:100 for IP) antibodies were purchased from Santa Cruz Biotechnology. Rabbit anti-Pirh2 (cat# ab189907; 1:1000 for IB) and anti-P53-AIP1 (cat# ab3678, RRID: AB_303996; 1:1000 for IHC) antibodies were purchased from Abcam. Rabbit anti-TIP60 (cat# 10827-1-AP, RRID: AB_2128431; 1:1000 for IB), anti-PGK2 (cat# 13686-1, RRID: AB_2161237; 1:1000 for IB), anti-PGAM1 (cat# 16126-1, RRID: AB_2160786; 1:1000 for IB), anti-PGAM2 (cat# 15550-1, RRID: AB_2299336; 1:1000 for IB), anti-HIPK2 (cat# 55408-1, RRID: AB_2881323; 1:1000 for IB), and mouse anti-tubulin (cat# 66031-1-Ig, RRID: AB_11042766; 1:20,000 for IB) antibodies were purchased from Proteintech. Rabbit anti-Ki67 (cat# MA5-14520, RRID: AB_10979488; 1:200 for IHC) and Alexa Fluor™ 488-conjugated, goat anti-mouse IgG (cat# R37120, RRID: AB_255654; 1:300 for IHC) and goat anti-rabbit IgG (cat# A-11034, RRID: AB_2576217; 1:300 for IHC) antibodies were purchased from Thermo. Mouse anti-NOXA (cat# OP180, RRID: AB_2268468; 1:1000 for IB) antibody was purchased from Calbiochem. Normal rabbit control IgG (cat# CR1; 1:100 for IP) was purchased from Sino Biological. HRP-conjugated, goat anti-mouse IgG (cat# 115-035-003, RRID: AB_10015289; 1:5000 for IB) and goat anti-rabbit IgG (cat# 111-035-003, RRID: AB_2313567; 1:5000 for IB) antibodies were purchased from Jackson ImmunoResearch.

Chemicals and assay kits

DMSO (cat# D2650), glucose (cat# G7021), CsCl (cat# 289329), NaHCO3 (cat# S5761), 3-PGA (cat# P8877), N-acetylglucosamine (GlcNAc; cat# A3286), Trizma® base (Tris; cat# T1503), NaCl (cat# S7653), EDTA (cat# E6758), EGTA (cat# E3889), SDS (cat# 436143), formaldehyde solution (formalin; cat# F8775), sodium pyrophosphate (cat# P8135), β-glycerophosphate (cat# 50020), 2-DG (cat# D8375), phenformin (cat# P7045), AICAR (cat# A9978), tamoxifen (cat# T5648), phosphate-buffered saline (PBS; cat# P5493), IPTG (cat# I6758), Na2HPO4 (cat# S7907), NaH2PO4 (cat# S8282), glycerol (cat# G5516), imidazole (cat# I5513), TritonTM X-100 (cat# T9284), L-glutathione reduced (GSH; cat# G4251), hydrazine (cat# 309400), glutamine (cat# 49419), NAD+ (cat# N3014), NADH (cat# N6005), Tween-20 (cat# P9416), polybrene (cat# H9268), lonidamine (cat# L4900), N-nitrosodiethylamine (DEN; cat# N0756), Necrostatin-1 (cat# N9037), GSK-872 (cat# 5303890001), Ferrostatin-1 (cat# SML0583), Erastin (cat# E7781), BSA (cat# A2153), Nonfat-Dried Milk bovine (cat# M7409), methoxyamine hydrochloride (cat# 89803), MTBSTFA (with 1% t-BDMCS; cat# M-108), hexane (cat# 34859), pyridine (cat# 270970), myristic-d27 acid (cat# 68698), bisBenzimide H 33342 trihydrochloride (Hoechst; cat# B2261), HIS-Select Nickel Affinity Gel (cat# P6611), EZview™ Red Anti-c-Myc Affinity Gel (cat# E6654) and corn oil (cat# C8267) were purchased from Sigma. rProtein A Sepharose Fast Flow (cat# 17127904) and Protein G Sepharose 4 Fast Flow (cat# 17061806) beads, and Superdex 200 Increase 10/300 GL (cat# 28990944) column were purchased from Cytiva. Nigericin (cat# HY-127019), SM-164 (cat# HY-15989) and Z-VAD (cat# HY-16658) were purchased from MedChemExpress. Doxycycline (cat# S4163) was purchased from Selleckchem. TNFα (cat# Z01001) was purchased from GenScript. Polyethylenimine (PEI; cat# 23966) was purchased from Polysciences. WesternBrightTM ECL and Peroxide solutions (cat# 210414-73) were purchased from Advansta. Protease inhibitor cocktail (cat# 70221) was purchased from Roche. ProLong™ Diamond Antifade Mountant (cat# P36970), LipofectamineTM 2000 (cat# 11668500), DMEM, high glucose (cat# 12800082), DMEM, no glucose (cat# 11966025), RPMI 1640 medium (cat# 11875093), RPMI 1640, no glucose (cat# 11879020), MEM non-essential amino acids solution (cat# 11140050), fetal bovine serum (cat# 10099141 C), penicillin-streptomycin (cat# 15140163), GlutaMAXTM (cat# 35050061), sodium pyruvate (cat# 11360070), 0.25% trypsin (cat# 15050057) and Prestained Protein MW Marker (cat# 26612) were purchased from Thermo. CCl4 was purchased from Macklin. Apo-BrdU In Situ DNA Fragmentation Assay Kit (cat# K401) was purchased from BioVision. Annexin V-FITC/PI Apoptosis Detection Kit (cat# 40302ES20) was purchased from Yeasen. [U-13C]-3-PGA (cat# P358008) was purchased from Toronto Research Chemicals.

Animals

All mouse experiments were approved by the Institutional Animal Care and the Animal Committee of Xiamen University (XMULAC20180028 and XMULAC20220050).

Wild-type C57BL/6 mice (000664) and Trp53-floxed mice (008462, established as described previously,80 and kindly provided by Dr. Anton Berns from University of Amsterdam) were obtained from The Jackson Laboratory. The Trp53-floxed mice were then crossed with Alb-CreERT2 mice to generate inducible liver-specific Trp53 knockout mice. Hepatic Trp53 was deleted by injecting intraperitoneally the mice with tamoxifen (dissolved in corn oil) at 200 mg/kg, 3 times a week. Mice with PHGDH, p53 and their mutant expression were generated by injection of AAVs carrying PHGDH, p53, or their mutants via the tail vein. Levels of the PHGDH or p53 proteins were analyzed at 4 weeks after the virus injection.

Mice were housed with free access to water and standard diet (65% carbohydrate, 11% fat, 24% protein) under specific pathogen-free conditions. The light was on from 8 a.m. to 8 p.m., with temperature kept at 21–24 °C, and humidity at 40%–70%. For CR, mice were individually caged for 1 month for precondition; each mouse was fed with 2.5 g of standard diet (70% of ad libitum food intake for a mouse at 3 months old and above) at 5 p.m. at each day. Male littermate controls were used throughout the study.

Human primary HCC cells

Human HCC cells were isolated from surgically removed liver tissues. Fresh tissues were minced, followed by digesting in 0.25% (w/v) trypsin supplemented with 0.5 mg/mL collagenase type IV for 10 min at 37 °C. Cells were then immediately plated (at 90% confluence) in collagen-coated 6-well plates in DMEM plus 10% FBS, 100 IU penicillin and 100 mg/mL streptomycin. After 6 h of attachment, the medium was refreshed, and cells were cultured for another 12 h before further use. This study was approved by the Human Research Ethics Committee of the Zhongshan Hospital, affiliated to Xiamen University (XMZSYYKY 2022-072), following the principles of the Declaration of Helsinki. Written informed consent was obtained from all participants.

Mouse HCC induction

The DEN/CCl4-induced mouse HCC model was established as described previously,62 with minor modifications. Briefly, mice at 4 weeks old and 5 weeks old were intraperitoneally injected with 80 mg/mL DEN twice, followed by intraperitoneally injected with 10% (v/v in corn oil) CCl4 at a dose of 5 mL/kg every week from 8 weeks old. AAVs carrying PHGDH and its mutants were intravenously injected every 8 weeks from 16 weeks. Liver samples were collected at 40 weeks old. CR was started at 12 weeks of age.

The MYC;Trp53−/− mouse HCC model was established as described previously,66 with minor modifications. Briefly, mice of 12 weeks old were used, and a mixture containing 30 μg of pX330-sg-p53,81 30 μg of PT3-EF1a-Myc, and 7.5 μg (4:1 ratio) of pCMV/SB10 transposase-encoding plasmids dissolved in 2 mL of 0.9% NaCl solution was prepared before the hydrodynamic tail-vein injection. For each mouse, a total volume of mixture corresponding to 10% of body weight was injected into the lateral tail vein in 5–7 s. Mice were then intraperitoneally injected with 10% (v/v in corn oil) CCl4 at a dose of 5 mL/kg every week beginning at 1 week after hydrodynamic tail-vein injection. AAVs carrying PHGDH and its mutants were intravenously injected once at 2 weeks after hydrodynamic tail-vein injection. Liver samples were collected at 8 weeks after the hydrodynamic tail-vein injection.

Histology

For hematoxylin & eosin (H&E) staining, liver tissues excised from blood-drained mice were cut into 0.5-cm3 cubes, and were fixed in 4% (v/v) paraformaldehyde for 24 h at room temperature, then transferred to embedding cassettes. The cassettes were then washed in running water for 12 h, followed by successive soaking each for 1 h in 70% ethanol (v/v in water), 80% ethanol, and 95% ethanol. The fixed tissues were further dehydrated in anhydrous ethanol for 1 h twice, followed by immersing in 50% xylene (v/v in ethanol) for 30 min, and two changes of xylene, 15 min each. Tissues were then immersed in 50% paraffin wax (58–60 °C; dissolved in equal volume of xylene) for 1 h, followed by two changes of paraffin wax, 1 h each. The dehydrated tissues were embedded in paraffin on a HistoCore Arcadia Paraffin Embedding Machine (Leica). Paraffin blocks were sectioned at a thickness of 5 μm, spread at 42 °C in a water bath for 3 min (HI1210, Leica), dried on an adhesion microscope slide at 42 °C on a heated flattening table (HI1220, Leica) for 12 h, and then incubated at 45 °C in a hot-air dryer for another 4 h. The sections were then de-paraffinized at 70 °C in a hot-air dryer for 4 h, followed by re-hydrating in the following order: two changes of xylene at 70 °C, 10 min each, two changes of anhydrous ethanol 5 min each, two changes of 95% ethanol 5 min each, one change each for 5 min of 80% ethanol, 70% ethanol, and 50% ethanol and then briefly in water. The sections were then stained in hematoxylin solution for 8 min, then washed in running water for 5 min, differentiated in 1% hydrochloric acid (in ethanol) for 30 s, washed in running water for 1 min, and immersed in 0.2% (v/v in water) ammonium hydroxide solution for 30 s, washed in running water for 1 min, and stained in eosin Y solution for 30 s. The stained sections were dehydrated in 70% ethanol for 5 min, twice in 95% ethanol 5 min each, twice in anhydrous ethanol 5 min each, two changes of xylene 15 min each. The stained sections were mounted with Canada balsam and visualized on an Axioscan 7 (Zeiss).

For IHC staining of phospho-p53 and apoptotic markers, liver tissues were fixed, dehydrated, embedded, sectioned and re-hydrated as in H&E staining, and were washed with water for three times, 5 min each at room temperature. The sections were then incubated in pre-heated (~95 °C) citrate antigen retrieval buffer (1 mM sodium citrate, pH 6.0, 0.05% (v/v) Tween-20) for 1.5 min, followed by cooling at room temperature for another 30 min. The sections were then washed with washing buffer (0.1% (v/v) Tween-20 in PBS) twice, 5 min each at room temperature, and then incubated in 10% (w/w) H2O2 (in methanol) solution at room temperature for 5 min, followed by washing with washing buffer for three times, 5 min each at room temperature. The sections were then incubated in 1% (w/v) BSA (diluted with PBS), at room temperature for 20 min. After draining, the liver sections were circled by a PAP pen (cat# Z377821, Sigma), followed by incubation with primary antibodies (diluted in 1% BSA solution) for 16 h at 4 °C in a dark, humidified chamber, followed by washing with washing buffer for 3 times, 5 min each at room temperature. The sections were then incubated with Alexa Fluor 488 goat anti-rabbit IgG or goat anti-mouse IgG (diluted in 1% BSA solution) for 30 min at room temperature in a dark, humidified chamber, followed by washing with washing buffer for 3 times, 5 min each at room temperature. After draining, sections were incubated with 0.4 μg/mL Hoechst solution for 15 min at room temperature, and were mounted with Canada balsam and visualized on an Axioscan 7 (Zeiss), except Supplementary information, Fig. S4b, c, e–g on a DM4 B (Leica).

The TUNEL/BrdU staining was performed as described previously,82 with minor modifications. Briefly, the liver tissues were fixed, dehydrated, embedded, sectioned and re-hydrated as in H&E staining, and were stained using the Apo-BrdU In Situ DNA Fragmentation Assay Kit according to the manufacturer’s instructions. Briefly, the re-hydrated sections were incubated in 0.85% for 5 min, PBS for 5 min, and 4% (v/v) formaldehyde for 15 min, all at room temperature. The sections were rinsed with PBS twice, 5 min each at room temperature, followed by incubating in 20 μg/mL of Proteinase K Solution (prepared by mixing 2 μL of 10 mg/mL Protease K with 998 μL of 100 mM Tris-HCl, pH 8.0 and 50 mM EDTA to generate a 1-mL solution) and rinsing with PBS, 5 min each at room temperature. The sections were then fixed with 4% (v/v) formaldehyde for 5 min at room temperature, followed by rinsing with PBS for 5 min at room temperature. The sections were washed with Wash Buffer (included in the Apo-BrdU In Situ DNA Fragmentation Assay Kit) twice, 5 min each at room temperature, and then incubated with DNA Labeling Solution (by mixing 10 μL of TdT Reaction Buffer, 0.75 μL of TdT Enzyme, 8 μL of Br-dUTP with 32.25 μL of double distilled water to generate a 51 μL solution) for 1 h at 37 °C in a dark, humidified chamber. The sections were then washed with PBS twice, 5 min each at room temperature, followed by incubating with Antibody Solution (by mixing 5 μL of anti-BrdU-FITC antibody and 95 μL of Rinse Buffer) for 0.5 h at room temperature in a dark, humidified chamber. The sections were then incubated with Propidium Iodide (PI)/RNase A solution for 0.5 h at room temperature in a dark, humidified chamber, and then washed twice with double distilled water, 5 min each at room temperature. Sections were mounted with ProLong Diamond Antifade Mountant, and were visualized on an Axioscan 7, except Supplementary Information, Fig. S4d on a DM4 B.

Histochemical images using DM4 B were processed by LAS X software (v.3.0.2.16120, Leica), and those on Axioscan 7 by Zen Blue 3.4 software (Zeiss). Images were formatted by Photoshop 2022 software (Adobe).

Plasmid

Full-length cDNAs used in this study were obtained either by PCR using cDNA from MEFs or by purchasing from Origene, Sino Biological or Genescript. Mutations of PHGDH and AXIN were performed by PCR-based site-directed mutagenesis using PrimeSTAR HS polymerase (Takara). pX330-sg-p53 (#59910, Addgene) plasmid was a kind gift from Dr. Tyler Jacks, PT3-EF1a-Myc (#92046, Addgene) from Dr. Xin Chen, pCMV/SB10 (#24551, Addgene) from Dr. Perry Hackett, and pAAV2/9n (#112865, Addgene) from Dr. James M. Wilson. Expression plasmids for various epitope-tagged proteins were constructed in the pcDNA3.3 vector (#K830001, Thermo) for transfection (ectopic expression), in the pBOBI vector for lentivirus packaging (stable expression), in the pLVX-IRES (#631849, Takara) for doxycycline-inducible expression, or in the pET-28a (#69864-3, Novagen) for bacterial expression. All expression plasmids constructed in this study have been deposited to Addgene (https://www.addgene.org/Sheng-cai_Lin/). The lentivirus-based vector pLV-H1-EF1a-puro (#SORT-B19, Biosettia) was used for the expression of siRNA in HEK293 cells and SK-Hep-1 cells, and the AAV-based vector pAAV2 for mouse liver. The control (siGFP) siRNA was constructed as described previously.83 All constructs were verified by sequencing (Invitrogen, China). The sequences for each siRNA are as follows: 5’-GGCGGTGAAACAGGATTTA-3’ for human HIPK2, 5’-GCTGTCCCAAGCATCAAAT-3’ for human PGK1, 5’-GGTGGTGGAATGGCTTATA-3’ for human PGK2, 5’-GGTCTCAATAAAGCAGAAA-3’ for human PGAM1, and 5’-CCCTACTACAACTCCATTA-3’ for human PGAM2. The siRNAs against human HK1, HK2, G6PD and GAPDH were constructed and validated as described previously.49,84Escherichia coli strain DH5α (cat# PTA-1977) was purchased from ATCC and Stbl3 (cat# C737303) from Thermo. All plasmids were amplified in E. coli strain DH5α, except those mutagenesis in Stbl3. All plasmids were verified by sequencing and were purified using the CsCl density gradient ultracentrifugation method.

Cell culture

In this study, no cell line used is on the list of known misidentified cell lines maintained by the International Cell Line Authentication Committee (https://iclac.org/databases/cross-contaminations/). HEK293T cells (cat# CRL-3216) and HEK293 cells (cat# CRL-1573) were purchased from ATCC, and SK-Hep-1 cells (cat# CL-0212) from Procell Life Science&Technology Co., Ltd. All cell lines were verified to be free of mycoplasma contamination and authenticated by STR sequencing.

HEK293T cells and HEK293 cells were maintained in DMEM (high glucose) supplemented with 3.7 g/L NaHCO3, 10% FBS, 100 IU penicillin, 100 mg/mL streptomycin at 37 °C in a humidified incubator containing 5% CO2. PEI at a final concentration of 10 μM was used to transfect HEK293T cells. The total DNA to be transfected for each plate was adjusted to the same amount by using the relevant empty vector. Transfected cells were harvested at 24 h after transfection. SK-Hep-1 cells were maintained in RPMI 1640 medium supplemented with MEM non-essential amino acids, sodium pyruvate, GlutaMAXTM, 10% FBS, 100 IU penicillin, and 100 mg/mL streptomycin at 37 °C in a humidified incubator containing 5% CO2.

Lentiviruses, including those for knockdown or stable expression, were packaged in HEK293T cells by transfection using Lipofectamine 2000. At 30 h post transfection, medium (DMEM supplemented with MEM non-essential amino acids; ~2 mL) was collected and centrifuged at 5000× g for 3 min at room temperature. The supernatant was then mixed with 10 μg/mL (final concentration; for HEK293) and 20 μg/mL polybrene (final concentration; for SK-Hep-1), followed by centrifuging at 3000× g for 30 min at room temperature (spinfection). Cells were incubated for another 12 h before further treatments.

The genes (PRKAA1, PRKAA2, PHGDH, p53 and AXIN) were deleted from HEK293 cells using the CRISPR-Cas9 system. Nucleotides were annealed to their complements containing the cloning tag aaac, and inserted into the back-to-back BsmB I restriction sites of lentiCRISPRv2 vector. The sequence for each sgRNA is as follows: 5’-GAAGATCGGCCACTACATTC-3’ for PRKAA1, 5’-GGCGGCTCTTTCAGCAGATT-3’ for PRKAA2, 5’-GAAGGGGAAATCTCTCACGG-3’ for PHGDH, 5’-TCTCGAAGCGCTCACGCCCA-3’ for p53, and 5’-GGGGTTGACTGGCTCCCGCC-3’ for AXIN. The constructs were then subjected to lentivirus packaging by transfecting 2 µg of relevant DNA using Lipofectamine 2000 transfection reagent into 105 HEK293T cells cultured in a well of a 6-well plate. At 30 h post transfection, the virus (~2 mL) was collected for infecting HEK293 cells at 40% confluence for 36 h. When cells were approaching to confluence, they were single-cell sorted into 96-well dishes. Clones were expanded and evaluated for knockout status by sequencing.

For glucose starvation, cells were rinsed twice with PBS, and then incubated in glucose-free DMEM or RPMI 1640 supplemented with 10% FBS and 1 mM sodium pyruvate for desired periods of time at 37 °C. For UV irradiation, cells were exposed to UV at a dose of 75 J/m2 on a UV crosslinker (CL-508; UVITEC) after removal of culture medium (DMEM) and plate lid. The irradiated cells were then cultured in DMEM for another 16 h before harvest.

Packaging and injection of AAV

AAVs were packaged in HEK293T cells using the protocol from Grieger et al. 85 In brief, cells used for in-house viral production were maintained in 150-mm dishes. A total of 7 μg of pAAV-RC2/9 (AAV2 inverted terminal repeat (ITR) vectors pseudo-typed with AAV9 capsid) plasmid, 21 μg of pAAV-helper plasmid and 7 μg of pAAV2 plasmid (carrying PHGDH or its mutant) were added to 4 mL of DMEM without phenol red, followed by mixing with 175 μL of PEI solution (1 mg/mL, pH 7.5). The mixtures were then incubated at room temperature for 20 min and then added to the dishes. At 60 h after transfection, cells were collected by scraping and centrifugation. The viral particles were purified from the pellet using an Optiprep gradient as previously described.85 The titers of purified AAV were determined by real-time qPCR as previously described.86 Viruses were stored at –80 °C before use, and were delivered to mice intravenously by lateral tail-vein injection. For each mouse, 1 × 1011 particles of virus, adjusted to 200 μL of final volume (with PBS, pH 7.4) was injected.

Immunoprecipitation and immunoblotting

For immunoprecipitation (IP) of endogenous p53, PHGDH and AXIN, a 10-cm dish of HEK293 cells, SK-Hep-1 cells (grown to 95% confluence), or HCC cells from 0.5 g of patient liver tissue were collected. Cells were lysed with 400 μL/dish of ice-cold lysis buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% (v/w) Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM β-glycerophosphate, with protease inhibitor cocktail), followed by sonication and centrifugation at 4 °C for 15 min. Cell lysates were incubated with respective antibodies overnight. Overnight protein aggregates were pre-cleared by centrifugation at 20,000× g for 10 min, and protein A/G beads (by mixing rProtein A Sepharose Fast Flow beads and Protein G Sepharose 4 Fast Flow beads in the 1:1 proportion, and then balanced with 100 times volume of lysis buffer twice) were then added into the lysate/antibody mixture (to 1% (v/v) final concentration) for another 1 h at 4 °C. The beads were centrifuged and washed with 100 times volume of lysis buffer for 3 times (by centrifuging at 2000× g) at 4 °C and then mixed with an equal volume of 2× SDS sample buffer and boiled for 10 min before immunoblotting.

For IP of ectopically expressed AXIN or PHGDH, HEK293T cells of a 6 cm-dish were used. Cells were transfected with different expression plasmids, and were collected at 24 h posttransfection. For IP of stably expressed AXIN or PHGDH, HEK293 cells of a 10 cm-dish were collected. Cells were lysed in 500 µL of ice-cold lysis buffer, followed by sonication and centrifugation at 4 °C for 15 min. Anti-HA (1:100) or anti-Myc (1:100) antibodies, along with protein A/G beads (1:100), were added into the supernatant and mixed for 4 h at 4 °C. The beads were washed with 200 times volume of lysis buffer for 3 times at 4 °C and then mixed with an equal volume of 2× SDS sample buffer and boiled for 10 min before immunoblotting.

To analyze the levels of p-p53, p-AMPKα and p-ACC in HEK293 cells, SK-Hep-1 cells and primary HCC cells, these cells were grown to 95% confluence in a well of a 6-well dish (except for HCC cells from 0.5 g of patient liver tissues) and were lysed with 250 μL of ice-cold lysis buffer. To analyze the levels of p-p53 and apoptotic markers in mouse livers, 100 mg of freshly excised tissue was lysed with ice-cold lysis buffer (10 μL/mg liver weight), followed by homogenization by a hand-hold homogenizer T 10 basic ULTRA-TURRAX® equipped with an S 10 N - 5 G dispersing tool, IKA. The lysates were sonicated and then centrifuged at 20,000× g for 10 min at 4 °C. After discarding the pellet, an equal volume of 2× SDS sample buffer was added into the supernatant. Samples were then boiled for 10 min before gel electrophoresis and immunoblotting.

All protein samples were subjected to immunoblotting on the same day of preparation without freeze-thaw cycle.

For immunoblotting, the SDS-polyacrylamide gels were prepared as described previously.86 Samples of less than 10 μL (except caspase-3, 20 μL) were loaded into wells, and the electrophoresis was run at 100 V by a Mini-PROTEAN Tetra Electrophoresis Cell (BIO-RAD). The thickness of gels used in this study was 1.0 mm, except for caspase-3, 1.5 mm. All samples were resolved on 8% resolving gels, except PGK1, PGK2, PGAM1, PGAM2, GAPDH and PIRH2 on 10%, caspase-3 on 13%, and PUMA, NOXA and BAX 15%. The resolved proteins were then transferred to the PVDF membrane (0.45 μm, cat# IPVH00010, Merck) as described previously.86 The blotted PVDF membrane was then blocked by 5% (w/v) BSA (for all antibodies against phosphorylated proteins) or 5% (w/v) non-fat milk (for all antibodies against total proteins) dissolved in TBST (40 mM Tris, 275 μM NaCl, 0.2% (v/v) Tween-20, pH 7.6) for 2 h on an orbital shaker at 60 rpm at room temperature, followed by rinsing with TBST for twice, 5 min each. The PVDF membrane was incubated with desired primary antibody overnight at 4 °C on an orbital shaker at 60 rpm, followed by rinsing with TBST for three times, 5 min each at room temperature, and then the secondary antibodies for 3 h at room temperature with gentle shaking. The secondary antibody was then removed, and the PVDF membrane was further washed with TBST for 3 times, 5 min each at room temperature. PVDF membranes were incubated in ECL mixture (by mixing equal volumes of ECL solution and Peroxide solution for 5 min), then life with Medical X-Ray Film (FUJIFILM). The films were then developed with X-OMAT MX Developer and Replenisher and X-OMAT MX Fixer and Replenisher solutions (Carestream) on a Medical X-Ray Processor (Carestream) using Developer (Model 002, Carestream). The developed films were scanned using a Perfection V850 Pro scanner (Epson) with an Epson Scan software (v.3.9.3.4), and were cropped using Photoshop 2022 software (Adobe). Levels of total proteins and phosphorylated proteins were analyzed on separate gels, and representative immunoblots are shown.

Determination of apoptosis by flow cytometry

Effects of low glucose on apoptosis were determined by a flow cytometry-based method using the Annexin V-FITC/PI Apoptosis Detection Kit. Briefly, HEK293 cells and SK-Hep-1 cells grown to 70%–80% confluence in a 6-well dish were washed with 3 mL of PBS (pre-heated to 37 °C) for 3 times, and then trypsinized with 0.25% trypsin, followed by centrifugation at 300× g for 5 min at 4 °C. Some 1 × 105 of trypsinized cells were washed with 0.5 mL of ice-cold PBS twice (by centrifuging at 300× g for 5 min at 4 °C). Cell suspensions were centrifuged at 300× g for 5 min at 4 °C, and then resuspended with 100 μL of Binding Buffer, followed by staining with 5 μL of Annexin V-FITC solution and 10 μL of PI solution for another 15 min at room temperature in the dark. Cell suspensions were then diluted with 400 μL of Binding Buffer, and then immediately subjected to flow cytometry analysis. Flow cytometry was performed on a Cytoflex LX (Beckman Coulter), with the 488-nm (50 mW) laser and the 525/40 filter used to excite and detect the fluorescence of Annexin V-FITC, and the 561-nm (30 mW) laser and 585/42 filter PI. Detector voltages were optimized using a modified voltage titration approach.87 Gating strategies used during the analysis were: (a) the FSC-A and SSC-A for selecting intact cells, and (b) FSC-H and FSC-width for excluding doublets (shown in Supplementary information, Fig. S3a). Gate boundaries were set either based on control samples, or followed density distributions based on best practices. Data were collected by the FACSDiva software (v8.0.2, BD Biosciences), followed by exporting as the FCS 3.1 format. The numbers of apoptotic cells (sum of Q2 and Q3, as indicated in each panel) were quantified with the FlowJo software (v10.4.0, BD Biosciences). During the analysis, a combination of manual gating and computational analysis approaches88 was used.

Measurement of glycolytic intermediates

Levels of glycolytic intermediates were analyzed by capillary electrophoresis-based mass spectrometry (CE-MS), and each measurement required cells collected from one 10-cm dish (60%–70% confluence) or 100 mg of HCC tissue. Cells were washed with 25 mL of 5% (m/v) mannitol solution (dissolved in water), and were instantly frozen in liquid nitrogen. After thawing, cells were then lysed with 1 mL of methanol containing IS1 (internal standard 1; 50 µM L-methionine sulfone, 50 µM D-campher-10-sulfonic acid, dissolved in water; 1:500 (v/v) added to the methanol and used to standardize the metabolite intensity and to adjust the migration time). For analysis of metabolites in the liver, the tissue was freshly excised by freeze-clamping, then washed in pre-cooled 5% mannitol solution and ground in 1 mL of methanol containing IS1. The lysate was then mixed with 1 mL of chloroform and 400 μL of water, followed by 20 s of vortexing. After centrifugation at 15,000× g for 15 min at 4 °C, 450 μL of aqueous phase was collected and was then filtrated through a 5 kDa cutoff filter (cat# OD003C34, PALL) by centrifuging at 12,000× g for 3 h at 4 °C. In parallel, quality control samples were prepared by combining 10 μL of the aqueous phase from each sample and then filtered alongside the samples. The filtered aqueous phase was then freeze-dried in a vacuum concentrator (a LABCONCO #7310037 centrifuge connected to a LABCONCO #7460037 cold trap and an EDWARDS nXDS15i pump) at 4 °C for 12 h, and then dissolved in 100 μL of water containing IS2 (50 µM 3-aminopyrrolidine dihydrochloride, 50 µM N,N-diethyl-2-phenylacetamide, 50 µM trimesic acid, 50 µM 2-naphtol-3,6-disulfonic acid disodium salt, dissolved in methanol; used to adjust the migration time). A total of 20 μL of re-dissolved solution was then loaded into an injection vial (cat# 9301-0978, Agilent Technologies; equipped with a snap cap (cat# 5042-6491, Agilent Technologies)). Before CE-MS analysis, the fused-silica capillary (cat# TSP050375, i.d. 50 µm × 80 cm; Polymicro Technologies) was installed in a CE/MS cassette (cat# G1603A, Agilent Technologies) on the CE system (Agilent Technologies 7100). The capillary was then pre-conditioned with Conditioning Buffer (25 mM ammonium acetate, 75 mM diammonium hydrogen phosphate, pH 8.5) for 30 min, followed by balancing with Running Buffer (50 mM ammonium acetate, pH 8.5; freshly prepared) for another 1 h. CE-MS analysis was run at anion mode, during which the capillary was washed by Conditioning Buffer, followed by injection of the samples at a pressure of 50 mbar for 25 s, and then separation with a constant voltage at –30 kV for another 40 min. Sheath Liquid (0.1 μM hexakis(1H, 1H, 3H-tetrafluoropropoxy)phosphazine, 10 μM ammonium trifluoroacetate, dissolved in methanol/water (50% v/v); freshly prepared) was flowed at 1 mL/min through a 1:100 flow splitter (Agilent Technologies 1260 Infinity II; actual flow rate to the MS: 10 μL/min) throughout each run. The parameters of mass spectrometer (Agilent Technologies 6545) were set as: (a) ion source: Dual AJS ESI; (b) polarity: negative; (c) nozzle voltage: 2000 V; (d) fragmentor voltage: 110 V; (e) skimmer voltage: 50 V; (f) OCT RFV: 500 V; (g) drying gas (N2) flow rate: 7 L/min; (h) drying gas (N2) temperature: 300 °C; (i) nebulizer gas pressure: 8 psig; (j) sheath gas temperature: 125 °C; (k) sheath gas (N2) flow rate: 4 L/min; (l) capillary voltage (applied onto the sprayer): 3500 V; (m) reference (lock) masses: m/z 1,033.988109 for hexakis(1H, 1H, 3H-tetrafluoropropoxy)phosphazine, and m/z 112.985587 for trifluoroacetic acid; (n) scanning range: 50–1100 m/z; and (n) scanning rate: 1.5 spectra/s. Data were collected using MassHunter LC/MS acquisition 10.1.48 (Agilent), and were processed using Qualitative Analysis B.06.00 (Agilent). For quantification of 3-PGA, [U-13C]-3-PGA was used to generate standard curves by plotting the ratios of detected labeled 3-PGA (areas) to the areas of IS1 and IS3, against the added concentrations of labeled 3-PGA. The amount of 3-PGA was then estimated according to equations generated from standard curves. The average cell volume was determined to be 2000 μm3 as described previously,89 and cell density 1.1 g/mL according to ref. 90

Measurement of intracellular amino acids

Levels of pyruvate and amino acids were determined by gas chromatography-mass spectrometry (GC-MS) as described previously,91 with minor modifications. Briefly, HEK293 cells from one 10-cm dish (60%–70% confluence) were collected for each measurement. Cells were incubated with DMEM (no glucose) supplemented with 1 mM sodium pyruvate, 10% FBS and 25 mM glucose or not (for starvation) for 2 h, and then lysed with 1 mL of 80% methanol (v/v in water) containing 10 µg/mL myristic-d27 acid as an internal standard, followed with 20 s of vortexing. After centrifugation at 15,000× g for 15 min at 4 °C, 800 µL of each supernatant (aqueous phase) was freeze-dried at 4 °C for 24 h. The lyophilized samples were then subjected to derivatization by vortexing for 1 min after mixing each with 50 µL of freshly prepared methoxyamine hydrochloride (20 mg/mL in pyridine), followed by incubation at 4 °C for 1 h. The mixtures were sonicated at 0 °C by bathing in ice slurry for 10 min, and were then incubated at 37 °C for 1.5 h, followed by mixing with 50 µL of MTBSTFA and incubated at 55 °C for 1 h. Before subjecting to GC-MS, samples were centrifuged at 15,000× g for 10 min, and some 60 μL of each supernatant was loaded into an injection vial (cat# 5182-0714, Agilent; with an insert (cat. HM-1270, Zhejiang Hamag Technology)) equipped with a snap cap (cat# HM-0722, Zhejiang Hamag Technology). GC was performed on an HP-5MS column (30 m × 0.25 mm i.d., 0.25 μm film thickness; cat# 19091S-433; Agilent) using a GC/MSD instrument (7890-5977B, Agilent). Briefly, the injector temperature of GC/MSD was set at 260 °C. The column oven temperature was first held at 70 °C for 2 min, then increased to 180 °C at the rate of 7 °C/min, then to 250 °C at 5 °C/min, then to 310 °C at 25 °C/min, where it was held for 15 min. The MSD transfer temperature was 280 °C. The MS quadrupole and source temperature were maintained at 150 °C and 230 °C, respectively. Measurements were performed in both a scan mode and a selected ion monitoring mode. The following m/z values were used for each compound: 174 for pyruvate; 260 for L-alanine; 418 for L-aspartic acid; 432 for L-glutamic acid; 246 for L-glycine; 258 for L-proline; 390 for L-serine; 200 for L-leucine and L-lsoleucine; 302 for L-phenylalanine; 302 for L-tyrosine; 244 for L-tryptophan; 302 for L-threonine; 406 for L-cysteine; 320 for L-methionine; 417 for L-asparagine; 431 for L-glutamine; 300 for L-lysine; and 440 for L-histidine. Data were collected using the MassHunter GC/MS Acquisition software (v.B.07.04.2260, Agilent). For quantification, peaks were extracted and integrated using GC-MS MassHunter Workstation Qualitative Analysis software (v.B.07.01SP1, Agilent).

Protein expression

The pET-28a-PHGDH, pET-28a-PHGDH-V261M and pET-28a-PSAT1 plasmids were transformed into the E. coli strain BL21 (DE3) (cat# EC0114, Thermo), followed by culturing in LB medium in a shaker at 200 rpm at 37 °C. The cultures of transformed cells were induced with 0.1 mM IPTG at an OD600 of 1.0. After incubating for another 12 h at 160 rpm at 16 °C, the cells were collected, followed by homogenization in a His binding buffer (50 mM sodium phosphate, pH 7.4, 150 mM NaCl, 1% Triton X-100, 5% glycerol, and 10 mM imidazole) on ice. The homogenates were then sonicated on ice, and were subjected to centrifugation at 150,000× g for 30 min at 4 °C, followed by incubating with Nickel Affinity Gel (pre-balanced with His binding buffer; 1:100) on a rotator for 2 h at 4 °C. The Nickel Affinity Gel was then washed with 100 times the volume of ice-cold His wash buffer (50 mM sodium phosphate, pH 7.4, 150 mM NaCl, and 20 mM imidazole), followed by incubating with His elution buffer (50 mM sodium phosphate, pH 7.4, 150 mM NaCl, and 250 mM imidazole) at 4 °C. Proteins were concentrated to ~3 mg/mL by ultrafiltration (UFC905096, Millipore) at 4 °C, then subjected to gel filtration (Superdex 200) balanced with a buffer containing 50 mM Tris-HCl, pH 7.4 and 150 mM NaCl at 4 °C.

Determination of PHGDH enzymatic activity

The enzymatic activity of PHGDH was determined as described previously,50 with minor modifications. Briefly, 10 μg of PHGDH and 5 µg of PSAT1 were pre-incubated in 100 μL of Reaction buffer (333 mM Tris-HCl, pH 9.0 and 1.7 mM EDTA, 3.3 mM GSH, 333 mM hydrazine, 10 mM glutamine and 2 mM NAD+) at 37 °C for 5 min in a well of a glass-bottom, 96-well microplate (cat# 3635, Corning). The reaction was initiated by pipetting 200 μL of 3-PGA solution (at desired concentrations; dissolved in Reaction Buffer) pre-warmed at 37 °C, into the well, followed by mixing and recording OD340 at 30-s intervals on a SpectraMax M5 microplate reader (Molecular Devices). The PHGDH activities were assessed with the initial velocities of NADH formation calculated according to the standard curve by plotting the values of OD340 against the added concentrations of NADH. All measurements were carried out in triplicate. Data were collected using the SoftMax Pro software (v.5.4.1.1, Molecular Devices) and exported to Graphpad Prism software (v.9.5.1, Graphpad) for further analysis.

Identification of p53-binding proteins

Twenty 10-cm dishes of HEK293 cells glucose-starved for 2 h were lysed with 10 mL of ice-cold lysis buffer, followed by sonication. After centrifugation, supernatants were incubated with 200 μL of mouse anti-p53 antibody or mouse control IgG antibody at 4 °C for 4 h. Protein aggregates were cleared by centrifugation at 20,000× g for 15 min, and 200 μL of protein A/G beads (pre-balanced with 100 times volume of lysis buffer twice) were then added into the lysate/antibody mixture for another 1 h at 4 °C. The beads were centrifuged and washed with 100 times volume of lysis buffer for 3 times (by centrifuging at 2000× g) at 4 °C and then mixed with a 1/5 volume of 5× SDS sample buffer and boiled for 10 min, followed by SDS-PAGE. The gels were stained with Coomassie Brilliant Blue R-250 dye, followed by decoloring. Gels were then excised, and were subjected to in-gel trypsin digestion and dried. Samples were analyzed on a nanoElute (Bruker) coupled to a timsTOF Pro (Bruker) equipped with a CaptiveSpray source. Peptides were dissolved in 5 μL of 0.1% formic acid (v/v) and were loaded onto a homemade C18 column (35 cm × 75 μm, ID of 1.9 μm, 100 Å). Samples were then eluted for 60 min with linear gradients of 3%–35% acetonitrile (v/v, in 0.1% formic acid) at a flow rate of 0.3 μL/min. MS data were acquired with a timsTOF Pro mass spectrometer (Bruker) operated in PASEF mode, and were analyzed using Peaks Studio software (X+, Bioinformatics Solutions). The mouse UniProt Reference Proteome database was used during data analysis.

Statistical analysis

Statistical analyzes were performed using Prism 9 (GraphPad Software). Each group of data was subjected to Kolmogorov-Smirnov test, Anderson–Darling test, D’Agostino–Pearson omnibus test or Shapiro–Wilk test for normal distribution when applicable. An unpaired two-tailed Student’s t-test was used to determine significance between two groups of normally distributed data. Welch’s correction was used for groups with unequal variances. An unpaired two-tailed Mann–Whitney test was used to determine significance between data without a normal distribution. For comparisons between multiple groups, an ordinary one-way or two-way ANOVA was used, followed by Tukey, Sidak, Dunnett or Dunn as specified in the figure legends. The assumptions of homogeneity of error variances were tested using F-test (P >  0.05). For comparison between multiple groups with two fixed factors, an ordinary two-way ANOVA was used, followed by Tukey’s or Sidak’s multiple comparisons test as specified in the figure legends. Geisser–Greenhouse’s correction was used where applicable. The adjusted means and SEM, or SD, were recorded when the analysis met the above standards. Differences were considered significant when P < 0.05, or P > 0.05 with large differences of observed effects (as suggested in refs. 92,93).