SAPS3 protein reduces AMPK activity and helps regulate metabolism during diet changes in male mice

We next sought to determine if the SAPS3-containing PP6 phosphatase complex dephosphorylates AMPK. First, we examined whether the association of SAPS3/PP6C with AMPK is responsive to glucose. Interestingly, binding of the SAPS3/PP6 complex with AMPK started to accumulate after glucose withdrawal and reached maximal level upon short-term glucose addback. The complex dissociated after the return of glucose for 3 h indicating the phosphatase complex is assembled when AMPK is phosphorylated, perhaps in preparation for nutrient recovery and subsequent dephosphorylation (Fig. 1f). We next tested whether SAPS3 had a preference for the phosphorylated form of AMPK and found that SAPS3 preferably associated with phospho-AMPK (Fig. 1g). To gain more insight into the dynamics of this complex, we used recombinant protein binding assays and found that SAPS3 was required for the association of PP6C and AMPK (Fig. 1h). These data suggest that SAPS3 is essential for bringing the PP6 catalytic subunit to AMPK.

Next, we investigated whether SAPS3 affects AMPK signaling. We found that knockdown of SAPS3 using shRNA significantly decreased dephosphorylation of both AMPK and its downstream substrate, ACC, upon glucose addback (Fig. 1i and Supplementary Fig. 1d, e). SAPS3 is part of a family that includes three members, SAPS1, SAPS2, and SAPS3. To identify whether these other family members play a role in the mediation of AMPK, we knocked down SAPS1, SAPS2, or SAPS3 and found only knockdown of SAPS3 was able to modulate AMPK dephosphorylation upon glucose addback (Fig. 1j and Supplementary Fig. 1f).

One of the main functions of the liver is to respond to low glucose levels by triggering gluconeogenesis, a response in part mediated by AMPK²¹. Thus, we asked whether loss of SAPS3 in the liver would affect the systemic response to fasting. We found that LKO mice were not affected by fasting glucose levels under normal dietary conditions (Fig. 3g). Interestingly, LKO mice had significantly lower levels of fasting blood glucose compared to FF mice under HFD (Fig. 3g). Moreover, LKO mice under HFD cleared blood glucose more efficiently than FF mice and performed significantly better in ITT (Fig. 3h and Supplementary Fig. 2g), again suggesting SAPS3 is important for maintaining metabolic homeostasis under HFD but not required when the metabolic environment is balanced. LKO mice also showed increased levels of the plasma C-peptide, a proportional readout for insulin sensitivity (Fig. 3i). Next, we investigated whether the function of SAPS3 is through regulating AMPK activation and found that SAPS3 deletion led to an increase in phospho-AMPK and phospho-ACC compared to FF under HFD (Fig. 3j and Supplementary Fig. 2h). Therefore, liver-specific SAPS3 knockout recapitulates the phenotype observed in SAPS3 whole body knock-out and is sufficient to hinder the onset of diet-induced liver steatosis, obesity, and insulin resistance.

Metabolic syndromes bring about diverse and severe changes to the transcriptional landscape. Indeed, HFD is reported to induce changes in gene expression profiles. With this in mind, we investigated the gene expression profile of the livers from mice lacking SAPS3 and fed HFD. We performed RNA-seq from liver tissues collected from FF and LKO mice following 16 weeks on HFD. The liver samples from FF and LKO mice under CD or HFD clustered by genotype, confirming that genotype-dependent transcriptional changes dominated the dataset. Interestingly, there was almost no difference in gene expression between FF and LKO livers under CD, again enforcing the idea that SAPS3 may be a metabolic stress-responsive protein. In contrast, HFD-induced differential gene expression is significant in SAPS3 knockout mice (Fig. 4c). Among 2205 genes that were significantly altered by HFD in FF mice, 1724 (78%) of these genes were completely reversed by the deletion of SAPS3, and 56 (3%) of these genes were partially reversed by SAPS3 knockout (Fig. 4d and Supplementary Fig. 4a). Together, we identified 1780 (81%) out of 2205 genes that were altered by HFD but reversed by SAPS3 deletion, suggesting that SAPS3 deletion hinders the transcriptional changes caused by HFD. To better understand the networks of genes regulated by SAPS3 during HFD, we used unbiased KEGG pathway analysis on the set of 1780 genes and found the most affected pathways by loss of SAPS3 were metabolic pathways that involve lipid handling, including Sphingolipid metabolism, biosynthesis of unsaturated fatty acids and fatty acid metabolism (Fig. 4e). Additionally, the PPAR pathway and adipocytokine pathway, which are activated by HFD^23,24, were decreased by SAPS3 knockout. Moreover, we used gene set enrichment analysis (GSEA) to unbiasedly explore the differential pathways in FF and LKO mice fed HFD. We found that CYP450 metabolism, the reported top-upregulated pathway in mice in response to HFD²⁵, was reversed by the deletion of SAPS3 (Fig. 4f). One of the most well-known consequences of HFD is cardiovascular diseases. Remarkably, the pathway involving pathogenesis of cardiovascular diseases was decreased by SAPS3 knockout (Fig. 4f). Importantly, and in line with our metabolomics analysis, adipogenesis and long-chain fatty acid synthesis were impaired by the loss of SAPS3 (Fig. 4g). When we investigated the expression of the gatekeeper enzymes for fatty acid synthase in the LKO and FF mice following HFD, we found decreased expression of fatty acid synthase (FAS) and acetyl-CoA carboxylase 2 (ACC2) further supporting the loss of SAPS3 blocks lipogenesis in the liver (Supplementary Fig. 4b). To further confirm the fatty acid synthesis was inhibited by the loss of SAPS3, we performed experiments using stable isotope-labeling. Using labeled palmitate tracing in WT and KO MEF cells, we found less stearate and oleate were synthesized from U-¹³C₁₆ palmitate when SAPS3 was lost, indicating synthesis of both saturated and unsaturated fatty acids is affected by SAPS3 deletion (Supplementary Fig. 4c). Thus, RNA expression profiling data demonstrate that the deletion of SAPS3 has a significant effect on gene expression under HFD and a limited influence on gene expression during balanced nutrition.

To further confirm if SAPS3 deletion regulates metabolic pathways through activation of AMPK, we knocked out AMPK α1 and α2 in WT and SAPS3 KO MEF cells using CRISPR- based gene editing (Fig. 5g). As above, we used ¹³C-glucose isotope tracer labeling to follow the fate of glucose to fatty acids. Deletion of AMPK significantly reduced glycolytic intermediates in WT cells as previously shown²⁹. As expected, the knockout of SAPS3 increased the concentration of glycolytic intermediates, which can be completely blocked by the deletion of AMPK (Fig. 5h). Meanwhile, loss of AMPK reversed the levels of labeled fatty acids that were reduced by SAPS3 knockout (Fig. 5i). Taken together, these results demonstrate that metabolic alterations in SAPS3 knockout cells were largely through activation of AMPK.

To further determine whether SAPS3 functions directly through hepatic activation of AMPK, we used Cre-dependent adeno-associated virus (AAV) expressing a dominant negative (DN) kinase dead (K45R) form of AMPK in the liver³⁰, then fed with HFD (Fig. 6g). Since LKO mice have liver-specific expression of Cre protein, AMPK-DN was only expressed in LKO mice, not in FF mice. The AMPK-DN expression was confirmed via Western blotting by reduced phospho-ACC in liver samples after HFD (Fig. 6h and Supplementary Fig. 6c). As a control, we found there was no effect on phospho-ACC in lung tissue (Supplementary Fig. 6d). We found the body weights of LKO mice with AAV-AMPK-DN were similar to FF mice under HFD (Fig. 6i) with similar food intakes (Supplementary Fig. 6e), suggesting that AMPK-DN suppresses the beneficial effects found in SAPS3 LKO mice in response to HFD. Moreover, AMPK-DN reverses the effect of SAPS3 knock-out on fatty acid synthesis and lipid accumulation in vivo (Fig. 6j). We also observed a similar level of liver hypertrophy and liver steatosis in LKO mice treated with AAV-AMPK-DN compared to FF mice (Fig. 6k, l and Supplementary Fig. 6f). When we investigated glucose and insulin sensitivity in these mice, we found under HFD, LKO mice with AMPK-DN reversed the sensitivity exhibited by SAPS3 LKO (Fig. 6m and Supplementary Fig. 6g) and also reversed the levels of fasting blood glucose (Supplementary Fig. 6h).

All studies involving mice were performed according to the Institutional Animal Care and Use Committee (IACUC)-approved protocols at the University of California, Irvine (IACUC protocol: AUP-20-159). We have complied with the relevant ethical considerations for animal research overseen by this committee. Ppp6r3^fl/fl C57BL/6 mice were generated by inGenious Targeting Laboratory. Ppp6r3^+/- mice were generated by microinjecting Cre recombinase RNA to ppp6r3^wt/fl embryos⁵⁵. WT and SAPS3 KO mice were generated by breeding ppp6r3^+/- mice. LKO mice were generated by breeding ppp6r3^fl/fl mice with C57BL/6^Alb-Cre mice obtained from the Jackson Laboratory. The following diets were used: Control diet (10 kcal% fat, D12450B, Research Diet Inc.); High Fat Diet (45 kcal% fat, D12451, and 32.5 kcal% fat, D12266, Research Diet Inc.). All mice were housed on a 12 h light/12 h dark cycle with free access to water and diet. Body weights were measured every week. Body fat and lean mass were measured in conscious animals using a whole-body composition analyzer (EchoMRI). For Compound C treatment, 3 mg/kg of compound C was intraperitoneally injected into the experimental group mice every day for two weeks. For AAV injection, pAAV-EF1a-Flex-DN-AMPK-K45R-T2A-mCherry plasmid⁵⁶ was amplified in recombination-deficient bacteria Stbl3, and Serotype 8 AAVs were packaged by the University of North Carolina Vector Core and the following titer was achieved (vg/ ml): 1.9 × 10¹².

293 T cells (ATCC® CRL-3216™) and HT1080 cells (ATCC® CCL-121™) were purchased from American Type Culture Collection (ATCC). SAPS3 KO and WT MEFs were generated from SAPS3 KO and WT mice. All cells were cultured in Dulbecco's modified Eagle's medium (DMEM) (10-017-CV, Corning) supplemented with 10% fetal bovine serum (FBS) (100−106, Gemini Bio-Products), 100 units/mL of penicillin, and 100 μg/mL of streptomycin (516106, Sigma). For glucose deprivation and add-back treatment, HT1080 cells were cultured in DMEM without glucose (11966025, Gibco), supplemented with 10% dialyzed FBS (FB-07, Omega) for 1 h, then the medium was replaced with the complete medium for 20 min. 293 T cells were cultured in DMEM without glucose, supplemented with 10% dialyzed FBS for 6 h, then the medium was replaced with the complete medium for 1 h or 3 h. For siRNA transfection, siSAPS1 (L-020420-01-0005, Horizon), siSAPS2 (L-021331-01-0005, Horizon) and siSAPS3 (L-014646-01-0005, Horizon) were transfected by Lipofectamine RNAiMAX transfection reagent (13778150, Invitrogen). To generate lentiviral particles, 293 T cells were cultured in a 6 cm dish and co-transfected with 1 μg control or shRNA vector, 0.5 μg pCMV-VSV-G, and 1 μg pCMV-dR8.2 dvpr by Lipofectamine 2000 (11668−019, Invitrogen) transfection reagent. The viral supernatants were collected and filtered. Cells were infected with the virus with 10 μg/ml polybrene.

A total of 293 T cells were transfected with plasmids encoding Flag-tagged AMPKα2 protein. Cells were deprived of glucose for 6 h and glucose was added back for 1 h, followed by lysis using RIPA buffer (150 mM KCl, 0.2% (v/v) NP-40, 10% (v/v) glycerol, 20 mM Tris at pH 7.5, 0.5 mM DTT) containing protease inhibitor complex (4693159001, Roche) and phosphatase inhibitor (1862495, Thermo fisher scientific). 20 μl anti-Flag beads (A2220, Sigma) were added to the cell lysates and rotated overnight. Beads were washed four times with cold lysis buffer and eluted with Flag peptide (F3290, Sigma) for 4 h. The elutes were separated by SDS-PAGE and visualized by Coomassie Blue staining (LC6025, Invitrogen). The binding proteins were identified by mass spectrometry analysis, performed by the Taplin Mass Spectrometry Facility at Harvard Medical School.

Cells were washed twice with PBS and fixed with 4% formaldehyde (Thermo Scientific, 28908) in PBS for 15 min at room temperature. Cells were washed again three times with PBS and permeabilized with 0.1% TritonX-100 (Fisher, BP151-500) in PBS at room temperature for 10 min. Then cells were washed three times with PBS and blocked with 3% BSA in PBS for 1 h. Thereafter, cells were incubated with anti-AMPKα (Thermo Scientific, MA537501, 1:50) and anti-SAPS3 (Thermo Scientific, 16944-1-AP, 1:100) antibodies overnight at 4 °C. Cells were washed three times with PBS and incubated with goat anti-rabbit 594 (Invitrogen, A11037, 1:500) and goat anti-mouse 488 (Invitrogen, A11029, 1:500) antibodies for 1 h at room temperature. After that, cells were washed three times and mounted with the Prolong Gold antifade reagent with DAPI (Invitrogen, P36931). The slides were visualized with Zeiss LSM 900 Airyscan.

IP was performed as previously described⁵⁷. Cells were cultured in 10 cm dishes in DMEM supplemented with 10% FBS. 80% confluent cells were washed twice with ice-cold PBS and lysed on ice using a cell scraper with lysis buffer (150 mM KCl, 0.2% (v/v) NP-40, 10% (v/v) glycerol, 20 mM Tris at pH 7.5, 0.5 mM DTT) containing protease inhibitor complex (4693159001, Roche). The protein concentrations of lysates were measured using the BCA Assay kit (23225, Thermo fisher scientific). 1 mg lysates were immunoprecipitated when rotating at 4֯C with 4 μg of antibodies or control IgG. A total of 20 μl Protein G agarose beads (15920010, Thermo fisher scientific), anti-Flag beads (A2220, Sigma), or anti-Myc beads (20169, Thermo fisher scientific) were added to each tube and rotated overnight. Beads were washed four times with cold lysis buffer, resuspended in SDS loading dye, and boiled for 5 min.

Tissues were homogenized, or cells grown in 6 cm dish were lysed on ice in lysis buffer (50 mM Tris-HCL [pH 7.4], 5 mM Sodium Fluoride, 5 mM Sodium Pyrophosphate, 1 mM EDTA, 1 mM EGTA, 250 mM Mannitol, 1% [v/v] Triton X-100) containing protease inhibitor complex (04693159001, Roche) and phosphatase inhibitor (1862495, Thermo fisher scientific). The protein concentrations were measured by using the BCA Assay kit (23225, Thermo fisher scientific). The lysates were boiled with NuPAGE LDS-PAGE sample buffer (1771559, Invitrogen) supplemented with 5% β-mercaptoethanol (M3148, Sigma) for 5 min. Equal amounts of protein were loaded on precast NuPAGE Bis-Tris Gels (NP0321BOX, Life Technologies) followed by transfer onto nitrocellulose membrane (1620115, Bio-Rad). The immunoblotting was performed with the following antibodies: anti-AMPKα (ab80039, Abcam, 1:1000), anti-pAMPKα (ab133448, Abcam, 1:1000), anti-pAMPK (4188, Cell signaling, 1:1000), anti-ACC (3662, Cell signaling, 1:1000), anti-pACC (3661, Cell signaling, 1:1000), anti-β-ACTIN (A1978, Sigma,1:5000), anti-GAPDH (2218, Cell signaling, 1:1000), anti-SAPS3 (A300-971A, Bethyl, 1:1000), anti-PP6C (A300-844A, Bethyl, 1:1000), anti-MYC (ab9106, abcam, 1:1000), anti-FLAG (F3165, Sigma, 1:1000), anti-SAPS3 (A300-972A, Bethyl, 1:1000), anti-SAPS1 (A300-968A, Bethyl, 1:1000), and anti-SAPS2 (A300-969A, Bethyl, 1:1000). All the uncropped images of blot results were shown in Supplementary Fig.. 7

PCR was performed using GoTaq® Master Mixes (M7122, Promega). PCR products were detected using agarose gel electrophoresis (1.5%). PCR primers are ppp6r3-F: 5'-TTCACACATACCCAGGAATCAGA-3'; R: 5'-GACTGCTGAGACACAAGGGC-3'. The PCR cycle parameters were as follows: 95 ˚C for 3 min; 40 cycles with denaturation at 95 ˚C for 10 sec, annealing at 50 ˚C for 30 sec.

Blood glucose levels were determined using an automated blood glucose reader (Accu-Check; Roche). Glucose tolerance tests were performed on mice that were fasted overnight. Blood was collected immediately before as well as 15, 30, 60, 90, and 120 min after intraperitoneal injection of glucose (1 g/kg body weight). For insulin tolerance tests, mice were fasted for 3 h and then injected with 0.5 U/kg body weight of insulin (91077 C, Sigma). Blood glucose was measured at 0, 15, 30, 60, 90, and 120 min.

After mice were euthanized, tissues were collected and fixed in 10% formalin. Formalin-fixed, paraffin-embedded blocks of tissues were used for haematoxylin and eosin (H&E) staining⁵⁷. 5 µm paraffin-embedded formalin-fixed slides were deparaffinized with xylene, and rehydrated by different concentrations of ethanol. Rinse the slides in distilled water and stain them in hematoxylin/Eosin. After dehydrating the slides in different concentrations of ethanol, slides were cleared in Xylene, and mounted with Cytoseal (Thermo fisher scientific). Image analysis was performed under a microscope (SeBa Laxco).

Mice were fasted for 4 h before blood was collected via cardiac puncture in BD vacutainer tubes (Fisher Scientific). The serum was obtained by centrifugation at 2000 x g for 15 min. Serum triglyceride (ab65336, Abcam), C-peptide (90050, Crystal Chem), alanine transaminase (ALT) (EALT-100, EnzyChrom), and aspartate transaminase (AST) (EASTR-100, BioAssay System) were measured according to manufacturers' instructions.

For the triglyceride test, 50 µl standard dilutions or samples (5 µl sample was adjusted volume to 50 µl with triglyceride assay buffer) were loaded into plates. Then, 50 µl of the reaction mixture was added to each standard, sample, and background control wells. The plate was mixed and incubated at room temperature for 60 min protected from light. After that, the absorbance was measured using a plate reader.

For the C-peptide test, 95 µl of sample diluent and 5 µl of sample or working standards were added to each well. The microplate was covered with the plate sealer and mixed for 10 sec. Then, the plate was incubated for 1 h at room temperature. The contents were aspirated, and the wells were washed with 300 µl of wash buffer six times followed by adding 100 µl anti-C peptide enzyme conjugate. The microplate was covered with the plate sealer and incubated for 1 h at room temperature. After that, contents were aspirated, and wells were washed with 300 µl of wash buffer six times. Immediately, 100 µl per well of the enzyme-substrate solution was dispensed. The plate was covered to let the mix react for 30 min at room temperature. In the end, the reaction was stopped by adding 100 µl of enzyme reaction stop solution, and the absorbance was measured using a plate reader.

For the ALT test, 20 µl of each sample was transferred to a 96 well plate. 200 µl working reagent (200 µl assay buffer, 5 µl cosubstrate, 1 µl LDH, and 4 µl reconstituted NADH) were added to the sample and standard wells. The plate was taped to mix and incubated at 37 ֯C. The plate was read at 5 min and 10 min. For the AST test, 20 µl of each sample was transferred to a 96 well plate. 200 µl working reagent (200 µl assay buffer, 1 µl cofactor, 1 µl enzyme mix, and 4 µl NADH) were added to the sample and standard wells. The plate was taped to mix and incubated at 37 ֯C. The plate was read at 5 min and 10 min.

A total of 3−10 mg of tissue samples were cut on dry ice and soaked in pre-cooled 80% methanol in HPLC-grade water. Samples were homogenized by Precellys 24 homogenizer using Precellys ceramic kit. After centrifuging at 17,000 x g at 4 ֯C for 10 min, the supernatant was transferred to a new tube and dried by speed vacuum. Liquid chromatography-mass spectrometry (LC-MS) was carried out at Duke University. The HPLC (Ultimate 3000 UHPLC) with an Xbridge amide column (Waters) is coupled to Q exactive plus hybrid quadrupole-orbitrap mass spectrometer (QE-MS) (Thermo Scientific) for compound separation and detection⁵⁸. 301 metabolites were detected from all of the samples Mass isotopomer distributions were determined by integrating metabolite ion fragments and corrected for natural abundance as described⁵⁹. MetaboAnalyst was used to analyze significantly changed metabolites and generate the heatmap and KEGG pathway analysis (www.metaboanalyst.ca/↗).

The pull-down was performed as previously described⁶⁰. 3 μg of recombinant Myc-PPP6R3 protein (TP328606, OriGene), GST-PPP6C protein (H00005537-P01, Novus), His-pAMPK (P48-10H-10, SignalChem) or His-AMPK (P48-14H-20, SignalChem) were mixed. The mixture was incubated at 37 ֯C for 1 h in 200 μl of assay buffer containing 25 mM Tris (pH8.0), 150 mM NaCl, 2 mM dithiothreitol, and 1 mg/ml BSA to block non-specific binding. Then, 20 μl His select Nickel affinity beads (P6611, Sigma) or 20 μl Protein G agarose beads with 4 μl anti-PP6C antibody were added to the assay buffer and rotated for 1 h at room temperature. After three washes with the assay buffer, the beads were resuspended in SDS loading dye, and examined by western blotting.

A total of 2 × 10⁵ cells WT or SAPS3 KO MEFs were plated on each well of a 6-well plate and cultured overnight. Then, the medium was replaced with 1 mL 10 mM glucose DMEM supplemented with 10% FBS. After culturing for 24 h, the medium was collected, and the cell number was counted by an automated cell counter (Bio-Rad). After cell debris was spun down, 0.8 mL medium was used to measure glucose by the Nova Biomedical BioProfile 100 with fresh 10 mM glucose DMEM as control. This analysis was as followed: glucose uptake = glucose in fresh medium (mM)-glucose in cultured medium (mM). All metabolite measurements were normalized based on cell number.

2×10⁵ cells were seeded in 6 cm plates containing DMEM supplemented with 10% FBS and cultured overnight. Cells were washed with PBS twice and cultured in glucose-free DMEM supplemented with 10% dialyzed FBS containing ¹³C₆-glucose (10 mM; Cambridge Isotope Laboratory). After culturing for 6 h which led to 85% labeling of the total glucose pool, the medium was aspirated at room temperature. Cells were washed by cold saline and put on dry ice. U-¹³C₁₆-palmitate was complexed to fatty-acid-free BSA (A8806, Sigma) in 6:1 molar ratio in PBS by rolling overnight at room temperature. When cells were at 80% confluence, the medium was replaced by DMEM (A1443001, Gibco) with 10 mM glucose, 2 mM glutamine, 10% dialyzed FBS and 0.1 mM [U-¹³C₁₆] palmitate. Cells were cultured in the labeled medium for 48 h which led to approximately 40% labeling of the total cellular palmitate pool.

For polar metabolites, 1 ml 80% methanol/water (HPLC grade) with norvaline as internal standard was added to cells. The plate was transferred to a −80 ˚C freezer and left for 15 min to further inactivate enzymes. Cells were then harvested by a silicone scraper and the whole-cell extract was transferred to a tube and centrifuged at 17000 x g for 10 min at 4 ˚C. The supernatant was transferred into tubes and dried by speed vacuum. Fifty microlitres of MOX (10 mg/ml in pyridine, 226904 Sigma) was added and the mixture was incubated at 42 ˚C for 1 h. After the samples were cooled down, 100 μl TBDMS (394882, Sigma) was added, and samples were incubated at 70 ˚C for 1 h. Then samples were transferred to GC vials and analyzed by Agilent 7820 A chromatography and Agilent 5977B mass spectrometer⁵⁷.

For nonpolar metabolites, 500 μl methanol, 500 μl chloroform, 200 μl H₂O and 10 μl 10 mM d31-C16:0 as internal standard was added to cells. The samples were vortexed and centrifuged at 14,000 xg for 5 min. The chloroform phase was dried and then derivatized to form fatty acid methyl esters (FAMEs) via addition of 500 μl 2% H₂SO₄ in methanol at 50 °C for 2 h followed by adding 100 μl saline and 500 μl hexane. After extraction, the hexane layer was dried and dissolved in 100 μl hexane. These samples were analyzed using a select FAME column.

5×10⁴ cells were seeded into seahorse XF24 microplates (100850, Agilent) and cultured for 6 h. Before incubation, the culture medium was removed and cells were washed with substrate limited medium (DMEM supplemented with 0.5 mM Glucose, 1 mM GlutaMAX (35050061, Thermo fisher scientific), 0.5 mM Carnitine (C0283, Sigma), and 1% FBS). Then, 300 μl substrate limited medium was added to each well and cultured overnight. On the next day, the substrate limited medium was replaced by 375 µl FAO buffer (111 mM NaCl, 4.7 mM KCl, 1.25 mM CaCl₂, 2 mM MgSO₄, and 1.2 mM NaH₂PO₄) supplemented with 2.5 mM glucose, 0.5 mM carnitine, and 5 mM HEPES. After 30 min, 37.5 µl etomoxir (40 μM final) (236020, Sigma) was added to the FAO buffer. 15 min after that, 87.5 µl palmitate: BSA or BSA control was added into the buffer, and XF cell mito stress assay (103015-100, Agilent) was initiated. 56 µl oligomycin (30 µM), 62 µl FCCP (20 µM), and 69 µl Rotnone/Antimycin A (20 µM) were added to the cartridge wells. OCR level was determined using Seahorse bioscience XF24 extracellular flux analyzer (Agilent) and each cycle of measurement involved mixing (3 min), waiting (2 min), and measuring (3 min) cycles. The data were analyzed in Seahorse Wave Desktop Software 2.4.

Two oligos were designed to knockout AMPKα1 and α2 using the software at http://crispor.tefor.net/crispor.py↗. F-5'-CACCGGAAGCAGAAGCACGACGGGC-3'; R-5'-AAACGCCCGTCGTGCTTCTGCTTC. LentiCRISPRv2 encoding Cas9 (52961, Addgene) was digested with BsmBI (R0580, Biolabs) and purified by DNA gel. The two oligos were ligated to the digested plasmid. SAPS3 KO and WT MEFs were infected with lentivirus containing lentiCRISPRv2-AMPKα1/α2-specific oligos and then selected by puromycin.

Total RNA was extracted using Trizol reagent. Reverse transcription reaction was performed using qScript cDNA Synthesis Kit (95047-100, Quanta Biosciences). qRT-PCR was performed in a CFX Connect Real-Time PCR Detection System (Bio-Rad) by using a reaction mixture with SYBR Green PCR Master Mix (95072, Quanta Biosciences). The PCR cycle parameters: 95 ˚C for 3 min; 40 cycles with denaturation at 95 ˚C for 10 sec, annealing at 55 ˚C for 30 sec. All the PCR amplification was performed in triplicate and repeated in three independent experiments. The relative quantities of genes were normalized to mouse 18 S RNA. The primers were as follows: ppp6r1 F: 5'- TGATCGCTTCCATCAGCTCC −3'; ppp6r1 R: 5'- GACCACGTGTAACCTCGTGT −3'; ppp6r2 F: 5'- ATCCATCCCCACCAGGATGA −3'; ppp6r2 R: 5'- CAGTCCTGCGACTCCAATGT −3'; ppp6r3 F: 5'- CTCCACAACCCAGGCAAGAT −3'; ppp6r3 R: 5'- TGCCGATCTTCTTCTTGCGA −3'; ampk α1 F: 5'- CGCAGACTCAGTTCCTGGAG −3'; ampk α1 R: 5'- CTTCACTTTGCCGAAGGTGC −3'; ampk α2 F: 5'- TCCTGAACACCTCAGCGTTC −3'; ampk α2 R: 5'- CTTCCGGTCAAAGAGCCAGT −3'; acc2 F: 5'- GAACCGGCTTCCTGGTTGTA −3'; acc2 R: 5'- TCCTCCCCTATGCCGAAAGA −3'; fas F: 5'- CAAGTGTCCACCAACAAGCG −3'; fas R: 5'- GGAGCGCAGGATAGACTCAC −3'.

All experiments were repeated independently at least three times with similar results. Results are shown as means; error bars represent standard deviation (s.d.). The unpaired two-tailed Student's t-test or one-way ANOVA was used to determine the statistical significance of differences between means (*p < 0.05, **p < 0.01, ***p < 0.001) and were calculated in Microsoft Excel or GraphPad Prism (v 9) software unless indicated separately.

Further information on research design is available in the linked to this article. Nature Portfolio Reporting Summary

Supplementary Information Description of Additional Supplementary Files Supplementary Data 1 Reporting Summary