Anti-apoptotic MCL-1 promotes long-chain fatty acid oxidation through interaction with ACSL1

Abstract

MCL-1 is essential for promoting the survival of many normal cell lineages and confers survival and chemoresistance in cancer. Beyond apoptosis regulation, MCL-1 has been linked to modulating mitochondrial metabolism, but the mechanism(s) by which it does so are unclear. Here, we show in tissues and cells that MCL-1 supports essential steps in long-chain (but not short-chain) fatty acid β-oxidation (FAO) through its binding to specific long-chain acyl-coenzyme A (CoA) synthetases of the ACSL family. ACSL1 binds to the BH3-binding hydrophobic groove of MCL-1 through a non-conventional BH3-domain. Perturbation of this interaction, via genetic loss of Mcl1, mutagenesis, or use of selective BH3-mimetic MCL-1 inhibitors, represses long-chain FAO in cells and in mouse livers and hearts. Our findings reveal how anti-apoptotic MCL-1 facilitates mitochondrial metabolism and indicate that disruption of this function may be associated with unanticipated cardiac toxicities of MCL-1 inhibitors in clinical trials.

Keywords: MCL-1; acyl-coenzyme A synthetase; apoptosis; fatty acid; metabolism; mitochondria; β-oxidation.

Figure 1.. Liver-specific Mcl1 deletion leads to accumulation of lipid droplets in the liver.

(A) Wild-type (Mcl1^wt) and Mcl1^F/F mice were injected with AAV-LP1-Cre to express Cre-recombinase in the liver. (B) Immunoblot analysis of indicated tissues from mice blotted for MCL-1. Each lane represents an individual animal. (C) Two weeks post injection, mice were subjected to 48-hour fasting (Fasted) or maintained on normal chow (Fed) and liver histology was observed. Black size bars indicate 1 cm and white bars indicate 25 mm. (D) Top Gene Ontology Biological Process pathways from gene set enrichment analysis between the Mcl1^L−/− and wild-type 48-hour fasted liver samples. (E) Transmission electron micrographs of control (Mcl1^wt) or deleted (Mcl1^L−/−) livers after 48-hour fast. Nucleus (N) and examples of lipid droplets (LD) are indicated. Size bars indicate 2 μm. (F) Blood triglycerides from Mcl1^wt (WT) or Mcl1^L−/− mice from fed and fasting conditions. Each bar represents mean measurements from 12 to 15 mice and error bars represent the standard error of mean (SEM). (G) Blood glucose levels from Mcl1^wt (WT) or Mcl1^L−/− mice from fed and fasting conditions. Each bar represents mean plus or minus SEM from 12-15 animals. Significance as determined by one-way ANOVA is indicated (**p<0.01). (H) Percent body weight of inguinal white adipose tissue (iWAT) obtained from indicated genotypes. Each bar represents mean from at least five animals. Significance by one-way ANOVA is indicated (**p<0.01).

Figure 2.. Loss of MCL-1 expression impairs long chain fatty acid oxidation.

Wild-type (WT) or Mcl1^F/F mice (Mcl1^L−/−) were intravenously injected with AAV-LP1-Cre to express Cre-recombinase in the liver. Two weeks post injection, mice were subjected to 48-hour fasting (Fasted) or maintained on normal chow (Fed). Average metabolite measurements from eight individual WT or Mcl1^L−/− liver samples per group from fed and fasting conditions for (A) β-hydroxybutyrate (βHB); (B) triacylglycerides (TAGs); and (C) free fatty acids (FFAs) evaluated by one-way ANOVA with α=0.05 (*p<0.05 and **p<0.01). Error bars represent the SEM. Lipid panel metabolomics on WT and Mcl1^L−/− fasted livers was performed for (D) FFAs of indicated carbon chain lengths and saturations, (E) TAGs of indicated carbon chain lengths and saturations, and (F) cholesterol esters (CEs) of indicated carbon chain lengths and saturations. Significance by two-way ANOVA with α=0.05 is indicated (*p<0.05). Scintillation counts per minute (CPM) per mg of WT or Mcl1^L−/− liver extracts for ¹⁴C-labeled fatty acid oxidation (FAO) for (G) ¹⁴C-palmitate and (H) ¹⁴C-hexanoate. Etomoxir (ETO) was added to repress long chain FAO. Data are average CPM/mg from at least three independent assays performed in triplicate and error bars represent the SEM. Significance by two-way ANOVA is indicated (**p<0.01, ****p<0.0001).

Figure 3.. ACSL family members interact with MCL-1 through a conserved BH3-like motif.

(A) Schematic of long chain FA import into the mitochondria. ACSL family members activate FA by ligation of CoA, which is exchanged for carnitine by CPT1 to allow shuttling into the matrix of the mitochondria. After import, CPT2 exchanges carnitine for CoA to drive mitochondrial β-oxidation. (B) Proteomic assessment of anti-FLAG or anti-GFP immunoprecipitation from Mcl1^L−/− livers reconstituted with FLAG-Mcl1 or GFP. Spectral counts of precipitations normalized to anti-GFP are presented. Statistical evaluation by t-test and p values are indicated. (C) Mcl1^F/F Rosa-ERCreT2 murine embryonic fibroblasts (MEFs) reconstituted with untagged Mcl1 or SPOT-Mcl1 were lysed and immunoprecipitated with an anti-SPOT resin. (D) Wild-type (WT) MEFs were immunoprecipitated with anti-GFP (control) or anti-MCL-1 resins. (E) WT mouse liver lysates were immunoprecipitated with anti-GFP (control) or anti-MCL-1 resins. (F) Human 293T cell lysates were immunoprecipitated with anti-GFP (control) or anti-MCL-1 resins. (G) Human 293T cells were transiently transfected with or without a FLAG-tagged human ACSL1 construct and subjected to immunoprecipitation with an anti-FLAG resin. (H) WT mouse liver lysates were immunoprecipitated with isotype matched normal rabbit IgG (control) or anti-ACSLl resins. For each of the immunoprecipitations (C–H), inputs (In) and eluates (E) were resolved, and blots probed for specified endogenous proteins. For each case, results are representative of three immunoprecipitations. (I) Mcl1^F/F Rosa-ERCreT2 murine embryonic fibroblasts (MEFs) reconstituted with untagged Mcl1, SPOT-Mcl1^wt, or SPOT-Mcl1^GRRL were lysed and immunoprecipitated with an anti-SPOT resin. Input (In) and eluates (E) were resolved, and blots probed for ACSL1, MCL-1, or BIM. Results are representative of at least three immunoprecipitations. (J) Alignment of amino acids from BH3-domains of BCL-2 family members (upper), non-BCL-2 family members (middle), and ACSL family members (bottom). Conserved residues highlighted by color: gray indicates hydrophobic residues, yellow indicates invariable Asp/Lys, and cyan indicates residues with smaller side chains. (K) WT mouse liver lysates were immunoprecipitated with anti-GFP (control) or anti-MCL-1 resins. In and E were resolved, and blots probed for endogenous ACSL family members, MCL-1, or BIM. Results are representative of three immunoprecipitations.

Figure 4.. Interaction of MCL-1 and ACSL1 dictates ACSL enzymatic activity in fatty acid oxidation (FAO).

(A) ACSL activity assessed from untreated (UT) Mcl1^wt (WT) or Mcl1^L−/− liver extracts. Triacsin C (TriC), an ACSL inhibitor, was added (10 μM) as control. Bars represent mean from four individual livers per genotype performed in triplicate plus or minus SEM. Significance by two-way ANOVA with α=0.05 is indicated (***p<0.001, ****p<0.0001). (B) Scintillation counts per minute (CPM) per mg of WT or Mcl1^L−/− liver extracts for complete ¹⁴C-labeled FAO of ¹⁴C-palmitate (¹⁴C-P. acid) or pre-activated ¹⁴C-palmitoylcarnitine (¹⁴C-P. carnitine). Data are average CPM/mg from four individual livers per genotype performed in triplicate plus or minus SEM. Significance by two-way ANOVA with α=0.05 is indicated (**p<0.01) or not significant (ns). (C) Seahorse assay of Mcl1^WT or Mcl1^−/− MEFs measuring oxygen consumption rates (OCR) in the presence or absence of etomoxir (ETO). Data are presented as ETO-OCR subtracted from total OCR at indicated respiration states: spare respiratory capacity (maximal OCR minus basal OCR) and maximal respiration (FCCP uncoupled OCR). Data are average OCR from three individual experiments plus or minus SEM. Significance by unpaired t-test (****p<0.0001). Assessment of complete ¹⁴CO₂ oxidation of (D) ¹⁴C-palmitate or (E) ¹⁴C-hexanoate in Mcl1^F/F Rosa-ERCreT2 MEFs reconstituted with empty vector (Empty), SPOT-Mcl1 (WT), or SPOT-Mcl1^GRRL (GRRL). Data represent average CPM from four independent assays plus or minus SEM. ETO was added at to repress long chain FAO. Significance by two-way ANOVA with α=0.05 is indicated (*p<0.05) or not significant (ns). (F) ACSL activity assessed from Mcl1^F/F Rosa-ERCreT2 MEFs reconstituted with empty vector (Empty), SPOT-Mcl1 (WT), or SPOT-Mcl1^GRRL (GRRL). TriC was added as control. Bars represent mean of four individual experiments plus or minus SEM. Significance by one-way ANOVA with α=0.05 is indicated (****p<0.0001). (G) Fluorescence microscopy of parental Mcl1^F/F Rosa-ERCreT2 MEFs or deleted cells reconstituted with empty vector, SPOT-Mcl1, or SPOT-Mcl1^GRRL that were pulsed with oleate (HBSS 0 hours), and then chased with minimal media (HBSS) for indicated time points. Lipid droplets were stained using BODIPY (green) and mitochondria were labeled with MitoTracker (red). Scale bars, 25 μm. (H) Quantification of oleate pulse-chase presented in G. Significance by one-way ANOVA with α=0.05 is indicated (****p<0.0001).

Figure 5.. ACSL family members interact with MCL-1 through its canonical BH3-binding groove.

(A) Ribbon diagram for the overlay of the structure of ACSL1 BH3–MCL-1 complex (green-grey) predicted by AlphaFold-Multimer and that of BIM BH3–MCL-1 complex (teal-grey) determined by X-ray crystallography (PDB ID 2PQK). The helix register is shifted by 2.5 Å. Peptide sequences used for ACSL1 BH3 and BIM BH3 are indicated. (B) 2D nuclear magnetic resonance spectral overlays for Apo-MCL-1 (orange) and with ACSL1 BH3-25mer (green). (C) Chemical shift perturbation (CSP) analysis for binding of 50 μM ¹⁵N-MCL-1 to 875 μM ACSL1 BH3 peptide plotted as a bar graph. The red dashed line represents average CSP plus standard deviation. (D) Mapping of CSPs represented as putty thickness (green) onto the structure of MCL-1 from the AlphaFold structure of ACSL1 BH3–MCL-1 complex (BH3 peptide is excluded for clarity). (E) 293T cells transfected with empty vector (EV) or epitope-tagged NOXA (HA-NOXA) were immunoprecipitated with anti-GFP (control) or anti-MCL-1 resins. For each immunoprecipitation, inputs (In) and eluates (E) were resolved, and blots probed for specified endogenous proteins or HA to detect HA-NOXA. Results are representative of three immunoprecipitations. (F) Scintillation counts per minute (CPM) per mg of MEFs transduced with empty vector (EV) or HA-NOXA, for ¹⁴C-palmitate. Data are average CPM/mg from three independent assays plus or minus SEM. Significance by unpaired t-test is indicated (***p<0.001).

Figure 6.. MCL-1 inhibitors disrupt ACSL1 binding and repress fatty acid oxidation (FAO).

293T cell lysates cultured with or without 500 nM (A) AZD5991, (B) S64315, (C) navitoclax (Navi), or (D) venetoclax (Ven) were immunoprecipitated with anti-GFP (control) or anti-MCL-1 resins. Input (In) and eluates (E) were resolved, and blots probed for endogenous ACSL1, MCL-1, or BIM. Results are representative of three immunoprecipitations. (E) Mcl1^F/F Rosa-ERCreT2 murine embryonic fibroblasts (MEFs) reconstituted with human MCL1, deleted, and treated with indicated BH3-mimetics at indicated doses (10-fold dilutions) for 16 hours and incubated with ¹⁴C-palmitate. Presented are mean scintillation counts per minute (CPM) for ¹⁴C-labeled FAO from three independent assays plus or minus SEM. Significance by two-way ANOVA with α=0.05 is indicated (*p<0.05). (F) Microscopy of Mcl1^F/F Rosa-ERCreT2 MEFs reconstituted with human MCL1 that were deleted and pulsed with oleate (HBSS 0 hour) and indicated BH3-mimetics (500 nM), and then chased with minimal media (HBSS) for indicated times. Lipid droplets were stained using BODIPY (green) and mitochondria were labeled with MitoTracker (red). Scale bars, 25 μm. (G) Quantification of oleate pulse-chase presented in F. Significance by one-way ANOVA with α=0.05 is indicated (****p<0.0001).

Figure 7.. Disruption of the interaction of ACSL1 with MCL-1 in cardiomyocytes represses fatty acid oxidation (FAO).

(A) Immunoblot analysis of heart tissues from wild-type (WT) orMc/7^H” “ mice blotted for MCL-1, ACSL family members, and β-tubulin (loading control). ACSL4 was not detected. Each lane represents an individual animal. (B) WT mouse heart lysates were immunoprecipitated with anti-GFP (control) or anti-MCL-1 resins. Input (In) and eluates (E) were resolved, and blots probed for endogenous ACSL1, MCL-1, or BIM. Results are representative of three immunoprecipitations. Mean scintillation counts per minute (CPM) per mg of WT or Mc/7^H” “ heart extracts for Re-labeled FAO for (C) ¹⁴C-palmitate and (D) ¹⁴C-hexanoate. Data are average CPM/mg from three independent assays plus or minus SEM. Significance by unpaired /-test is indicated (*p<0.05). (E) Total mean ATP content of WT or Mcl1^H−− hearts. Six hearts per genotype were measured; significance by unpaired /-test is indicated (*p<0.05). Scintillation CPM/mg of (Bax and Bak deleted) DKO or (Mcl1, Bax, and Bak deleted) TKO heart mitochondria for ¹⁴C-labeled FAO. Counts for complete oxidation of (F) ¹⁴C-palmitate and (G) ¹⁴C-hexanoate. Etomoxir (ETO) was added to repress long chain FAO. Data are average CPM/mg from three independent heart isolates plus or minus SEM. Significance by two-way ANOVA is indicated (*p<0.05, **p<0.01). (H) Heart lysates were incubated with or without 1 μM AZD5991 and immunoprecipitated with anti-GFP (control) or anti-MCL-1 resins. In and E were resolved, and blots probed for endogenous ACSL1, MCL-1, or BIM. Results are representative of three immunoprecipitations. Mice were injected with vehicle or S63845 for 5 days after which isolated heart mitochondria were incubated with (I) ¹⁴C-palmitate or (J) ¹⁴C-hexanoate. Average CPM/mg for ¹⁴C-labeled FAO from assays performed from seven mice per treatment group plus or minus SEM. Significance by unpaired t-test (*p<0.01).