A drug-like molecule engages nuclear hormone receptor DAF-12/FXR to regulate mitophagy and extend lifespan

Abstract

Autophagy-lysosomal function is crucial for maintaining healthy lifespan and preventing age-related diseases. The transcription factor TFEB plays a key role in regulating this pathway. Decreased TFEB expression is associated with various age-related disorders, making it a promising therapeutic target. In this study, we screened a natural product library and discovered mitophagy-inducing coumarin (MIC), a benzocoumarin compound that enhances TFEB expression and lysosomal function. MIC robustly increases the lifespan of Caenorhabditis elegans in an HLH-30/TFEB-dependent and mitophagy-dependent manner involving DCT-1/BNIP3 while also preventing mitochondrial dysfunction in mammalian cells. Mechanistically, MIC acts by inhibiting ligand-induced activation of the nuclear hormone receptor DAF-12/FXR, which, in turn, induces mitophagy and extends lifespan. In conclusion, our study uncovers MIC as a promising drug-like molecule that enhances mitochondrial function and extends lifespan by targeting DAF-12/FXR. Furthermore, we discovered DAF-12/FXR as a previously unknown upstream regulator of HLH-30/TFEB and mitophagy.

Extended Data Fig. 1 |. MIC enhances TFEB/HLH-30 expression and lysosomal activity.

a, Day-2 HLH-30::GFP (MAH235) images 24 h post-treatment (20 °C) with DMSO or MIC. On 5 min starvation, animals show 100% nuclear localization (arrowheads). Assay was independently repeated three times (n ≥ 30 animals). Scale: 10 μm. b, Punctae count (muscles) of day-2 lgg-1p::gfp::lgg-1 (DA2123) post 24 h DMSO/MIC treatment (20 °C). Image highlighting LGG-1 punctae (arrow). Assay was independently repeated three times (n = 48 and 32 animals). Graph show mean ± s.d. (p-values via two-tailed unpaired t-test). Scale: 10 μm. c, Punctae count (pharynx) of day-2 lgg-1p::gfp::lgg-1 (DA2123) post 24 h DMSO/MIC treatment (20 °C). Image highlighting LGG-1 punctae (arrow). Assay was independently repeated three times (n = 28 and 32 animals). Graph show mean ± s.d. ns ≥ 0.05 (p-values via two-tailed unpaired t-test). Scale: 10 μm. d, Punctae count (neurons) in day-2 rgef-1p::gfp::lgg-1 (MAH242) post 24 h DMSO/MIC treatment (20 °C). Image highlighting LGG-1 punctae (arrow). Assay was independently repeated three times (n = 30 and 31 animals). Graph show mean ± s.d. (p-values via two-tailed unpaired t-test). Scale: 10 μm. e, Lysosomal area changes in day-2 WT(N2) 24 h post-treatment (20 °C) via lysotracker red. Assay was independently repeated three times with ≥15 animals/repeat (n = 311 and 571 lysosomes). The box plot shows the 25th–75th percentiles, median line, and min–max whiskers. p = 0.0004, (p-values via two-tailed unpaired t-test). f, Day-2 plmp1::lmp-1::GFP (RT258) images 24 h post-treatment (20 °C). Assay was independently repeated three times (n = 19 and 22 animals). The box plot shows the 25th–75th percentiles, median line, and min–max whiskers. p < 0.0001, (p-values via two-tailed unpaired t-test). Scale: 10 μm. g, Day-2 unc-54p: SEP::mCherry::laat-1 (arrow) images 24 h post-treatment (20 °C). Graph show mean ± s.d. Assay was independently repeated three times (n = 9 and 7 animals). p = 0.0066, (p-values via two-tailed unpaired t-test). Scale: 10 μm.

Extended Data Fig. 2 |. MIC modulates activity of a nuclear hormone receptor DAF-12/FXR.

a, mRNA of lysosomal and mitophagy genes, day-1 WT(N2) and daf-12(rh61rh411) (20 °C). Assay was independently repeated four times. Graph show mean ± s.d. (p-values via two-tailed unpaired t-test). b, WT(N2) and daf-12(rh61rh411), day-1 (20 °C); Left: lysotracker intensity, (n = 311 and n = 325 lysosomes), p < 0.0001. Right: lysosomal area, (n = 311 and n = 321 lysosomes), p = 0.0013. Assay was independently repeated three times with ≥15 animals/repeat. (P-value via two-tailed unpaired t-test). Scale: 10 μm. c, WT(N2) and daf-12(rh61rh411) 24 h post-DMSO/MIC treatment, day-2 (20 °C); lysotracker intensity, Left: (n = 272 and n = 617 lysosomes), p < 0.0001; Right: (n = 325 and n = 477 lysosomes), p < 0.0001. Assay was independently repeated three times with ≥15 animals/repeat. (p-value via two-tailed unpaired t-test). d. mRNA of lysosomal genes, day-2 WT(N2) and hlh-30(tm1978) (20 °C). Assay was independently repeated five times. Graph show mean ± s.e.m. (p-values via two-tailed unpaired t-test). Right: lysotracker intensity, hlh-30(tm1978) 24 h post-DMSO/MIC treatment, day-2 (20 °C), (n = 447 and n = 156 lysosomes), p < 0.0001. Assay was independently repeated three times with ≥15 animals/repeat. (P-value via two-tailed unpaired t-test). The box plots (b, c and d) represent the 25th–75th percentiles, the line depicts the median and the whiskers show the min–max values. e, DAF-12 activity post-dafachronic acid (ligand) doses quantified by one-hybrid reporter assay. Graph show mean ± s.d. Assay was independently repeated four times, except for ligand 50 nM that is, two, Ligand200 (p = 0.0045) and Ligand500 (p = 0.0003), (P-value via one-way ANOVA with Tukey’s multiple comparisons test). f, Schematic depicts decadiene A (DFA) action. DFA induces HLH-30 expression. HLH-30::GFP (MAH235), day-2 animals post-DFA/DMSO treatment (20 °C) probed with GFP antibody, Actin (ACT) as loading control. Assay was independently repeated two times. g, mRNA of lysosomal genes, day-2 WT(N2) and daf-12(rh61rh411) (20 °C). Assay was independently repeated two times. (p-values via two-tailed unpaired t-test).

Extended Data Fig. 3 |. MIC-induced mitophagy enhances mitochondrial health.

a. Image depicts co-localization (arrowheads) in day-2 reporter animals treated with DMSO, MIC, or FCCP at 20 °C, expressing mCherry::LGG-1 and mitochondrial GFP. FCCP is used as positive control. Right: Spearman’s co-localization coefficient. Assay was independently repeated four times. Graph show mean ± s.d. MIC (p = 0.0464) and FCCP (p = 0.0003), (p-value via two-tailed unpaired t-test). Scale: 10 μm. b. mRNA of mitophagy genes in day-2 WT(N2) and daf-12(rh61rh411) mutants 24 h post DMSO or MIC (20 °C). Assay was independently repeated four times. Graph show mean ± s.d. daf-12 vs WT, for dct-1 (p = 0.0059) and drp-1, (p = 0.0476), WT (DMSO vs MIC), for dct-1 (p = 0.0071) and for drp-1 (p = 0.0326), (p-value by unpaired t-test). c. mRNA of mitophagy genes in day-2 WT(N2) and hlh-30(tm1978) mutants 24 h post DMSO or MIC (20 °C). hlh-30 vs WT, for dct-1 (p = 0.0099), pink-1 (p = 0.0277). WT (DMSO vs MIC), for dct-1 (p = 0.0004) and hlh-30 (DMSO vs MIC) for pink-1 (p = 0.0163). Assay was independently repeated three times. Graph show mean ± s.d., (p-value by unpaired t-test). d. Quantification of mitophagy in muscle-specific mito-rosella reporter upon RNAi treatment, performed at day-2, 24 h post DMSO/MIC (20 °C). Assay was independently repeated three times (n = 45, 44, 34 and 26 animals). Graph show mean ± s.d. control RNAi, MIC vs DMSO (p = 0.0020), (p-value via one-way ANOVA with Tukey’s multiple comparison test). e. Survival of DMSO/MIC treated day-5 pdr-1(gk448) mutants 5 h (20 °C) post-rotenone. Assay was independently repeated three times. Graph show mean ± s.d., p = 0.0226 (p-value via two-tailed unpaired t-test). f. Oxygen consumption rate, basal (B) and max (M) of day-2 WT(N2) and hlh-30(tm1978) mutants 24 h post DMSO/MIC (20 °C). Assay was independently repeated four times. Graph show mean ± s.e.m. (B vs M) for; WT-DMSO (p = 0.0169), WT-MIC (p < 0.0001), hlh-30-DMSO (p = 0.0269), hlh-30-MIC (p = 0.0005); (max-WT-DMSO vs max-WT-MIC, p < 0.0001), (max-WT-MIC vs max-hlh-30-MIC, p < 0.0001), (p-value via two-way ANOVA with Sidak’s multiple comparison test).

Extended Data Fig. 4 |. Lifespan extending effect of MIC.

a, Survival curve of WT(N2) animals, treatment (median days); N2-DMSO (22) and N2-MIC (29), ****p < 0.0001 (p-value via Mantel-Cox long-rank test). MIC exposure from days 1–12 (15 °C) of adulthood. Detailed statistical information for all lifespans is presented in Supplementary Table 1. b, Survival curve of WT(N2) animals, treatment (median days); N2-DMSO (17) and N2-MIC (17), MIC exposure (15 °C) during development only (egg to young adult). c, Left: Survival curve of WT(N2) and hlh-30(tm1978) animals on dafadine A (DFA) post-development, N2-DMSO (17) and N2-DFA (19); p = 0.0021, hlh-30-DMSO (15) and hlh-30-DFA (12); p < 0.0001, (p-value via Mantel-Cox long-rank test). Assay was independently repeated three times. Right: Survival of day-2 WT(N2) and hlh-30(tm1978), 3 h post-FCCP (20 °C). Assay was independently repeated two times, ≥30 animals/repeat. WT-DMSO vs WT-DFA 20 μM; (p = 0.0113), hlh-30-DMSO vs WT-DFA 40 μM; (p = 0.0116), (p-values via two-way ANOVA with Sidak’s multiple comparisons test). d, Survival curve of pdr-1(gk448) animals, pdr-1-DMSO (15) and pdr-1-MIC (17), (p = 0.0009) (p-value via Mantel-Cox long-rank test). Assay was independently repeated two times. MIC exposure from days 1–7 (20 °C) of adulthood. e, Median lifespan shown for conditions/genotypes post treatment. N2 (D-20 °C) indicates MIC treatment only during development. ‘Life-long’ means exposure throughout the life starting from day-of adulthood. Each dot depicts median lifespan from a plate of ~50 animals. Lifespan change is averaged across trials. DMSO vs MIC for N2(15 °C)-life-long; (p = 0.0149), N2(20 °C)-life-long; (p = 0.0400), N2(15 °C)-day1–15; (p < 0.0001), N2(20 °C)-day1–7; (<0.0001) and mev-1(20 °C)-day1–7; (p = 0.0020), (p-value via two-tailed unpaired t-test). Detailed statistical information for all lifespans is presented in Supplementary Table 1.

Extended Data Fig. 5 |. Mitophagy inducer urolithin A modulates DAF-12/FXR activity.

a, Human FXR activity was measured post-treatment with GW4064 and MIC. Assay was independently repeated four times. Graph show mean ± s.d. GW4064 (p < 0.0001), (P-value via one-way ANOVA with Tukey’s multiple comparisons test). b, left: UA and MIC structure is shown with benzocoumarin rings depicted in grey. Right: Human FXR activity was measured post-treatment with GW4064 and (+/− UA). Assay was independently repeated four times. Graph show mean ± s.d. GW4064 vs ctrl (p < 0.0001) and GW4064-UA50 (p = 0.0012). (p-value via two-way ANOVA with Tukey’s multiple comparisons test). c, DAF-12 activity post-dafachronic acid treatment, alone or with UA, was quantified by one-hybrid reporter assay. Graph show mean ± s.d. Assay was independently repeated three times. DMSO + Ligand (p = 0.0005), Ligand+ UA30 (p = 0.0076) Ligand +UA50 (p = 0.0069), (P-value via one-way ANOVA with Tukey’s multiple comparisons test).

Fig. 1 |. MIC enhances TFEB/HLH-30 expression and lysosomal activity.

a, Workflow of the natural product library (NPL) screen to identify TFEB inducers. b, TFEB promoter activity 24 h after MIC or DMSO treatment in N27 cells was tested via luciferase reporter assay. P = 0.0169 (two-tailed unpaired t-test). c, mRNA expression of Tfeb and targets 24 h after MIC (P values determined by two-way ANOVA with Sidak’s multiple comparisons test). Assay (b and c) was independently repeated twice. d, TFEB promoter activity 24 h after MIC tested by luciferase reporter assay. Assay was independently repeated four times. 10 μM (P = 0.1092), 20 μM (P = 0.0012), 30 μM (P < 0.0001) and 50 μM (P < 0.0001) (P values determined by one-way ANOVA with Tukey’s multiple comparisons test). e, Day 2 HLH-30::GFP animals (MAH235) 24 h (20 °C) after treatment. Assay was independently repeated three times (n = 10). The box plot shows the 25th–75th percentiles, median line and minimum–maximum whiskers. P < 0.0001 (P value determined by two-tailed unpaired t-test). Scale, 10 μm. f, HLH-30 protein levels after treatment analyzed in day 2 (20 °C) HLH-30::GFP (MAH235) using antibody against GFP. Tubulin (TUB) was used as loading control. Assay was independently repeated three times. Graph shows mean ± s.d. 24 h (P = 0.0070) (P value determined by one-way ANOVA with Tukey’s multiple comparison test). g, mRNA of hlh-30 and lysosomal genes 24 h (20 °C) after treatment in day 2 WT(N2). Assay was independently repeated twice (P values determined by two-way ANOVA with Sidak’s multiple comparisons test). h, Lysosomal acidity 24 h (20 °C) after treatment in day 2 WT(N2) analyzed using LysoTracker Red staining. Assay was independently repeated three times (n = 22 and 25). The box plot shows the 25th–75th percentiles, median line and minimum–maximum whiskers. P < 0.0001 (P value determined by two-tailed unpaired t-test). Scale, 10 μm. i, Day 5 MAH235 probed with GFP and CPL-1/CTSD antibody, which recognizes its pro and mature (M) form (20 °C). TUB was used as loading control. Assay was independently repeated four times. The box plot shows the 25th–75th percentiles, median line and minimum–maximum whiskers. HLH-30 expression (P = 0.0002), M/pro-CATD (P = 0.0125) (P value determined by two-tailed unpaired t-test). j, Day 5 nuc-1:mCherry probed with GFP antibody, which recognizes mCHERRY (20 °C). Assay was independently repeated six times. The box plot shows the 25th–75th percentiles, median line and minimum–maximum whiskers. P = 0.0027 (P value determined by two-tailed unpaired t-test). NS, not significant.

Fig. 2 |. MIC modulates activity of the NHR DAF-12/FXR.

a, JASPAR CORE identified DAF-12 binding sites (AGT(A/G)CA) on hlh-30 gene’s promoter. Hochbaum et al.ʼs ChIP-seq data verified 12 DAF-12 sites within ~3.6 kb promoter of hlh-30. b, hlh-30 mRNA expression in day 2 WT(N2) and daf-12(rh61rh411) mutants 24 h after DMSO or MIC treatment (20 °C). Graph shows mean ± s.d. Assay was independently repeated three times. WT (P = 0.0191) (P value determined by one-way ANOVA with Tukey’s multiple comparisons test). c, Schematic depicts a subset of DAF-12 ChIP-seq positive genes further quantified by qRT–PCR. Supplementary Table 3 lists other targets and binding sites. d, Lysosomal genes’ mRNA in day 2 WT(N2) and daf-12(rh61rh411) mutants 24 h after DMSO or MIC treatment (20 °C). Assay was independently repeated six times. Graph shows mean ± s.e.m. (P values determined by unpaired two-tailed t-test). e, DAF-12 activity after dafachronic acid treatment, alone or with MIC, was quantified by one-hybrid reporter assay. The graph on the right shows DAF-12 activity. Graph shows mean ± s.d. Assay was independently repeated three times. DMSO-Ligand (P = 0.0001), Ligand-MIC30 (P = 0.0354) and Ligand-MIC50 (P = 0.0030) (P values determined by one-way ANOVA with Tukey’s multiple comparisons test). f. myo-2 promoter activity in day 2 pmyo-2::GFP (OK39) after DMSO, MIC or DFA treatment (20 °C) was analyzed by immunoblot against GFP. Assay was independently repeated four times. Graph shows mean ± s.d. GFP intensity was normalized to tubulin (TUB). MIC (P = 0.0005) and DFA (P = 0.0093) (P values determined by two-tailed unpaired t-test). Ctrl, control; NS, not significant.

Fig. 3 |. MIC-induced mitophagy enhances mitochondrial health.

a, GFP-tagged mitochondria in intestines, day 2 GL382 (20 °C). Assay was independently repeated four times, showing mean ± s.d. P = 0.0095 (P value determined by two-tailed unpaired t-test). Scale, 10 μm. b, drp-1 mRNA, day 2 WT(N2) (20 °C). Assay was independently repeated three times, showing mean ± s.d. P = 0.0008 (P value determined by two-tailed unpaired t-test). c, Survival, day 2 WT(N2) 3 h after FCCP. Assay was independently repeated two times, ≥30 animals per repeat (P values determined by two-way ANOVA with Sidak’s test). d. Survival, day 5 WT(N2), 5 h after rotenone. Assay was independently repeated three times; data were analyzed from n = 8 wells, ≥15 animals per well to account for variability, showing mean ± s.d. (P values determined by two-way ANOVA with Sidak’s test). e. mito-Rosella:mitophagy represented by pH-sensitive GFP to pH-insensitive dsRed ratio. mito-Rosella reporter, day 2, 24 h after treatment (20 °C). f, Neurons (n = 58 and 150 animals), P < 0.0001. g, Muscles (n = 30 and 27 animals), P = 0.0042. Both assays were independently repeated three times, showing mean ± s.d (P value determined by two-tailed unpaired t-test). Scale, 10 μm. h, Mitochondrially targeted dFP::TOMM-7 (DLM14) upon lysosomal entry releases mFP, indicating mitophagy. Protein levels were quantified by GFP antibody, day 2 animals (20 °C). Tubulin (TUB) was used as loading control. Normalized mFP (that is, mFP/dFP::TOMM-7). Assay was independently repeated four times, showing mean ± s.d., P = 0.0131 (P value determinedd by two-tailed unpaired t-test). i, mito-Rosella reporter (muscle), day 2, 24 h after treatment (20 °C). Assay was independently repeated three times (n = 45, 44, 34, 26, 53 and 55 animals), showing mean ± s.d. Control-RNAi, MIC versus DMSO (P = 0.0020), daf-12 versus control-RNAi (DMSO) (P < 0.0001) (P values determined by one-way ANOVA with Tukey’s test). j, Survival, day 5 WT(N2) and daf-12(rh61rh411), 5 h (20 °C) after rotenone. Assay was independently repeated three times; data were analyzed from n = 8, 8, 5 and 5 wells, ≥15 animals per well to account for variability, showing mean ± s.d. DMSO-WT versus MIC-WT (P = 0.0004) and MIC-WT versus MIC-daf-12 (P < 0.0001) (P values determined by one-way ANOVA with Tukey’s test). k, Survival, day 5 WT(N2) and dct-1(tm376), 5 h (20 °C) after rotenone. Assay was independently repeated three times; data were analyzed from n = 12 wells, ≥15 animals per well to account for variability, showing mean ± s.d. DMSO-WT versus MIC-WT (P < 0.0001) and MIC-WT versus MIC-dct-1 (P < 0.0001) (P values determined by one-way ANOVA with Tukey’s test). l, OCR basal (B) and max (M), day 2 WT(N2), daf-12(rh61rh411), dct-1(tm376), after 24 h (20 °C). Assay was independently repeated at least three times (n = 4, 5, 3, 3 and 3 independent repeats), showing mean ± s.e.m. MIC-WT (P < 0.0001) (P value determined by two-way ANOVA with Sidak’s test). NS, not significant.

Fig. 4 |. MIC extends lifespan and reduces neuropathologies in C. elegans.

a–i, Survival curve of a, WT(N2) animals, treatment (median days); N2-DMSO (16) and N2-MIC (21), P < 0.0001 (P value determined by Mantel–Cox log-rank test). MIC exposure from days 1–7 (20 °C) of adulthood for all lifespans unless mentioned otherwise. Detailed statistical information for all lifespans is presented in Supplementary Table 1. b, daf-12(rh61rh411) animals, daf-12-DMSO (10) and daf-12-MIC (10). c, hlh-30(tm1978) animals, hlh-30-DMSO (12) and hlh-30-MIC (12). d, lmp-1(nr2045) animals, lmp-1-DMSO (21) and lmp-1-MIC (21). e, vha-3(ok1501) animals, vha-3-DMSO (16) and vha-3-MIC (16). f, dct-1(tm376) animals, dct-1-DMSO (15) and dct-1-MIC (15). g, drp-1(tm1108) animals, drp-1-DMSO (19) and drp-1-MIC (19). h, gas-1( fc21) animals, gas-1-DMSO (15) and gas-1-MIC (15). i, mev-1(kn1) animals, mev-1-DMSO (12) and mev-1-MIC (10), P < 0.0001. Assay was independently repeated six (a), four (b), three (c–g) and two (h,i) times. j, Motor function in temperature-sensitive myo-3p::Aβ_1–42 (CL4176) assessed for paralysis after MIC treatment (25 °C). Assay was independently repeated two times, ≥90 animals. MIC-20 μM (P = 0.0219) and MIC-50 μM (P = 0.0015) (P values determined by two-tailed unpaired t-test). k, Motor function in day 8 unc-54p::αSynuclein::YFP (NL5901) assessed by thrashing assay after MIC treatment (20 °C). Assay was independently repeated three times (n = 30, 30 and 22 animals). The box plot shows the 25th–75th percentiles, median line and minimum–maximum whiskers. MIC-20 μM (P = 0.0014) and MIC-50 μM (P = 0.0099) (P values determined by two-tailed unpaired t-test). l, Survival curve of snb-1p::hTDP-43 (CL6049) animals (20 °C). Median lifespan (days): CL6049-DMSO (12), CL6049-MIC 20 μM (15), P = 0.0017, and CL6049-MIC 50 μM (17), P < 0.0001. Assay was independently repeated two times (P values determined by Mantel–Cox log-rank test). NS, not significant.

Fig. 5 |. Conserved therapeutic potential of MIC in mammalian cells.

a, Human FXR activity. Graph shows mean ± s.d. (P value determined by two-way ANOVA with Tukey’s multiple comparisons test). b, mRNA of lysosomal and mitophagy genes, after 24 h in C2C12 myoblasts. Graph shows mean ± s.d. (P value determined by two-tailed unpaired t-test). Assays (a and b) were independently repeated four times. c, Lysosomal acidity after 24 h measured by LysoTracker Red. Left, histogram with MFI, ~50,000 cells with bafilomycin A1 (Baf) used as positive control. Right, fold changes. Baf (P < 0.0001) and MIC (P = 0.0006) (P values determined by one-way ANOVA with Dunnett’s multiple comparisons test). d, Lysosomal activity after 24 h, C2C12 cells (images) quantified by DQ-BSA (right). Baf was used as control. The box plot shows the 25th–75th percentiles, median line and minimum–maximum whiskers. n = 64 and 76 cells, P < 0.0001 (P value determined by two-tailed unpaired t-test). Scale, 200 μm. e, Mitochondrial activity after 24 h in C2C12 by MitoTracker Red. Left, histogram with MFI, ~50,000 cells with UA as positive control. Right, fold changes. MIC-20μM (P = 0.0090) and UA (P = 0.0012) (P values deterrmined by one-way ANOVA with Dunnett’s multiple comparison test). f, Mitochondrial ROS after 24 h of drug in C2C12 measured by MitoSOX Red, 4-h rotenone exposure (10 μM). UA was used as positive control. Scatter plot, ~50,000 cells; gating shows percent of cells positive for MitoSOX Red. g, Mitochondrial ROS after 24 h in C2C12. Graph shows mean ± s.d. Ctrl-rotenone (P = 0.0011) and MIC-rotenone (P = 0.0348) (P values determined by one-way ANOVA with Tukey’s multiple comparison test). Assays (c–e and g) were independently repeated two times with triplicates. h, Cell viability after 24 h in C2C12. Assay was independently repeated four times. Graph shows mean ± s.d. Rotenone-MIC0 (P < 0.0001), rotenone-MIC20 (P = 0.0005) and Rotenone-MIC30 (P = 0.0002) (P values determined by one-way ANOVA with Tukey’s multiple comparison test). i, OCR of C2C12, basal (B) and maximum (M) 24 h after DMSO, -MIC or -guggulusterone (zGG), 10 μM and 20 μM at 37 °C. Assay was independently repeated at least four times (DMSO, n = 4; MIC10, n = 4; MIC20, n = 5; zGG10, n = 5; zGG20, n = 5). The box plot shows the 25th–75th percentiles, median line and minimum–maximum whiskers. MIC 10 μM (P = 0.0365) and zGG 20 μM (P = 0.0092) (P values determined by two-way ANOVA with Sidak’s multiple comparison test). Ctrl, control; Mt., mitochondrial; NS, not significant.