Dysfunction in the mitochondrial Fe-S assembly machinery leads to formation of the chemoresistant truncated VDAC1 isoform without HIF-1α activation

Abstract

Biogenesis of iron-sulfur clusters (ISC) is essential to almost all forms of life and involves complex protein machineries. This process is initiated within the mitochondrial matrix by the ISC assembly machinery. Cohort and case report studies have linked mutations in ISC assembly machinery to severe mitochondrial diseases. The voltage-dependent anion channel (VDAC) located within the mitochondrial outer membrane regulates both cell metabolism and apoptosis. Recently, the C-terminal truncation of the VDAC1 isoform, termed VDAC1-ΔC, has been observed in chemoresistant late-stage tumor cells grown under hypoxic conditions with activation of the hypoxia-response nuclear factor HIF-1α. These cells harbored atypical enlarged mitochondria. Here, we show for the first time that depletion of several proteins of the mitochondrial ISC machinery in normoxia leads to a similar enlarged mitochondria phenotype associated with accumulation of VDAC1-ΔC. This truncated form of VDAC1 accumulates in the absence of HIF-1α and HIF-2α activations and confers cell resistance to drug-induced apoptosis. Furthermore, we show that when hypoxia and siRNA knock-down of the ISC machinery core components are coupled, the cell phenotype is further accentuated, with greater accumulation of VDAC1-ΔC. Interestingly, we show that hypoxia promotes the downregulation of several proteins (ISCU, NFS1, FXN) involved in the early steps of mitochondrial Fe-S cluster biogenesis. Finally, we have identified the mitochondria-associated membrane (MAM) localized Fe-S protein CISD2 as a link between ISC machinery downregulation and accumulation of anti-apoptotic VDAC1-ΔC. Our results are the first to associate dysfunction in Fe-S cluster biogenesis with cleavage of VDAC1, a form which has previously been shown to promote tumor resistance to chemotherapy, and raise new perspectives for targets in cancer therapy.

Fig 1. Depletion of proteins of the mitochondrial ISC assembly machinery leads to the formation of enlarged mitochondria without HIF-1α activation.

HeLa cells were transfected with either negative control (NC), or iscu-, nfs1-, mfrn2- or hsc20-siRNA for 6 days. (A) Confocal microscopy after CMXRos staining to visualize mitochondria (upper panels). Scale bar: 10 μm. (lower panels) Higher magnification of the part of the upper panel image delineated by a white square. (B) 6 days after transfection, total protein extracts were analyzed by immunoblotting using VDACs poly antibody detecting all isoforms of VDAC, and antibodies against ISCU or HSC20. β-Actin was used as loading control. The mRNA levels of nfs1 and mfrn2 were determined by RT-qPCR 24 h after transfection. Data are normalized to 18S ribosomal rRNA levels and represented as a percentage of NC ± S.D. *, p < 0.001 (n = 3). (C) Cells were transfected with the specified siRNA for 6 days or treated with desferrioxamine (DFO) for 16 h. Immunoblotting was carried out to determine HIF-1α, CAIX, ISCU, NFS1 and HSC20 protein levels in total extracts. β-Actin is shown as loading control.

Fig 2. Iron depletion and nitric oxide stress induce the accumulation of the truncated VDAC1 form.

(A) Total protein extracts of HeLa cells treated for 16 h with FAC or DFO were analyzed by immunoblotting using VDACs poly antibody. Prohibitin was used as loading control. (B) Total protein extracts of HeLa cells treated with DMSO (control), DFO or SIH for the indicated times were analyzed by immunoblotting using VDACs poly antibody. β-Actin antibody was used as loading control. (C) HepG2 cells were treated for 24 h with DMSO (control) or SIH. Immunoblotting was carried out using VDACs poly antibody on total protein extracts (TE), and on mitochondrial (Mito) and cytosolic (Cyto) fractions. mitoNEET and NUBP1 were used as mitochondrial and cytosolic markers, respectively. (D). HeLa cells were treated for 16 h with DMSO (control), SIH or DETA-NO (NO). Total protein extracts were analyzed by immunoblotting using VDACs poly antibody and anti-HIF-1α and -CAIX antibodies. β-Actin was used as loading control. (E) HeLa cells were grown in normoxia (Nx) or 1% O₂ hypoxia (Hx) conditions for 3 days, or grown in normoxia conditions and treated for 24 h with DMSO (Control, Co), SIH or DETA-NO (NO), or grown in normoxia (21% O₂) conditions and transfected with either negative control (NC) or iscu-siRNA for 3 days. Total protein extracts were analyzed by immunoblotting using antibodies against the N-terminus of VDAC1 isoform (VDAC1-Nter), the three VDAC isoforms (VDACs poly), ISCU and CAIX. (F) HeLa cells were grown in hypoxia (Hx, 1% O₂) conditions and transfected with iscu- or NC-siRNA for 3 days, or grown in normoxia (Nx, 21% O₂) and treated or not with DFO for 16 h. Total proteins were analyzed by western blot using VDACs poly antibody and anti-HIF-1α, -CAIX, -ISCU antibodies. β-Actin was used as loading control.

Fig 3. Impact of hypoxia on proteins of the mitochondrial ISC assembly machinery.

(A) Total protein extracts from HeLa cells grown in normoxia (Nx, 21% O₂) or hypoxia (Hx, 1% O₂) conditions for the indicated times were analyzed by immunoblotting using VDACs poly antibody and anti-CAIX, -ISCU, -FXN, -NFS1, -HSC20 antibodies. β-Actin was used as loading control. (right panel) Bar graph represents the amount of the indicated proteins relative to ß-actin level determined by quantification of n = 3 immunoblot analysis using the Odyssey System Imager. Mean and standard deviation of 3 independent experiments are shown (* p < 0.05, n = 3). (B) HeLa cells were either untreated or treated with CoCl₂ for 2 days. Total proteins were analyzed by western blotting using anti-HIF-1α, -ISCU, -FXN antibodies. Vinculin was used as loading control. (C) HeLa cells grown in normoxia (Nx, 21% O₂) or hypoxia (Hx, 1% O₂) conditions for 3 days and mRNA levels of iscu, fxn and vegf A were determined by RT-qPCR and normalized to 18S ribosomal rRNA levels. The bar graph presents the ratio between the relative level of each mRNA under hypoxic and normoxic conditions with a logarithmic scale. Mean and standard deviation of 6 independent experiments are shown (* p < 0.05, n = 6). (D) Total protein extracts from HeLa cells grown under normoxic (Nx, 21% O₂) or hypoxic (Hx, 1% O₂) conditions for the indicated times were analyzed by immunoblotting using anti-CIAPIN1, -NUBP1, -NARFL antibodies. β-Actin was used as loading control.

Fig 4. Protein depletion in the mitochondrial ISC machinery confers cell protection against STS-induced apoptosis.

(A) HeLa cells were transfected with negative control (NC) or iscu-, nfs1-, mfrn2- or hsc20-siRNA for 6 days. Cells treated with staurosporine (STS) for 4 h were used as positive control for apoptosis induction. After staining with Hoechst 33342, cells were analyzed by epifluorescence microscopy. Scale bar: 100 μm. (B) HeLa cells were either transfected with NC or iscu-siRNA for 3 days or treated with STS for 4 h. Total protein extracts were analyzed by immunoblotting using antibodies against cleaved caspase 3 (C3), PARP1 and ISCU. Vinculin was used as loading control. (C) Cells were transfected with negative control (NC) or, nfs1-, mfrn2- or hsc20-siRNA for 6 days or STS-treated for 4 h. Total protein extracts were analyzed by immunoblotting using antibodies against cleaved caspase-3, PARP1, NFS1 and HSC20. Vinculin was used as loading control. (D) Schematic representation of the protocol used. Phase 1—HeLa cells were seeded the day before the transfection with mfrn2-, iscu- or NC siRNA and incubated for 6 days (144 hours) under normoxic conditions (21% O₂) with a transfection every 3 days. Phase 2 –Cells were either untreated or treated with STS for 4 h. (E and F) Total protein extracts were analyzed by immunoblotting using antibodies against cleaved caspase-3 (C3) and ISCU. β-Actin was used as loading control. (Lower panel) The bar graph presents the ratio between the amounts of cleaved C3 over ß-actin determined by quantification of the immunoblot. Mean and standard deviation of n = 3 (iscu-siRNA) and n = 4 (mfrn2-siRNA) independent experiments (* p < 0.05).

Fig 5. Knock-down of CISD2 induces the accumulation of the truncated VDAC1 form.

(A) Total protein extracts of either non-transfected (-) or transfected HeLa cells for 3 days with scramble (NC)-, cisd2-, or nfs1-siRNA were analyzed by immunoblotting using antibodies against VDACs poly, CISD2, and α-tubulin, which was used as loading control. Eight independent experiments were performed and a representative western blot is shown. (B) Total mRNA was extracted from HeLa cells transfected with scramble (NC)-, cisd2-, or nfs1-siRNA. Twenty-four hours after transfection, the mRNA levels of cisd2 and nfs1 were determined by quantitative RT-qPCR. Data were normalized to gapdh mRNA levels. Means ± standard deviation of n = 4 independent experiments are presented (*** p < 0.001).

Fig 6. Downregulation of mitochondrial Fe-S cluster assembly leads to C-terminal VDAC1 truncation and subsequent resistance to pro-apoptotic treatments.

Red and black arrows represent links unveiled in the present study and in previous studies [17,18], respectively. Hypoxia stabilizes and activates HIF-1α (1) and leads to drastic changes in mitochondrial morphology [17] and VDAC1-ΔC accumulation [18](2). In the present study, we showed that hypoxia leads to the downregulation of NFS1, FXN and ISCU, three components of the mitochondrial ISC core machinery (3). Whether this regulation directly involves HIF-1α remains to be explored (4). Subsequently, deficit in Fe-S cluster biogenesis leads to the loss of the mitochondrial network, appearance of enlarged mitochondria and accumulation of VDAC1-ΔC. We identified the MAM-anchored Fe-S protein CISD2 as a key protein for the cellular fate of VDAC1 (5). We still cannot exclude that CISD2 is the only Fe-S protein involved in this process (6). Interestingly, when components of the ISC core machinery were silenced by siRNA in hypoxia, VDAC1-ΔC levels were increased compared to hypoxia alone.