SIRT4 increases cell recycling through ULK1 without involving HDAC6 or OPA1

The following reagents were obtained from the indicated companies: CoCl₂ (Roth, Karlsruhe, Germany), Bafilomycin A₁ (BafA₁) and tubacin (Biomol, Hamburg, Germany), CCCP (carbonyl cyanide 3‐chlorophenylhydrazone; Sigma‐Aldrich), oligomycin and rapamycin (Biomol and Bio‐Techne GmbH, Wiesbaden, Germany), and MYLS22 (Biozol, Hamburg, Germany). Primary antibodies used in immunoblot analysis are listed in Table S1. Anti‐mouse (700 nm; LI‐COR IRDye #926‐32 213) or anti‐rabbit (800 nm; LI‐COR IRDye #926‐68 072) were used as secondary antibodies.

Parental and SIRT4 wild‐type/mutant expressing HEK293 cell lines were cultured at 37 °C and 5% CO₂ in DMEM (Dulbecco's modified Eagle medium) containing high glucose (4.5 g·L⁻¹; Thermo Fisher Scientific, Karlsruhe, Germany) with 10% fetal bovine serum (FBS; Thermo Fisher Scientific) and penicillin (100 units·mL⁻¹)/streptomycin (100 μg·mL⁻¹) (Genaxxon Bioscience, Ulm, Germany). Original HEK293 cells were obtained from the German Collection of Microorganisms and Cell Cultures GmbH (DSMZ: ACC 305; Cellosaurus: CVCL_0045). HEK293‐eGFP, HEK293‐SIRT4‐eGFP, and HEK293‐SIRT4(H161Y)‐eGFP cell lines as well as HEK293 cell lines stably expressing Myc‐Flag, SIRT4‐Myc‐Flag, or SIRT4(H161Y)‐Myc‐Flag have been described previously [17, 21] and were grown under selection with 400 μg·mL⁻¹ Geneticin/G418 (Genaxxon Bioscience). HEK293 cell lines were regularly authenticated by morphological growth, flow cytometry, and immunoblot analysis. Experiments were performed with mycoplasma‐free cell lines.

HEK293 cell lines were subjected to mitophagy induction in a chemical hypoxia/pseudohypoxia model [4, 25] using CoCl₂ treatment at a concentration of 400 μm for 24 h or as indicated. Alternatively, mitophagy was triggered by mitochondrial membrane depolarization upon combined treatment with CCCP (5 μm) and oligomycin (5 μm) for 24 h. General autophagy was induced by rapamycin treatment (1.25 μm, 24 h). Inhibitors against HDAC6 (tubacin) and OPA1 (MYLS22) were used for a (co)treatment period of 24 h at concentrations of 2 and 50 μm, respectively. BafA₁, an inhibitor of autophagosome–lysosome fusion, was used for a (co)treatment period of 4 h at a concentration of 0.5 μm.

Total cell lysates were generated using lysis buffer containing 0.3% CHAPS (3‐[(3‐cholamidopropyl) dimethylammonio]‐1‐propanesulfonate), 50 mm Tris/HCl (pH 7.4), 150 mm NaCl, 1 mm Na₃VO₄, 10 mm NaF, 1 mm EDTA, 1 mm EGTA, 2.5 mm Na₄O₇P₂, and 1 μm DTT. The cOmplete™ protease inhibitor cocktail (Sigma‐Aldrich) was used to prevent the degradation of proteins in the lysates. The latter were cleared by centrifugation (11 000 g at 4 °C for 20 min), and the protein concentration of the supernatants (total cell lysates) was determined using the Bradford assay (Roth). Relative quantification of protein levels (as compared to α‐tubulin or β‐actin protein loading controls) was performed by imagej‐based densitometric analysis of specific immunoblot signals.

The generation of HEK293 cells with downregulated SIRT4 expression due to stable expression of a SIRT4‐specific shRNA has been described [25].

mRNAs were isolated using the NucleoSpin(R) RNA Kit from Macherey‐Nagel (Düren, Germany). QPCR‐based quantification of mRNAs (50 ng mRNA per sample) was performed as instructed by the manufacturer using the iTaq™ Universal SYBR Green one‐step kit from Bio‐Rad (Dreieich, Germany). Primers used were as follows: SIRT4 (forward): 5′‐GTGGATGCTTTGCACACCAAGG‐3′, SIRT4 (reverse): 5′‐GGTTCAGGACTTGGAAACGCTC‐3′; GAPDH (forward) 5′‐ACCACCCTGTTGCTGTAGCCA‐3′, GAPDH (reverse) 5′‐GTCTCCTCTGACTTCAACAGC‐3′. Relative gene expression levels were calculated using the 2−ΔΔCt method [30].

Data are presented as mean ± SD. Multiple comparisons were analyzed by one‐way or two‐way analysis of variance (ANOVA) followed by Tukey's post hoc test to identify group differences in variance analysis using the graphpad prism software (GraphPad Software, Boston, MA, USA). Statistical significance was set at the level of P ≤ 0.05 (*P ≤ 0.05, **P ≤ 0.01, and ***P ≤ 0.001).

First of all, we performed a comparative characterization of the role of SIRT4 in the regulation of mitophagy/autophagy caused by different stressors, in particular CoCl₂‐induced pseudohypoxia that activates HIF1α and upregulates LC3B‐II (Fig. S1). We employed HEK293 cells stably expressing eGFP (vector control), wild‐type SIRT4‐eGFP, or its catalytically inactive mutant SIRT4(H161Y)‐eGFP (Fig. S2A,B) [17, 25]. Interestingly, among the cell lines analyzed, HEK293‐SIRT4(HY)‐eGFP cells showed strongly reduced LC3B‐II levels upon CoCl₂ treatment for 24 h (Fig. 1A,B), whereas the levels of LCB‐I were comparable between the genotypes analyzed (Fig. 1C). Similar results were obtained for LC3B‐II when treating these cell lines with CoCl₂ for 36 h (Fig. S3). We extended these findings to flow cytometry‐based measurement of autophagic flux using the GFP‐LC3‐RFP‐LC3ΔG reporter probe [26], for which HEK293 cells stably expressing SIRT4‐myc‐Flag or its catalytically inactive mutant SIRT4(H161Y)‐myc‐Flag were generated (Fig. S2C,D). The results of these experiments were consistent with the LC3B‐II immunoblot analyses, observing the weakest CoCl₂‐triggered reduction in the GFP‐LC3/RFP‐LC3ΔG ratio (i.e., the lowest autophagic flux induction) for SIRT4(H161Y)‐myc‐Flag expressing cells (Fig. 1D). Moreover, flow cytometric analyses using mitochondria‐specific readouts showed that the CoCl₂ triggered increase in mitophagic activity (mt‐mKEIMA probe) was lowest in SIRT4(H161Y) expressing cells, concomitant with the strongest relative increase in mitochondrial content (Mitomark Green I probe) (Fig. 1E,F). Of note, the SIRT4(H161Y)‐mediated inhibition of autophagic flux was also observed in GFP‐LC3/RFP‐LC3ΔG ratio measurements upon CCCP/oligomycin‐triggered inhibition of mitochondrial membrane potential, resulting in mitophagy, as well as upon general autophagy induced by rapamycin‐mediated mTORC1 inhibition (Fig. S4A,B). The latter was confirmed through effective downregulation of pS6, a target of the mTORC1–p70S6‐kinase (p70S6K) axis (Fig. S4C). Taken together, expression of the dominant‐negative SIRT4(H161Y), and this at lower protein levels as compared to ectopic wild‐type SIRT4 (Fig. S2A,C), impairs autophagic flux upon CoCl₂‐induced pseudohypoxia, mitochondrial membrane depolarization, or rapamycin treatment.

The autophagy regulator HDAC6 interacts with SIRT4 [21, 22] and is also found at increased levels upon expression of SIRT4 or SIRT4(H161Y) (Fig. 2A,B). In contrast to eGFP expressing controls, SIRT4 and SIRT4(HY) expressing cells maintained high HDAC6 protein levels and did not downregulate HDAC6 upon CoCl₂ treatment (Fig. 2C). Concomitantly, upregulated HDAC6 caused downregulation of acetylated α‐tubulin (acK40), which is a major HDAC6 target (Fig. 2D,E). Acetylated α‐tubulin (acK40) is required for efficient microtubule dynamics and transport processes in connection with autophagy/mitophagy. Given that pharmacological inhibition of HDAC6 can restore autophagy in various experimental model systems [31, 32, 33, 34], we next addressed the effect of the HDAC6 inhibitor tubacin on autophagic flux in SIRT4(H161Y) expressing cells. As expected, tubacin treatment resulted in clear upregulation of acetylated α‐tubulin (K40) levels in all genotypes analyzed (Fig. S5). However, at the same time, inhibition of HDAC6 failed to rescue CoCl₂‐induced LC3B‐II increase, autophagic flux, and mitophagy in HEK293‐SIRT4(H161Y) cells (Fig. 3). Of note, co‐treatment with tubacin and CCCP/oligomycin (Fig. S6A), but not tubacin and rapamycin (Fig. S7A), revealed a slight and significant rescue in autophagic flux as measured through the flow cytometric GFP‐LC3/RFP‐LC3ΔG ratio method. We conclude that the increased HDAC6 levels are not decisive for the impaired LC3B‐II/autophagic flux in SIRT4/SIRT4(H161Y) expressing HEK293 cells upon induction of mitophagy or autophagy.

SIRT4 interacts with the mitochondrial GTPase OPA1 and stabilizes its inner‐membrane bound long form (OPA1‐L) [17] that promotes mitochondrial fusion and thereby potentially counteracts mitophagy as compared to OPA1‐S [19, 35]. Upon CoCl₂ treatment, HEK293 cells expressing SIRT4(H161Y) maintained the highest OPA1‐L levels and therefore OPA1‐L/OPA1‐S ratios as analyzed over a period of 36 h (Fig. 4A,B). Next, we tested if the increased OPA1‐L may affect LC3B‐II upregulation. Therefore, HEK293 SIRT4(H161Y) cells were subjected to co‐treatment with the OPA1 inhibitor MYLS22 [36]. As indicated in Fig. 4C,D, MYLS22 reduced the OPA1‐L/OPA1‐S ratio in SIRT4(H161Y) expressing cells, although there was no additive/synergistic effect with CoCl₂ co‐treatment. However, although MYLS22 alone did upregulate LC3B‐II levels, co‐treatment using MYLS22 and CoCl₂ failed to significantly rescue LC3B‐II (Fig. 5A,B), autophagic flux, and mitophagy (Fig. 5C,D) in CoCl₂‐treated HEK293‐SIRT4(HY) cells. In addition, co‐treatment using MYLS22 and CCCP/oligomycin (Fig. S6B) or MYLS22 and rapamycin (Fig. S7B) did not significantly increase autophagic flux measured by the GFP‐LC3/RFP‐LC3ΔG ratio‐based method. Thus, our data indicate that altered mitochondrial dynamics is not the reason for the impaired LC3B‐II/autophagic flux in SIRT4(H161Y) expressing HEK293 cells upon induction of mitophagy or autophagy.

Execution of autophagy depends on fusion of autophagosomes with lysosomes that can be inhibited by the V‐ATPase and SERCA inhibitor BafA₁ [37, 38, 39]. BafA₁ treatment is widely used to address the basis for increased LC3B‐II levels and autophagic flux, that is, to differentiate between increased biosynthesis/modification of LC3B‐I to LC3B‐II vs. increased auto‐lysosomal degradation of LC3B‐II [39]. Interestingly, under basal conditions, SIRT4(H161Y) expressing HEK293 cells displayed an approximately three‐fold increase in LC3B‐II upon BafA₁ treatment for a period of 4 h (Fig. 6). However, upon combined CoCl₂ treatment, BafA₁ failed to significantly increase LC3B‐II levels (Fig. 6), indicating that an accelerated degradation of LC3B‐II by the autolysosome does not account for the observed phenotype. These results suggest that the stress‐induced generation of LCB‐II is inhibited upon SIRT4(H161Y) expression at an earlier stage, that is, at the LC3B‐I to LC3BII conversion or even at the initiation of autophagy.

Major upstream regulators of autophagy include mTORC1 that affects the central ULK1‐driven autophagy initiation complex in a negative manner [7, 8, 9, 11, 40, 41]. Given the interaction of SIRT4 with the AMPK/mTORC1 pathway [42, 43, 44], we asked whether the early regulatory phase of autophagy may be affected by ectopic expression of the dominant‐negative mutant SIRT4(H161Y). Therefore, we analyzed the relative levels of ULK1 phosphorylated at serine residues Ser758 and Ser638 that both represent mTORC1 substrates [7, 12, 45], associated with inhibition of ULK1 function and hence decreased initiation of autophagy. As indicated in Fig. 7, combined immunoblotting and densitometric analysis revealed that CoCl₂ treatment resulted in downregulation of ULK1^pS758 and ULK^pS638 levels in HEK293 cells that express eGFP or SIRT4‐eGFP. In contrast, in SIRT4(H161Y)‐GFP expressing HEK293 cells, CoCl₂ failed to reduce the amount of both ULK1^pS758 and ULK^pS638 as compared to the untreated controls. In conclusion, our findings suggest that expression of SIRT4(H161Y) negatively interferes with ULK1 dephosphorylation in response to stressors that induce autophagy.