Sirtuin 3 regulates mitochondrial protein acetylation and metabolism in tubular epithelial cells during renal fibrosis

Proximal tubular epithelial cells (TECs) demand high energy and rely on mitochondrial oxidative phosphorylation as the main energy source. However, this is disturbed in renal fibrosis. Acetylation is an important post-translational modification for mitochondrial metabolism. The mitochondrial protein NAD+-dependent deacetylase sirtuin 3 (SIRT3) regulates mitochondrial metabolic function. Therefore, we aimed to identify the changes in the acetylome in tubules from fibrotic kidneys and determine their association with mitochondria. We found that decreased SIRT3 expression was accompanied by increased acetylation in mitochondria that have separated from TECs during the early phase of renal fibrosis. Sirt3 knockout mice were susceptible to hyper-acetylated mitochondrial proteins and to severe renal fibrosis. The activation of SIRT3 by honokiol ameliorated acetylation and prevented renal fibrosis. Analysis of the acetylome in separated tubules using LC–MS/MS showed that most kidney proteins were hyper-acetylated after unilateral ureteral obstruction. The increased acetylated proteins with 26.76% were mitochondrial proteins which were mapped to a broad range of mitochondrial pathways including fatty acid β-oxidation, the tricarboxylic acid cycle (TCA cycle), and oxidative phosphorylation. Pyruvate dehydrogenase E1α (PDHE1α), which is the primary link between glycolysis and the TCA cycle, was hyper-acetylated at lysine 385 in TECs after TGF-β1 stimulation and was regulated by SIRT3. Our findings showed that mitochondrial proteins involved in regulating energy metabolism were acetylated and targeted by SIRT3 in TECs. The deacetylation of PDHE1α by SIRT3 at lysine 385 plays a key role in metabolic reprogramming associated with renal fibrosis.


INTRODUCTION
Chronic kidney disease (CKD) is characterized by progressive kidney dysfunction of at least three months duration, and it affects about 10% of adults worldwide [1,2]. Several pathologies including diabetes nephropathy (DN), chronic glomerulonephritis, hypertension, chronic tubulointerstitial nephritis, and polycystic kidney diseases (PKD) [3] lead to CKD. Regardless of the contributing pathology, irreversible fibrosis is the main feature of kidney failure in CKD [4,5]. Extensive investigation has confirmed that renal interstitial fibrosis is not only a common histomorphological change associated with end-stage renal disease (ESRD) but also a key factor in determining renal progressive failure in CKD. Therefore, the pathogenesis of renal interstitial fibrosis has become a hot topic in the field of renal diseases.
Proximal tubular epithelial cells (TECs) are abundant in mitochondria and depend on fatty acid oxidation (FAO) and mitochondrial oxidative phosphorylation (OXPHOS) as energy sources to maintain the structure and functions of TECs, including excretion, secretion, and reabsorption [6,7]. Mitochondrial homeostasis and energy metabolism in TECs play central roles in the development of renal disease as interstitial fibrosis progresses [7][8][9]. Recent studies have shown that the main causes of interstitial fibrosis are defective FAO and OXPHOS [10,11]. Defective FAO in TECs is often associated with metabolic reprogramming to glycolysis during the progression of acute kidney injury (AKI), PKD, and DN [12][13][14][15][16][17]. However, the mechanism of metabolic reprogramming in TECs during the progression of renal fibrosis remains unclear.
N-lysine acetylation is a post-translational modification of histones that have been widely studied in transcriptional regulation. Proteomic analyses over the past decade have shown that non-histone proteins are frequently acetylated and are a major component of acetylation in mammalian cells. Non-histone acetylation is involved in various physiological processes, such as gene transcription, DNA damage and repair, cell division, signal transduction, protein folding, autophagy, and metabolism. Sirtuins are NAD + -dependent deacetylases involved in metabolism, stress response, and longevity. Among them, SIRT3 is a mitochondrial deacetylase that mediates the activity of many metabolic enzymes involved in the mitochondrial electron transport chain (ETC), TCA cycle, glycolysis, fatty acid metabolism, and ATP synthesis. Roles of SIRT3 have been identified in AKI [18][19][20], DN [14,[21][22][23], aging [21,24,25], and kidney tumors [26]. The role of SIRT3 in acetylation in the kidneys is mainly based on the acetylomes of organs such as the liver, and muscle. A quantitative map of the acetylome is required to understand the diverse mechanisms of acetylation in kidneys during the progress of renal fibrosis.
Here, we describe changes in the acetylome profile of tubules in fibrotic kidneys. Further, we identified altered acetylated proteins in tubules from fibrotic kidneys and determined their association with mitochondria. We hypothesized that SIRT3 is a major regulator of mitochondrial acetylation and metabolic homeostasis. Our findings will further our understating of the role of SIRT3 in metabolic reprogramming associated with renal fibrosis.

Identification of SIRT3 regulated acetylation in human CKD
We analyzed SIRT3 expression in kidney biopsy specimens from 14 patients with CKD and various degrees of renal fibrosis using quantitative real-time polymerase chain reaction (qRT-PCR) (Supplementary Table 1). The SIRT3 mRNA level was decreased in the hyper-fibrosis group with the area of matrix deposition over 20% analysis by Masson staining compared with the hypo-fibrosis group with the area of matrix deposition below 20% (Fig. 1A). We also evaluated SIRT3 expression and acetylation profiles by immunostaining. Figure 1B shows that SIRT3 protein was predominantly localized in renal TECs in normal kidneys, but decreased in fibrotic areas in kidney biopsies from patients with a variety of CKD. Furthermore, the amount of acetylation was low in normal kidneys and increased in samples of fibrotic kidneys (Fig. 1C).
Loss of SIRT3 leads to mitochondrial protein hyperacetylation in tubules in renal fibrosis We created C57BL/6J mice with unilateral ureteral obstruction (UUO) to assess the role of SIRT3 in renal fibrosis. The mRNAs level of sirtuins were all decreased in tubules from mice with UUO on postoperative day (POD) 1 ( Supplementary Fig. 1A). The decrease of Sirt3 mRNA was most significant. Western blot analysis showed that the expression of SIRT3 was decreased in a time-dependent manner along with increased fibronectin (FN) and collagen I expression after UUO ( Supplementary Fig. 1B, Fig. 1D, E), resulting in global elevated protein acetylation (Fig. 1F, G). Because SIRT3 is a mitochondrial deacetylase, we isolated mitochondria from tubules after UUO. Increased acetylation of mitochondrial proteins were detected from tubules from mice with UUO ( Fig. 1H, I, Supplementary Fig. 1C). Immunostaining further confirmed decreased SIRT3 and increased acetylation in mitochondria ( Fig. 1J, K, Supplementary Fig. 1D).
We treated C57BL/6J male mice with honokiol (HKL), an activator of SIRT3, to determine the role of SIRT3 in renal fibrosis [27,28], and assessed changes in acetylation and metabolism in mitochondria. HKL restored SIRT3 expression, and significantly reduced mitochondrial acetylation in tubules ( Fig. 2A-D). We then investigated whether SIRT3 activation correlates with tubular damage and renal fibrosis. HKL preserved the expression of proximal tubular cell marker including Na/K-ATPase, aquaporin 1 (AQP1), and E-cadherin, attenuated tubular atrophy, and reduced the accumulation of extracellular matrix proteins indicated by collagen I, FN expression, induced by UUO (Fig. 2E-J). We further analyzed mitochondrial acetylation profiles in Sirt3 knockout (KO) mice ( Fig. 3A-C). Sirt3 KO did not affect the expression of other SIRTs ( Supplementary Fig. 2A). The amount of acetylation in mitochondria was mildly increased in Sirt3 KO mice (Supplementary Fig. 1B). However, Sirt3 KO mice developed obvious mitochondrial acetylation and severe fibrosis after UUO in contrast to wild-type (WT) control littermates ( Fig. 3D-I).

SIRT3 regulates mitochondrial acetylation in vitro
Transforming growth factor-β (TGF-β) is the master regulator of renal fibrosis [29]. We next explored whether SIRT3 mediates mitochondrial acetylation in proximal TECs in response to TGF-β. We found that incubating primary TECs with TGF-β1 for 1 h resulted in decreased SIRT3 expression and increased mitochondrial acetylation (Fig. 4A-E). Consistent with the result in vivo, SIRT3 activation by HKL also reduced mitochondrial acetylation and blocked phenotypic changes in TECs induced by TGF-β1 ( Fig. 4F-I). We blocked SIRT3 with 3-(1H-1,2,3-triazol-4-yl)pyridine (3-TYP) [30,31] and assessed the effects of disrupting SIRT3regulated mitochondrial acetylation. 3-TYP exerted the same phenotypic transformation as TGF-β1 described above (Fig. 4J-M). We investigated the effects of damping Sirt3 expression with small interfering (si) RNA against Sirt3 on acetylation, to determine the role of SIRT3 in acetylation. The reduced SIRT3 expression in TECs resulted in increased mitochondrial acetylation (Fig. 4N-O) and induced TEC injury (Fig. 4P-Q), which coincided with the effects of 3-TYP. These data in vitro showed that SIRT3 mediates the increased mitochondrial acetylation in TECs associated with profibrotic stimulation.
Pro-fibrotic stimulation substantially alters tubular mitochondrial protein acetylation We next examined acetylome and proteome responses to fibrosis in tubules. A comparison of the tubular acetylome of mice with and without UUO ( Supplementary Fig. 3A) found that 1870 of 7059 acetylation sites in 895 of 1863 proteins changed in UUO mice ( Supplementary Fig. 3B). Most proteins were multi-acetylated with 58.03% having two or more AcK sites ( Supplementary Fig.  3C). These findings considerably exceeded previous estimations of acetylated proteins and sites in the kidneys [32][33][34]. Most of the changed proteins were hyper-acetylated and in fact, acetylation at 1789 sites in 822 proteins increased more than two-fold in mice with UUO (Fig. 5A). We evaluated these proteins according to their GO cellular compartment annotations to determine the subcellular localization of the modified proteins. Most of the altered acetylated proteins resided in the cytoplasm, which contained 35.85% of all changed proteins. Mitochondria accounted for 26.68% of the changed acetylated proteins (Fig. 5B). We analyzed the altered acetylated proteins in mitochondria from fibrotic kidneys. The protein acetylation profiles in mitochondria were similar to the global changes. We found that 565 sites in 220 of 238 (92.44%) changed mitochondrial proteins were increased >2fold and accounted for 96.91% of all acetylated sites (P < 0.01) (Fig. 5C). The mitochondrial proteins were multi-acetylated with 57.4% having at least two AcK sites (Fig. 5D). Pathway enrichment analysis of the hyper-acetylated mitochondrial proteins revealed that the main pathways were regulated by lysine acetylation such as oxidative phosphorylation (313 sites in 47 proteins), the TCA cycle (281 sites in 18 proteins), pyruvate metabolism (194 sites in 16 proteins), fatty acid metabolism (192 sites in 15 proteins), lysine degradation (187 sites in 12 proteins), and ketone body metabolism (58 sites in 5 proteins) (Fig. 5E, Supplementary  Fig. 3D). Because metabolic activities in proximal TECs rely primarily on mitochondrial FAO to generate ATP for energy. We identified lysine acetylation sites on mitochondrial proteins involved in most of the enzymatic processes in mitochondrial metabolic pathways ( Supplementary Fig. 3E). We then examined acetylation sites in proteins involved in the TCA cycle, FAO, and ETC. We found that three sites in PDHE1α in the TCA cycle, one in carnitine palmitoyltransferase 1 (CPT1) and three in ATP synthase subunit O (ATP5O) in the ETC were increased >2-fold (P < 0.01) (Fig. 5F). To further confirm the acetylation of mitochondrial proteins, we immunoprecipitated tubular mitochondrial lysates isolated from sham-operated and UUO mice using an anti-AcK antibody and western blotted them against PDHE1α, CPT1a, and ATP5O. The findings confirmed that mitochondrial proteins in the tubules from the fibrotic kidneys were hyper-acetylated (Fig. 5G). These data overall showed that pro-fibrotic stimulation results in the hyperacetylation of mitochondrial proteins associated with metabolic pathways.

Mitochondrial metabolism is regulated by SIRT3 via acetylation
Previous findings have suggested that SIRT3 regulates the acetylation of mitochondrial proteins. We further investigated the acetylation of PDHE1α, CPT1a, and ATP5O in tubules from mice treated with HKL and Sirt3 KO mice. HKL blocked the hyperacetylation of PDHE1α, CPT1a, and ATP5O in mice with UUO ( Fig. 5H, I). On the contrary, Sirt3 KO mice with or without UUO had more acetylated PDHE1α, CPT1a, and ATP5O than WT mice (Fig. 5J, K, Supplementary Fig. 2C). Taken together, these results indicate that SIRT3 can counteract the effects of acetylation on mitochondrial metabolism during the progression of renal fibrosis.
SIRT3 regulates PDH enzymatic activity via PDHE1α deacetylation Cellular component GO enrichment analysis of upregulated acetylated proteins in tubules collected from mice with UUO revealed that pyruvate dehydrogenase complex (PDHC) was the most obviously enriched GO term (Fig. 6A). We previously immunoprecipitated tubular mitochondrial lysates using anti-AcK D, E Western blots of SIRT3 expression in fibrotic kidneys compared with sham on a postoperative day (POD) 1 (D). Relative abundance of SIRT3 expression, α-tubulin served as the standard (*P < 0.05; n = 3) (E). F, G Western blots of proteins from tubules isolated from kidneys of UUO and sham-operated mice at day 1 using an anti-acetylated-lysine antibody. α-tubulin served as the standard (*P < 0.05; n = 3). H, I Western blots of acetylated lysine in tubular mitochondrial lysates from sham-operated and UUO mice on POD 1. SDHA served as the standard (*P < 0.05; n = 3). J, K Representative immunostaining and immunohistochemical findings of SIRT3 and acetyl-lysine expression in kidneys. The bar indicates 50 μm. and found hyperacetylated PDHE1α in fibrotic kidneys (Fig. 5G). The results of several studies have suggested that PDHE1α acetylation results in PDHE1α phosphorylation and the inhibition of pyruvate dehydrogenase complex (PDC) activity [35]. The enzyme activity of PDH and western blots of proteins in tubules separated from fibrotic kidneys also confirmed that increased lysine PDHE1α acetylation was accompanied by reduced PDH enzyme activity and increased PDHE1α phosphorylation (Fig. 6B-D). Activation SIRT3 by HKL could Fig. 2 Honokiol reduces mitochondrial acetylation and renal fibrosis. A, B: Amount of SIRT3 protein in kidney tissues from UUO mice injected with or without honokiol (HKL). α-tubulin served as the standard (*P < 0.05, † P < 0.05; n = 3). C, D Western blots of acetyl-lysine in tubular mitochondrial lysates from UUO mice injected with or without HKL on POD 1. SDHA served as the standard (*P < 0.05, † P < 0.05; n = 3). E, F Western blots of E-cadherin, Na/K-ATPase, AQP1, FN, and collagen I in UUO mice injected with or without HKL on POD 7. α-tubulin served as the standard (*P < 0.05, † P < 0.05; n = 3). G-J Representative images of H&E (G) (Star indicated tubular atrophy), Masson (H) (Arrow indicated collagen deposition), and immunohistochemical staining for FN (I) and collagen I (J) in kidneys on POD 7 after UUO in mice injected with honokiol. The bar indicates 50 μm.
Y. Zhang et al. inhibit PDHE1α phosphorylation and restore PDH enzyme activity in mice with UUO ( Fig. 6E-G). On the contrary, Sirt3 KO mice with UUO had more PDHE1α phosphorylation and reduced PDH enzyme activity ( Fig. 6H-J).
We investigated the role of SIRT3 on the regulation of PDHE1α acetylation in TECs incubated with TGF-β1 in vitro. We found that increased PDHE1α acetylation was accompanied by reduced PDH enzyme activity and increased PDHE1α phosphorylation (Fig. 7A-D). The activation of SIRT3 by HKL repressed PDHE1α acetylation, and HKL restored phosphorylated PDHE1α, PDH enzyme activity, ATP levels, and lactate production ( Fig. 7A-F). We inhibited SIRT3 activity using 3-TYP and inhibited SIRT3 expression via Sirt3 siRNA transfection. Each of 3-TYP or Sirt3 siRNA transfection increased PDHE1α acetylation, increased PDHE1α phosphorylation, reduced PDH enzyme activity, decreased ATP levels, and increased lactate production (Fig. 7G-R). These data support the notion that SIRT3 functions in PDHE1α deacetylation and regulates a metabolic switch in TECs during the development of renal fibrosis.
Sites where acetylation regulates PDHE1α activity Mass spectrometry-based proteomic analyses have identified three potentially acetylated sites in PDHE1α (Fig. 8A). The K149 and K385 of PDHE1α are conserved from mice to humans (Fig. 8B). We generated an acetyl-deficient K-R mutant of PDHE1α at K149 (149R), K267 (267R), and K385 (385R) to understand more about these acetylation sites. Western blotting with anti-AcK antibody showed that K385 was the major acetylation site of PDHE1α in TECs incubated with TGF-β1 (Fig. 8C, E). The activity of PDH and PDHE1α phosphorylation remained unchanged in TECs with the K385R mutation stimulated with TGF-β1 (Fig. 8D, F). Figure 8G shows the MS/MS spectrum of the acetylated K385 site. We then explored whether SIRT3 regulates acetylation at the K385 site. Wildtype (WT) and 385R mutant plasmid of PDHE1α were transfected with Sirt3 siRNA in TECs. K385 mutation damped the acetylation of PDHE1α induced by Sirt3 knockdown (Fig. 8H, J). The PDH activity was higher and PDHE1α phosphorylation was lower in TECs with K385 mutation (Fig. 8I-K). These results suggest that the acetylation of lysine 385 in PDHE1α regulated by SIRT3 represents a key element in the function of PDH in TECs under pro-fibrotic stimulation (Fig. 8L).

DISCUSSION
To our knowledge, the present study is the first to analyze the spectrum of the acetylome of TECs in fibrotic kidneys and show that most mitochondrial proteins were hyper-acetylated. Our findings support the notion that SIRT3 is a crucial regulator of mitochondrial protein acetylation and metabolic homeostasis. The deacetylation of PDHE1α by SIRT3 provides the primary link between glycolysis and the TCA cycle in TECs during the development of renal fibrosis.
Renal TECs are rich in mitochondria, which rely on oxidative phosphorylation (OXPHOS) to generate ATP. Reduced OXPHOS in proximal TECs contributes to AKI, DN and interstitial fibrosis [10]. Lysine acetylation is a novel post-translational modification of energy metabolism. Numerous enzymes involved in energy metabolism are acetylated in many metabolic diseases [36]. The discovery of hyper-acetylated renal mitochondrial protein provides a new avenue for understanding mitochondrial dysfunction in the development of renal fibrosis. We found that 1789 unique AcK sites in 822 proteins were hyper-acetylated in TECs from UUO, compared with sham-operated mice. Mitochondrial proteins are highlighted as prominent targets of tubular hyper-acetylation (~26%) in fibrotic kidneys. Almost all mitochondrial proteins involved in FAO, TCA cycle, and the ETC were hyper-acetylated. Similar acetylation profiles have been identified in diabetic kidneys and ethanol-induced nephropathy [33,34]. Acetylation neutralizes the positive charge of lysine and impairs the catalytic activity of most mitochondrial enzymes. Acetylation reduces the activity of long-chain acyl-CoA dehydrogenase that catalyzes the PDHE1α or CPT1a or ATP5O served as the standard (*P < 0.05, † P < 0.05; n = 3). J, K Tubular mitochondria extracts from Sirt3 KO and WT mice with UUO on POD 1 immunoprecipitated with anti-acetyl-lysine antibody and analyzed using anti-PDHE1α, anti-CPT1a, and anti-ATP5O. PDHE1α or CPT1a or ATP5O served as the standard (*P < 0.05; n = 3).
first FAO in mitochondria and results in the excessive accumulation of acylcarnitines and triglycerides [37,38]. The acetylation of proteins such as PDHE1α, ATP5O, and CPT1a in the TCA cycle and the ETC results in reduced activity and disrupted oxidative phosphorylation, decreased ATP production, aberrant lipid deposition, and increased glycolysis indicated dysfunctional mitochondrial metabolism in tubules from fibrotic kidneys.
The major mitochondrial protein deacetylase is SIRT3. Mitochondrial proteins were remarkably hyper-acetylated in mice lacking SIRT3 [39][40][41]. The expression of SIRT3 was decreased in a time-dependent manner after UUO. However, mitochondrial acetylation levels were increased sharply after UUO, peaked at day 1, and then gradually decreased. TECs are rich in mitochondria, which is the main source of acetyl-CoA [7]. The decrease of acetylation level at day 7 after UUO may be due to tubular atrophy and mitochondrial dysfunction [42,43].
Acetylome and target protein analyses have shown that SIRT3 functions in the regulation of OXPHOS, FAO, and the TCA cycle. Many studies have shown that SIRT3 is expressed predominantly in mitochondria-rich tissues such as the liver, muscle, heart, brain, and kidney [44]. Furthermore, SIRT3 is prominently expressed in renal cortical tubular cells from younger mice and gradually Fig. 6 Acetylation of PDHE1α is increased in tubules from fibrotic kidneys. A Cellular component GO enrichment analysis of upregulated acetylated proteins in tubules separated from UUO and sham-operated mice. Bubble chart shows top 20 significantly enriched terms (vertical axis, functional category; horizontal axis, Log2 transformed ratio of upregulated acetylated proteins to identified proteins. Colored circles indicate significant p values; circle size indicates a number of differential proteins in terms. B, C Western blots of phosphorylated PDHE1α and total PDHE1α expression in tubules from UUO and sham-operated mice on POD 1. PDHE1α served as the standard (*P < 0.05; n = 3). D Enzyme activity of PDH in tubular mitochondria from UUO and sham-operated mice on POD 1 (*P < 0.05 vs. sham; n = 4); E, F Western blots of phosphorylated PDHE1α and total PDHE1α expression in tubules from UUO and sham-operated mice injected with or without HKL on POD 1. PDHE1α served as the standard (*P < 0.05, † P < 0.05; n = 3). G Enzyme activity of PDH in tubular mitochondria from UUO and sham-operated mice injected with or without HKL on POD 1. (*P < 0.05, † P < 0.05; n = 5-6). H, I Western blots of phosphorylated PDHE1α and total PDHE1α expression in tubules from Sirt3 KO and WT mice with UUO on POD 1. PDHE1α served as the standard (*P < 0.05; n = 3). J Enzyme activity of PDH in tubular mitochondria from Sirt3 KO and WT mice with UUO on POD 1. (*P < 0.05; n = 6). decreases over time [45], as well as with AKI [18,19,46], DN [14,23,47], and hypertension nephropathy [48]. Targeting SIRT3 can effectively reduce mitochondrial fragmentation, reduce the degree of renal damage, and accelerate the recovery of renal function, suggesting that SIRT3 could promote recovery from kidney injury by maintaining the stability of mitochondrial structure and function [18,19,46]. The consequences of Sirt3 KO in mice are that they develop fibrosis in the kidneys, liver, and lungs at the age of 15 months [25]. Pro-fibrotic stimulation reduces renal SIRT3 expression, increases mitochondrial protein acetylation, and disrupts mitochondrial metabolism. Because SIRT3 overexpression could rescue tubule damage and ameliorate renal fibrosis, we believe that reducing SIRT3 and increasing mitochondrial protein acetylation is a novel mechanism of tubular disturbance in the development of renal fibrosis.
Cellular component GO enrichment findings revealed that the PDC was the most obviously acetylated target in fibrotic kidneys. The PDC functions as a metabolic center that catalyzes pyruvate into acetyl-CoA, thus linking glycolysis to the TCA cycle. The metabolic profile of renal TECs during the development of renal fibrosis switches from OXPHOS to glycolysis, which is relevant to SIRT3 expression and activity. Lactic acid has been shown to activate myofibroblasts [49] and promote the transformation of M2 macrophages [50]. Inhibition of PDH activity leads to metabolic transformation to glycolysis, resulting in increased lactic acid production, which may promote the progressing of renal fibrosis. Thus, we postulate that reduced SIRT3 expression and hyper-acetylated PDC could be the mechanism of metabolic reprogramming in fibrotic TECs.
or K385 could alter the catalytic function of PDH. We found that K385R was resistant to TGF-β1 and that Sirt3 knockdown resulted in the inhibition of PDH enzyme activity. These results suggested that SIRT3 regulates PDHE1α activity via lysine 385 deacetylation in TECs.
In summary, our data showed that mitochondrial proteins involved in regulating energy metabolism were acetylated and targeted by SIRT3 in TECs. The deacetylation of PDHE1α at lysine 385 by SIRT3 plays a key role in metabolic reprogramming in renal fibrosis.

MATERIALS AND METHODS Human kidney tissue handling
We collected biopsy specimens from patients at the kidney disease center of the Second Affiliated Hospital of Nanjing Medical University. All patients provided written informed consent to participate in the study, the protocol of which was implemented in accordance with the 1975 Declaration of Helsinki (2013 amendment), and approved by the Medical Experiment Ethics Committee at the Second Affiliated Hospital of Nanjing Medical University (2015KY018). The patients were assigned to groups based on whether they had <20% or ≥20% fibrotic areas determined by Masson staining. The kidney biopsy samples were further analyzed by qRT-PCR and immunohistochemical staining for SIRT3 (cat: 2627, Cell Signaling Technology [CST] Inc., Danvers, MA, USA) and acetyl-lysine (cat: 9441, CST). Total RNA of human kidney biopsy sample was prepared using RNeasy FFPE Kits (cat: 73504, Qiagen GmbH, Hilden, Germany).

Mice and animal models
Male C57BL/6J mice purchased from Nanjing Medical University Experimental Animal Center were housed according to the guidelines of the Institutional Animal Care and Use Committee at Nanjing Medical University. Renal fibrosis was induced in mice in vivo by UUO. HKL was administered intraperitoneally (i.p.) to UUO mice at doses of 0.2 and 1 mg/ kg body weight (bw) for 1 or 7 consecutive days.
The Sirt3 KO mice were provided by Professor Jiahe Xie (Gannan Medical University, Jiangxi, China). Male Sirt3 KO mice and littermate WT mice were used in experiments at the age of 6 weeks. Renal injury was generated by UUO in Sirt3 KO and WT mice. All animals were randomly assigned into different groups (n: [5][6][7][8]. No blinding was done.

Mitochondria isolation
Mitochondria isolated from mouse renal tubular using the Tissue Mitochondria Isolation Kit (cat: C3606, Beyotime, China), and mitochondria of cultured PTCs were isolated by using the Cell Mitochondria Isolation Kit (cat: C3601, Beyotime) according to the manufacturer's protocol.

Immunoprecipitation and immunoblotting
The isolated mitochondria were resuspended in RIPA buffer in the presence of protease inhibitors. For immunoprecipitation, preclear lysate by adding appropriate control IgG (normal rabbit IgG, cat: 12-370, Millipore), together with Protein A/G PLUS-Agarose (cat: sc-2003, Santa Cruz, Biotech, CA, USA), incubate at 4°C for 1 h. After centrifuging, transfer supernatant to a fresh microcentrifuge tube. Add PDH-E1α antibody (cat: ab168379, Abcam) or acetyl-lysine antibody (cat: 9814, CST), incubate at 4°C overnights. Add Protein A/G PLUS-Agarose and incubate at 4°C for 4 h, collect immunoprecipitants by centrifugation and discard the supernatant. Wash pellet 4 times with PBS buffer, bound proteins were eluted by boiling in SDS sample buffer, resolved by SDS-PAGE, and then subjected to Western analysis.

LC-MS/MS analysis
Tubules separated from mice under sham-or UUO-operation on POD 1 were ground by liquid nitrogen for acetylome analysis perform by PTM-Biolabs (HangZhou, China). The Acetylation quantitative study was carried out Acetylation enrichment method and high-resolution liquid chromatography-mass spectrometry (LC-MS)/MS. Normalization with protein quantification to remove the effect of protein expression on modification abundance was used for subsequent bioinformatics analysis. Screening of the differentially modified sites followed the following criteria: twofold change threshold, t test p value < 0.05.
The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD021357.

Statistical analysis
Data are presented as means ± standard error of the mean. Data were statistically analyzed using SigmaStat software (Jandel Scientific, San Rafael, CA, USA). Between-group comparisons were assessed using oneway analyses of variance, followed by Student-Newman-Keuls tests. Values with P < 0.05 were considered statistically significant.

DATA AVAILABILITY
All data generated or analyzed during this study are included in this published article and its supplementary information files. The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.