www.nature.com/cddis ARTICLE OPEN Acox2 is a regulator of lysine crotonylation that mediates hepatic metabolic homeostasis in mice Yuan Zhang1,2,8, Yuling Chen3,8, Zhao Zhang4,8, Xiang Tao1, Sha Xu 3, Xinyan Zhang1, Tinatin Zurashvili5, Zhouping Lu2, ✉ ✉ ✉ José Ramon Bayascas6, Liping Jin 2 , Jianyuan Zhao 7 and Xiangyu Zhou 1 © The Author(s) 2022 Acyl-CoA oxidase 2 (Acox2) is an enzyme involved in peroxisomal bile acid synthesis and branched-chain fatty acid degradation. Acox2 knockout (−/−) mice spontaneously developed liver cancer with marked lymphocytic infiltrate. Tandem-affinity purification coupled with mass spectrometry analysis revealed that Acox2 interacted with methylcrotonoyl-CoA carboxylase followed by co- immunoprecipitation confirmation. Here we reported that non-histone lysine crotonylation (Kcr) levels were downregulated in Acox2−/− mice livers. Interestingly, Kcr signals were concentrated in the nucleus of tumor cells but mostly located in the cytoplasm of adjacent normal liver cells of Acox2−/− mice. Quantitative analysis of the global crotonylome further revealed that 54% (27/50) of downregulated non-histone Kcr sites were located in mitochondrial (11/50) and peroxisomal (17/50) enzymes including Ehhadh, Scp2, Hsd17b4, Crot, Etfa, Cpt1a, Eci1/2, Hadha, Etfdh, and Idh2. Subsequent site-directed mutagenesis and transcriptome analysis 1234567890();,: revealed that Ehhadh K572cr might have site-specific regulatory roles by downregulating TOP3B expression that lead to increased DNA damage in vitro. Our findings suggested Acox2 is a regulator of Kcr that might play critical role on hepatic metabolic homeostasis. Cell Death and Disease (2022)13:279 ; https://doi.org/10.1038/s41419-022-04725-9 INTRODUCTION Lysine crotonylation (Kcr) has emerged as a key posttransla- Beta-oxidation chain shortening of very-long-chain fatty acids, tional modification that plays regulatory roles in many biological 2-methyl-branched fatty acids, and the CoA esters of C27 bile processes [10–12]. Interestingly, a recent study reported that both acid intermediates di- and trihydroxycoprostanoic acid occur in ACOX1 and ACOX3 were key crotonyl-CoA-producing enzymes, the peroxisomes [1, 2]. There are four enzymatic steps in each and deletion of ACOX3 in hESCs led to reduced Kcr, which cycle of peroxisomal β-oxidation, including oxidation, hydra- impaired mesoderm or endoderm differentiation [13]. Short-chain tion, dehydrogenation, and thiolytic cleavage [3]. Acyl-CoA enoyl–CoA hydratase (ECHS1), an enzyme involved in mitochon- oxidases (ACOXs) catalyze the first step. Three ACOX1/2/3 with drial fatty acid beta-oxidation, controls intracellular crotonyl-CoA different substrate specificities are found in humans [4]. ACOX1 and maintains the maturity and homeostasis of cardiomyocytes is responsible for the oxidation of very long straight-chain fatty through histone crotonylation [14]. These findings suggested that acids [5], ACOX2 is the only acyl-CoA oxidase involved in bile mitochondrial and peroxisomal fatty acid oxidation was important acid biosynthesis, and both ACOX2 and ACOX3 are involved in for crotonyl-CoA production. On the other hand, the classic the degradation of branched-chain fatty acids [4, 6]. ACOX2 is histone acetyltransferases p300, pCAF, and MOF also have histone predominantly expressed in the liver and kidney, whereas crotonyltransferase activity [15, 16], whereas the classic histone ACOX3 has relatively lower expression levels in the liver. In the deacetylases HDAC1/2/3 and SIRT1/2/3 can remove Kcr under peroxisomes, D -bifunctional enzyme (HSD17B4) or different conditions [15, 17]. Many non-histone proteins are also L-bifunctional enzyme (EHHADH) catalyze the second and third modified by crotonylation [18]. However, the functional elucida- steps of β-oxidation, respectively [7]. The last step in beta- tion of non-histone Kcr is in its infancy. oxidation can be performed by multiple thiolases, including We previously reported that Acox2 deficiency is associated with acetyl-CoA acyltransferase 1 (ACAA1), and the sterol carrier a metabolic switch in primary malignant cardiac tumor progres- protein x (SCPx or SCP2) [8]. Then shortened fatty acids are sion [19, 20]. ACOX2 homozygous stop-gain (p.Y69*) and missense transported to the cytosol for mitochondrial uptake through the (p.R225W) mutations have been identified in patients with action of carnitine octanoyltransferase (CROT) [9]. transaminase elevation and liver fibrosis [21, 22]. We generated 1 Obstetrics and Gynecology Hospital of Fudan University, School of Life Sciences, Fudan University, 200011 Shanghai, China. 2Shanghai Key Laboratory of Maternal Fetal Medicine, Department of Assisted Reproduction, Shanghai First Maternity and Infant Hospital, Tongji University School of Medicine, 201204 Shanghai, China. 3State Key Lab of Genetic Engineering, School of Life Sciences, Fudan University, 200438 Shanghai, China. 4School of Basic Medicine, Fudan University Shanghai Medical College, 200032 Shanghai, China. 5Department of Biochemistry, David Tvildiani Medical University, Tbilisi 0159, Georgia. 6Departament de Bioquímica i Biologia Molecular, Institut de Neurociències, Universitat Autònoma de Barcelona, Barcelona 08193, Spain. 7Institute for Developmental and Regenerative Cardiovascular Medicine, MOE-Shanghai Key Laboratory of Children’s Environmental Health, Xinhua Hospital, Shanghai Jiao Tong University School of Medicine, 200092 Shanghai, China. 8These authors contributed equally: Yuan Zhang, Yuling Chen, Zhao Zhang. ✉email: jinliping@tongji.edu.cn; zhaojy@fudan.edu.cn; husq04@163.com Received: 26 October 2021 Revised: 22 February 2022 Accepted: 11 March 2022 Official journal of CDDpress Y. Zhang et al. 2 an Acox2 knockout mice model that exhibited the typical clinical Kcr changes are associated with liver cancer progression in symptoms of fatty liver disease at the age of 4 months [23]. Acox2−/− mice Notably, nicotinamide N-methyltransferase (NNMT), a critical To investigate the potential correlation between Kcr levels and regulator of metabolism and epigenetics [24, 25], and its product HCC progression, we examined Kcr changes in tumor and adjacent 1-methylnicotinamide were uniformly downregulated in both normal cells of Acox2−/− mice. A previously well-characterized serum and liver tissues of Acox2−/− mice [23]. Since Acox2 and rabbit pan-anti-Kcr antibody (PTM-501) was utilized to perform Acox3 have similar function in peroxisomal β-oxidation; therefore, immunofluorescence staining [13, 28]. To further clarify the hepatic Kcr was presumed to be downregulated in Acox2−/− mice. specificity of this Kcr antibody, we firstly evaluated the distribution In this study, we perform integrative analysis and find that Acox2 and levels of Kcr in two cultured cell lines including iPSC (induced is a master regulator for protein post-translational modifications pluripotent stem cells) and HepG2. We observed strong positive that involved in the metabolic homeostasis. signals in both nucleus and cytoplasm of iPSCs (Fig. S2A), which is consistent with the findings from the previous study [13]. Meanwhile, we found that the Kcr signals are mostly concentrated RESULTS in the nucleus of HepG2 cells (Fig. S2A). Since KATs and HDACs Acox2 loss induces liver cancer in mice can modulate histone Kcr levels [29, 30], HepG2 cells were then We generated an Acox2 knockout C57BL/6n mouse model using treated with p300 inhibitors and HDACIs, respectively. Compared the CRISPR/Cas9 gene editing system. A deletion of 15,146 bp with to cells without treatment (Ctrl), we identified decreased Kcr levels a span from exon 3 to exon 13 was identified (Supplementary Fig. in cells treated with C646 (2 μm) whereas Kcr signals were S1A) in Acox2 knockout (-/-) mice. Importantly, Hematoxylin and upregulated in cells treated with FK228 (0.25 μM) by immunoblot- eosin (H&E) staining revealed that 28.57% (6/21) Acox2−/− mice ting (Fig. S2B) and immunofluorescence assay (Fig. S2C). These spontaneously developed liver cancer without high-fat diet findings suggested that pan-anti-Kcr antibody provided adequate induction or small molecule treatment at the postnatal age of sensitivity for detecting Kcr changes in Acox2−/− mice. In this 8–9 months (Fig. 1A–C and S1B), whereas wild type (WT) mice did study, we found that Kcr levels were more concentrated in the not develop such tumors. Inflammatory infiltration is very popular nucleus of tumor cells (Figs. 3A, B and S3) whereas adjacent in the liver tissues of Acox2−/− mice. We performed immunohis- normal liver cells that harbor most of Kcr signals in the cytoplasm tochemical stains for immune cell markers (CD3, CD20, CD138, and through confocal immunofluorescence microscopy. By contrast, CD68) to determine the immune cells types and immunocyte levels of Ehhadh were markedly downregulated in the tumor infiltration characteristics. The predominant cell-type is lesions of Acox2−/− mice (Fig. 3A), which further validates the B-lymphocytes CD20, followed by T-lymphocytes CD3 and plasma specificity of the fluorescence signal for the anti-Kcr antibody. In cells CD138, but is almost negative for macrophages CD68 (Fig. addition, Kcr fluorescence signal intensity in cytoplasm of normal 1D). There was a co-appearance of CD3 and CD20. Compared with liver cells in Acox2−/− mice was significantly decreased when WT mice, we also observed positive staining for Alpha-fetoprotein compared with WT mice (Fig. S4A). Since Acox2−/− mice displayed (AFP) in Acox2−/− mice (Fig. 1E). Heterogeneity was identified in bile duct proliferation (Fig. S4B), we also evaluate the possible Kcr tumor lesion of Acox2−/− mice liver (Fig. S1B). In addition, the changes in areas of ductular reaction. Staining for Sox9 and CK7 ACOX2 expression level in humans is significantly downregulated was used to outline bile ducts and ductular reaction as previous in HCC than in normal liver specimens (fold-change = −3.19; p- studies [31–33]. We found that Kcr levels were increased in areas value = 1.91 × 10−7) (Fig. 1F) and decreases gradually in malignant of ductular reaction when compared with adjacent liver tissues stages when compared with the early stages (p-value = 0.00096) (Figs. 3C and S4C). Co-localizations of Kcr signal and ductal (Fig. 1G). These findings are further supported by low ACOX2 markers (CK7 and Sox9) were observed by double immunostain- expression in human HCC is correlated with a decrease in OS ing (Figs. 3C and S4D). These findings suggest that Kcr might play (overall survival, p-value = 0.036) (Fig. 1H) and DFS (disease free important roles in liver cancer progression. survival, p-value = 0.048) (Fig. 1I) when comparing to high ACOX2 expression as observed on analyzing data from TCGA portal. Downregulation of non-histone Kcr in Acox2 mice Since Acox2 and Acox3 have overlapped function on peroxisomal Acox2 Interacts with methylcrotonoyl-CoA carboxylase β-oxidation, we postulated that Kcr levels in Acox2−/− mice might We next used tandem-affinity purification coupled with mass be changed. To address this concern, an overview of four different spectrometry (TAP-MS) analysis to determine the interactome classic protein post-translational modifications, including ubiqui- of ACOX2 and its functions in MHCC97H and HEK-293T cell tination (Fig. 4A), acetylation (Fig. 4B), succinylation (Fig. 4C), and lines, respectively (Supplementary Tables S1 and S2). Since crotonylation (Fig. 4D), was obtained through western blot ACOX2 works as a homodimer that is essential for biological analysis using the corresponding pan antibody. We did not function, ACOX2 was also identified as an interacting partner in observe significant differences in acetylation, ubiquitination, and both cell lines, which further validate the data reliability of TAP- succinylation between liver tissues of Acox2−/− mice and their WT MS analysis. Approximately 40% (12/30) of the top 30 littermates. However, non-histone lysine crotonylation (Kcr) was interacting proteins for Acox2 overlapped in MHCC97H and partially downregulated in the liver tissue of Acox2−/− mice when HEK-293T cells (Fig. 2A). In particular, four classic biotin- compared with wild-type mice (Fig. 4D). By contrast, a mild dependent carboxylases [26], including acetyl-CoA carboxylase increase in histone Kcr was observed in Acox2−/− mice. In 1 (ACACA) and 2 (ACACB), and methylcrotonoyl-CoA carbox- addition, decreased non-histone Kcr was also detected in the ylase subunit alpha (MCCA) and beta chain (MCCB), propionyl- kidney tissues of Acox2−/− mice (Fig. 4F) whereas acetylation (Fig. CoA carboxylase subunit alpha (PCCA) and subunit beta (PCCB), 4E) was not significantly changed. and pyruvate carboxylase (PC), have protein–protein interac- tions with ACOX2. The interaction between ACOX2 and MCCC2 Histone (H2b) K86cr was upregulated in Acox2−/− mice (MCCB) was further confirmed by co-immunoprecipitation in To precisely determine hepatic Kcr changes caused by Acox2 HEK-293T cells (Fig. 2B). MCCC1 and MCCC2 could form a knockout, global crotonylome analysis was performed using the heterodimer and catalyze the carboxylation of 3- liver tissues of Acox2−/− (n = 3) and WT littermates (n = 3) that methylcrotonyl-CoA to 3-methylglutaconyl-CoA [27], and intra- were collected at four months of age (Fig. 5A). Principal cellular crotonyl-CoA could directly regulate Kcr [12]. In component analysis (PCA) separated the six liver tissue samples addition, ACOX2 has protein interaction with DNA topoisome- into two distinct categories (Fig. 5B). All 2324 quantifiable rases including TOP2A and TOP2B in MHCC97H cells (Table S1). crotonylation sites in 848 proteins were normalized according to Cell Death and Disease (2022)13:279 Y. Zhang et al. 3 Fig. 1 Acox2 loss induces HCC in mice. A Percentage distribution of liver cancer was compared between Acox2−/− (n = 21) and WT (n = 17) mice at the age of 8–9 months. B The representative image of HE stained cross-sections of liver tissue in Acox2−/− and WT mice. Scale bar, 500 μm. C Three tumor lesions (I–III) from a single cross-section of Acox2−/− mouse liver are boxed to show the enlarged region (IV–VI) as indicated. Scale bars, 100 μm in upper panel and 30 μm in bottom panel. D Staining was performed using HE and immunostaining against CD3, CD20, CD68, and CD138 on liver paraffin sections of Acox2−/− or WT mice. Representative images were displayed as indicated. Scale bar, 100 μm. E Immunohistochemical staining of formalin-fixed paraffin-embedded section of mouse liver tissues using AFP antibody. Representative images were displayed as indicated. Scale bar, 20 μm. F, G ACOX2 expression level in humans is significantly downregulated in HCC than in normal liver specimens (fold-change = −3.193; p-value = 1.91 × 10−7) (F) and decreases gradually in malignant stage (III) when compared with the early stage (I) (p-value = 0.00096) (G) by analyzing data from TCGA portal. H, I Low ACOX2 expression in human HCC is correlated with a decrease in OS (overall survival, p-value = 0.036) (H) and DFS (disease free survival, p-value = 0.048) (I) when comparing to high ACOX2 expression as observed on analyzing data from TCGA portal. the corresponding quantitative proteomics analysis for each Crotonylome analysis identified a total of 23 Kcr sites on sample. We found that Kcr is prevalent in mouse liver and almost histone proteins, including histone (H)1.4 (Hist1h1e), H2A every enzyme that catalyzes intermediate metabolism, including (H2afj, H2afx, Hist2h2ac, and Hist2h2bb), H2B (Hist1h2bc and fatty acid oxidation, the TCA cycle, gluconeogenesis, the urea Hist1h2bp), H3.3 (H3f3a), and H4 (Hist1h4a). Among them, cycle, glycogen metabolism, and glycolysis, was crotonylated. We 60.87% (14/23) of Kcr sites were quantifiable (Table 1). We next analyzed the flanking sequences of these sites and found found that H2bc4 (K86cr) was the only site that could meet our that aspartic acid (D) and phenylalanine (F) were overrepresented, filter strategy (fold-change > 1.4, p < 0.05) that is significantly whereas glutamine (Q), arginine (R), and cysteine (C) were less upregulated in Acox2−/− mice. However, classic H2b clustered than expected at the −1 and + 1 positions surrounding the lysine histone 4 (H2bc4) (K12cr) was not significantly changed in crotonylation sites (Fig. 5C). Acox2−/− mice in our study. These findings suggest that histone Cell Death and Disease (2022)13:279 Y. Zhang et al. 4 Fig. 2 Acox2 interacts with methylcrotonoyl-CoA carboxylase. A Overlapped Acox2-interacting proteins in MHCC97H and HEK293T cells by TAP-MS analysis were displayed. B Anti-Myc immunoprecipitation assays were performed in HEK293T cells overexpressing Flag-tagged ACOX2 and Myc-tagged MCCC2 constructs followed by immunoblotting with the indicated antibodies. Specificity of antibodies was evaluated as indicated (left panel). MCCC2 methylcrotonoyl-CoA carboxylase subunit 2, MS mass spectrometry, TAP tandem affinity purification. and non-histone Kcr might be differentially regulated in “peroxisomal part,” and “mitochondrion” were the top 3 enriched Acox2−/− mice livers. terms for CC analysis (Fig. 5G). These findings reveal the key regulatory roles of non-histone Kcr on peroxisomal and mitochon- Downregulated non-histone Kcr sites on metabolic enzymes drial metabolism in mouse liver (Fig. 5H). For non-histone proteins, a total of 50 downregulated Kcr sites on 38 proteins and 3 upregulated Kcr sites on 3 proteins were Validation of Kcr changes on Ehhadh and Crot identified in Acox2−/− mice (fold-change > 1.4; p < 0.05) when In our study, the most significantly changed Kcr sites were compared with the WT group (Fig. 5D and Table S3), which further mainly located at key enzymes in peroxisomal beta-oxidation confirmed the downregulation of non-histone Kcr in Acox2−/− including Ehhadh, Crot, Scp2, Hsd17b4, Acaa1a and Abcd3. mice. Approximately 32.0% (16/50) of downregulated Kcr sites Interestingly, protein levels of these enzymes carrying a number were located in the proteins of peroxisomal β-oxidation pathway of downregulated Kcr sites were significantly upregulated in including Crot (K[lys]33, K69, K143, K226, K384, and K489), Hsd17b4 Acox2−/− mice (Fig. 6A, B). All six Kcr sites in Crot, including K33, (K57, K68, and K81), Ehhadh (K329, K344, K572), Scp2 (K211 and K524), K69, K489, K384, K226, and K132, were downregulated by 0.391-, Acaa1a (K292), and Abcd3 (K529) (Fig. 5E). Moreover, 22.0 % (11/50) 0.426-, 0.494-, 0.522-, 0.544-, and 0.664-fold, respectively. of downregulated Kcr sites were located in mitochondrial proteins, Ehhadh K329, K344, and K572 decreased by 0.284-, 0.448-, and including Hadha (K190), Eci1 (K229), Eci2 (K90), Cpt1a (K180), Otc 0.621-fold, respectively, whereas K335 did not significantly (K80), Idh2 (K67), Slc25a22 (K188), Slc25a10 (K253), Etfa (K139), Etfdh change (fold-change = 0.906) (Fig. 6C). A total of 17 Kcr sites (K152), and Hmgcs2 (K327) (Fig. 5E). Analysis of Biological Process were identified in Scp2. Sequencing alignment indicated that (BP) and Cellular Component (CC) of downregulated Kcr proteins Ehhadh K329 and K572 (Fig. 6D) and Crot K69, K384 and K489 (Fig. by Gene Ontology (GO) revealed that “cellular lipid catabolic 6E) were highly conserved across species, including zebrafish or process,” “lipid oxidation,” and “lipid modification” were the top 3 xenopus. We then generated Crot or Ehhadh mutants in which enriched terms for BP analysis (Fig. 5F), and “peroxisome,” lysine Kcr sites were simultaneously mutated to arginine by site- Cell Death and Disease (2022)13:279 Y. Zhang et al. 5 Fig. 3 Kcr levels are associated with HCC progression in Acox2−/− mice. A Fluorescent immunohistochemistry using DAPI, anti-Ehhadh and pan anti-Kcr were performed on three slides of a single tumor focus. Scale bar, 25 μm. B Immunohistochemical staining showed Kcr signals were concentrated in the nucleus of tumor cells (upper panel) but mainly located in the cytoplasm of adjacent normal liver cells (bottom panel) in Acox2−/− mice, Scale bar, 3 μm. C Kcr levels were increased in the regions of bile duct proliferation in Acox2−/− mice by double immunostaining with Kcr and a ductal marker CK7. Co-localization of Kcr signal with CK7 was observed and indicated by white arrow. Representative images were displayed. Scale bar, 10 μm (upper) and 3 μm (bottom). Cell Death and Disease (2022)13:279 Y. Zhang et al. 6 Fig. 4 Downregulation of non-histone Kcr in Acox2−/− mice. A–D Overview of anti-ubiquitin (A), -succinyl lysine (B), -acetyl lysine (C), and –crotonyl lysine (D) in liver tissues of Acox2−/− (n = 3) and WT (n = 3) mice using corresponding pan-antibodies by western blots. E, F Overview of anti-acetyl lysine (E), and –crotonyl lysine (F) in kidney tissues of Acox2−/− (n = 3) and WT (n = 3) mice using corresponding pan- antibodies by western blots. specific mutagenesis to mimic a decrotonylated status. Croto- (Fig. 6F–I), which further validate the data reliability of nylation of Ehhadh (Fig. 6F, G) and Crot (Fig. 6H, I) was further crotonylome analysis. confirmed by using immunoprecipitation assays in HEK-293T and HepG2 cell lines overexpressing Myc-tagged Ehhahdh or Site-specific regulatory role of Ehhadh Kcr on the Crot recombinant proteins, respectively. Mutating lysine to transcriptome in vitro arginine in Kcr sites of Crot (K69R and K384R) or Ehhadh (K329R, Ehhadh K329cr (fd = 0.284) and K335cr (fd = 0.906) were K344R and K572R) reduces Kcr levels to various degrees in vitro differentially regulated in Acox2−/− mice, which implicates the Cell Death and Disease (2022)13:279 Y. Zhang et al. 7 Fig. 5 Global crotonylome analysis of liver tissues in Acox2−/− and WT mice. A Volcano plots showing the significantly changed Kcr sites (FDR < 1%, p < 0.05). Up- and downregulated Kcr sites are highlighted in red and blue, respectively. B Score plots of principal component analysis (PCA) of Acox2−/− and WT crotonylome. C Motif analysis of all identified crotonylated sites. D Compared with the wild type (WT) group, 50 downregulated Kcr sites on 38 proteins and 3 upregulated Kcr sites on 3 proteins were identified in Acox2−/− mice (fold-change >1.4; FDR < 1%; p < 0.05). E Venn diagram showing 32% (16/50) and 22% (11/50) of downregulated Kcr sites were located in the proteins involved in peroxisomal and mitochondrial metabolism, respectively. F, G Analysis of Biological Process (F) and Cellular Component (G) of downregulated Kcr proteins by Gene Ontology (GO) enrichment (Fisher’s exact test; p < 0.05). Top selected pathways are shown in each category as indicated. H Schematic representation of downregulated hepatic Kcr on enzymes involved in fatty acid oxidation and TCA cycle in Acox2−/− mice. Proteins carrying downregulated Kcr sites are marked in brown. Cell Death and Disease (2022)13:279 Y. Zhang et al. 8 Table 1. 14 Quantifiable Kcr sites on histone proteins were identified in Acox2−/− mice by global crotonylome. Gene name Histone protein description Amino acid Kcr sites Kcr Ratio (Acox2−/− _vs_ WT) p-value H2afx Histone H2AX K 134 0.968 0.43172 H3f3a Histone H3.3 K 80 0.911 0.38218 Hist1h1e Histone H1.4 K 106 0.939 0.2908 Hist1h2bc Histone H2B type 1-C/E/G K 35 0.96 0.39846 Hist1h2bc Histone H2B type 1-C/E/G K 21 0.969 0.84268 Hist1h2bc Histone H2B type 1-C/E/G K 86 1.511 0.005361 Hist1h2bc Histone H2B type 1-C/E/G K 12 0.985 0.92898 Hist1h2bc Histone H2B type 1-C/E/G K 121 1.025 0.73184 Hist1h2bc Histone H2B type 1-C/E/G K 47 1.108 0.132937 Hist1h2bp Histone H2B type 1-P K 12 1.139 0.0129956 Hist1h4a Histone H4 K 78 0.888 0.112897 Hist2h2ac Histone H2A type 2-C K 37 0.954 0.73574 Hist2h2ac Histone H2A type 2-C K 96 0.907 0.062754 Hist2h2bb Histone H2B type 2-B K 21 1.015 0.82528 Bold values represent the significantly upregulated Kcr site (fold-change >1.4, p-value < 0.05). possible site-specific regulatory role of Ehhadh Kcr on biological DISCUSSION processes. To mimic the upregulation of Ehhadh and investi- Our study highlights the oncogenic property of Acox2 deficiency gate the site-specific effects of Kcr changes, HEK-293T cells in tumor progression as those of our previous [19] and other were transfected with control (Ctrl) and Ehhadh WT or mutants studies [39]. Acox2 could be a critical regulator for metabolic (K to R) recombinant Myc-tagged plasmids, as indicated. homeostasis through post-translational modifications, probably Overexpression of Ehhadh was validated through western Kcr. Meanwhile, our results demonstrated that Kcr is very blotting. We next performed RNA-seq based transcriptome prevalent in liver and virtually every enzyme in glycolysis, the analysis with three biological replicates for each sample. tricarboxylic acid (TCA) cycle, and fatty acid metabolism was Subsequent PCA analysis based on 3 biological replicates crotonylated. Importantly, most of downregulated non-histone Kcr effectively separated K572R from Ehhadh WT or the other sites in Acox2−/− mice were located in metabolic enzymes, which mutants (Fig. S5A). Hierarchical clustering results and heatmap extend the scope of metabolism regulation by Kcr to an extent representation of DEGs profiles are shown (Figs. 7A and S5B). Of comparable to that of other major posttranslational modifications note, Top3b (DNA topoisomerase III beta) was significantly such as acetylation [40] and ubiquitination [41]. In our study, downregulated in cells expressing Ehhadh K572R when lymphocytes were diffusely positive for CD20 in Acox2−/− mice. compared with WT. GO enrichment analysis revealed that The characteristic of the tumor-infiltrating leukocytes could “DNA topoisomerase type I (single strand cut, ATP-indepen- impact clinical outcome in HCC [42]. Although the role of dent) activity” was the most significantly enriched term on site- tumor-infiltrating B cells remains controversial, two previous specific DEGs from K572R probably due to the downregulation studies reported that margin-infiltrating CD20(+) B cells correlate of DNA topoisomerase III beta (TOP3B) (Fig. 7B). Subsequent with favorable prognosis in HCC [43] and CD20(+) showed a trend cDNA amplification (Fig. 7C), real-time (RT) quantitative PCR towards better DFS [44]. Interaction between tumor-infiltrating B (Fig. 7D), and western blot (Fig. 7E) uniformly confirmed that cells and T cells controls the progression of HCC [45]. These TOP3B expression was downregulated in cells transfected with findings support the anti-tumor activity of CD20(+) B cells in HCC. Ehhadh K572R when compared with Ehhadh WT and other A recent study found that histone crotonylation is crucial for mutants. These findings suggest that non-histone Kcr has site- endoderm differentiation by regulating metabolic switch [13]. specific regulatory roles in biological processes. Differentiation of pluripotent embryonic stem cells to mesoderm Previous studies found that loss of TOP3B leads to increased and endoderm triggers a metabolic switch from glycolysis to DNA damage and impaired recovery after replication stress oxidative phosphorylation [46]. This metabolic switch is believed [34, 35]. We utilized a anti-gammaH2AX antibody, a sensitive to be important to release stem cells from pluripotency and indicator of DNA damage and replication stress, to perform promote differentiation [47]. Interestingly, this switch is the immunofluorescence staining (Fig. 8A) and Western blot (Fig. reverse process for the classic Warburg effect [48]. Thus, Kcr 8B). We found that gammaH2AX formation is enhanced in cells changes caused by Acox2 deficiency might be directly involved in expressing Ehhadh K572R, when compared with the WT and the the metabolic switch that is essential for tumor transformation in other mutants. ATM is the major kinase involved in the liver tissues of Acox2−/− mice. phosphorylation of H2AX in response to DNA double-strand Several key peroxisomal enzymes, including Ehhadh, Crot, breaks (DSBs) [36–38]. Thus, we next evaluate the phosphoryla- Scp2, and Hsd17b4, carrying decreased Kcr sites were uniformly tion levels of ATM-CHK2 pathway and found that phosphoryla- upregulated in Acox2−/− mice. Studies have found that tion of ATM (Ser1981) and CHK2 (Thr68) were significantly acetylation can directly affect enzyme stability [49]. For upregulated in cells expressing Ehhadh K572R when compared example, the PEPCK13KacR mutant was more stable than the with the WT (Fig. 8B). Furthermore, compared with WT mice, wild type, whereas the PEPCK13KacQ mutant remained unstable Top3b protein levels were significantly downregulated in the [49]. Acetylation and crotonylation have similar mechanisms liver of Acox2−/− mice by Western Blot (Fig. 8C, D) (p-value < and are both involved in metabolic regulation. Since histone 0.01, n = 6). These findings suggested that non-histone Kcr acetyltransferases p300, pCAF, and MOF also have histone changes caused by Acox2 knockout might be involved in DNA crotonyltransferase activity [15, 16], whereas the classic histone damage responses. deacetylases HDAC3 and SIRT1/2/3 can remove Kcr under Cell Death and Disease (2022)13:279 Y. Zhang et al. 9 different conditions [15, 17], we postulated that changes at escaped genes in postmeiotic spermatogenic cells [12]. In the non-histone Kcr levels were involved in stability regulation of present study, we identified an upregulated Kcr site at H2b (K86cr) metabolic enzymes. whereas H2b (K12cr) was not significantly changed in Acox2−/− A recent study found that CDYL-regulated histone Kcr, mainly mice. Thus, investigating the site-specific biological function of H2b (K12cr), controls the expression of sex chromosome-linked H2b (K86cr) on liver homeostasis by generating point knock-in Cell Death and Disease (2022)13:279 Y. Zhang et al. 10 Fig. 6 Downregulated non-histone Kcr sites on metabolic enzymes. A Key peroxisomal enzymes including Ehhadh, Scp2, Crot, and Hsd17b4 carrying downregulated Kcr sites were significantly upregulated (fd > 1.5) in Acox2−/− mice. B Upregulation of Ehhadh expression in Acox2−/− mice was verified through western blotting. C Heatmap showing profiles of all quantifiable Kcr sites in Crot, Ehhadh, and Scp2 before or after normalization according to quantitative proteomics analysis. D, E Sequence alignment of identified Kcr sites in Ehhadh (D) or Crot (E) in different species including xenopus and zebrafish. F, G Replacing lysine with arginine at Kcr sites lead to decreased Ehhadh Kcr levels in 293T (F) and HepG2 (G) cells as indicated. Cells were transfected with Myc-tagged WT and point mutants of Ehhadh. Cell lysates were immunoprecipitated with an anti-Myc antibody followed by immunoblotting with a pan-Kcr antibody. H, I Replacing lysine with arginine at Kcr sites leads to decreased Crot Kcr levels in 293T cells (H) and HepG2 cells (I). Representative image from three independent experiments was displayed. mutation would be a potential avenue for understanding these C57BL/6n female mice (7–8 weeks old) were used as embryo donors. results. Briefly, fertilized embryos (zygotes) were collected from the oviducts, and In this study, many downregulated Kcr sites in Acox2−/− mice Cas9 mRNA (100 ng/μL) plus sgRNA (50 ng/μL) targeting Acox2 were mixed were located in proteins involved in mitochondrial beta-oxidation, and injected into the cytoplasm of fertilized eggs with both pronuclei including Hadha and Eci1/2, as well as the TCA cycle, including visible in Chatot–Ziomek–Bavister medium. The injected zygotes were then cultured in Quinn’s Advantage cleavage medium (In Vitro Fertiliza- Idh2. Thus, Kcr might provide a vehicle for the cross-talk between tion, Inc.) containing SCR7 (50 μm; TOCRIS Bioscience, Bristol, UK) for ~24 h, peroxisome and mitochondria, which is critical for metabolic and 18–20 two-cell-stage embryos were transferred into the oviduct of a regulation. A recent study found that CROT promotes fatty acid pseudo-pregnant ICR female mouse at 0.5 dpc (days postcoitus). All mice metabolism and is involved in mitochondrial dysfunction during had access to food and water. All experiments were performed in vascular calcification [50]. Thus, dysregulated Kcr sites in accordance with Health Guide for the Care and Use of Laboratory Animals. peroxisomal enzymes could be developed as a potential Genomic DNA was extracted from tail tips, and PCR-based genotyping was therapeutic target for human disease. performed. Sequences of sgRNA and primers are provided in Supplemen- In this study, Top3b expression is downregulated in the cells tary data. For all studies, littermate WT and Acox2−/− mice derived from expressing Ehhadh K572R as well as in the liver tissues of Acox2−/ heterozygous parents were used. − mice. Loss of TOP3B was recently reported to increase DNA damage and genome instability [34, 35]. DNA damage sensor Immunofluorescence and immunohistochemistry gamma-H2AX is increased in preneoplastic lesions of HCC [51]. For paraffin sections, dewaxing, hydration, and antigen retrieval steps were Mutating the putative Ku70 phosphorylation sites in increased performed by following standard protocol. After blocking (10% normal goat serum [NGS]), slides were incubated with the primary antibodies as gamma-H2AX and spontaneous induction of HCC [52]. Phosphor- following: anti-pan-Kcr (PTM Biolabs 501 and 502, 1:1000 dilution); anti- ylation of Chk2, a central effector of DNA damage response, at Ehhadh (Santa Cruze sc-393123, 1:200 dilution); anti-CK7(Abcam ab181598, threonine68 (Thr68) leads to Chk2 activation and Patients with 1:2000 dilution); anti-Sox9 (Santa Cruze sc-166505, 1:100 dilution) over- HCC had increased CHK2 mRNA [53]. Interestingly, a recent study night at 4 °C. The slides were then incubated with the secondary antibody by Carloni et al found that CHK2 overexpression enhances (Goat anti-rabbit or -mouse IgG, Alexa Fluor 488 or 594; Thermo Fisher chromosomal instability and HCC progression [54]. In addition, a Scientific, 1:1000 dilution) for 2 h at 25 °C, followed by staining with DAPI protein called TAZ in pre-tumor NASH-hepatocytes promotes DNA (Thermo Fisher Scientific)/PBS for 6 min. Confocal imaging was performed damage and thereby induce hepatocarcinogenesis [55]. These using an SP8 system (Leica), and images were processed using the Leica AF findings suggested that downregulated Top3b might cause software suite. For immunohistochemistry, anti-CD3 (99940, 1:150 dilution), genome instability that is benefit for tumorgenesis in the liver anti-CD20 (70168, 1:500 dilution) and anti-CD68 (97778, 1:200 dilution) antibodies are from CST and anti-CD138 (Abcam ab128936, 1:500 dilution). tissues of Acox2−/− mice. On the other hand, non-histone Kcr was Sections were incubated with biotinylated secondary antibodies and also downregulated in the kidney tissues of Acox2−/− mice. horseradish peroxidase avidin D (HRP, Vector). After three 5-min washes of Interestingly, quantitative proteomic analysis indicated that PBS, the sections were developed with DAB kit (Vector) as standard protein level of MRE11, a double-strand break repair nuclease, procedures. was significantly upregulated in the kidney tissues of Acox2−/− mice (data not shown). Besides self-activation by autophosphor- Tandem affinity purification ylation at S1981, ATM was activated via the MRE11 complex [56]. Tandem affinity purification (TAP)-MS analysis was performed as described Interestingly, MRE11 complex could interact with both ATM and [60]. Briefly, MHCC97H and HEK-293T cells (ATCC®ACS-4500™) were grown RPA1 [57, 58]. A recent study by Yu et al. found that RPA1 Kcr plays in DMEM High Glucose (4.5 g/L) (Gibco™), supplemented with 10% FBS a key role in homologous recombination DNA repair [12]. Thus, (Gibco), 100 IU/mL penicillin and 100 μg/mL streptomycin at 37 °C in a 5% upregulated MRE11 might contribute to the DNA damages CO2 atmosphere. Cells were then transfected with pMCB-SBP-Flag-ACOX2 responses in the kidney tissues of Acox2−/− mice. Over all, our containing a puromycin resistance marker. Cells were lysed on ice in 0.1% findings provide additional evidences to support the role of non- NP40 buffer (20 mM Tris-HCl [pH 8.0], 150 mM NaCl) and centrifuged to histone Kcr on DNA damage responses. remove cell debris; the resulting cell lysates were incubated with streptavidin-conjugated beads (GE Healthcare) for 3 h at 4 °C. The The main limitation of our study is that we could not precisely precipitates were washed three times with 0.1% NP40 buffer, 2 times determine the mechanism of different changes in Kcr for histone with ddH2O, and three times with 50 mmol/L NH4CO3 and subjected to and non-histone proteins in Acox2−/− mice. Thus, some unknown tryptic digestion at 37 °C overnight. The peptides in the supernatant were mechanism might regulate this biological process. We postulated collected by centrifugation, dried in a speed vacuum (Eppendorff), and that the imbalanced distribution of crotonyl-CoA in the nucleus dissolved in NH4CO3 buffer containing 0.1% formic acid and 5% and cytoplasm is the main cause for this phenomenon. Overall, acetonitrile before being subjected to mass spectrometry. our results suggest that Acox2 is a regulator for protein post- translational modifications, mainly Kcr, that play important roles Kcr peptide enrichment on metabolic homeostasis and liver cancer progression in mice. To enrich Kcr modified peptides, tryptic peptides were dissolved in NETN buffer (100 mM NaCl, 1 mM EDTA, 50 mM Tris-HCl, and 0.5% NP-40 [pH 8.0]) and then incubated with antibody beads (PTM Bio Inc., Hangzhou) as MATERIALS AND METHODS described [12] at a ratio of 15-μL beads/mg protein at 4 °C overnight. The Generation of Acox2 knockout mice antibody beads were washed four times with NETN buffer and two times Acox2 knockout (KO) mice were generated by Gempharmatech Co., Ltd. with ddH2O. The Kcr peptides were then eluted by adding the elution (Nanjing, China) based on the CRISPR-Cas9 technique, as described [59]. buffer with 0.1% trifluoroacetic acid. For LC-MS/MS analysis, the resulting Cell Death and Disease (2022)13:279 Y. Zhang et al. 11 Fig. 7 Site-specific regulatory role of Ehhadh K572cr on the transcriptome in vitro. A Hierarchical clustering and Heatmap showing the profiles of differentially expressed genes (DEGs) in 293T cells expressing Ehhadh K572R when compared with Ehhadh WT. B Gene Ontology enrichment analysis revealed that “DNA topoisomerase type I (single strand cut. ATP-independent) activity” was the most significantly enriched term in the K572R vs WT group. C, D mRNA levels of TOP3B were further confirmed by cDNA PCR (C) and quantitative real-time PCR (D). *p < 0.05; ***p < 0.001 (one-way ANOVA test, n = 4). E Western blot showed protein levels of TOP3B was downregulated whereas expression of FH and HK2 was not affected in cells overexpression K572R when compared with WT and other mutants. Representative image from three independent experiments was displayed. peptides were desalted with C18 ZipTips (Millipore) according to the Site-directed mutagenesis and immunoprecipitation manufacturer’s instructions. pCMV6-Ehhadh (Myc-DDK-tagged) and -Crot (Myc-DDK-tagged) mouse ORF clones were purchased from OriGene Technologies Inc (Rockville, USA). The QuickChange site-directed mutagenesis kit (Stratagene; Agilent Tandem mass tagging proteomic analysis Technologies) was used to substitute the identified lysine Kcr sites with Tandem mass tagging-based proteomics analysis was supported by Jingjie arginine and generate four mutant Ehhadh variants (K329R, K335R, K344R, PTM Biolabs (Hangzhou, China) as described [61]. The detailed procedures and K572R) and five mutant Crot variants (K33R, K69R, K143R, K384R, and are provided as Supplementary data. For GO analysis, proteins were K489R), which were confirmed by Sanger sequencing and compared to a classified using GO annotation into three categories: biological process, reference sequence (RefSeq accession number: NM_023733 for Crot and cellular compartment, and molecular function. For each category, a two- NM_023737 for Ehhadh). Then HepG2 or 293T cells were transfected with tailed Fisher’s exact test was employed to test the enrichment of Myc-tagged Crot and Ehhadh constructs followed by anti-Myc immuno- differentially modified proteins against all identified proteins. GO with a precipitaion as above. corrected p < 0.05 was considered significant and GO term enrichment was performed using DAVID 6.8. RNA sequencing Transcriptome libraries were prepared using a TruSeq RNA Sample Co-immunoprecipitation Preparation kit (Illumina) and then sequenced using the Illumina HiSeq PE pcDNA3.1-MCCC2-Myc (NM_022132) and pEnter-ACOX2-Flag-His platform at 2 × 151-bp read lengths as previously [62]. The expression level of (NM_003500) expression vectors were obtained from WZ Biosciences each transcript was calculated according to fragments per kilobase of exon Inc. (Jinan, China) followed by Sanger sequencing validation. HEK-293T per million mapped reads. Differentially expressed genes (DEGs) between was transfected with Myc-tagged MCCC2 and Flag-tagged ACOX2 groups were selected based on a log fold-change >2 and a false discovery using Lipofectamine 3000 (Thermo Fisher). After 36 h, cells were lysed rate FDR < 0.05. Gene ontology (GO) functional-enrichment and Kyoto in Pierce IP Lysis Buffer (87787, Thermo Fisher) supplemented with Encyclopedia of Genes and Genomes pathway analyses were performed proteinase inhibitor. After centrifugation, 200 μg lysates were incu- using GOATOLLS and KOBAS. DEGs were considered significantly enriched bated with anti-Myc-conjugated agarose beads (20168, Thermo Fisher) when the p value was less than 0.05 following Bonferroni’s correction. for immunoprecipitation following standard protocol. Cell Death and Disease (2022)13:279 Y. Zhang et al. 12 Fig. 8 Non-histone Kcr changes caused by Acox2 knockout might be involved in DNA damage responses. A Immunofluorescence staining with anti-gammaH2AX antibody in cells expressing Ehhadh WT or (K-R) mutants. B Immunoblot analysis of gammaH2AX and pan- or phospho-ATM and CHK2 in cells expressing Ehhadh WT or (K-R) mutants. Beta-actin served as a loading control. C, D Top3b protein levels were significantly downregulated in the livers of Acox2−/− mice by western Blot (C). Quantification data of Top3b expression level when normalized against Beta-actin (D). **p value < 0.01 (two-tailed Student’s t test; n = 6). A.U arbitrary units. Representative image from three independent experiments was displayed. cDNA amplification and RT-qPCR significance of differences between WT and mutant groups using Total RNA was isolated from HEK-293T cells transfected with mouse Myc- GraphPad Prism software v. 5.01 (GraphPad Software, Inc., La Jolla, CA, tagged Ehhadh WT or mutant constructs using the RNAprep pure Cell Kit USA). Differences were significant when p values were <0.05. (DP430, Tiangen, Beijing, China) followed by first-strand cDNA synthesis. cDNA was then amplified using PCR with TOP3B primers followed by Sanger sequencing. RT-qPCR was performed using the FastFire qPCR DATA AVAILABILITY PreMix (SYBR Green) (FP207, TIANGEN) on an ABI StepOnePlus instrument. The RNA-sequencing data have been deposited in the Gene Expression Omnibus The delta-delta-Ct (ddCt) algorithm was utilized to analyze the relative (GEO) repository under accession number GSE194202. Other data that support the change in gene expression by using the housekeeping gene GAPDH as the fndings of this study are available from the corresponding authors upon reasonable internal reference. The data are presented as the mean ± SD of four request. independent experiments; each sample was assayed in triplicate in each experiment. RT-qPCR primer sequences are available in Supplementary data. REFERENCES 1. Reddy JK, Hashimoto T. Peroxisomal beta-oxidation and peroxisome proliferator- Immunoblotting activated receptor alpha: an adaptive metabolic system. Annu Rev Nutr. After undergoing electrophoresis and membrane transfer, the proteins 2001;21:193–230. were probed with one of the following primary antibodies in either 5% 2. Poirier Y, Antonenkov VD, Glumoff T, Hiltunen JK. Peroxisomal beta-oxidation–a nonfat milk or 5% bovine serum albumin: anti-pan Kcr (PTM, 501, 502), metabolic pathway with multiple functions. Biochim Biophys Acta. anti-pan Kac (PTM, 101), anti-pan Ksu (PTM, 401), anti-pan ubiquitin (PTM, 2006;1763:1413–26. 1107), anti-ATM (Abways, CY7229), anti-phospho-ATM (Ser1981) (Abways, 3. Houten SM, Denis S, Argmann CA, Jia Y, Ferdinandusse S, Reddy JK, et al. Per- CY5111), anti-CHK2 (CST, 6334), anti-phospho-CHK2 (Thr68) (CST, 2197), oxisomal L-bifunctional enzyme (Ehhadh) is essential for the production of anti-H2AX (CST, 7631), anti-gammaH2AX (CST, 9718), anti-GAPDH (CST, medium-chain dicarboxylic acids. J Lipid Res. 2012;53:1296–303. 8844), anti-Ehhadh (Santa Cruz, sc-393123), TOP3B (Santa Cruz, sc-137238), 4. Ferdinandusse S, Denis S, van Roermund CWT, Preece MA, Koster J, Ebberink MS, anti-FH (4567, CST), anti-G6PD (12263, CST), anti-HK2 (2867, CST), anti-Flag et al. A novel case of ACOX2 deficiency leads to recognition of a third human (CST, 8146), and anti-Myc (CST, 2278). peroxisomal acyl-CoA oxidase. Biochim Biophys Acta Mol Basis Dis. 2018;1864:952–8. 5. Raas Q, Saih FE, Gondcaille C, Trompier D, Hamon Y, Leoni V, et al. A microglial Statistics cell model for acyl-CoA oxidase 1 deficiency. Biochim Biophys Acta Mol Cell Biol Data are presented as individual samples and mean ± SEM. Student’s two- Lipids. 2019;1864:567–76. tailed unpaired t-test or one-way ANOVA was used to determine statistical Cell Death and Disease (2022)13:279 Y. Zhang et al. 13 6. Vanhove GF, Van Veldhoven PP, Fransen M, Denis S, Eyssen HJ, Wanders RJ, et al. 31. Antoniou A, Raynaud P, Cordi S, Zong Y, Tronche F, Stanger BZ, et al. Intrahepatic The CoA esters of 2-methyl-branched chain fatty acids and of the bile acid bile ducts develop according to a new mode of tubulogenesis regulated by the intermediates di- and trihydroxycoprostanic acids are oxidized by one single transcription factor SOX9. Gastroenterology. 2009;136:2325–33. peroxisomal branched chain acyl-CoA oxidase in human liver and kidney. J Biol 32. Tulasi DY, Castaneda DM, Wager K, Hogan CB, Alcedo KP, Raab JR, et al. Sox9EGFP Chem. 1993;268:10335–44. defines biliary epithelial heterogeneity downstream of yap activity. Cell Mol 7. Ferdinandusse S, Denis S, Overmars H, Van Eeckhoudt L, Van Veldhoven PP, Gastroenterol Hepatol. 2021;11:1437–62. Duran M, et al. Developmental changes of bile acid composition and conjugation 33. Paku S, Dezso K, Kopper L, Nagy P. Immunohistochemical analysis of cytokeratin in L- and D-bifunctional protein single and double knockout mice. J Biol Chem. 7 expression in resting and proliferating biliary structures of rat liver. Hepatology. 2005;280:18658–66. 2005;42:863–70. 8. Martin GG, Atshaves BP, Landrock KK, Landrock D, Schroeder F, Kier AB. Loss of L- 34. Yang Y, McBride KM, Hensley S, Lu Y, Chedin F, Bedford MT. Arginine methylation FABP, SCP-2/SCP-x, or both induces hepatic lipid accumulation in female mice. facilitates the recruitment of TOP3B to chromatin to prevent R loop accumula- Arch Biochem Biophys. 2015;580:41–9. tion. Mol Cell. 2014;53:484–97. 9. Le Borgne F, Ben Mohamed A, Logerot M, Garnier E, Demarquoy J. Changes in 35. Zhang T, Wallis M, Petrovic V, Challis J, Kalitsis P, Hudson DF. Loss of TOP3B leads carnitine octanoyltransferase activity induce alteration in fatty acid metabolism. to increased R-loop formation and genome instability. Open Biol. 2019;9:190222. Biochem Biophys Res Commun. 2011;409:699–704. 36. Burma S, Chen BP, Murphy M, Kurimasa A, Chen DJ. ATM phosphorylates histone 10. Tan M, Luo H, Lee S, Jin F, Yang JS, Montellier E, et al. Identification of 67 histone H2AX in response to DNA double-strand breaks. J Biol Chem. 2001;276:42462–7. marks and histone lysine crotonylation as a new type of histone modification. 37. Smith J, Tho LM, Xu N, Gillespie DA. The ATM-Chk2 and ATR-Chk1 pathways in Cell 2011;146:1016–28. DNA damage signaling and cancer. Adv Cancer Res. 2010;108:73–112. 11. Jiang G, Li C, Lu M, Lu K, Li H. Protein lysine crotonylation: past, present, per- 38. Collins PL, Purman C, Porter SI, Nganga V, Saini A, Hayer KE, et al. DNA double- spective. Cell Death Dis. 2021;12:703. strand breaks induce H2Ax phosphorylation domains in a contact-dependent 12. Yu H, Bu C, Liu Y, Gong T, Liu X, Liu S, et al. Global crotonylome reveals CDYL- manner. Nat Commun. 2020;11:3158. regulated RPA1 crotonylation in homologous recombination-mediated DNA 39. Zhang Q, Zhang Y, Sun S, Wang K, Qian J, Cui Z, et al. ACOX2 is a prognostic repair. Sci Adv. 2020;6:eaay4697. marker and impedes the progression of hepatocellular carcinoma via PPARα 13. Fang Y, Xu X, Ding J, Yang L, Doan MT, Karmaus PWF, et al. Histone crotonylation pathway. Cell Death Dis. 2021;12:15. promotes mesoendodermal commitment of human embryonic stem cells. Cell 40. Menzies KJ, Zhang H, Katsyuba E, Auwerx J. Protein acetylation in metabolism - Stem Cell. 2021;28:748–63.e7. metabolites and cofactors. Nat Rev Endocrinol. 2016;12:43–60. 14. Tang X, Chen XF, Sun X, Xu P, Zhao X, Tong Y, et al. Short-chain enoyl-CoA 41. Popovic D, Vucic D, Dikic I. Ubiquitination in disease pathogenesis and treatment. hydratase mediates histone crotonylation and contributes to cardiac home- Nat Med. 2014;20:1242–53. ostasis. Circulation. 2021;143:1066–9. 42. Pagès F, Galon J, Dieu-Nosjean MC, Tartour E, Sautès-Fridman C, Fridman WH. 15. Sabari BR, Zhang D, Allis CD, Zhao Y. Metabolic regulation of gene expression Immune infiltration in human tumors: a prognostic factor that should not be through histone acylations. Nat Rev Mol Cell Biol. 2017;18:90–101. ignored. Oncogene. 2010;29:1093–102. 16. Sabari BR, Tang Z, Huang H, Yong-Gonzalez V, Molina H, Kong HE, et al. Intra- 43. Shi JY, Gao Q, Wang ZC, Zhou J, Wang XY, Min ZH, et al. Margin-infiltrating CD20 cellular crotonyl-CoA stimulates transcription through p300-catalyzed histone (+) B cells display an atypical memory phenotype and correlate with favorable crotonylation. Mol Cell. 2015;58:203–15. prognosis in hepatocellular carcinoma. Clin Cancer Res. 2013;19:5994–6005. 17. Fellows R, Denizot J, Stellato C, Cuomo A, Jain P, Stoyanova E, et al. Microbiota 44. Schoenberg MB, Hao J, Bucher JN, Miksch RC, Anger HJW, Mayer B, et al. Peri- derived short chain fatty acids promote histone crotonylation in the colon vascular tumor-infiltrating leukocyte scoring for prognosis of resected hepato- through histone deacetylases. Nat Commun. 2018;9:105. cellular carcinoma patients. Cancers (Basel). 2018;10:389. 18. Xu W, Wan J, Zhan J, Li X, He H, Shi Z, et al. Global profiling of crotonylation on 45. Garnelo M, Tan A, Her Z, Yeong J, Lim CJ, Chen J, et al. Interaction between non-histone proteins. Cell Res. 2017;27:946–9. tumour-infiltrating B cells and T cells controls the progression of hepatocellular 19. Zhou X, Wang H. ACOX2 deficiency in primary malignant cardiac tumors. Proc carcinoma. Gut. 2017;66:342–51. Natl Acad Sci USA. 2017;114:E3590–E3591. 46. Cliff TS, Wu T, Boward BR, Yin A, Yin H, Glushka JN, et al. MYC controls human 20. Zhou X, Xu M, Zeng W, Chen Z, Lu G, Gong Y, et al. Combined effects of FH pluripotent stem cell fate decisions through regulation of metabolic flux. Cell (E404D) and ACOX2 (R409H) cause metabolic defects in primary cardiac malig- Stem Cell. 2017;21:502–16. e9 nant tumor. Cell Death Discov. 2018;4:18. 47. Gu W, Gaeta X, Sahakyan A, Chan AB, Hong CS, Kim R, et al. Glycolytic metabolism 21. Vilarinho S, Sari S, Mazzacuva F, Bilgüvar K, Esendagli-Yilmaz G, Jain D, et al. plays a functional role in regulating human pluripotent stem cell state. Cell Stem ACOX2 deficiency: A disorder of bile acid synthesis with transaminase elevation, Cell. 2016;19:476–90. liver fibrosis, ataxia, and cognitive impairment. Proc Natl Acad Sci USA. 48. Koppenol WH, Bounds PL, Dang CV. Otto Warburg’s contributions to current 2016;113:11289–93. concepts of cancer metabolism. Nat Rev Cancer. 2011;11:325–37. 22. Monte MJ, Alonso-Peña M, Briz O, Herraez E, Berasain C, Argemi J, et al. ACOX2 49. Zhao S, Xu W, Jiang W, Yu W, Lin Y, Zhang T, et al. Regulation of cellular meta- deficiency: An inborn error of bile acid synthesis identified in an adolescent with bolism by protein lysine acetylation. Science. 2010;327:1000–4. persistent hypertransaminasemia. J Hepatol. 2017;66:581–8. 50. Okui T, Iwashita M, Rogers MA, Halu A, Atkins SK, Kuraoka S, et al. CROT (carnitine 23. Zhang Y, Lu Z, Zeng W, Zhao J, Zhou X. Two sides of NNMT in alcoholic and O-octanoyltransferase) is a novel contributing factor in vascular calcification via nonalcoholic fatty liver development. J Hepatol. 2021;74:1250–3. promoting fatty acid metabolism and mitochondrial dysfunction. Arterioscler 24. Sperber H, Mathieu J, Wang Y, Ferreccio A, Hesson J, Xu Z, et al. The metabolome Thromb Vasc Biol. 2021;41:755–68. regulates the epigenetic landscape during naive-to-primed human embryonic 51. Matsuda Y, Wakai T, Kubota M, Osawa M, Takamura M, Yamagiwa S, et al. DNA stem cell transition. Nat Cell Biol. 2015;17:1523–35. damage sensor γ-H2AX is increased in preneoplastic lesions of hepatocellular 25. Eckert MA, Coscia F, Chryplewicz A, Chang JW, Hernandez KM, Pan S, et al. carcinoma. ScientificWorldJournal. 2013;2013:597095. Proteomics reveals NNMT as a master metabolic regulator of cancer-associated 52. Saha J, Bae J, Wang SY, Lu H, Chappell LJ, Gopal P, et al. Ablating putative fibroblasts. Nature. 2019;569:723–8. Ku70 phosphorylation sites results in defective DNA damage repair and 26. Feldman GL, Wolf B. Deficient acetyl CoA carboxylase activity in multiple car- spontaneous induction of hepatocellular carcinoma. Nucleic Acids Res. boxylase deficiency. Clin Chim Acta. 1981;111:147–51. 2021;49:9836–50. 27. Baumgartner MR, Dantas MF, Suormala T, Almashanu S, Giunta C, Friebel D, et al. 53. Lulli M, Del Coco L, Mello T, Sukowati C, Madiai S, Gragnani L, et al. DNA damage Isolated 3-methylcrotonyl-CoA carboxylase deficiency: evidence for an allele- response protein CHK2 regulates metabolism in liver cancer. Cancer Res. specific dominant negative effect and responsiveness to biotin therapy. Am J 2021;81:2861–73. Hum Genet. 2004;75:790–800. 54. Carloni V, Lulli M, Madiai S, Mello T, Hall A, Luong TV, et al. CHK2 overexpression 28. Wei W, Liu X, Chen J, Gao S, Lu L, Zhang H, et al. Class I histone deacetylases are and mislocalisation within mitotic structures enhances chromosomal instability major histone decrotonylases: evidence for critical and broad function of histone and hepatocellular carcinoma progression. Gut. 2018;67:348–61. crotonylation in transcription. Cell Res. 2017;27:898–915. 55. Wang X, Zeldin S, Shi H, Zhu C, Saito Y, Corey KE, et al. TAZ-induced Cybb 29. Huang H, Wang DL, Zhao Y. Quantitative crotonylome analysis expands the roles contributes to liver tumor formation in non-alcoholic steatohepatitis. J Hepatol. of p300 in the regulation of lysine crotonylation pathway. Proteomics. 2018;18: 2021. https://doi.org/10.1016/j.jhep.2021.11.031. e1700230. 56. Wang Z, Gong Y, Peng B, Shi R, Fan D, Zhao H, et al. MRE11 UFMylation promotes 30. Kelly RDW, Chandru A, Watson PJ, Song Y, Blades M, Robertson NS, et al. Histone ATM activation. Nucleic Acids Res. 2019;47:4124–35. deacetylase (HDAC) 1 and 2 complexes regulate both histone acetylation and 57. Lavin MF. ATM and the Mre11 complex combine to recognize and signal DNA crotonylation in vivo. Sci Rep. 2018;8:14690. double-strand breaks. Oncogene 2007;26:7749–58. Cell Death and Disease (2022)13:279 Y. Zhang et al. 14 58. Robison JG, Elliott J, Dixon K, Oakley GG. Replication protein A and the Mre11. COMPETING INTERESTS Rad50.Nbs1 complex co-localize and interact at sites of stalled replication forks. J The authors declare no competing interests. Biol Chem. 2004;279:34802–10. 59. Chen W, Zhang Y, Shen L, Zhu J, Cai K, Lu Z, et al. Biallelic DNAH9 mutations are identified in Chinese patients with defective left-right patterning and cilia-related ADDITIONAL INFORMATION complex congenital heart disease. Hum Genet. 2022. https://doi.org/10.1007/ Supplementary information The online version contains supplementary material s00439-021-02426-5. Epub ahead of print. available at https://doi.org/10.1038/s41419-022-04725-9. 60. Wang D, Wang F, Shi KH, Tao H, Li Y, Zhao R, et al. Lower circulating folate induced by a fidgetin intronic variant is associated with reduced congenital heart Correspondence and requests for materials should be addressed to Liping Jin, disease susceptibility. Circulation. 2017;135:1733–48. Jianyuan Zhao or Xiangyu Zhou. 61. Wang Z, Yang X, Liu C, Li X, Zhang B, Wang B, et al. Acetylation of PHF5A modulates stress responses and colorectal carcinogenesis through alternative Reprints and permission information is available at http://www.nature.com/ splicing-mediated upregulation of KDM3A. Mol Cell. 2019;74:1250–63. e6 reprints 62. Zhang Y, Chen W, Zeng W, Lu Z, Zhou X. Biallelic loss of function NEK3 mutations deacetylate α-tubulin and downregulate NUP205 that predispose individuals to Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims cilia-related abnormal cardiac left-right patterning. Cell Death Dis. 2020;11:1005. in published maps and institutional affiliations. ACKNOWLEDGEMENTS This work was supported by grants from National Key Research and Development Open Access This article is licensed under a Creative Commons Project (2018YFA0801300), the National Natural Science Foundation of China Attribution 4.0 International License, which permits use, sharing, (81870285 and 82171845), Shanghai Rising-Star Program (18QA1403500), the adaptation, distribution and reproduction in any medium or format, as long as you give Shanghai Science and Technology International Exchange and Cooperation Fund appropriate credit to the original author(s) and the source, provide a link to the Creative (grant no.18410741900). Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly AUTHOR CONTRIBUTIONS from the copyright holder. To view a copy of this license, visit http://creativecommons. X.Z., J.Z., and L.J. conceived the concept, designed and supervised the experiments. X. org/licenses/by/4.0/. Z., Y.Z., Y.C., X.T, S.X., X.Z., and Z.L. performed the experiments. Z.Z. and T.Z. analyzed data. X.Z, J.Z., and J.R.B interpreted data. X.Z. wrote the manuscript. All authors approved the final version of the paper. © The Author(s) 2022 Cell Death and Disease (2022)13:279