The cellular acetyl-CoA levels play important roles in the sustenance of histone acetylation [
10]. To further delineate the effect of SEC63-mediated ACLY stabilization, we analyzed histone acetylation levels after SEC63 depletion. A significant decline in global acetylated levels of histones H3 was observed after SEC63 depletion. Moreover, this effect was not found after SEC63 ablation in ACLY-depleted cells (Fig.
6J). Thus, these findings indicate that upregulation of ACLY by SEC63 promotes acetylation of H3 through increasing cellular acetyl-CoA level.
ACLY-related histone acetylation participates in modulating the transcription of downstream genes. After ER stress, IRE1α is activated to induce the expression of XBP1s, which is an active transcription factor. XBP1s upregulates the expression of UPR target genes, such as ER chaperones (
HSPA5,
HSP90B1, and
CALR) and lipid metabolism (
SCD1,
DGAT2, and
FASN) [
39]. It’s unknown whether ACLY is involved in modulating UPR target genes. To address this issue, we analyzed the effect of SEC63 and ACLY on UPR target genes. As shown in Fig.
7D, depletion of SEC63 significantly reduced the expression of the target genes under ER stress conditions. Moreover, knockdown of ACLY abrogated this effect. Concurringly, H3 acetylation was decreased at the promoter region of these genes after SEC63 knockdown and this effect was abolished when ACLY was silenced (Fig.
7E). Together, these findings reveal that upregulation of ACLY by SEC63 specifically offers sufficient acetyl-CoA to alter the epigenetic profile for supporting UPR.
Intriguingly, we observed decreased H3 acetylation levels at the promoter region of
SNAI1 in SEC63 depleted cells compared to those in control cells (Fig.
7E), which is consistent with reduced expression of Snail1 in SEC63 depleted cells (Fig.
6E). These results indicate that SEC63 epigenetically modulates expression of Snail1. To elucidate the underlying mechanism, we identified the SEC63-interacting protein upon ER stress. Pathway analysis revealed that SEC63 interacting proteins were mainly enriched in metabolic pathway, protein processing in ER, and adhesion-related proteins. Among the adhesion-related protein, SMAD3 is of interest to us due to its highest scores in mass spectrometry (Supplementary Fig. S
6C). Moreover, it has been found that SMAD3 promotes Snail1 transcription [
40]. SEC63 failed to regulate Snail1 expression in SMAD3-depleted cells (Fig.
7F and Supplementary Fig. S
6D), indicating that SEC63 upregulates Snail1 in SMAD3-dependent way. The acetylation of SMAD3 at K20 and K117 in nucleus is critical for enhancing SMAD3 mediated transcriptional activity [
41]. We hypothesized that SEC63 contributed to increase the acetylation of SMAD3 by stabilizing ACLY to offer acetyl-CoA. To test this possibility, we confirmed the binding between SEC63 and SMAD3 and the interaction was enhanced upon ER stress (Fig.
7G). Moreover, the acetylation of SMAD3 was dramatically increased upon ER stress and SEC63 knockdown decreased the acetylation of SMAD3 (Fig.
7H), implying the importance of SEC63 on acetylation of SMAD3 under ER stress. Previous studies show that IRE1α/STAT3 signaling increases phosphorylation of SMAD3 after ER stress [
42,
43], which is indispensable for nuclear translocation of SMAD3 [
44]. SEC63 had little effect on phosphorylation of SMAD3, suggesting that SEC63 mainly regulates SMAD3 in nucleus. Since lysine acetyltransferase 6A (KAT6A) is the critical acetyltransferase for SMAD3 [
41], we tested the roles of SEC63 or ACLY knockdown on the acetylation of SMAD3 when cells were treated with KAT6A inhibitor (WM1119). The results revealed that the knockdown of SEC63 or ACLY had no impact on the SMAD3 acetylation level after KAT6A was inhibited (Fig.
7I and Supplementary Fig. S
6E). Thus, SEC63 contributes to increase the acetylation of SMAD3 by binding to SMAD3 and stabilizing ACLY to offer acetyl-CoA. Furthermore, SEC63 knockdown reduced the interaction of SMAD3 with the promoter of
SNAI1 (Fig.
7J). After the construction of acetylation-defective SMAD3 2KR (K20R and K117R, mutation lysine to arginine) mutant and acetylation-mimic SMAD3 2KQ (K20Q and K117Q, mutation lysine to glutamine) mutant, we observed 2KR mutant decreased the binding of SMAD3 to the promoter of
SNAI1 and 2KQ increased the binding. Importantly, SEC63 only affected the binding of wild-type SMAD3 to the promoter of
SNAI1 (Fig.
7K), suggesting that SEC63-associated acetylation of SMAD3 modulates SMAD3-mediated transcription of
SNAI1. Together, these data suggest that SEC63-associated acetylation of SMAD3 epigenetically promotes the expression of Snail1 to facilitate HCC metastasis in the nucleus upon ER stress.
Next, we analyzed the relations of SEC63, ACLY and SMAD3 in HCC tissues. The expression of SEC63 is positively associated with ACLY or SMAD3 expression (Supplementary Fig. S
6F). Co-upregulation of SEC63/ACLY/SMAD3 expression predicted poorer prognosis compared to the rest patients in TCGA-LIHC cohort (Supplementary Fig S
6G). Data analysis from ICGC-LIRI-JP cohort also suggested similar phenomena (Supplementary Fig. S
6H). These findings indicate that the metastasis caused by SEC63/ACLY/SMAD3 interactions in the nucleus is critical for cancer cell adaption to ER stress as well as contributes to the poor outcome of HCC patients.