Perm1 regulates cardiac energetics as a downstream target of the histone methyltransferase Smyd1

Abstract

The transcriptional regulatory machinery in mitochondrial bioenergetics is complex and is still not completely understood. We previously demonstrated that the histone methyltransferase Smyd1 regulates mitochondrial energetics. Here, we identified Perm1 (PPARGC-1 and ESRR-induced regulator, muscle specific 1) as a downstream target of Smyd1 through RNA-seq. Chromatin immunoprecipitation assay showed that Smyd1 directly interacts with the promoter of Perm1 in the mouse heart, and this interaction was significantly reduced in mouse hearts failing due to pressure overload for 4 weeks, where Perm1 was downregulated (24.4 ± 5.9% of sham, p<0.05). Similarly, the Perm1 protein level was significantly decreased in patients with advanced heart failure (55.2 ± 13.1% of donors, p<0.05). Phenylephrine (PE)-induced hypertrophic stress in cardiomyocytes also led to downregulation of Perm1 (55.7 ± 5.7% of control, p<0.05), and adenovirus-mediated overexpression of Perm1 rescued PE-induced downregulation of estrogen-related receptor alpha (ERRα), a key transcriptional regulator of mitochondrial energetics, and its target gene, Ndufv1 (Complex I). Pathway enrichment analysis of cardiomyocytes in which Perm1 was knocked-down by siRNA (siPerm1), revealed that the most downregulated pathway was metabolism. Cell stress tests using the Seahorse XF analyzer showed that basal respiration and ATP production were significantly reduced in siPerm1 cardiomyocytes (40.7% and 23.6% of scrambled-siRNA, respectively, both p<0.05). Luciferase reporter gene assay further revealed that Perm1 dose-dependently increased the promoter activity of the ERRα gene and known target of ERRα, Ndufv1 (Complex I). Overall, our study demonstrates that Perm1 is an essential regulator of cardiac energetics through ERRα, as part of the Smyd1 regulatory network.

Fig 1. Perm1 is a downstream target of Smyd1.

(A-B) RNA-seq of siSmyd1 cardiomyocytes. A volcano plot shows significantly downregulated (green) and upregulated (red) genes. Perm1, indicated by a yellow circle, is one of the genes most significantly downregulated by siSmyd1 (Panel A). Box plots show the decreased mRNA levels of Perm1 in cardiomyocytes in which Smyd1 expression was significantly reduced by siSmyd1 as compared with control cells that were treated with scrambled-siRNA (scr-siRNA) (n = 4/group, Panel B). (C) qRT-PCR showing a significant reduction in Perm1 in cardiomyocytes in which Smyd1 was knocked-down by siRNA (n = 6/group). Statistics were performed using the t-test with 2 tails. *p<0.05. (D) Previous ChIP-seq studies have established enrichment of H3K4me3 within the promoter region of the Perm1 locus (WashU EpiGenome Database), which was targeted for qPCR reaction (“Region 2”). As a negative control, the ~2k bp upstream of the transcription start site (TSS) was targeted (“Region 1”), where small H3K4me3 peaks were observed. (E) ChIP-PCR showing no significant interaction of Smyd1 with Region 1 (left), while Smyd1 does bind to Region 2 in the hearts of WT mice. This interaction was significantly reduced in the TAC heart (TAC +) (right) (n = 4 in sham, n = 4 in TAC). Statistics were performed using the t-test with 2 tails. *: p<0.05. Error bars are ±SEM.

Fig 2. Perm1 is downregulated in the failing heart.

(A-C) Downregulation of Perm1 and Smyd1 in mouse hearts subjected to TAC surgery for 4 weeks, where left ventricular ejection fraction (LVEF) and fractional shortening (LVFS) were significantly decreased as compared with sham-operated hearts (n = 4/group for Perm1 expression, n = 8-10/group for Smyd1 expression) (also see S1 Table). Consistent with previous studies in the mouse skeletal muscle [10, 25, 26], Western blotting analysis shows two isoforms of Perm1 at 90 kDa and 100 kDa in the mouse heart (Panel B). (D-F) Downregulation of Perm1 and Smyd1 in patients with heart failure, where LVEF was significantly decreases as compared with donor hearts (n = 7 in donor group, n = 9–10 in heart failure patients, also see S2 Table). Statistics were performed using the t-test with 2 tails. *: p<0.05. Error bars are ±SEM.

Fig 3. Perm1 positively regulates metabolism in cardiomyocytes.

(A-B) Western blot analysis showing a significant reduction in the endogenous protein levels of Perm1, ERRα, and PGC-1α in NRVMs that were transfected with siPerm1, as compared with control cells that were treated with scrambled-siRNA (scr-siRNA) (n = 4/group). (C-F) Illumina-based RNA-seq showing that differentially expressed genes in siPerm1 NRVMs as compared with control include 801 upregulated genes and 736 downregulated genes (n = 4/group, p<0.05, Panel C, also see S2 Dataset). Enrichment analysis of significantly regulated genes for the Reactome terms biological pathways shows that the pathway most affected by Perm1 knockout is metabolism (Panel D, S3 Dataset). Pie charts representing the major metabolic pathways of differentially regulated genes involved in metabolism in siPerm1 cardiomyocytes (Panel E, S4 Dataset). Heat maps of genes comprising the pie chart presented in Panel E display the dictionary of changes by Perm1 knockout (Panel F). (a) Biosynthesis of amino acid, (b) Sphingolipid metabolism and signaling pathway, (c) Carbon metabolism and glycolysis, (d) Fatty acid metabolism, (e) Glutathione metabolism, (f) Glucosaminoglycan biosynthesis, (g) Phosphatidylinositol signaling system, (h) TCA cycle / electron transport chain (ETC). (G) qRT-PCR showing the expression of genes involved in energetics in Perm1-knockdown (siPerm1) led to downregulation of energetics-related genes as compared to in control cells (n = 3-5/group, ±SEM). (H) Cell Mito Stress Test was performed using a Seahorse XF96 analyzer (n = 8/group, ±SD). O₂ consumption rate (OCR) was measured in the presence of glucose as a major respiration substrate. (I) Results were normalized by nuclear staining as described in our publication [2]. Error bars are ±SEM. Statistics were performed using the t-test with 2 tails. *p<0.05.

Fig 4. Adenovirus-mediated overexpression of Perm1 (Ad-Perm1) enhances energetics in cardiomyocytes, partially mediated by Smyd1.

(A) qRT-PCR showing the expression of genes involved in energetics (n = 4/group). Statistics were performed using the t-test with 2 tails. *p<0.05. (B) Cell Mito Stress Test was performed using a Seahorse XF96 analyzer in NRVMs (n = 8/group, ±SD). O₂ consumption rate (OCR) was measured in the presence of glucose as a major respiration substrate. (C) Results were normalized by nuclear staining as described in our publication [2]. Error bars are ±SEM. Statistics were performed using the t-test with 2 tails. *p<0.05, (D) A Venn diagram showing the genes downregulated in both siSmyd1 and siPerm1 NRVMs that were detected in RNA-seq (top). The overlaping genes were further analyzed by KEGG pathway analysis (bottom). (E) qRT-PCR shows that siSmyd1 downregulated CPT2 and Ndufa8, both of which were completely rescued by Perm1overexpression (Ad-Perm1). In contrast, Ndufv1 expression was not affected by siSmyd1, suggesting that Ndufv1 is not a Smyd1 target gene. Ndufv1 was upregulated by Ad-Perm1 in the presene and absence of siSmyd1 (n = 3-5/group, t-test). *: p<0.05 compared with control, #: p<0.05 compared with siSmyd1. Statistics were performed using one-way ANOVA. Error bars are ±SEM.

Fig 5. Perm1 is downregulated in cardiomyocytes under cellular hypertrophic stress.

(A) RT-PCR showing Perm1 downregulation in NRVMs treated with phenylephrine (PE) (n = 3/group *:p<0.05 by one-way ANOVA). (B) Immunostaining of NRVMs showing the localization of Perm1 in the nucleus and the peripheral nuclear region (Top). After incubation with PE for 24hr, Perm1 was less localized in the nucleus and showed a diffuse expression pattern. Cardiac muscle-specific anti-troponinC (cTnc) was used to distinguish cardiomyocytes from fibroblasts. Scale bar: 20 μm. (n = 3/group). (C-F) RT-PCR (Panel C) and Western blot analysis (Panels D-F) showing that Perm1 overexpression prior to 48h incubation with PE (group Perm1+/ PE+) either completely or partially rescues downregulation of some metabolic genes during PE-induced hypertrophic stress. The control (group Perm1-/PE-) was obtained by treating NRVMs with adenovirus-null for 48h without incubating with PE. The data for group Perm1-/PE- and group Perm1+/PE- in Panel C is the same as shown in Fig 4A. Note that the expression levels of the Complex II and III subunits were quantified using individual antibodies againt SDHB and UQCRC2 in separate blots due to being adjacent to the band from the relatively highly abundant Complex IV subunit. Statistics were performed using one-way ANOVA. *: p<0.05 compared with Perm1-/PE- (control); §: p<0.05 compared with Perm1+/PE+; #: p<0.05 compared with Perm1-/PE-. Error bars are ±SEM.

Fig 6. Perm1 activates the promoters of ERRα and its target gene through the ERRE.

(A) Luciferase reporter assays showing that Perm1 increases ERRE activity (TCAAGGTCA) in a dose-dependent manner, whereas mutant ERRE (mERRE, TCAGAATCA) is not be activated by Perm1. (B) Perm1 activates the ERRE within the Ndufv1 promoter (a subunit of Complex I, an ERRα target gene) in a dose-dependent manner. Note that the ERRE in mouse Ndufv1 promoter is TGAAGGTGA. (C) Perm1 does not increase the promoter activity of PGC-1α but rather represses the gene (n = 5/group). Statistics were performed using the t-test with 2 tails (*p<0.05). Error bars are ±SEM.