The splicing regulator SLU7 is required to preserve DNMT1 protein stability and DNA methylation

Abstract

Gene expression is finely and dynamically controlled through the tightly coordinated and interconnected activity of epigenetic modulators, transcription and splicing factors and post-translational modifiers. We have recently identified the splicing factor SLU7 as essential for maintaining liver cell identity and genome integrity and for securing cell division both trough transcriptional and splicing mechanisms. Now we uncover a new function of SLU7 controlling gene expression at the epigenetic level. We show that SLU7 is required to secure DNMT1 protein stability and a correct DNA methylation. We demonstrate that SLU7 is part in the chromatome of the protein complex implicated in DNA methylation maintenance interacting with and controlling the integrity of DNMT1, its adaptor protein UHRF1 and the histone methyl-transferase G9a at the chromatin level. Mechanistically, we found that SLU7 assures DNMT1 stability preventing its acetylation and degradation by facilitating its interaction with HDAC1 and the desubiquitinase USP7. Importantly, we demonstrate that this DNMT1 dependency on SLU7 occurs in a large panel of proliferating cell lines of different origins and in in vivo models of liver proliferation. Overall, our results uncover a novel and non-redundant role of SLU7 in DNA methylation and present SLU7 as a holistic regulator of gene expression.

Graphical Abstract

SLU7 is required to maintain DNMT1 stability and secure global DNA methylation.

Figure 1.

SLU7 knockdown compromises DNA methylation. (A) SLU7 knockdown induces a decrease in 5-methyl-cytosine (5-mC) content. DNA methylation levels were analyzed by ELISA 72 h after transfection with siSLU7 or control siGL in the human hepatoma cell lines PLC/PRF/5 and HepG2, the colorectal carcinoma cell line HCT116 and the lung cancer cell line H358. The percentage of 5-mC content is indicated. *P < 0.05, **P < 0.01, ***P < 0.001. Western blot analysis confirms SLU7 knockdown. ACTIN was used as loading control. At least three independent experiments in duplicate were performed for each cell line. (B) Representative images of 5-methyl cytosine (5-mC) immunodetection (in green) in PLC/PRF/5 (top) and HepG2 (bottom) cells 72 h after transfection with siSLU7 or control siGL. Nuclei were stained with DAPI (in blue). Scale bar: 10 μm. (C, D) Scatter plots of reduced representation bisulfite sequencing (RRBS) DNA methylation β-values of siGL versus siDNMT1 (C) or siGL versus siSLU7 (D) PLC/PRF/5 cells 72 h after transfection. Fold change methylation differences higher than 0.6 were highlighted within a green triangle (hypomethylated) or a red triangle (hypermethylated). Graphs represent the percentage of hypomethylated (FC < 0,6) and hypermethylated (FC > 0,6) CpGs. (E) Number of single CpGs, genes with CpGs and CpGs within CpG islands (CGIs) found hypomethylated (FC < 0.6) in PLC/PRF/5 cells 72 h after transfection with siDNMT1 (left) or siSLU7 (right) compared to control siGL cells. (F) Venn diagram of the overlap between genes (top) or CpGs (bottom) found hypomethylated (FC < 0,6) in siSLU7 and siDNMT1 PLC/PRF/5 transfected cells when compared with control siGL cells.

Figure 2.

SLU7 knockdown activates transcription of genes silenced by DNA methylation. (A) Graphs reporting DNA methylation levels (β-values obtained by Reduced Representation Bisulfite Sequencing (RRBS) analysis) of TRIM47, TFF3 and DLG3 gene regions found hypomethylated in both SLU7 (siSLU7, in red) and DNMT1 (siDNMT1, in green) knockdown PLC/PRF/5 cells compared to control siGL (in blue) cells 72 h after transfection. (B) Representative methylation-specific PCR (MSP) results for DNA methylation analysis of TRIM47, TFF3 and DLG3 regions shown in (A). U: unmethylated and M: methylated. (C) Expression by RT-qPCR of TRIM47 and DLG3 in PLC/PRF/5 cells 72 h after transfection with siSLU7 or control siGL. (D, E) DNA methylation analysis by MSP of RASSF1A and GNMT promoter regions (D) and RASSF1A and GNMT expression by RT-qPCR (E) in PLC/PRF/5 cells 72 h after transfection with siSLU7, siDNMT1 or control siGL. U: unmethylated and M: methylated. (F, G) DNA methylation analysis by MSP of GNMT promoter region (F) and GNMT expression by RT-qPCR (G) in HepG2 and HCT-116 cells 72 h after transfection with siSLU7, siDNMT1 or control siGL. U: unmethylated and M: methylated. (H) Expression by RT-qPCR of cancer testis antigens (MAGE-A1 and MAGE-A7), endogenous retroviruses (envFc2) and imprinted genes (H19) in PLC/PRF/5, HepG2, HCT116 and H358 cells 72 h after transfection with siSLU7 or control siGL. RPLP0 expression was used as housekeeping gene in (C, E, G, H). *P < 0.05, **P < 0.01, ***P < 0.001. All experiments were performed at least three times with biological duplicates per condition.

Figure 3.

SLU7 knockdown decreases DNMT1 protein levels in proliferating cells. (A) Representative Western blots (top) of DNMT1 and SLU7 in PLC/PRF/5 cells 24, 48 and 72 h after transfection with siSLU7 or control siGL. ACTIN expression was shown as loading control. SLU7 and DNMT1 protein levels were quantified using Image J software from at least three independent experiments with biological duplicates per condition (bottom). (B) Histograms reporting SLU7 and DNMT1 mRNA levels by real time PCR in the same experiments analyzed in panel A. RPLP0 expression was used as housekeeping gene in (B). *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 4.

SLU7 is required for DNMT1 protein stability in a wide panel of cell lines from different origins. (A) Western blots of DNMT1 and SLU7 in HepG2, H358, A549, HCT116, HeLa, MCF7, U87 and A375 cancer cell lines 48h after transfection with siSLU7 or control siGL. ACTIN expression was shown as loading control. (B, C) pcDNA6.1A empty plasmid (pC) or the corresponding plasmid to express SLU7 (pSLU7) were transiently transfected for 48 h in PLC/PRF/5, HepG2 and HCT116 cells followed by Western blot (B) and real time PCR (C) analyses of DNMT1 and SLU7. ACTIN expression was shown as loading control in (B). RPLP0 expression was used as housekeeping gene in (C). (D) PLC/PRF/5 and HEK293T cells transfected with siGL or siSLU7 for 48 h and treated with 10 μM MG-132 for the last 20 h were harvested to analyze DNMT1 and SLU7 expression by Western blot. HCT116 cells were transfected with siGL or siSLU7 for 36 h and treated with 20 μM MG-132 for the last 12 h before harvesting. The expression of ACTIN was analyzed as loading control. DNMT1 protein levels were quantified using ImageJ software from at least three independent experiments with biological duplicates per condition (bottom). *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 5.

SLU7 is part of the methylation complex DNMT1/UHRF1/G9a and it is required for its stability on the chromatin. (A) SLU7 and DNMT1 were coimmunoprecipitated with each other in PLC/PRF/5, HCT116, MCF7, H358 and HEK293T cells. Pre-immune IgG immunoprecipitation and Inputs were shown as control. (B) Chromatin-bound cell extracts from PLC/PRF/5 cells were immunoprecipitated with anti-SLU7 and anti-DNMT1 antibodies or pre-immune IgG and blotted with DNMT1, SLU7, G9a and UHRF1 antibodies. Western blot on inputs from the different subcellular fractions were shown as control. (C) Chromatin extract for PLC/PRF/5 cells 48 h after transfection with siSLU7 were analyzed by western blot with antibodies against DNMT1, SLU7, G9a and UHRF1. HISTONE 3 and PCNA were used as loading control. (D) DNMT1, SLU7, G9a and UHRF1 protein levels were analyzed by western blot in PLC/PRF/5 and HCT116 cells 48 h after transfection with control siGL, siSLU7 siDNMT1, siG9a and siUHRF1. ACTIN was used as loading control.

Figure 6.

SLU7 knockdown induces DNMT1 acetylation and impairs HDAC1-DNMT1 interaction. (A) Representative Western blots of DNMT1 and SLU7 in HCT116 DNMT1^–/– cells 24 and 48 h after transfection with siSLU7. ACTIN expression was shown as loading control. (B) Histograms reporting SLU7 and DNMT1 mRNA levels by real time qPCR from three independent experiments including those shown in panel A. RPLP0 expression was used as housekeeping gene. *P < 0.05, **P < 0.01. (C) SLU7 knockdown leads to increase acetylation of endogenous DNMT1. Protein lysates from 36h control siGL and siSLU7 transfected PLC/PRF/5 and HEK293T cells treated with 10 μM MG-132 for the last 20 h were immunoprecipitated with anti-DNMT1 and blotted with anti-acetyl-lysine (Ac-K), DNMT1 and USP7 antibodies. DNMT1, USP7 and ACTIN are shown as control in the cell inputs. (D) DNMT1 protein decrease was detected by Western blot analysis in PLC/PRF/5 cells 48 h after SLU7 and HDAC1 knockdown. (E) DNMT1, SLU7 and HDAC1 protein levels were analyzed by Western blot in PLC/PRF/5 cells 36 h after transfection with control siGL or siSLU7 and pcDNA3.1 empty plasmid or the corresponding plasmid to express HDAC1-flag. ACTIN was used as loading control. (F) Chromatin-bound cell extracts from PLC/PRF/5 cells were immunoprecipitated with anti-SLU7 and anti-DNMT1 antibodies or pre-immune IgG and blotted with DNMT1, SLU7 and HDAC1 antibodies. (G) Extracts from PLC/PRF/5 cells 36 h after siGL and siSLU7 transfection were subjected to DNMT1 immunoprecipitation and tested by Western blot with antibodies against DNMT1, SLU7 and HDAC1. Immunoprecipitation with pre-immune IgG and inputs were used as controls.

Figure 7.

SLU7 is required to ensure DNMT1 stability in vivo. (A) Slu7 and Dnmt1 mRNA levels in the liver of mice 24, 48 and 72 h after 2/3 partial hepatectomy (PH) or control surgery (SHAM) 21 days after injection of adenoassociated virus to inhibit liver SLU7 expression (AAV-shSLU7) or control virus (AAV-Ren) (n = 5 mice per group). RPLP0 expression was used as housekeeping gene. (B) Western blot detection of DNMT1 and SLU7 protein levels in the liver of AAV-shSLU7 and AAV-Ren mice Sham operated or 48 and 72 h after 2/3 partial hepatectomy (PH). ACTIN expression was shown as loading control. (C) Western blot detection of DNMT1 and SLU7 protein levels in the liver of AAV-shSLU7 and AAV-Ren mice after CCl4 induction of acute or chronic liver damage. GAPDH expression was shown as loading control. (D) SLU7 immunoprecipitation in vivo after 2/3 partial hepatectomy (PH) and CCl₄-mediated induction of acute or chronic liver damage. Protein lysates were immunoblotted with anti-SLU7, anti-DNMT1 and anti-HDAC1 antibodies. Immunoprecipitation with pre-immune IgG and inputs were shown as control.

Figure 8.

Schematic representation of the mechanisms regulated by SLU7 involved in DNA methylation maintenance. See text for details.