Direct binding of CoREST1 to SUMO-2/3 contributes to gene-specific repression by the LSD1/CoREST1/HDAC complex

Abstract

Posttranslational modification of transcription factors by the small ubiquitin-related modifier SUMO is associated with transcriptional repression, but the underlying mechanisms remain incompletely described. We have identified binding of the LSD1/CoREST1/HDAC corepressor complex to SUMO-2. Here we show that CoREST1 binds directly and noncovalently to SUMO-2, but not SUMO-1, and CoREST1 bridges binding of the histone demethylase LSD1 to SUMO-2. Depletion of SUMO-2/3 conjugates led to transcriptional derepression, reduced occupancy of CoREST1 and LSD1, and changes in histone methylation and acetylation at some, but not all, LSD1/CoREST1/HDAC target genes. We have identified a nonconsensus SUMO-interaction motif (SIM) in CoREST1 required for SUMO-2 binding, and we show that mutation of the CoREST1 SIM disrupted SUMO-2 binding and transcriptional repression of some neuronal-specific genes in nonneuronal cells. Our results reveal that direct interactions between CoREST1 and SUMO-2 mediate SUMO-dependent changes in chromatin structure and transcription that are important for cell-type-specific gene expression.

Figure 1

CoREST1 binds to SUMO-2 (A) CoREST1 binds to SUMO-2 directly and non-covalently in vitro. Purified recombinant (His)₆-CoREST1 was mixed with GST, GST-SUMO-1, GST-SUMO-2, or GST-SUMO-2-K2A (K33A, K35A). Proteins retained on Glutathione Sepharose beads were visualized using anti-6xHis antibody to detect (His)₆-CoREST1 or Coomassie staining to detect GST or GST tagged SUMO variants. (B) Cell lysates from HEK293T cells transfected with 3xFLAG-CoREST1 or 3xFLAG vector were immunoprecipitated with anti-FLAG Agarose beads. Bound proteins were subjected to Western Blot (WB) analyses with antisera to SUMO-1 or SUMO-2/3, as indicated. Anti-FLAG WB shows the expression of 3xFLAG-CoREST1 and the anti-β-Actin WB indicates equal loading. Asterisks indicate positions of IgG heavy and light chains. 4% of input material was loaded.

Figure 2

CoREST1 bridges LSD1 binding to SUMO-2 Purified recombinant (His)₆-LSD1 was incubated with GST (G), GST-SUMO-2 (S2) or GST-CoREST1 (C). Purified recombinant (His)₆-CoREST1 was included in the binding reaction where indicated (lanes 2 and 4). Proteins retained on Glutathione Sepharose beads were visualized using anti-6xHis antibody to detect (His)₆-LSD1 and (His)₆-CoREST1 or Coomassie staining to detect GST ,GST-SUMO-2 or GST-CoREST1. Input=10%.

Figure 3

Mapping the SUMO-2 interaction domain in CoREST1 (A) Summary of SUMO-2 binding by CoREST1 deletion mutants. Relative affinity of SUMO-2 binding by CoREST1 deletion mutants is indicated on the right based on in vitro protein binding assays in Figures 3B, 3C and Supplemental Figure 2. The positions of indicated CoREST1 domains are: ELM2, 100-164; first SANT domain, 188-236; first Coiled Coil (CC), 241-268; second CC, 331-366; second SANT domain, 379-427. (B) Representative binding of CoREST1 deletions to SUMO-2 in vitro. (His)₆ tagged CoREST1 full length (WT) or truncated proteins bearing the indicated amino acids were incubated with GST (G), GST-SUMO-2 (S2), or GST-LSD1 (L). Bound proteins were visualized by WB with anti-6xHis antibody (top) or Coomassie staining (bottom). Input=10%. (C) CoREST1 residues 255-275 are required for binding to SUMO-2. Deletions of CoREST1 acidic stretch (d255-263), linker (d264-270), hydrophobic core (d270-d275) or a combined deletion (d255-275) were assayed for binding to GST-SUMO-2 as described above.

Figure 4

An unusual SUMO-2 interaction motif in CoREST1 (A) CoREST3 binds to LSD1 but does not bind to SUMO-2 in vitro. Purified recombinant (His)₆-CoREST1 or (His)₆-CoREST3 were assayed for binding to GST (G), GST-SUMO-2 (S2), or GST-LSD1 (L) as described in Figure 3. (B) Sequence alignment of a region of CoREST1 (RCOR1) and CoREST3 (RCOR3) proteins from Homo sapiens (hs), Mus musculus (mm), and Xenopus tropicalis (xt). Acidic stretch and hydrophobic core of CoREST1 are boxed. CoREST1 hydrophobic residues (IIV) mutated in further assays are shaded. (C) Purified recombinant wild type CoREST1 (WT) or the indicated mutants (IIV to AAA or IIV to NSY) were assayed for binding to GST (G), GST-SUMO-2 (S2), or GST-LSD1 (L) as described in Figure 3. (D) Sequence alignment of the indicated regions of RCOR1, FIP1L1 and RBBP4. Hydrophobic residues mutated in further assays are shaded. (E) Purified recombinant (His)₆-FIP1L1, (His)₆-RBBP4, and control (His)₆-UBC9 were assayed for binding to GST (G), GST-SUMO-1 (S1), or GST-SUMO-2 (S2) as described in Figure 1. Proteins were visualized by WB using anti-FIP1L1 or anti-6xHis antibodies (top) or Coomassie staining (bottom). Input=2% for FIP1L1 and RBBP4; 10% for UBC9. (F) Purified recombinant (His)₆-FIP1L1 WT or AAA mutant (LVL to AAA), (His)₆-RBBP4 WT or AAA mutant (III to AAA) were assayed for binding to GST (G) or GST-SUMO-2 (S2) as described above. Input=2%.

Figure 5

Repression of some CoREST1 target genes is SUMO-2/3 dependent (A) CoREST1 RNAi reduces CoREST1 levels. Protein extracts prepared from HeLa cells stably expressing control scrambled RNAi (scRNAi) or CoREST1 RNAi were analyzed by WB with the indicated antibodies. (B) De-repression of neuronal specific genes by RNAi-mediated knockdown of CoREST1 or LSD1. RNA isolated from HeLa cells stably expressing control scRNAi, CoREST1 RNAi or LSD1 RNAi was subjected to RT-qPCR to examine the mRNAs levels of the CoREST1/LSD1 target genes SCN1A, SCN2A2 and SCN3A. Error bars of SEMs were generated from multiple experiments (n≥2). Asterisks denote statistically significant values relative to scRNAi (t test, **p<0.05, ***p<0.01). (C) SUMO-2/3 de-conjugation by SUMO-specific protease SENP3 overexpression. Total protein lysates from HeLa cells transfected with 3xFLAG-SENP3-dN, catalytically inactive 3xFLAG-SENP3-dN-C/S or 3xFLAG vector were analyzed with antisera to SUMO-1 or SUMO-2/3, as indicated. Anti-FLAG WB shows the expression of 3xFLAG-SENP3 and the anti-β-Actin WB indicates equal loading. (D) Depletion of SUMO-2/3 conjugates leads to de-repression of SCN1A and SCN3A but not SCN2A2. SCN1A, SCN2A2, and SCN3A mRNA levels were determined by RT-qPCR from HeLa cells transfected with 3xFLAG-SENP3-dN, 3xFLAG-SENP3-dN-C/S or 3xFLAG vector. Error bars of SEMs were generated from multiple experiments (n≥4). One-way ANOVA analyses showed that the SCN1A and SCN3A transcript levels in 3xFLAG-SENP3-dN over-expression cells were significantly higher than those over-expressing catalytically inactive 3xFLAG-SENP3-dN-C/S or 3xFLAG vector (*p<0.1).

Figure 6

SUMO-2/3 regulates CoREST1/LSD1 occupancy and histone modification status at specific promoters (A) SCN1A and SCN3A promoters are occupied by both CoREST1 and SUMO-2/3 while SCN2A2 promoter is occupied by CoREST1 but not SUMO-2/3 in HeLa cells. Chromatin immunoprecipitation (ChIP) assays were performed using anti-CoREST1, anti-SUMO-2/3 or control IgG and PCR was performed with primers specific for SCN1A, SCN2A2 and SCN3A promoters or RNA polymerase II polypeptide A exon (POLR2A ex). Input=1%. (B) Quantification of CoREST1, LSD1 and SUMO-2/3 occupancy on the SCN1A, SCN2A2 and SCN3A promoters by ChIP-qPCR. Error bars of SEMs were generated from multiple experiments (n≥2). Asterisks denote statistically significant occupancy of the protein at tested promoters relative to occupancy at POLR2A exon (t test, *p<0.1, **p<0.05, ***p<0.01). (C) Over-expression of 3xFLAG-SENP3-dN is coupled with dissociation of SUMO-2/3, CoREST1 and LSD1 and elevated levels of acetylated histone 3 (AcH3) and di-methylated histone 3 lysine 4 (diMeH3K4) at SCN1A and SCN3A. HeLa cells transfected with 3xFLAG-SENP3-dN, 3xFLAG-SENP3-dN-C/S or 3xFLAG vector were subject to ChIP assays with the indicated antibodies. Relative fold values indicate occupancies relative to 3xFLAG vector set at 1. Error bars of SEMs were generated from duplicate experiments. Asterisks denote statistically significant values in 3xFLAG-SENP3-dN over-expression samples compared to 3xFLAG-SENP3-dN-C/S or 3xFLAG vector samples by one-way ANOVA analysis. (*p<0.1, **p<0.05, ***p<0.01).

Figure 7

SUMO-2 interacting motif in CoREST1 is required for transcriptional repression of SCN1A and SCN3A (A) Schematic map of rescue plasmid construction. cDNA encoding CoREST1-WT or CoREST1-AAA resistant to CoREST1-RNAi, and a U6 promoter (U6-p) driven CoREST1 shRNA cassette were cloned in a pMSCV based retroviral vector. Expression of puromycin resistance gene (Puro^r) was controlled by an internal ribosomal entry sequence (IRES). U6-t, RNA polym Perase III transcription terminator. (B) CoREST1-mediated repression of SCN1A and SCN3A depends on its SUMO-2 interacting motif. mRNA was isolated from HeLa cells stably expressing the indicated control (scRNAi) or CoREST1 RNAi alone or with CoREST1-RNAi resistant WT (WT-res) or AAA (AAA-res) and levels of the SCN1A, SCN2A2, SCN3A and CoREST1 transcripts were determined by RT-qPCR. Error bars of SEMs were generated from multiple experiments (n≥3). Asterisks denote statistically significant values relative to scRNAi (t test, **p<0.05, ***p<0.01). (C) Recruitment of CoREST1 to SCN1A and SCN3A, but not SCN2A2, promoters depends on its SUMO-2 interaction motif. Anti-CoREST1 ChIP analyses were performed using CoREST1 WT rescue (WT-res) or mutant rescue (AAA-res) stable HeLa cells and quantitated by qPCR. Fold change relative to control HeLa cells is indicated. Error bars of SEMs were generated from duplicate experiments. Asterisks denote statistically significant values relative to WT-res (t test, *p<0.1). (D) Model illustrating SUMO-2/3-dependent (left panel) and SUMO-independent (right panel), promoter-specific transcriptional repression by the LSD1/CoREST1/HDAC co-repressor complex. On the left, the CoREST1 SUMO interaction motif binds a SUMO-2/3 (S) modified transcription factor (TF) or chromatin associated protein. On the right, a distinct region of CoREST1 is shown binding to REST, as an example of a SUMO-independent mechanism of recruitment.