Metastasis-related methyltransferase 1 (Merm1) represses the methyltransferase activity of Dnmt3a and facilitates RNA polymerase I transcriptional elongation

Metastasis-related methyltransferase 1 (Merm1) represses the methyltransferase activity of Dnmt3a... Abstract Stimulatory regulators for DNA methyltransferase activity, such as Dnmt3L and some Dnmt3b isoforms, affect DNA methylation patterns, thereby maintaining gene body methylation and maternal methylation imprinting, as well as the methylation landscape of pluripotent cells. Here we show that metastasis-related methyltransferase 1 (Merm1), a protein deleted in individuals with Williams–Beuren syndrome, acts as a repressive regulator of Dnmt3a. Merm1 interacts with Dnmt3a and represses its methyltransferase activity with the requirement of the binding motif for S-adenosyl-L-methionine. Functional analysis of gene regulation revealed that Merm1 is capable of maintaining hypomethylated rRNA gene bodies and co-localizes with RNA polymerase I in the nucleolus. Dnmt3a recruits Merm1, and in return, Merm1 ensures the binding of Dnmt3a to hypomethylated gene bodies. Such interplay between Dnmt3a and Merm1 facilitates transcriptional elongation by RNA polymerase I. Our findings reveal a repressive factor for Dnmt3a and uncover a molecular mechanism underlying transcriptional elongation of rRNA genes. Merm1, gene body methylation, Dnmt3a, transcriptional elongation, rRNA genes Introduction Methylation of cytosine bases at CpG dinucleotide is an important regulatory modification in eukaryotic genomes. Three catalytically active DNA methyltransferases (Dnmts), Dnmt1, Dnmt3a, and Dnmt3b, have been identified in mammals (Bestor et al., 1988; Yen et al., 1992; Okano et al., 1998; Xie et al., 1999). Dnmt1 functions predominantly in maintaining DNA methylation patterns (Gruenbaum et al., 1982; Leonhardt et al., 1992), while Dnmt3a and Dnmt3b are responsible for de novo methylation of unmethylated and hemimethylated DNA (Hsieh, 1999; Okano et al., 1999). Several members of the Dnmt3 family, including DNA methyltransferase 3-like protein Dnmt3L and Dnmt3b isoforms without sequence motifs necessary for methyltransferase activity, have been shown to act as stimulatory factors to facilitate DNA methyltransferase activity (Chédin et al., 2002; Suetake et al., 2004; Duymich et al., 2016). Dnmt3L has been identified as a regulatory protein of Dnmt3a and Dnmt3b that stimulates de novo methylation, thereby maintaining gene body methylation, maternal methylation imprinting, and the methylation landscape of pluripotent cells (Bourc’his et al., 2001; Neri et al., 2013). Removal of Dnmt3L prevents methylation of sequences that are normally maternally methylated and causes developmental defects in mouse embryos (Bourc’his et al., 2001). Catalytically inactive Dnmt3b isoforms are able to collaborate with Dnmt3a to restore DNA methylation, especially in gene bodies in differentiated cells (Duymich et al., 2016). In addition, Uhrf1 has been recognized as a Dnmt1 accessory protein that binds to hemimethylated CpG and recruits Dnmt1 to ensure the DNA methylation pattern during DNA replication or repression of its direct target genes (Bostick et al., 2007; Sharif et al., 2007). Regulatory proteins for DNA methyltransferases play critical roles in controlling DNA methylation and gene expression patterns, and identification of new regulatory proteins for both stimulatory and repressive regulators of DNA methyltransferases will provide further insight into mechanisms of gene expression and mammalian development, as well as human disease. Metastasis-related methyltransferase 1 (Merm1) was originally identified as Wbscr22. It is one of 26–28 genes that are deleted from 7q11.23 in Williams–Beuren syndrome, which is characterized by distinctive facial features, mental retardation, hypercalcemia, and hypertension (Doll and Grzeschik, 2001). Merm1 is conserved in yeast and human. This gene encodes a protein containing a nuclear localization signal and an S-adenosyl-L-methionine (SAM) binding motif typical of methyltransferases. Initially, Merm1 was shown to be implicated in methylation and processing of 18S rRNA, and its stability was found to be regulated by TRMT112 through the ubiquitin–proteasome pathway (Figaro et al., 2012; Õunap et al., 2013; Tafforeau et al., 2013; Haag et al., 2015; Õunap et al., 2015; Zorbas et al., 2015). Further investigation revealed that Merm1 plays a critical role in epigenetic regulation of gene expression in mammals. Merm1 regulates glucocorticoid receptor recruitment to the genome and mediates subsequent histone modification to maintain open chromatin (Jangani et al., 2014), suggesting that it plays a role in chromatin-based gene expression. In addition, Merm1 promotes cancer metastasis by inhibiting Zac1-mediated p53-dependent apoptosis, in which Merm1 methylates histone H3 lysine 9 (H3K9) at the Zac1 locus, thereby producing a transcriptionally repressive chromatin environment (Nakazawa et al., 2011). Thus, Merm1 is able to drive epigenetic alterations to affect gene activities. Merm1 itself does not have active histone or DNA methyltransferase activities (Nakazawa et al., 2011); therefore, its role in gene expression may be to function as a recruiter or modulator for epigenetic enzymes. Here we report for the first time that Merm1 acts as a repressive regulator of Dnmt3a and participates in regulating gene body methylation and RNA polymerase I (Pol I) transcription elongation at rRNA genes. We found that Dnmt3a binds to unmethylated rDNA gene bodies because of inhibition of its DNA methyltransferase activity by Merm1. Through direct binding to Dnmt3a, Merm1 maintains gene body hypomethylation, leading to co-occupancy with Dnmt3a, Pol I, and upstream binding factor (UBF) at unmethylated gene bodies. This complex is capable of ensuring efficient Pol I elongation on rDNA templates. Thus, Merm1 is a repressive regulator of Dnmt3a that participates in generating unmethylated gene bodies that allow transcriptional elongation by Pol I. Results Merm1 interacts with Dnmt3a and inhibits its DNA methyltransferase activity To assess the biological functions of Merm1, co-immunoprecipitation was performed to explore its binding partners. Surprisingly, Dnmt3a and Merm1 immunoprecipitated each other in vivo (Figure 1A). Furthermore, co-immunoprecipitation assays utilizing nuclear extracts from 293T cells harboring ectopic co-expression of HA-tagged Merm1 and Myc-tagged Dnmts confirmed that Merm1 precipitated Dnmt3a and Dnmt3b (Figure 1B). To assess the physical interaction between these two proteins, recombinant GST-tagged human Merm1 and Dnmts were prokaryotically purified and subjected to pull-down assays. Merm1 interacted directly with Dnmt3a and Dnmt3b, but not Dnmt1 (Figure 1C). These results demonstrated that Merm1 interacts directly with Dnmt3a and Dnmt3b. Figure 1 View largeDownload slide Merm1 binds to Dnmt3a and inhibits its DNA methyltransferase activity. (A) Dnmt3a interacts with Merm1 in vivo. Whole-cell extracts from HEK293T (293T) cells were incubated with antibodies against Merm1 or Dnmt3a. Co-immunoprecipitated proteins were monitored by immunoblotting using the indicated antibodies. (B) Merm1 directly interacts with Dnmt3a and Dnmt3b in vitro. Purified GST-Merm1 and GST-Dnmts were incubated with an antibody against Merm1. Co-precipitation of Dnmts was assessed by immunoblotting. Arrows or asterisks indicate the position of Dnmt3a or Dnmt3b, respectively. (C) Merm1 interacts with both Dnmt3a and Dnmt3b. Nuclear extracts from 293T cells expressing HA-Merm1 and Myc-Dnmts were incubated with an antibody against the HA epitope. Co-precipitated Dnmts were monitored by immunoblotting. Ten percent of the input is shown. (D) Merm1 inhibits the DNA methyltransferase activity of Dnmt3a. In vitro DNA methyltransferase activity assays were performed using purified proteins and synthetic substrate DNA. Error bars represent SD (n = 3). (E) Schematic illustration of the HA-tagged full-length (HA-Merm1), four glycines (4 G) to four arginines (4 R) substitution mutation (HA-Merm1-4G-4R*), N-terminal domain (HA-Merm1-N), and C-terminal domain (HA-Merm1-C) truncation mutants of Merm1. The numbers are the amino acid sequence numbers of Merm1. I, IV, IX, and X are conserved motifs of AdoMet-MTases. NLS, nuclear localization signal. (F) The N-terminal domain of Merm1 interacts with Dnmt3a in vitro. Whole-cell extracts from 293T cells expressing either the HA-tagged N-terminal domain or C-terminal domain of Merm1 were incubated with immobilized GST–Dnmt3a fusion proteins. Bound proteins were analyzed by immunoblotting using an anti-HA antibody. (G) The N-terminal domain of Merm1 binds to Dnmt3a. Whole-cell extracts from 293T cells expressing either the HA-tagged N-terminal domain or C-terminal domain of Merm1 were incubated with an antibody against Dnmt3a. Co-precipitated proteins were monitored by immunoblotting using an anti-HA antibody. (H) Mutation of the predicted SAM binding motif weakens the interaction between Merm1 and Dnmt3a. Nuclear extracts from 293T cells expressing HA-Merm1-4G-4R* and Myc-Dnmt3a were incubated with antibodies against HA or Myc. Co-precipitated proteins were monitored by immunoblotting. (I) Both the N-terminal and C-terminal domains of Merm1 are required for efficient inhibition of the methyltransferase activity of Dnmt3a. Whole-cell extracts from 293T cells expressing the indicated HA-tagged proteins were used to perform DNA methyltransferase activity assays. *P < 0.1, **P < 0.01, ***P < 0.001, two-tailed unpaired Student’s t-tests were performed. Figure 1 View largeDownload slide Merm1 binds to Dnmt3a and inhibits its DNA methyltransferase activity. (A) Dnmt3a interacts with Merm1 in vivo. Whole-cell extracts from HEK293T (293T) cells were incubated with antibodies against Merm1 or Dnmt3a. Co-immunoprecipitated proteins were monitored by immunoblotting using the indicated antibodies. (B) Merm1 directly interacts with Dnmt3a and Dnmt3b in vitro. Purified GST-Merm1 and GST-Dnmts were incubated with an antibody against Merm1. Co-precipitation of Dnmts was assessed by immunoblotting. Arrows or asterisks indicate the position of Dnmt3a or Dnmt3b, respectively. (C) Merm1 interacts with both Dnmt3a and Dnmt3b. Nuclear extracts from 293T cells expressing HA-Merm1 and Myc-Dnmts were incubated with an antibody against the HA epitope. Co-precipitated Dnmts were monitored by immunoblotting. Ten percent of the input is shown. (D) Merm1 inhibits the DNA methyltransferase activity of Dnmt3a. In vitro DNA methyltransferase activity assays were performed using purified proteins and synthetic substrate DNA. Error bars represent SD (n = 3). (E) Schematic illustration of the HA-tagged full-length (HA-Merm1), four glycines (4 G) to four arginines (4 R) substitution mutation (HA-Merm1-4G-4R*), N-terminal domain (HA-Merm1-N), and C-terminal domain (HA-Merm1-C) truncation mutants of Merm1. The numbers are the amino acid sequence numbers of Merm1. I, IV, IX, and X are conserved motifs of AdoMet-MTases. NLS, nuclear localization signal. (F) The N-terminal domain of Merm1 interacts with Dnmt3a in vitro. Whole-cell extracts from 293T cells expressing either the HA-tagged N-terminal domain or C-terminal domain of Merm1 were incubated with immobilized GST–Dnmt3a fusion proteins. Bound proteins were analyzed by immunoblotting using an anti-HA antibody. (G) The N-terminal domain of Merm1 binds to Dnmt3a. Whole-cell extracts from 293T cells expressing either the HA-tagged N-terminal domain or C-terminal domain of Merm1 were incubated with an antibody against Dnmt3a. Co-precipitated proteins were monitored by immunoblotting using an anti-HA antibody. (H) Mutation of the predicted SAM binding motif weakens the interaction between Merm1 and Dnmt3a. Nuclear extracts from 293T cells expressing HA-Merm1-4G-4R* and Myc-Dnmt3a were incubated with antibodies against HA or Myc. Co-precipitated proteins were monitored by immunoblotting. (I) Both the N-terminal and C-terminal domains of Merm1 are required for efficient inhibition of the methyltransferase activity of Dnmt3a. Whole-cell extracts from 293T cells expressing the indicated HA-tagged proteins were used to perform DNA methyltransferase activity assays. *P < 0.1, **P < 0.01, ***P < 0.001, two-tailed unpaired Student’s t-tests were performed. Bioinformatic sequence alignments showed that Merm1 contains several domains (I, IV, IX, and X) that are conserved in canonical Dnmts such as Dnmt1, Dnmt3a, and Dnmt3b (Supplementary Figure S1A). Similar to Dnmt3L, Merm1 does not contain a complete set of canonical DNA methyltransferase domains, and knockdown of Merm1 did not affect histone modification levels, indicating that it is not a histone or DNA methyltransferase enzyme (Supplementary Figures S1B and S2). Given that Merm1 interacts with Dnmt3a and Dnmt3b, we reasoned that it may be involved in modulating the DNA methyltransferase activity of known Dnmts. The regulatory effects of Merm1 on the activity levels of DNA methyltransferase Dnmt3a and Dnmt3b were assessed via the 3H-labeled S-adenosyl methionine (SAM) addition method in vitro. The liquid scintillation analyzer results revealed that Merm1 inhibited the methyltransferase activity of Dnmt3a without affecting that of Dnmt3b, although Merm1 and Dnmt3b were capable of direct interaction (Figure 1D and Supplementary Figure S3). Merm1 has a critical methyltransferase-related domain in its N-terminal and C-terminal regions. The N-terminal domain possesses a conserved motif (motif I) that consists of four glycines (4G), which are predicted to be responsible for binding of the methyl group donor SAM and cofactors (Figure 1E and Supplementary Figure S1A) (Nakazawa et al., 2011). To determine which domain of Merm1 is essential for its interaction with Dnmt3a and inhibitory effect on Dnmt3a activity, two truncation mutants (Merm1-N and Merm1-C) of HA-tagged Merm1 were generated and transiently overexpressed in 293T cells. The Merm1-N mutant was efficiently pulled down by the purified GST–Dnmt3a fusion protein, whereas relatively less of the Merm1-C mutant was precipitated by Dnmt3a (Figure 1F). Similarly, immunoprecipitation analysis using specific Dnmt3a antibodies showed that the Merm1-N mutant co-precipitated Dnmt3a with higher affinity than that with which it co-precipitated the Merm1-C mutant (Figure 1G). These results suggest that Merm1-N is required for the interaction between Merm1 and Dnmt3a. Considering that Merm1-N contains the SAM binding motif of Merm1, we constructed a substitution mutant (Merm1-4G4R*) with a disrupted SAM binding motif structure, which impaired SAM binding (Nakazawa et al., 2011), to test whether this motif contributed to the interaction between Merm1 and Dnmt3a. Co-immunoprecipitation analysis with Myc-tagged Dnmt3a, HA-tagged Merm1, or the 4G4R* mutant showed that the 4G4R* mutant significantly weakened the interaction between Merm1 and Dnmt3a (Figure 1H). Consistent with this finding, the inhibitory effect of Merm1 on the methyltransferase activity of Dnmt3a was mildly alleviated in the HA-Merm1-4G4R* substitution mutant and severely disrupted in the HA-Merm1-N truncation mutant (Figure 1H). Notably, although the C-terminal domain was not indispensable in the binding of Merm1 to Dnmt3a, it was required for efficient repression of the methyltransferase activity of Dnmt3a (Figure 1I), suggesting that the integrity of Merm1 is a prerequisite for its inhibitory effect on Dnmt3a. Taken together, these findings suggest that Merm1 interacts with Dnmt3a and inhibits its DNA methyltransferase activity in a manner requiring the SAM binding motif at the N-terminus of Merm1. Merm1 is capable of maintaining hypomethylated rDNA gene bodies Next, we tested whether the regulatory effect of Merm1 on Dnmt3a is involved in regulating gene expression. Previously, both Merm1 and Dnmt3a have been shown to localize in the nucleolus, and Dnmt3a has been shown to associate directly with rDNA (Majumder et al., 2006). Therefore, we assessed whether Dnmt3a and Merm1 function together to regulate the epigenetic state of rDNA and thereby regulate rDNA transcription. To monitor the binding of Merm1 to rDNA, we performed chromatin immunoprecipitation (ChIP) assays comparing Merm1 occupancy with that of Pol I. Merm1 binding was observed throughout the rDNA repeats, but it was enriched at rDNA gene bodies (Figure 2A). Although the overall level of rDNA-associated Merm1 was lower than that of Pol I, it was much higher than that of IgG (Supplementary Figure S4). In addition, association of Merm1 with gene bodies was significantly inhibited due to decreased expression of Merm1 caused by deleting one allele of Merm1 (designated as Merm1+/−) via CRISPR/Cas9 genome editing (Figure 2B), underscoring the specificity of the ChIP assays. Of note, we were unable to obtain homozygous deletion of Merm1 (Merm1−/−) in cells or mice, likely because of lethal effects caused by the complete loss of Merm1. Consistent with this finding, RNAi-mediated knockdown of the Merm1 ortholog in Caenorhabditis elegans was embryonic-lethal (Piano et al., 2002). Figure 2 View largeDownload slide Merm1 maintains unmethylated rDNA gene bodies. (A) Merm1 is associated with rDNA. ChIP assays were performed to show the rDNA occupancy of Merm1 and Pol I (RPA194) in 293T cells. The scheme indicates the positions of the primers used to amplify regions of rDNA. The primer sequences are listed in Supplementary Table S2. Error bars denote SD (n = 3). (B) ChIP assays were conducted in parallel with chromatin from wild-type 293T control cells (Ctrl) or Merm1+/− cells (#1; #2) (left panel). Error bars denote SD (n = 3). Protein levels of Merm1 in wild-type 293T (Ctrl) and Merm1-knockout cells (#1; #2) were tested by immunoblotting (right panel). Tubulin served as a loading control. (C) Loss of Merm1 increases gene body methylation. Bisulfite sequencing was performed to assess the DNA methylation status of rDNA gene body regions (+8017/+8224, the same as region C indicated in A) in 293T and Merm1+/− cells. Methylated and unmethylated CpG sites are shown as black and white circles, respectively. The numbers refer to the positions of the outermost CpG residues. (D) Analysis of the results in C showed the percentage of methylated and unmethylated gene bodies upon depletion of Merm1. (E) Depletion of Merm1 leads to a significant increase in DNA methylation in gene body regions, but not IGS. NIH3T3 cells were infected with retroviruses encoding Merm1-specific shRNA (shMerm1) or control shRNA (shCtrl) for 5 days. Upper panel: 5mC MeDIP-seq profiles for rDNA. Lower panel: overlay tracks of 5mC MeDIP-seq data from control and Merm1-depleted NIH3T3 cells. The locations of the 18S and 28S rRNA coding regions are shown at the bottom. *P < 0.1, **P < 0.01, ***P < 0.001, two-tailed unpaired Student’s t-tests were performed. Figure 2 View largeDownload slide Merm1 maintains unmethylated rDNA gene bodies. (A) Merm1 is associated with rDNA. ChIP assays were performed to show the rDNA occupancy of Merm1 and Pol I (RPA194) in 293T cells. The scheme indicates the positions of the primers used to amplify regions of rDNA. The primer sequences are listed in Supplementary Table S2. Error bars denote SD (n = 3). (B) ChIP assays were conducted in parallel with chromatin from wild-type 293T control cells (Ctrl) or Merm1+/− cells (#1; #2) (left panel). Error bars denote SD (n = 3). Protein levels of Merm1 in wild-type 293T (Ctrl) and Merm1-knockout cells (#1; #2) were tested by immunoblotting (right panel). Tubulin served as a loading control. (C) Loss of Merm1 increases gene body methylation. Bisulfite sequencing was performed to assess the DNA methylation status of rDNA gene body regions (+8017/+8224, the same as region C indicated in A) in 293T and Merm1+/− cells. Methylated and unmethylated CpG sites are shown as black and white circles, respectively. The numbers refer to the positions of the outermost CpG residues. (D) Analysis of the results in C showed the percentage of methylated and unmethylated gene bodies upon depletion of Merm1. (E) Depletion of Merm1 leads to a significant increase in DNA methylation in gene body regions, but not IGS. NIH3T3 cells were infected with retroviruses encoding Merm1-specific shRNA (shMerm1) or control shRNA (shCtrl) for 5 days. Upper panel: 5mC MeDIP-seq profiles for rDNA. Lower panel: overlay tracks of 5mC MeDIP-seq data from control and Merm1-depleted NIH3T3 cells. The locations of the 18S and 28S rRNA coding regions are shown at the bottom. *P < 0.1, **P < 0.01, ***P < 0.001, two-tailed unpaired Student’s t-tests were performed. Given that Merm1 was enriched at gene bodies, its regulatory role on epigenetic modifications on gene bodies was assessed. Merm1 affected rDNA gene body methylation in wild-type and Merm1+/− cells. Bisulfite sequencing analysis showed extensive gene body methylation at rDNA gene bodies, and depletion of Merm1 significantly increased the proportion of methylated gene bodies from 43% to 78% (Figure 2C and D). To verify that Merm1 regulates gene body methylation, methylated DNA immunoprecipitation sequencing (MeDIP-seq) assays were performed. Knockdown of Merm1 increased the level of gene body methylation, but DNA methylation at intergenic spacers (IGS) remained unchanged (Figure 2E), which confirmed the validity of the MeDIP-seq assay. These results indicate that Merm1 maintains hypomethylated rRNA gene bodies. Merm1 is recruited by Dnmt3a and ensures the binding of Dnmt3a to hypomethylated gene bodies via co-occupation with Pol I and UBF The finding that Merm1 inhibits the activity of Dnmt3a and maintains hypomethylated rDNA gene bodies raises the possibility that both Dnmt3a and Merm1  bind to hypomethylated gene bodies. Indeed, the ChIP-bisulfite assays revealed that Merm1 was associated exclusively with unmethylated gene bodies, whereas Dnmt1 was only associated with hypermethylated rDNA, and Pol I and UBF bound to unmethylated rDNA as expected (Figure 3A and B). However, Dnmt3a bound to both methylated and unmethylated gene bodies, but was preferentially associated with the latter (Figure 3C). Thus, Pol I, Merm1, and Dnmt3a are all restricted to unmethylated gene bodies, suggesting the possibility that these proteins may co-occupy gene bodies. Sequential-ChIP assays confirmed that Merm1 co-occupied with Dnmt3a, Pol I, and UBF, but not with Dnmt1 (Figure 3D), suggesting the co-occupancy on hypomethylated rDNA gene bodies coordinates rDNA transcription elongation. Figure 3 View largeDownload slide Merm1 ensures the binding of Dnmt3a to unmethylated gene bodies via co-occupation with UBF and Pol I. (A) Merm1 is associated with hypomethylated rDNA, similar to Pol I and UBF (upper panel), while Dnmt1 is associated with hypermethylated rDNA (lower panel). ChIP-enriched DNA from 293T cells using the indicated antibodies was used to perform bisulfite sequencing. Human rDNA sequences (+8017/+8224) were amplified, cloned, and sequenced. Methylated and unmethylated CpG sites are shown as black and white circles, respectively. The numbers refer to the positions of the outermost CpG residues. (B) Analysis of the results in A showed the percentage of methylated and unmethylated gene bodies associated with Pol I, UBF, Merm1, and Dnmt1. (C) Dnmt3a is preferentially associated with unmethylated rDNA. (D) Sequential-ChIP assays show that Merm1 co-localizes with Dnmt3a, Pol I, and UBF. Cross-linked chromatin from 293T cells was immunoprecipitated with an antibody against Merm1 (first ChIP), followed by precipitation with antibodies against Dnmt1, Dnmt3a, Pol I (RPA194), and UBF (second ChIP). The co-precipitated DNA was analyzed by PCR. (E) Reduction of Merm1 abundance increases Dnmt3a occupancy at gene bodies. ChIP assays performed using primers amplifying the 28S rRNA coding region were used to show the relative rDNA occupancy of Dnmt3a in 293T and Merm1+/− cells. Error bars represent SD (n = 3). (F) Analysis of the results in E showed increased binding of Dnmt3a to methylated gene bodies in Merm1 knockdown cells. (G) Loss of Dnmt3a reduces the occupancy of Merm1 at gene bodies. ChIP assays were performed using primers amplifying the 28S rRNA coding region to show the relative rDNA occupancy of Merm1 in 293T and Dnmt3a−/− cells. (H) Dnmt3a is associated exclusively with hypermethylated gene bodies after reduction of the abundance of Merm1. ChIP-enriched DNA from 293T or Merm1+/− cells obtained using an antibody against Dnmt3a was used to perform bisulfite sequencing. Error bars represent SD (n = 3). *P < 0.1, two-tailed unpaired Student’s t-tests were performed. Figure 3 View largeDownload slide Merm1 ensures the binding of Dnmt3a to unmethylated gene bodies via co-occupation with UBF and Pol I. (A) Merm1 is associated with hypomethylated rDNA, similar to Pol I and UBF (upper panel), while Dnmt1 is associated with hypermethylated rDNA (lower panel). ChIP-enriched DNA from 293T cells using the indicated antibodies was used to perform bisulfite sequencing. Human rDNA sequences (+8017/+8224) were amplified, cloned, and sequenced. Methylated and unmethylated CpG sites are shown as black and white circles, respectively. The numbers refer to the positions of the outermost CpG residues. (B) Analysis of the results in A showed the percentage of methylated and unmethylated gene bodies associated with Pol I, UBF, Merm1, and Dnmt1. (C) Dnmt3a is preferentially associated with unmethylated rDNA. (D) Sequential-ChIP assays show that Merm1 co-localizes with Dnmt3a, Pol I, and UBF. Cross-linked chromatin from 293T cells was immunoprecipitated with an antibody against Merm1 (first ChIP), followed by precipitation with antibodies against Dnmt1, Dnmt3a, Pol I (RPA194), and UBF (second ChIP). The co-precipitated DNA was analyzed by PCR. (E) Reduction of Merm1 abundance increases Dnmt3a occupancy at gene bodies. ChIP assays performed using primers amplifying the 28S rRNA coding region were used to show the relative rDNA occupancy of Dnmt3a in 293T and Merm1+/− cells. Error bars represent SD (n = 3). (F) Analysis of the results in E showed increased binding of Dnmt3a to methylated gene bodies in Merm1 knockdown cells. (G) Loss of Dnmt3a reduces the occupancy of Merm1 at gene bodies. ChIP assays were performed using primers amplifying the 28S rRNA coding region to show the relative rDNA occupancy of Merm1 in 293T and Dnmt3a−/− cells. (H) Dnmt3a is associated exclusively with hypermethylated gene bodies after reduction of the abundance of Merm1. ChIP-enriched DNA from 293T or Merm1+/− cells obtained using an antibody against Dnmt3a was used to perform bisulfite sequencing. Error bars represent SD (n = 3). *P < 0.1, two-tailed unpaired Student’s t-tests were performed. Significantly, ChIP-bisulfite assays carried out in Merm1+/− cells revealed that upon depletion of Merm1, Dnmt3a was no longer bound to hypomethylated rDNA gene bodies, but rather was exclusively associated with hypermethylated rDNA gene bodies. However, the association between Pol I and hypomethylated gene bodies was unchanged in Merm1+/− cells (Figure 3E and F; Supplementary Figure S5). These results suggest that association of Dnmt3a with hypomethylated rDNA gene bodies was caused by the repressive effect of Merm1 on Dnmt3a, because depletion of Merm1 relieved this inhibitory effect and allowed Dnmt3a to bind to hypermethylated gene bodies. Moreover, neither loss of Dnmt3a nor concurrent disruption of Merm1 and Dnmt3a (Merm1+/−/Dnmt3a−/−, also denoted as DKO) decreased gene body methylation (Supplementary Figure S6), indicating that depletion of Dnmt3a leads to disassociation of the Merm1−Dnmt3a complex and departure of Merm1 from gene bodies. Indeed, depletion of Dnmt3a impaired the association between Merm1 and rDNA (Figure 3G), whereas knockdown of Merm1 enhanced the binding of Dnmt3a to rDNA (Figure 3H), suggesting that Dnmt3a guides Merm1 to rDNA gene bodies. Taken together, these findings demonstrate that Merm1 is recruited by Dnmt3a, thereby establishing hypomethylated gene bodies where Dnmt3a, Pol I, and UBF show co-occupancy. Merm1 co-localizes with Pol I and UBF Given that Merm1, UBF, and Pol I bind to and co-occupy unmethylated gene bodies, we assumed that Merm1 may interact and co-localize with Pol I and UBF. Indeed, we found that Merm1 interacted with RNA Pol I and UBF, but not with productive preinitiation complex components TAFI95 and TIF-IA, which are restricted to rDNA promoters (Figure 4A). To further assess the interactions among Merm1, Pol I, and UBF, we investigated the sub-cellular localization of Merm1 using indirect immunofluorescence microscopy. Immunofluorescent staining of Merm1 with a commercial antibody showed that endogenous Merm1 was observed throughout the nucleus in 293T cells, with a portion co-localized with Pol I and UBF (Supplementary Figure S7A). Furthermore, Merm1 fusion proteins with an HA or GFP tag were generated and transiently transfected into 293T, U2OS, and NIH3T3 cells. The fluorescent signals were captured using a spinning-disk confocal laser scanning microscope or an inverted immunofluorescent microscope, after which the signals were analyzed. HA-tagged Merm1 was most pronounced in the nucleoli of 293T cells, where it partially co-localized with Pol I and UBF (Supplementary Figure S7B). The GFP-tagged Merm1 signal was also enriched in the nucleoli of 293T and NIH3T3 cells, where it partially co-localized with Pol I and UBF (Supplementary Figure S7C and D). Moreover, analysis of the fluorescent signals derived from 3D stacks of GFP-tagged Merm1 at different angles confirmed that Merm1 was associated with Pol I and UBF in U2OS cells (Figure 4B and C; Supplementary Movies S1 and S2). Taken together, these results confirm that Merm1 interacts and co-localizes with Pol I and UBF. Figure 4 View largeDownload slide A fraction of Merm1 interacts and co-localizes with Pol I and UBF within nucleoli. (A) Merm1 interacts with the Pol I transcription machinery. Cell lysates were incubated with an anti-Merm1 antibody, and co-precipitated proteins were monitored by immunoblotting. Ten percent of the input is shown. (B) Merm1 partially co-localizes with Pol I and UBF in nucleoli. Fluorescent images of GFP-tagged Merm1, Pol I, and UBF in human U2OS cells are displayed. Scale bar, 2 μm. (C) 3D isosurface images of Merm1, Pol I, and UBF from two different angles were generated with image stacks using Volocity 6.3. Figure 4 View largeDownload slide A fraction of Merm1 interacts and co-localizes with Pol I and UBF within nucleoli. (A) Merm1 interacts with the Pol I transcription machinery. Cell lysates were incubated with an anti-Merm1 antibody, and co-precipitated proteins were monitored by immunoblotting. Ten percent of the input is shown. (B) Merm1 partially co-localizes with Pol I and UBF in nucleoli. Fluorescent images of GFP-tagged Merm1, Pol I, and UBF in human U2OS cells are displayed. Scale bar, 2 μm. (C) 3D isosurface images of Merm1, Pol I, and UBF from two different angles were generated with image stacks using Volocity 6.3. Interplay between Merm1 and Dnmt3a facilitates Pol I elongation The finding that Merm1 is enriched at gene bodies and interacts with Pol I and UBF, but not with promoter preinitiation complex components, suggests that the interplay between Dnmt3a and Merm1 may regulate rDNA transcription elongation. Therefore, the level of pre-rRNA synthesis in Merm1+/−, Dnmt3a−/−, and DKO cells was assessed by qRT-PCR. DKO cells displayed severely impaired rDNA transcription in comparison with that of Merm1+/− or Dnmt3a−/− cells (Supplementary Figure S8). Short 5′-fluorouridine (FUrd) pulse-labeling of DKO cells revealed drastic decreases in the rates of rRNA synthesis and proliferation (Figure 5A and B), suggesting that interaction between Dnmt3a and Merm1 facilitated rDNA transcription. However, examination of the occupancy of Pol I at rDNA in the wild-type and DKO cells using ChIP assays produced the unexpected result that, in contrast with compromised pre-rRNA synthesis, simultaneous loss of Merm1 and Dnmt3a led to increased occupancy of Pol I at gene bodies (Figure 5C). Whole-genome scale ChIP-seq analysis of Pol I confirmed that concurrent depletion of Merm1 and Dnmt3a increased Pol I enrichment at rDNA gene bodies (Figure 5D). These findings suggested that the severe inhibition of transcription by DKO cells may be caused by retardation of Pol I elongation and accumulation at gene bodies. To test this hypothesis, an actinomycin D (AMD) release assay was performed to compare Pol I elongation in wild-type and DKO cells. Cells were treated with a low concentration of AMD for 1 h to stop rDNA transcription and trap Pol I at rDNA promoters, after which the medium containing AMD was replaced with new medium without AMD to reinitialize movement of Pol I from promoters to gene bodies (Figure 5E). Following the AMD treatment and re-initialization of Pol I movement, Pol I enrichment at rDNA gene bodies was tested. AMD significantly decreased enrichment of Pol I at rDNA gene bodies, but gradual re-occupancy of Pol I at rDNA gene bodies was observed in 293T and DKO cells following the removal of medium containing AMD (Figure 5F). However, DKO cells exhibited a much slower re-occupancy rate of Pol I at rDNA gene bodies in comparison with that of wild-type cells (Figure 5F), suggesting that simultaneous deletion of Merm1 and Dnmt3a led to retardation of Pol I movement along gene bodies. Taken together, these findings demonstrate that Merm1 and Dnmt3a synergistically facilitate rDNA transcription by supporting Pol I elongation at rDNA gene bodies. Figure 5 View largeDownload slide Merm1 and Dnmt3a are required for Pol I transcription elongation. (A) Simultaneous knockout of one allele of Merm1 and two alleles of Dnmt3a impairs rDNA transcription. Merm1+/−/Dnmt3a−/− cells (Merm1+/−/Dnmt3a−/− #1 and Merm1+/−/Dnmt3a−/− #2, denoted as DKO #1 and DKO #2 below) were mixed with 293T cells, labeled for 15 min with 2 mM FUrd, and stained with antibodies against Dnmt3a and BrdU. Merm1+/−/Dnmt3a−/− cells are encircled. The bar diagram shows results obtained by subjecting 300 Merm1+/−/Dnmt3a−/− cells and 293T cells to FUrd staining. Scale bar, 5 μm. (B) Simultaneous depletion of Merm1 and Dnmt3a (DKO #1 and DKO #2) further inhibits cell proliferation in comparison with the inhibitory effect induced by depletion of Merm1 (Merm1+/−#1) or Dnmt3a (Dnmt3a−/− #1) alone. Cells were counted every 24 h. (C) Depletion of Merm1 or Dnmt3a impairs the association of Pol I with rDNA, while simultaneous depletion of these proteins increases Pol I occupancy. ChIP assays were performed using primers amplifying the 28S rRNA coding region to show the relative rDNA occupancy of Pol I in the indicated cells. Error bars represent SD (n = 3). (D) Pol I occupancy at rDNA gene body regions is increased in Merm1+/−/Dnmt3a−/− cells. Upper panel: Pol I ChIP-seq profiles for rDNA. Lower panel: overlay tracks of Pol I ChIP-seq data from the indicated cell lines. The locations of the 18S and 28S rRNA coding regions are shown at the bottom. (E) Scheme of the AMD release assay. 293T and Merm1+/−/Dnmt3a−/− cells (DKO #1 and DKO #2) were treated with 0.05 μg/ml AMD for 1 h, followed by replacement of the medium with AMD-free medium to allow transcription for 8, 16, or 24 h. Groups that were not subjected to AMD treatment (No AMD) or subjected to 1 h of AMD treatment (AMD 1 h) were used as controls. (F) The AMD release assay shows that simultaneous depletion of Merm1 and Dnmt3a reduces the transcription elongation rate of Pol I. ChIP assays were performed using primers that amplify the 28S rRNA coding region to show the relative rDNA occupancy of Pol I in the indicated treated cells. Error bars represent SD (n = 3). *P < 0.1, **P < 0.01, ***P < 0.001, ****P < 0.0001, two-tailed unpaired Student’s t-tests were performed. Figure 5 View largeDownload slide Merm1 and Dnmt3a are required for Pol I transcription elongation. (A) Simultaneous knockout of one allele of Merm1 and two alleles of Dnmt3a impairs rDNA transcription. Merm1+/−/Dnmt3a−/− cells (Merm1+/−/Dnmt3a−/− #1 and Merm1+/−/Dnmt3a−/− #2, denoted as DKO #1 and DKO #2 below) were mixed with 293T cells, labeled for 15 min with 2 mM FUrd, and stained with antibodies against Dnmt3a and BrdU. Merm1+/−/Dnmt3a−/− cells are encircled. The bar diagram shows results obtained by subjecting 300 Merm1+/−/Dnmt3a−/− cells and 293T cells to FUrd staining. Scale bar, 5 μm. (B) Simultaneous depletion of Merm1 and Dnmt3a (DKO #1 and DKO #2) further inhibits cell proliferation in comparison with the inhibitory effect induced by depletion of Merm1 (Merm1+/−#1) or Dnmt3a (Dnmt3a−/− #1) alone. Cells were counted every 24 h. (C) Depletion of Merm1 or Dnmt3a impairs the association of Pol I with rDNA, while simultaneous depletion of these proteins increases Pol I occupancy. ChIP assays were performed using primers amplifying the 28S rRNA coding region to show the relative rDNA occupancy of Pol I in the indicated cells. Error bars represent SD (n = 3). (D) Pol I occupancy at rDNA gene body regions is increased in Merm1+/−/Dnmt3a−/− cells. Upper panel: Pol I ChIP-seq profiles for rDNA. Lower panel: overlay tracks of Pol I ChIP-seq data from the indicated cell lines. The locations of the 18S and 28S rRNA coding regions are shown at the bottom. (E) Scheme of the AMD release assay. 293T and Merm1+/−/Dnmt3a−/− cells (DKO #1 and DKO #2) were treated with 0.05 μg/ml AMD for 1 h, followed by replacement of the medium with AMD-free medium to allow transcription for 8, 16, or 24 h. Groups that were not subjected to AMD treatment (No AMD) or subjected to 1 h of AMD treatment (AMD 1 h) were used as controls. (F) The AMD release assay shows that simultaneous depletion of Merm1 and Dnmt3a reduces the transcription elongation rate of Pol I. ChIP assays were performed using primers that amplify the 28S rRNA coding region to show the relative rDNA occupancy of Pol I in the indicated treated cells. Error bars represent SD (n = 3). *P < 0.1, **P < 0.01, ***P < 0.001, ****P < 0.0001, two-tailed unpaired Student’s t-tests were performed. Discussion Over the past few decades, numerous studies have focused on elucidating the mechanisms underlying DNA methylation and the manner in which it controls gene transcription. With regard to rRNA genes, Dnmt1, Dnmt3a, and Dnmt3b have been shown to localize in the nucleolus and associate with rDNA, implying that they have essential roles in epigenetic control of rDNA transcription (Majumder et al., 2006). It is widely acknowledged that DNA methylation in promoters silences rRNA genes, while active promoters are characterized by DNA hypomethylation (Grummt and Längst, 2013). Dnmt1- and/or Dnmt3b-mediated DNA methylation on rDNA promoters can impair the assembly of the transcription initiation complex, including recruitment of Pol I and UBF (McStay and Grummt, 2008). In contrast, genetic disruption of Dnmt1 and Dnmt3b leads to hypomethylation and activation of rDNA promoters (Majumder et al., 2006; Gagnon-Kugler et al., 2009; Grummt and Längst, 2013). However, neither ectopic Dnmt3a nor decreased expression of Dnmt3a affected rDNA promoter activity (Majumder et al., 2006), indicating that Dnmt3a plays a role distinct from those of Dnmt1 and Dnmt3b. Indeed, unlike Dnmt1 or Dnmt3b, Dnmt3a is not required for the maintenance of methylation at heterochromatic repeat regions (Okano et al., 1999). Deletion of Dnmt1 or Dnmt3b leads to early embryonic lethality, while Dnmt3a-null mice appeared to be grossly normal at birth, but showed postnatal developmental defects and died prematurely, suggesting that Dnmt3a regulates euchromatic expression and postnatal development (Li et al., 1992). More importantly, Dnmt3a has been found to occupy the gene bodies of a large cohort of neurogenic genes, for which it facilitates transcription by functionally antagonizing polycomb repression. These findings reveal that Dnmt3a may maintain active transcription elongation of genes critical for development (Wu et al., 2010). In this study, we revealed that Dnmt3a is involved in the modulation of gene body methylation and transcription elongation of rRNA genes. Thus, our results and others demonstrate Dnmt3a-dependent transcription elongation of Pol I- and Pol II-transcribed genes. In addition, our study identified a repressive regulator of Dnmt3a, nucleolar protein Merm1, which cooperated with Dnmt3a to modulate the methylation status of gene bodies, thereby regulating their capacity for rDNA transcription elongation. Several regulators have been shown to enhance the activity of DNA methyltransferases and play critical roles in maintaining DNA methylation patterns (Chédin et al., 2002; Duymich et al., 2016). Our results reveal that Merm1 acts as a negative regulator of Dnmt3a. Although Merm1 contains a SAM binding motif, it does not possess DNA methyltransferase activity or the conserved catalytic domain common to DNA methyltransferases. Remarkably, Merm1 negatively regulates the DNA methyltransferase activity of Dnmt3a to generate unmethylated rDNA templates, thereby allowing Pol I transcription elongation. Inhibition of methyltransferase activity by Merm1 resulted in association of Dnmt3a with unmethylated gene bodies. Under these circumstances, Dnmt3a could not exert its rDNA methyltransferase activity, and Dnmt3a knockout did not further affect gene body methylation. However, loss of Merm1 restored the DNA methyltransferase activity of Dnmt3a and rendered Dnmt3a able to methylate rDNA gene bodies, thus enhancing rDNA gene body methylation. Our findings provide new insight into the manner in which Dnmt3a and Merm1 interact to modulate the methylation state of gene bodies and control gene transcription elongation. Previously, Merm1 and its yeast homolog Bud23 were identified as 18S rRNA base methyltransferases that play critical roles in pre-rRNA processing (Figaro et al., 2012; Õunap et al., 2013; Tafforeau et al., 2013; Létoquart et al., 2014; Haag et al., 2015; Õunap et al., 2015; Zorbas et al., 2015). It is clear that Merm1 regulates pre-rRNA processing, but it may have additional functions in mammalian cells. First, although mammalian Merm1 can facilitate rRNA methylation, such modification does not seem to be required for pre-processing (Zorbas et al., 2015). Furthermore, human Merm1 and Bud23 have only 47% amino acid similarity. Moreover, yeast does not possess cytosine DNA methyltransferases, so yeast DNA is unmethylated. These findings suggest that human Merm1 and Bud23 have similar, but probably not identical functions (Nakazawa et al., 2011; Jangani et al., 2014; Õunap et al., 2015). Indeed, Merm1 has been shown to drive alterations of epigenetic modification at promoters in mammalian cells (Jangani et al., 2014), and here we verified that Merm1 acts as a repressive regulator of Dnmt3a and activates rDNA transcription. Thus, in addition to its role in pre-rRNA processing, modulation of epigenetic states by Merm1 is required for rDNA transcription. It is not surprising that Merm1 plays dual roles in the processes of rDNA transcription and pre-rRNA processing. Some other proteins, such as Dnmt1, nucleolin, and fibrillarin, have well-characterized roles in multiple steps of rRNA processing, including regulation of rDNA transcription, pre-rRNA splicing, and assembly of ribosomal particles (Egyhazi et al., 1988; Grummt, 1998; Roger et al., 2002). With regard to Dnmt1, the extensive loss of DNA methylation at rDNA loci caused by Dnmt1 knockout reduced rDNA transcription and led to defective rRNA processing (Gagnon-Kugler et al., 2009), suggesting that the ribosomal synthesis rate is controlled by regulating the rates of transcription and subsequent rRNA maturation. In conclusion, we have shown that Merm1 interacts with Pol I and UBF and regulates the activity of Dnmt3a, thereby collaborating to control rDNA transcription elongation. Depletion of Merm1 increased gene body methylation and consequently decreased Pol I binding. Simultaneous knockout of one allele of Merm1 and two alleles of Dnmt3a did not affect gene body methylation, but rather retarded Pol I elongation, resulting in the accumulation of Pol I on rDNA templates. This finding suggests that both Merm1 and Dnmt3a are required for transcriptional elongation of rDNA (Figure 6). In addition, depletion of either Merm1 or Dnmt3a produced similar changes in the expression levels of many Pol II-transcribed genes (Supplementary Figure S9), suggesting that Merm1 and Dnmt3a cooperate to regulate Pol I- and Pol II-mediated gene transcription. Therefore, our results may reveal a general cellular mechanism, in which the generation of unmethylated gene bodies by a repressive regulator of DNA methyltransferases allows transcription elongation. There are alternative pathways that lead to gene bodies demethylation, such as that involving the TET family of demethylases (Santi et al., 1983; He et al., 2011). Therefore, it is essential to explore the functional relationships between TET proteins, Merm1, and Dnmt3a to reveal how they may cooperate to control gene body methylation and transcriptional elongation. Figure 6 View largeDownload slide Proposed model depicting the roles of Merm1 and Dnmt3a in regulating Pol I elongation. (A) Merm1 is recruited by Dnmt3a and inhibits the DNA methyltransferase activity of Dnmt3a, resulting in co-occupancy of Dnmt3a and Pol I at hypomethylated gene bodies, thus allowing Pol I elongation. (B) Upon depletion of Merm1, the methyltransferase activity of Dnmt3a is restored, leading to binding of Dnmt3a to methylated gene bodies and dissociation of Pol I. (C) In the case of simultaneous knockout of one allele of Merm1 and two alleles of Dnmt3a (arrows), loss of these two proteins leads to retardation of elongation and accumulation of Pol I on gene bodies. Figure 6 View largeDownload slide Proposed model depicting the roles of Merm1 and Dnmt3a in regulating Pol I elongation. (A) Merm1 is recruited by Dnmt3a and inhibits the DNA methyltransferase activity of Dnmt3a, resulting in co-occupancy of Dnmt3a and Pol I at hypomethylated gene bodies, thus allowing Pol I elongation. (B) Upon depletion of Merm1, the methyltransferase activity of Dnmt3a is restored, leading to binding of Dnmt3a to methylated gene bodies and dissociation of Pol I. (C) In the case of simultaneous knockout of one allele of Merm1 and two alleles of Dnmt3a (arrows), loss of these two proteins leads to retardation of elongation and accumulation of Pol I on gene bodies. Materials and methods Cell culture U2OS, NIH3T3, 293T, and the derived knockout or knockdown cells were all cultured at 5% CO2, 37°C in Dulbecco’s modified Eagle’s medium (DMEM, Gibco) supplemented with 10% (v/v) fetal bovine serum (FBS). For growth curve, wild-type 293T and knockout cells were in inoculums of 1 × 106, and NIH3T3 and the corresponding knockdown cells were in inoculums of 0.5 × 106 in 100-mm culture dish. Cells were counted with hemocytometer every day and cumulative cell numbers were calculated and presented. Plasmids cDNA encoding human Merm1 (GenBank accession number NM_017528) was PCR amplified with specific primers and the amplicons were ligated into pEGFP-C1, pGEX-4T-1, and pCMV-3 × HA to generate recombinant expression constructs. cDNA encoding mouse Merm1 (GenBank accession number NM_025375.3) was cloned into pEGFP-C1. Transcript variants of Merm1 to produce Merm1-4G4R, Merm1-N, and Merm1-C were cloned and inserted into pCMV-3 × HA to generate recombinant expression vectors. The primers used to generate recombinant constructs are listed in Supplementary Table S1. pcDNA3/Myc-Dnmt1 (Addgene plasmid # 36939), pcDNA3/Myc-Dnmt3a (Addgene plasmid # 35521), and pcDNA3/Myc-Dnmt3b1 (Addgene plasmid # 35522) are gifts from Arthur Riggs (Chen et al., 2005; Li et al., 2006). Antibodies Anti-Dnmt3a (sc-20703), anti-Merm1 (sc-135322), anti-RPA194 (sc-28714), anti-UBF (sc-9131X), and anti-c-Myc (sc-40X) were purchased from Santa Cruz. Anti-α-tubulin (T6199), anti-HA (H9658), and anti-Bromodeoxyuridine (B2531) were obtained from Sigma. Anti-5-methylcytosine (BI-MECY-0100) was from Eurogentec. Anti-H3K4me3 (04-745) and anti-H3K27me3 (07-449) were from Upstate. Anti-Merm1 (ab97911), anti-H3K9me3 (ab8898), anti-H3K36me3 (ab9050), anti-H4K20me3 (ab9053), and anti-H3 (ab1791) were from Abcam. Secondary antibodies IRDye800CW goat anti-mouse IgG (926-32210) and IRDye800CW goat anti-rabbit IgG (926-32211) were purchased from LI-COR. Targeting strategy of knockdown and knockout For knockdown in NIH3T3 cells, the shRNA targeting mouse Merm1 (5′-TCGCAACTCACGGATGATT-3′) was cloned into the retroviral vector pQXCIP. Plat-E cell was transfected with the recombinant pQXCIP constructs to package the retrovirus. The packaged virus was collected and applied to infect NIH3T3 cells. Cells were selected with puromycin for 5 days. For knockout in 293T cells, the CRISPR/Cas9 system was utilized. Recombinant pcDNA3.1-NLS-Cas9-NLS containing human Cas9 and pUC19 containing specific guide RNA were co-transfected into 293T cells. Single clone of 293T was selected and cultured. Cells were harvested to extract genomic DNA and the DNA was followed by Sanger sequencing and DNAMAN alignment analysis. The sequence information of guide RNAs is provided in Supplementary Table S1. Immunofluorescence Cells grown on coverslips were washed with 1× PBS and fixed in 4% paraformaldehyde for 10 min at room temperature. For FUrd labeling, cells were labeled with 2 mM of FUrd for 5 or 15 min before fixation. After washing three times with 1× PBS (5 min each), cells were permeabilized with 0.5% Triton X-100 for 15 min. Then, cells were washed with PBS for three times (5 min each) and were blocked with 1% BSA in 1× PBS (w/v) for 30 min at room temperature. After washing three times with PBS (5 min each), cells were incubated with primary antibodies diluted in PBST (1× PBS + 0.1% Tween-20) at 4°C overnight. After incubation, cells were washed with PBST (1× PBS + 0.1% Triton X-100) for four times. Then cells were incubated with appropriate secondary antibodies (dilution at 1:400, Invitrogen) for 2 h, washed with 0.1% PBST (Triton X-100) and stained with 2 ng/μl of DAPI for 5 min. The coverslips were sealed and the images were recorded with immunofluorescent microscope. The obtained FUrd signals were quantified with Image J software (National Institutes of Health). For immunofluorescence of endogenous Merm1, we used antibodies from Abcam (ab97911) at a concentration of 20 μg/ml. For immunofluorescence of GFP-tagged or HA-tagged Merm1, fluorescent signals were observed under the spinning-disk confocal laser scanning microscope (Nikon) and analyzed with Volocity 6.3 (PerkinElmer). ChIP and real-time quantitative PCR ChIP assays were performed as described (Zhou et al., 2002). Cross-linked chromatin was sonicated to yield fragments of 400−500 base pairs (bp). Then, the chromatin was pre-cleared, 10% input were reserved and the rest were incubated with respective antibodies. Immunoprecipitated proteins were captured with protein A/G Sepharose beads. After reversal of the crosslinking with the existence of RNase A and digestion of proteins with proteinase K, the DNA was purified by phenol–chloroform extraction or chelex-100 resin and amplified by real-time and quantitative PCR (RT-qPCR). The RT-qPCR was run on Light Cycler (Roche) with 2× SYBR Premix Ex Taq (RR420A, Takara). The conditions were set up according to the tutorial manual provided by the manufacturer. The PCR results were acquired and analyzed by normalization to the 10% input. The ChIP primers used in this study are listed in Supplementary Table S2. Co-immunoprecipitation and western blotting assays To prepare nuclear extract, nuclei were lysed in IP buffer (20 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% (v/v) Triton X-100) for 30 min at 4°C and cleared by centrifugation (16000 g, 15 min, 4°C). To prepare whole-cell extract, cells were lysed with RIPA buffer (10 mM Tris-HCl, pH 8.0, 1 mM EDTA, 0.5 mM EGTA, 1% Triton X-100, 0.1% sodium deoxycholate, 0.1% SDS, 140 mM NaCl and add 1 mM PMSF before use) for 1 min and the suspension was diluted 10 times with RIPA buffer without detergent, after incubation for 30 min at 4°C, the lysate was cleared by centrifugation (16000 g, 15 min, 4°C). Then, the lysate was reserved for 10% input and the rest were incubated with respective antibodies at 4°C overnight. Then the protein A/G beads were added to capture the proteins. After washing four times with IP buffer (for nuclear extract) or 0.1× RIPA buffer (for whole-cell extract), the beads carried immunoprecipitated proteins were boiled with 2× SDS loading buffer (80 mM Tris, pH 6.8, 10% glycerol, 2% SDS, 20% β-mercaptoethanol) and centrifuged at 12000 g for 5 min at room temperature. The supernatant and the input were analyzed by western blotting. The protein was subject to SDS-PAGE, transferred to nitrocellulose membrane and incubated with primary antibodies of interest at 4°C overnight. Primary antibodies used were mentioned above. The protein blots were then washed three times in PBST (1× PBS + 0.1% Tween-20) and incubated with diluted secondary antibodies for 2 h at room temperature. And then the blots were washed three times with PBST before visualized under odyssey infrared imaging system (Odyssey, LI-COR). DNA methylation assays To monitor CpG methylation, genomic DNA or ChIP-enriched DNA was treated with sodium bisulfite using EpiTect Bisulfite Kit (Qiagen). Briefly, the genomic DNA was subject to bisulfite conversion reaction. The converted DNA was then purified and amplified by PCR with specific primers. The amplicons were electrophoresed by agarose gel and the purified amplicons were ligated into the pEasy-T5 vector. The recombinant pEasy-T5 constructs were transduced into competent Escherichiacoli, which were plated and incubated at 37°C. At least 15 independent colonies of the E. coli were selected and sequenced. The sequencing results were analyzed by BiQ analyzer (Max Planck Institut Informatik). The primer information is provided in Supplementary Table S2. In vitro DNA methyltransferase activity assays For DNA methyltransferase activity assays, 0.1 μg purified His-tagged Dnmt3a or Dnmt3b was incubated with 0−0.2 μg purified GST-tagged Merm1 or His-tagged Dnmt3L at 37°C for 30 min in 10 μl buffer containing 20 mM Hepes, pH 7.5, 50 mM NaCl, 0.5 mg/ml BSA, 0.5 μCi tritiated SAM (3H-S-adenosyl methionine, PerkinElmer), and 100 ng substrate DNA. After the reaction, DNA was recovered and counted in a scintillation counter. Sequence of the substrate DNA is 5′-GATCGCCGATGCGCGAATCGCGATCGATGCGAT-3′. In addition, quantification of DNA methyltransferase activity was also performed using the EpiQuik DNA Methyltransferase Activity/Inhibition Assay Kit (Epigentek). Methylated DNA immunoprecipitation and next-generation sequencing MeDIP DNA libraries were prepared using the NEBNext DNA Library Prep Master Mix Set for Illumina (E6040, NEB). Briefly, fragmented DNA was end repaired and ligated with adapter. After size selection and purification using AMPure XP Beads (A6380, Beckman), the recovered DNA was denatured and incubated with the specific antibody against 5mC. Then, the immunoprecipitated fragments were captured with protein A/G agarose beads. After digestion of proteins with proteinase K, the DNA fragments were extracted and amplified by adaptor-mediated PCR. Libraries were sequenced using the Illumina HiSeq 2000 sequencing platform. The sequencing reads were aligned to mm9 using bowtie2 under the default parameters, then PCR duplicated reads were removed by Picard (MarkDuplicates). Samtools was used to select properly mapped reads (samtools view -Sb -h -f 2). Finally, peaks were detected by macs (–nomodel –nolambda -w –space = 30). ChIP-Seq Samples were processed for sequencing using NEBNext DNA Library Prep Master Mix Set for Illumina (E6040, NEB). In brief, ChIP-enriched DNA fragments were end repaired and ligated with adapter. After size selection using AMPure XP Beads, the library was amplified by adaptor-mediated PCR. Libraries were sequenced using the Illumina HiSeq 2000 sequencing platform. The alignment step was the same as MeDIP-seq, but the peaks calling were using masc under the default parameters (Landt et al., 2012). Supplementary material Supplementary material is available at Journal of Molecular Cell Biology online. Acknowledgements We thank I. Grummt (German Cancer Research Center, Germany) for the valuable discussions and advice. We thank Ruiqian Li and Ping Zhu (Biodynamic Optical Imaging Center, Peking University, China) for epigenome analysis, Guohong Li (Institute of Biophysics, Chinese Academy of Sciences, China) for assistance with DNA methylation assays. Funding This work was supported by the National Natural Science Foundation of China (31471205, 31171255, 91219101, and 81330009) and the National Basic Research Program of China (2013CB530700). Conflict of interest none declared. Author contributions L.Z. and G.L. designed and performed cellular, molecular, biochemical, and animal experiments; X.H., W.X., Y.G., J.F., and L.Z. contributed to cellular and biochemical experiments; C.Z. contributed to epigenome analysis; T.N. and J.G. provided critical reagents and contributed to data analysis; W.T. and T.N conceived the project and designed experiments; W.T., L.Z., and G.L wrote the manuscript. References Bestor, T., Laudano, A., Mattaliano, R., et al.  . ( 1988). Cloning and sequencing of a cDNA encoding DNA methyltransferase of mouse cells: the carboxyl-terminal domain of the mammalian enzymes is related to bacterial restriction methyltransferases. J. Mol. Biol.  203, 971– 983. Google Scholar CrossRef Search ADS PubMed  Bostick, M., Kim, J.K., Estève, P.-O., et al.  . ( 2007). UHRF1 plays a role in maintaining DNA methylation in mammalian cells. Science  317, 1760– 1764. Google Scholar CrossRef Search ADS PubMed  Bourc’his, D., Xu, G.-L., Lin, C.-S., et al.  . ( 2001). Dnmt3L and the establishment of maternal genomic imprints. Science  294, 2536– 2539. Google Scholar CrossRef Search ADS PubMed  Chen, Z.X., Mann, J.R., Hsieh, C.L., et al.  . ( 2005). Physical and functional interactions between the human DNMT3L protein and members of the de novo methyltransferase family. J. Cell. Biochem.  95, 902– 917. Google Scholar CrossRef Search ADS PubMed  Chédin, F., Lieber, M.R., and Hsieh, C.-L. ( 2002). The DNA methyltransferase-like protein DNMT3L stimulates de novo methylation by Dnmt3a. Proc. Natl Acad. Sci. USA  99, 16916– 16921. Google Scholar CrossRef Search ADS   Doll, A., and Grzeschik, K.H. ( 2001). Characterization of two novel genes, WBSCR20 and WBSCR22, deleted in Williams-Beuren syndrome. Cytogenet. Cell Genet.  95, 20– 27. Google Scholar CrossRef Search ADS PubMed  Duymich, C.E., Charlet, J., Yang, X., et al.  . ( 2016). DNMT3B isoforms without catalytic activity stimulate gene body methylation as accessory proteins in somatic cells. Nat. Commun.  7, 11453. Google Scholar CrossRef Search ADS PubMed  Egyhazi, E., Pigon, A., Chang, J.-H., et al.  . ( 1988). Effects of anti-C23 (nucleolin) antibody on transcription of ribosomal DNA in Chironomus salivary gland cells. Exp. Cell Res.  178, 264– 272. Google Scholar CrossRef Search ADS PubMed  Figaro, S., Wacheul, L., Schillewaert, S., et al.  . ( 2012). Trm112 is required for Bud23-mediated methylation of the 18S rRNA at position G1575. Mol. Cell. Biol.  32, 2254– 2267. Google Scholar CrossRef Search ADS PubMed  Gagnon-Kugler, T., Langlois, F., Stefanovsky, V., et al.  . ( 2009). Loss of human ribosomal gene CpG methylation enhances cryptic RNA polymerase II transcription and disrupts ribosomal RNA processing. Mol. Cell  35, 414– 425. Google Scholar CrossRef Search ADS PubMed  Gruenbaum, Y., Cedar, H., and Razin, A. ( 1982). Substrate and sequence specificity of a eukaryotic DNA methylase. Nature  295, 620– 622. Google Scholar CrossRef Search ADS PubMed  Grummt, I. ( 1998). Regulation of mammalian ribosomal gene transcription by RNA polymerase I. Prog. Nucleic Acid Res. Mol. Biol.  62, 109– 154. Google Scholar CrossRef Search ADS   Grummt, I., and Längst, G. ( 2013). Epigenetic control of RNA polymerase I transcription in mammalian cells. Biochim. Biophys. Acta  1829, 393– 404. Google Scholar CrossRef Search ADS PubMed  Haag, S., Kretschmer, J., and Bohnsack, M.T. ( 2015). WBSCR22/Merm1 is required for late nuclear pre-ribosomal RNA processing and mediates N7-methylation of G1639 in human 18S rRNA. RNA  21, 180– 187. Google Scholar CrossRef Search ADS PubMed  He, Y.-F., Li, B.-Z., Li, Z., et al.  . ( 2011). Tet-mediated formation of 5-carboxylcytosine and its excision by TDG in mammalian DNA. Science  333, 1303– 1307. Google Scholar CrossRef Search ADS PubMed  Hsieh, C.-L. ( 1999). In vivo activity of murine de novo methyltransferases, Dnmt3a and Dnmt3b. Mol. Cell. Biol.  19, 8211– 8218. Google Scholar CrossRef Search ADS PubMed  Jangani, M., Poolman, T.M., Matthews, L., et al.  . ( 2014). The methyltransferase WBSCR22/Merm1 enhances glucocorticoid receptor function and is regulated in lung inflammation and cancer. J. Biol. Chem.  289, 8931– 8946. Google Scholar CrossRef Search ADS PubMed  Landt, S.G., Marinov, G.K., Kundaje, A., et al.  . ( 2012). ChIP-seq guidelines and practices of the ENCODE and modENCODE consortia. Genome Res.  22, 1813– 1831. Google Scholar CrossRef Search ADS PubMed  Leonhardt, H., Page, A.W., Weier, H.-U., et al.  . ( 1992). A targeting sequence directs DNA methyltransferase to sites of DNA replication in mammalian nuclei. Cell  71, 865– 873. Google Scholar CrossRef Search ADS PubMed  Li, E., Bestor, T.H., and Jaenisch, R. ( 1992). Targeted mutation of the DNA methyltransferase gene results in embryonic lethality. Cell  69, 915– 926. Google Scholar CrossRef Search ADS PubMed  Li, H., Rauch, T., Chen, Z.X., et al.  . ( 2006). The histone methyltransferase SETDB1 and the DNA methyltransferase DNMT3A interact directly and localize to promoters silenced in cancer cells. J. Biol. Chem.  281, 19489– 19500. Google Scholar CrossRef Search ADS PubMed  Létoquart, J., Huvelle, E., Wacheul, L., et al.  . ( 2014). Structural and functional studies of Bud23–Trm112 reveal 18S rRNA N7-G1575 methylation occurs on late 40S precursor ribosomes. Proc. Natl Acad. Sci. USA  111, E5518– E5526. Google Scholar CrossRef Search ADS   Majumder, S., Ghoshal, K., Datta, J., et al.  . ( 2006). Role of DNA methyltransferases in regulation of human ribosomal RNA gene transcription. J. Biol. Chem.  281, 22062– 22072. Google Scholar CrossRef Search ADS PubMed  McStay, B., and Grummt, I. ( 2008). The epigenetics of rRNA genes: from molecular to chromosome biology. Annu. Rev. Cell Dev. Biol.  24, 131– 157. Google Scholar CrossRef Search ADS PubMed  Nakazawa, Y., Arai, H., and Fujita, N. ( 2011). The novel metastasis promoter Merm1/Wbscr22 enhances tumor cell survival in the vasculature by suppressing Zac1/p53-dependent apoptosis. Cancer Res.  71, 1146– 1155. Google Scholar CrossRef Search ADS PubMed  Neri, F., Krepelova, A., Incarnato, D., et al.  . ( 2013). Dnmt3L antagonizes DNA methylation at bivalent promoters and favors DNA methylation at gene bodies in ESCs. Cell  155, 121– 134. Google Scholar CrossRef Search ADS PubMed  Okano, M., Bell, D.W., Haber, D.A., et al.  . ( 1999). DNA methyltransferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development. Cell  99, 247– 257. Google Scholar CrossRef Search ADS PubMed  Okano, M., Xie, S., and Li, E. ( 1998). Cloning and characterization of a family of novel mammalian DNA (cytosine-5) methyltransferases. Nat. Genet.  19, 219– 220. Google Scholar CrossRef Search ADS PubMed  Piano, F., Schetter, A.J., Morton, D.G., et al.  . ( 2002). Gene clustering based on RNAi phenotypes of ovary-enriched genes in C. elegans. Curr. Biol.  12, 1959– 1964. Google Scholar CrossRef Search ADS PubMed  Roger, B., Moisand, A., Amalric, F., et al.  . ( 2002). Repression of RNA polymerase I transcription by nucleolin is independent of the RNA sequence that is transcribed. J. Biol. Chem.  277, 10209– 10219. Google Scholar CrossRef Search ADS PubMed  Santi, D.V., Garrett, C.E., and Barr, P.J. ( 1983). On the mechanism of inhibition of DNA-cytosine methyltransferases by cytosine analogs. Cell  33, 9– 10. Google Scholar CrossRef Search ADS PubMed  Sharif, J., Muto, M., Takebayashi, S.-I., et al.  . ( 2007). The SRA protein Np95 mediates epigenetic inheritance by recruiting Dnmt1 to methylated DNA. Nature  450, 908– 912. Google Scholar CrossRef Search ADS PubMed  Suetake, I., Shinozaki, F., Miyagawa, J., et al.  . ( 2004). DNMT3L stimulates the DNA methylation activity of Dnmt3a and Dnmt3b through a direct interaction. J. Biol. Chem.  279, 27816– 27823. Google Scholar CrossRef Search ADS PubMed  Tafforeau, L., Zorbas, C., Langhendries, J.-L., et al.  . ( 2013). The complexity of human ribosome biogenesis revealed by systematic nucleolar screening of pre-rRNA processing factors. Mol. Cell  51, 539– 551. Google Scholar CrossRef Search ADS PubMed  Wu, H., Coskun, V., Tao, J., et al.  . ( 2010). Dnmt3a-dependent nonpromoter DNA methylation facilitates transcription of neurogenic genes. Science  329, 444– 448. Google Scholar CrossRef Search ADS PubMed  Xie, S., Wang, Z., Okano, M., et al.  . ( 1999). Cloning, expression and chromosome locations of the human DNMT3 gene family. Gene  236, 87– 95. Google Scholar CrossRef Search ADS PubMed  Yen, R.-W.C., Vertino, P.M., Nelkin, B.D., et al.  . ( 1992). Isolation and characterization of the cDNA encoding human DNA methyltransferase. Nucleic Acids Res.  20, 2287– 2291. Google Scholar CrossRef Search ADS PubMed  Zhou, Y., Santoro, R., and Grummt, I. ( 2002). The chromatin remodeling complex NoRC targets HDAC1 to the ribosomal gene promoter and represses RNA polymerase I transcription. EMBO J.  21, 4632– 4640. Google Scholar CrossRef Search ADS PubMed  Zorbas, C., Nicolas, E., Wacheul, L., et al.  . ( 2015). The human 18S rRNA base methyltransferases DIMT1L and WBSCR22-TRMT112 but not rRNA modification are required for ribosome biogenesis. Mol. Biol. Cell  26, 2080– 2095. Google Scholar CrossRef Search ADS PubMed  Õunap, K., Käsper, L., Kurg, A., et al.  . ( 2013). The human WBSCR22 protein is involved in the biogenesis of the 40S ribosomal subunits in mammalian cells. PLoS One  8, e75686. Google Scholar CrossRef Search ADS PubMed  Õunap, K., Leetsi, L., Matsoo, M., et al.  . ( 2015). The stability of ribosome biogenesis factor WBSCR22 is regulated by interaction with TRMT112 via ubiquitin-proteasome pathway. PLoS One  10, e0133841. Google Scholar CrossRef Search ADS PubMed  © The Author(s) (2018). Published by Oxford University Press on behalf of Journal of Molecular Cell Biology, IBCB, SIBS, CAS. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Molecular Cell Biology Oxford University Press

Metastasis-related methyltransferase 1 (Merm1) represses the methyltransferase activity of Dnmt3a and facilitates RNA polymerase I transcriptional elongation

Loading next page...
 
/lp/ou_press/metastasis-related-methyltransferase-1-merm1-represses-the-2MZog08v4C
Publisher
Oxford University Press
Copyright
© The Author(s) (2018). Published by Oxford University Press on behalf of Journal of Molecular Cell Biology, IBCB, SIBS, CAS. All rights reserved.
ISSN
1674-2788
eISSN
1759-4685
D.O.I.
10.1093/jmcb/mjy023
Publisher site
See Article on Publisher Site

Abstract

Abstract Stimulatory regulators for DNA methyltransferase activity, such as Dnmt3L and some Dnmt3b isoforms, affect DNA methylation patterns, thereby maintaining gene body methylation and maternal methylation imprinting, as well as the methylation landscape of pluripotent cells. Here we show that metastasis-related methyltransferase 1 (Merm1), a protein deleted in individuals with Williams–Beuren syndrome, acts as a repressive regulator of Dnmt3a. Merm1 interacts with Dnmt3a and represses its methyltransferase activity with the requirement of the binding motif for S-adenosyl-L-methionine. Functional analysis of gene regulation revealed that Merm1 is capable of maintaining hypomethylated rRNA gene bodies and co-localizes with RNA polymerase I in the nucleolus. Dnmt3a recruits Merm1, and in return, Merm1 ensures the binding of Dnmt3a to hypomethylated gene bodies. Such interplay between Dnmt3a and Merm1 facilitates transcriptional elongation by RNA polymerase I. Our findings reveal a repressive factor for Dnmt3a and uncover a molecular mechanism underlying transcriptional elongation of rRNA genes. Merm1, gene body methylation, Dnmt3a, transcriptional elongation, rRNA genes Introduction Methylation of cytosine bases at CpG dinucleotide is an important regulatory modification in eukaryotic genomes. Three catalytically active DNA methyltransferases (Dnmts), Dnmt1, Dnmt3a, and Dnmt3b, have been identified in mammals (Bestor et al., 1988; Yen et al., 1992; Okano et al., 1998; Xie et al., 1999). Dnmt1 functions predominantly in maintaining DNA methylation patterns (Gruenbaum et al., 1982; Leonhardt et al., 1992), while Dnmt3a and Dnmt3b are responsible for de novo methylation of unmethylated and hemimethylated DNA (Hsieh, 1999; Okano et al., 1999). Several members of the Dnmt3 family, including DNA methyltransferase 3-like protein Dnmt3L and Dnmt3b isoforms without sequence motifs necessary for methyltransferase activity, have been shown to act as stimulatory factors to facilitate DNA methyltransferase activity (Chédin et al., 2002; Suetake et al., 2004; Duymich et al., 2016). Dnmt3L has been identified as a regulatory protein of Dnmt3a and Dnmt3b that stimulates de novo methylation, thereby maintaining gene body methylation, maternal methylation imprinting, and the methylation landscape of pluripotent cells (Bourc’his et al., 2001; Neri et al., 2013). Removal of Dnmt3L prevents methylation of sequences that are normally maternally methylated and causes developmental defects in mouse embryos (Bourc’his et al., 2001). Catalytically inactive Dnmt3b isoforms are able to collaborate with Dnmt3a to restore DNA methylation, especially in gene bodies in differentiated cells (Duymich et al., 2016). In addition, Uhrf1 has been recognized as a Dnmt1 accessory protein that binds to hemimethylated CpG and recruits Dnmt1 to ensure the DNA methylation pattern during DNA replication or repression of its direct target genes (Bostick et al., 2007; Sharif et al., 2007). Regulatory proteins for DNA methyltransferases play critical roles in controlling DNA methylation and gene expression patterns, and identification of new regulatory proteins for both stimulatory and repressive regulators of DNA methyltransferases will provide further insight into mechanisms of gene expression and mammalian development, as well as human disease. Metastasis-related methyltransferase 1 (Merm1) was originally identified as Wbscr22. It is one of 26–28 genes that are deleted from 7q11.23 in Williams–Beuren syndrome, which is characterized by distinctive facial features, mental retardation, hypercalcemia, and hypertension (Doll and Grzeschik, 2001). Merm1 is conserved in yeast and human. This gene encodes a protein containing a nuclear localization signal and an S-adenosyl-L-methionine (SAM) binding motif typical of methyltransferases. Initially, Merm1 was shown to be implicated in methylation and processing of 18S rRNA, and its stability was found to be regulated by TRMT112 through the ubiquitin–proteasome pathway (Figaro et al., 2012; Õunap et al., 2013; Tafforeau et al., 2013; Haag et al., 2015; Õunap et al., 2015; Zorbas et al., 2015). Further investigation revealed that Merm1 plays a critical role in epigenetic regulation of gene expression in mammals. Merm1 regulates glucocorticoid receptor recruitment to the genome and mediates subsequent histone modification to maintain open chromatin (Jangani et al., 2014), suggesting that it plays a role in chromatin-based gene expression. In addition, Merm1 promotes cancer metastasis by inhibiting Zac1-mediated p53-dependent apoptosis, in which Merm1 methylates histone H3 lysine 9 (H3K9) at the Zac1 locus, thereby producing a transcriptionally repressive chromatin environment (Nakazawa et al., 2011). Thus, Merm1 is able to drive epigenetic alterations to affect gene activities. Merm1 itself does not have active histone or DNA methyltransferase activities (Nakazawa et al., 2011); therefore, its role in gene expression may be to function as a recruiter or modulator for epigenetic enzymes. Here we report for the first time that Merm1 acts as a repressive regulator of Dnmt3a and participates in regulating gene body methylation and RNA polymerase I (Pol I) transcription elongation at rRNA genes. We found that Dnmt3a binds to unmethylated rDNA gene bodies because of inhibition of its DNA methyltransferase activity by Merm1. Through direct binding to Dnmt3a, Merm1 maintains gene body hypomethylation, leading to co-occupancy with Dnmt3a, Pol I, and upstream binding factor (UBF) at unmethylated gene bodies. This complex is capable of ensuring efficient Pol I elongation on rDNA templates. Thus, Merm1 is a repressive regulator of Dnmt3a that participates in generating unmethylated gene bodies that allow transcriptional elongation by Pol I. Results Merm1 interacts with Dnmt3a and inhibits its DNA methyltransferase activity To assess the biological functions of Merm1, co-immunoprecipitation was performed to explore its binding partners. Surprisingly, Dnmt3a and Merm1 immunoprecipitated each other in vivo (Figure 1A). Furthermore, co-immunoprecipitation assays utilizing nuclear extracts from 293T cells harboring ectopic co-expression of HA-tagged Merm1 and Myc-tagged Dnmts confirmed that Merm1 precipitated Dnmt3a and Dnmt3b (Figure 1B). To assess the physical interaction between these two proteins, recombinant GST-tagged human Merm1 and Dnmts were prokaryotically purified and subjected to pull-down assays. Merm1 interacted directly with Dnmt3a and Dnmt3b, but not Dnmt1 (Figure 1C). These results demonstrated that Merm1 interacts directly with Dnmt3a and Dnmt3b. Figure 1 View largeDownload slide Merm1 binds to Dnmt3a and inhibits its DNA methyltransferase activity. (A) Dnmt3a interacts with Merm1 in vivo. Whole-cell extracts from HEK293T (293T) cells were incubated with antibodies against Merm1 or Dnmt3a. Co-immunoprecipitated proteins were monitored by immunoblotting using the indicated antibodies. (B) Merm1 directly interacts with Dnmt3a and Dnmt3b in vitro. Purified GST-Merm1 and GST-Dnmts were incubated with an antibody against Merm1. Co-precipitation of Dnmts was assessed by immunoblotting. Arrows or asterisks indicate the position of Dnmt3a or Dnmt3b, respectively. (C) Merm1 interacts with both Dnmt3a and Dnmt3b. Nuclear extracts from 293T cells expressing HA-Merm1 and Myc-Dnmts were incubated with an antibody against the HA epitope. Co-precipitated Dnmts were monitored by immunoblotting. Ten percent of the input is shown. (D) Merm1 inhibits the DNA methyltransferase activity of Dnmt3a. In vitro DNA methyltransferase activity assays were performed using purified proteins and synthetic substrate DNA. Error bars represent SD (n = 3). (E) Schematic illustration of the HA-tagged full-length (HA-Merm1), four glycines (4 G) to four arginines (4 R) substitution mutation (HA-Merm1-4G-4R*), N-terminal domain (HA-Merm1-N), and C-terminal domain (HA-Merm1-C) truncation mutants of Merm1. The numbers are the amino acid sequence numbers of Merm1. I, IV, IX, and X are conserved motifs of AdoMet-MTases. NLS, nuclear localization signal. (F) The N-terminal domain of Merm1 interacts with Dnmt3a in vitro. Whole-cell extracts from 293T cells expressing either the HA-tagged N-terminal domain or C-terminal domain of Merm1 were incubated with immobilized GST–Dnmt3a fusion proteins. Bound proteins were analyzed by immunoblotting using an anti-HA antibody. (G) The N-terminal domain of Merm1 binds to Dnmt3a. Whole-cell extracts from 293T cells expressing either the HA-tagged N-terminal domain or C-terminal domain of Merm1 were incubated with an antibody against Dnmt3a. Co-precipitated proteins were monitored by immunoblotting using an anti-HA antibody. (H) Mutation of the predicted SAM binding motif weakens the interaction between Merm1 and Dnmt3a. Nuclear extracts from 293T cells expressing HA-Merm1-4G-4R* and Myc-Dnmt3a were incubated with antibodies against HA or Myc. Co-precipitated proteins were monitored by immunoblotting. (I) Both the N-terminal and C-terminal domains of Merm1 are required for efficient inhibition of the methyltransferase activity of Dnmt3a. Whole-cell extracts from 293T cells expressing the indicated HA-tagged proteins were used to perform DNA methyltransferase activity assays. *P < 0.1, **P < 0.01, ***P < 0.001, two-tailed unpaired Student’s t-tests were performed. Figure 1 View largeDownload slide Merm1 binds to Dnmt3a and inhibits its DNA methyltransferase activity. (A) Dnmt3a interacts with Merm1 in vivo. Whole-cell extracts from HEK293T (293T) cells were incubated with antibodies against Merm1 or Dnmt3a. Co-immunoprecipitated proteins were monitored by immunoblotting using the indicated antibodies. (B) Merm1 directly interacts with Dnmt3a and Dnmt3b in vitro. Purified GST-Merm1 and GST-Dnmts were incubated with an antibody against Merm1. Co-precipitation of Dnmts was assessed by immunoblotting. Arrows or asterisks indicate the position of Dnmt3a or Dnmt3b, respectively. (C) Merm1 interacts with both Dnmt3a and Dnmt3b. Nuclear extracts from 293T cells expressing HA-Merm1 and Myc-Dnmts were incubated with an antibody against the HA epitope. Co-precipitated Dnmts were monitored by immunoblotting. Ten percent of the input is shown. (D) Merm1 inhibits the DNA methyltransferase activity of Dnmt3a. In vitro DNA methyltransferase activity assays were performed using purified proteins and synthetic substrate DNA. Error bars represent SD (n = 3). (E) Schematic illustration of the HA-tagged full-length (HA-Merm1), four glycines (4 G) to four arginines (4 R) substitution mutation (HA-Merm1-4G-4R*), N-terminal domain (HA-Merm1-N), and C-terminal domain (HA-Merm1-C) truncation mutants of Merm1. The numbers are the amino acid sequence numbers of Merm1. I, IV, IX, and X are conserved motifs of AdoMet-MTases. NLS, nuclear localization signal. (F) The N-terminal domain of Merm1 interacts with Dnmt3a in vitro. Whole-cell extracts from 293T cells expressing either the HA-tagged N-terminal domain or C-terminal domain of Merm1 were incubated with immobilized GST–Dnmt3a fusion proteins. Bound proteins were analyzed by immunoblotting using an anti-HA antibody. (G) The N-terminal domain of Merm1 binds to Dnmt3a. Whole-cell extracts from 293T cells expressing either the HA-tagged N-terminal domain or C-terminal domain of Merm1 were incubated with an antibody against Dnmt3a. Co-precipitated proteins were monitored by immunoblotting using an anti-HA antibody. (H) Mutation of the predicted SAM binding motif weakens the interaction between Merm1 and Dnmt3a. Nuclear extracts from 293T cells expressing HA-Merm1-4G-4R* and Myc-Dnmt3a were incubated with antibodies against HA or Myc. Co-precipitated proteins were monitored by immunoblotting. (I) Both the N-terminal and C-terminal domains of Merm1 are required for efficient inhibition of the methyltransferase activity of Dnmt3a. Whole-cell extracts from 293T cells expressing the indicated HA-tagged proteins were used to perform DNA methyltransferase activity assays. *P < 0.1, **P < 0.01, ***P < 0.001, two-tailed unpaired Student’s t-tests were performed. Bioinformatic sequence alignments showed that Merm1 contains several domains (I, IV, IX, and X) that are conserved in canonical Dnmts such as Dnmt1, Dnmt3a, and Dnmt3b (Supplementary Figure S1A). Similar to Dnmt3L, Merm1 does not contain a complete set of canonical DNA methyltransferase domains, and knockdown of Merm1 did not affect histone modification levels, indicating that it is not a histone or DNA methyltransferase enzyme (Supplementary Figures S1B and S2). Given that Merm1 interacts with Dnmt3a and Dnmt3b, we reasoned that it may be involved in modulating the DNA methyltransferase activity of known Dnmts. The regulatory effects of Merm1 on the activity levels of DNA methyltransferase Dnmt3a and Dnmt3b were assessed via the 3H-labeled S-adenosyl methionine (SAM) addition method in vitro. The liquid scintillation analyzer results revealed that Merm1 inhibited the methyltransferase activity of Dnmt3a without affecting that of Dnmt3b, although Merm1 and Dnmt3b were capable of direct interaction (Figure 1D and Supplementary Figure S3). Merm1 has a critical methyltransferase-related domain in its N-terminal and C-terminal regions. The N-terminal domain possesses a conserved motif (motif I) that consists of four glycines (4G), which are predicted to be responsible for binding of the methyl group donor SAM and cofactors (Figure 1E and Supplementary Figure S1A) (Nakazawa et al., 2011). To determine which domain of Merm1 is essential for its interaction with Dnmt3a and inhibitory effect on Dnmt3a activity, two truncation mutants (Merm1-N and Merm1-C) of HA-tagged Merm1 were generated and transiently overexpressed in 293T cells. The Merm1-N mutant was efficiently pulled down by the purified GST–Dnmt3a fusion protein, whereas relatively less of the Merm1-C mutant was precipitated by Dnmt3a (Figure 1F). Similarly, immunoprecipitation analysis using specific Dnmt3a antibodies showed that the Merm1-N mutant co-precipitated Dnmt3a with higher affinity than that with which it co-precipitated the Merm1-C mutant (Figure 1G). These results suggest that Merm1-N is required for the interaction between Merm1 and Dnmt3a. Considering that Merm1-N contains the SAM binding motif of Merm1, we constructed a substitution mutant (Merm1-4G4R*) with a disrupted SAM binding motif structure, which impaired SAM binding (Nakazawa et al., 2011), to test whether this motif contributed to the interaction between Merm1 and Dnmt3a. Co-immunoprecipitation analysis with Myc-tagged Dnmt3a, HA-tagged Merm1, or the 4G4R* mutant showed that the 4G4R* mutant significantly weakened the interaction between Merm1 and Dnmt3a (Figure 1H). Consistent with this finding, the inhibitory effect of Merm1 on the methyltransferase activity of Dnmt3a was mildly alleviated in the HA-Merm1-4G4R* substitution mutant and severely disrupted in the HA-Merm1-N truncation mutant (Figure 1H). Notably, although the C-terminal domain was not indispensable in the binding of Merm1 to Dnmt3a, it was required for efficient repression of the methyltransferase activity of Dnmt3a (Figure 1I), suggesting that the integrity of Merm1 is a prerequisite for its inhibitory effect on Dnmt3a. Taken together, these findings suggest that Merm1 interacts with Dnmt3a and inhibits its DNA methyltransferase activity in a manner requiring the SAM binding motif at the N-terminus of Merm1. Merm1 is capable of maintaining hypomethylated rDNA gene bodies Next, we tested whether the regulatory effect of Merm1 on Dnmt3a is involved in regulating gene expression. Previously, both Merm1 and Dnmt3a have been shown to localize in the nucleolus, and Dnmt3a has been shown to associate directly with rDNA (Majumder et al., 2006). Therefore, we assessed whether Dnmt3a and Merm1 function together to regulate the epigenetic state of rDNA and thereby regulate rDNA transcription. To monitor the binding of Merm1 to rDNA, we performed chromatin immunoprecipitation (ChIP) assays comparing Merm1 occupancy with that of Pol I. Merm1 binding was observed throughout the rDNA repeats, but it was enriched at rDNA gene bodies (Figure 2A). Although the overall level of rDNA-associated Merm1 was lower than that of Pol I, it was much higher than that of IgG (Supplementary Figure S4). In addition, association of Merm1 with gene bodies was significantly inhibited due to decreased expression of Merm1 caused by deleting one allele of Merm1 (designated as Merm1+/−) via CRISPR/Cas9 genome editing (Figure 2B), underscoring the specificity of the ChIP assays. Of note, we were unable to obtain homozygous deletion of Merm1 (Merm1−/−) in cells or mice, likely because of lethal effects caused by the complete loss of Merm1. Consistent with this finding, RNAi-mediated knockdown of the Merm1 ortholog in Caenorhabditis elegans was embryonic-lethal (Piano et al., 2002). Figure 2 View largeDownload slide Merm1 maintains unmethylated rDNA gene bodies. (A) Merm1 is associated with rDNA. ChIP assays were performed to show the rDNA occupancy of Merm1 and Pol I (RPA194) in 293T cells. The scheme indicates the positions of the primers used to amplify regions of rDNA. The primer sequences are listed in Supplementary Table S2. Error bars denote SD (n = 3). (B) ChIP assays were conducted in parallel with chromatin from wild-type 293T control cells (Ctrl) or Merm1+/− cells (#1; #2) (left panel). Error bars denote SD (n = 3). Protein levels of Merm1 in wild-type 293T (Ctrl) and Merm1-knockout cells (#1; #2) were tested by immunoblotting (right panel). Tubulin served as a loading control. (C) Loss of Merm1 increases gene body methylation. Bisulfite sequencing was performed to assess the DNA methylation status of rDNA gene body regions (+8017/+8224, the same as region C indicated in A) in 293T and Merm1+/− cells. Methylated and unmethylated CpG sites are shown as black and white circles, respectively. The numbers refer to the positions of the outermost CpG residues. (D) Analysis of the results in C showed the percentage of methylated and unmethylated gene bodies upon depletion of Merm1. (E) Depletion of Merm1 leads to a significant increase in DNA methylation in gene body regions, but not IGS. NIH3T3 cells were infected with retroviruses encoding Merm1-specific shRNA (shMerm1) or control shRNA (shCtrl) for 5 days. Upper panel: 5mC MeDIP-seq profiles for rDNA. Lower panel: overlay tracks of 5mC MeDIP-seq data from control and Merm1-depleted NIH3T3 cells. The locations of the 18S and 28S rRNA coding regions are shown at the bottom. *P < 0.1, **P < 0.01, ***P < 0.001, two-tailed unpaired Student’s t-tests were performed. Figure 2 View largeDownload slide Merm1 maintains unmethylated rDNA gene bodies. (A) Merm1 is associated with rDNA. ChIP assays were performed to show the rDNA occupancy of Merm1 and Pol I (RPA194) in 293T cells. The scheme indicates the positions of the primers used to amplify regions of rDNA. The primer sequences are listed in Supplementary Table S2. Error bars denote SD (n = 3). (B) ChIP assays were conducted in parallel with chromatin from wild-type 293T control cells (Ctrl) or Merm1+/− cells (#1; #2) (left panel). Error bars denote SD (n = 3). Protein levels of Merm1 in wild-type 293T (Ctrl) and Merm1-knockout cells (#1; #2) were tested by immunoblotting (right panel). Tubulin served as a loading control. (C) Loss of Merm1 increases gene body methylation. Bisulfite sequencing was performed to assess the DNA methylation status of rDNA gene body regions (+8017/+8224, the same as region C indicated in A) in 293T and Merm1+/− cells. Methylated and unmethylated CpG sites are shown as black and white circles, respectively. The numbers refer to the positions of the outermost CpG residues. (D) Analysis of the results in C showed the percentage of methylated and unmethylated gene bodies upon depletion of Merm1. (E) Depletion of Merm1 leads to a significant increase in DNA methylation in gene body regions, but not IGS. NIH3T3 cells were infected with retroviruses encoding Merm1-specific shRNA (shMerm1) or control shRNA (shCtrl) for 5 days. Upper panel: 5mC MeDIP-seq profiles for rDNA. Lower panel: overlay tracks of 5mC MeDIP-seq data from control and Merm1-depleted NIH3T3 cells. The locations of the 18S and 28S rRNA coding regions are shown at the bottom. *P < 0.1, **P < 0.01, ***P < 0.001, two-tailed unpaired Student’s t-tests were performed. Given that Merm1 was enriched at gene bodies, its regulatory role on epigenetic modifications on gene bodies was assessed. Merm1 affected rDNA gene body methylation in wild-type and Merm1+/− cells. Bisulfite sequencing analysis showed extensive gene body methylation at rDNA gene bodies, and depletion of Merm1 significantly increased the proportion of methylated gene bodies from 43% to 78% (Figure 2C and D). To verify that Merm1 regulates gene body methylation, methylated DNA immunoprecipitation sequencing (MeDIP-seq) assays were performed. Knockdown of Merm1 increased the level of gene body methylation, but DNA methylation at intergenic spacers (IGS) remained unchanged (Figure 2E), which confirmed the validity of the MeDIP-seq assay. These results indicate that Merm1 maintains hypomethylated rRNA gene bodies. Merm1 is recruited by Dnmt3a and ensures the binding of Dnmt3a to hypomethylated gene bodies via co-occupation with Pol I and UBF The finding that Merm1 inhibits the activity of Dnmt3a and maintains hypomethylated rDNA gene bodies raises the possibility that both Dnmt3a and Merm1  bind to hypomethylated gene bodies. Indeed, the ChIP-bisulfite assays revealed that Merm1 was associated exclusively with unmethylated gene bodies, whereas Dnmt1 was only associated with hypermethylated rDNA, and Pol I and UBF bound to unmethylated rDNA as expected (Figure 3A and B). However, Dnmt3a bound to both methylated and unmethylated gene bodies, but was preferentially associated with the latter (Figure 3C). Thus, Pol I, Merm1, and Dnmt3a are all restricted to unmethylated gene bodies, suggesting the possibility that these proteins may co-occupy gene bodies. Sequential-ChIP assays confirmed that Merm1 co-occupied with Dnmt3a, Pol I, and UBF, but not with Dnmt1 (Figure 3D), suggesting the co-occupancy on hypomethylated rDNA gene bodies coordinates rDNA transcription elongation. Figure 3 View largeDownload slide Merm1 ensures the binding of Dnmt3a to unmethylated gene bodies via co-occupation with UBF and Pol I. (A) Merm1 is associated with hypomethylated rDNA, similar to Pol I and UBF (upper panel), while Dnmt1 is associated with hypermethylated rDNA (lower panel). ChIP-enriched DNA from 293T cells using the indicated antibodies was used to perform bisulfite sequencing. Human rDNA sequences (+8017/+8224) were amplified, cloned, and sequenced. Methylated and unmethylated CpG sites are shown as black and white circles, respectively. The numbers refer to the positions of the outermost CpG residues. (B) Analysis of the results in A showed the percentage of methylated and unmethylated gene bodies associated with Pol I, UBF, Merm1, and Dnmt1. (C) Dnmt3a is preferentially associated with unmethylated rDNA. (D) Sequential-ChIP assays show that Merm1 co-localizes with Dnmt3a, Pol I, and UBF. Cross-linked chromatin from 293T cells was immunoprecipitated with an antibody against Merm1 (first ChIP), followed by precipitation with antibodies against Dnmt1, Dnmt3a, Pol I (RPA194), and UBF (second ChIP). The co-precipitated DNA was analyzed by PCR. (E) Reduction of Merm1 abundance increases Dnmt3a occupancy at gene bodies. ChIP assays performed using primers amplifying the 28S rRNA coding region were used to show the relative rDNA occupancy of Dnmt3a in 293T and Merm1+/− cells. Error bars represent SD (n = 3). (F) Analysis of the results in E showed increased binding of Dnmt3a to methylated gene bodies in Merm1 knockdown cells. (G) Loss of Dnmt3a reduces the occupancy of Merm1 at gene bodies. ChIP assays were performed using primers amplifying the 28S rRNA coding region to show the relative rDNA occupancy of Merm1 in 293T and Dnmt3a−/− cells. (H) Dnmt3a is associated exclusively with hypermethylated gene bodies after reduction of the abundance of Merm1. ChIP-enriched DNA from 293T or Merm1+/− cells obtained using an antibody against Dnmt3a was used to perform bisulfite sequencing. Error bars represent SD (n = 3). *P < 0.1, two-tailed unpaired Student’s t-tests were performed. Figure 3 View largeDownload slide Merm1 ensures the binding of Dnmt3a to unmethylated gene bodies via co-occupation with UBF and Pol I. (A) Merm1 is associated with hypomethylated rDNA, similar to Pol I and UBF (upper panel), while Dnmt1 is associated with hypermethylated rDNA (lower panel). ChIP-enriched DNA from 293T cells using the indicated antibodies was used to perform bisulfite sequencing. Human rDNA sequences (+8017/+8224) were amplified, cloned, and sequenced. Methylated and unmethylated CpG sites are shown as black and white circles, respectively. The numbers refer to the positions of the outermost CpG residues. (B) Analysis of the results in A showed the percentage of methylated and unmethylated gene bodies associated with Pol I, UBF, Merm1, and Dnmt1. (C) Dnmt3a is preferentially associated with unmethylated rDNA. (D) Sequential-ChIP assays show that Merm1 co-localizes with Dnmt3a, Pol I, and UBF. Cross-linked chromatin from 293T cells was immunoprecipitated with an antibody against Merm1 (first ChIP), followed by precipitation with antibodies against Dnmt1, Dnmt3a, Pol I (RPA194), and UBF (second ChIP). The co-precipitated DNA was analyzed by PCR. (E) Reduction of Merm1 abundance increases Dnmt3a occupancy at gene bodies. ChIP assays performed using primers amplifying the 28S rRNA coding region were used to show the relative rDNA occupancy of Dnmt3a in 293T and Merm1+/− cells. Error bars represent SD (n = 3). (F) Analysis of the results in E showed increased binding of Dnmt3a to methylated gene bodies in Merm1 knockdown cells. (G) Loss of Dnmt3a reduces the occupancy of Merm1 at gene bodies. ChIP assays were performed using primers amplifying the 28S rRNA coding region to show the relative rDNA occupancy of Merm1 in 293T and Dnmt3a−/− cells. (H) Dnmt3a is associated exclusively with hypermethylated gene bodies after reduction of the abundance of Merm1. ChIP-enriched DNA from 293T or Merm1+/− cells obtained using an antibody against Dnmt3a was used to perform bisulfite sequencing. Error bars represent SD (n = 3). *P < 0.1, two-tailed unpaired Student’s t-tests were performed. Significantly, ChIP-bisulfite assays carried out in Merm1+/− cells revealed that upon depletion of Merm1, Dnmt3a was no longer bound to hypomethylated rDNA gene bodies, but rather was exclusively associated with hypermethylated rDNA gene bodies. However, the association between Pol I and hypomethylated gene bodies was unchanged in Merm1+/− cells (Figure 3E and F; Supplementary Figure S5). These results suggest that association of Dnmt3a with hypomethylated rDNA gene bodies was caused by the repressive effect of Merm1 on Dnmt3a, because depletion of Merm1 relieved this inhibitory effect and allowed Dnmt3a to bind to hypermethylated gene bodies. Moreover, neither loss of Dnmt3a nor concurrent disruption of Merm1 and Dnmt3a (Merm1+/−/Dnmt3a−/−, also denoted as DKO) decreased gene body methylation (Supplementary Figure S6), indicating that depletion of Dnmt3a leads to disassociation of the Merm1−Dnmt3a complex and departure of Merm1 from gene bodies. Indeed, depletion of Dnmt3a impaired the association between Merm1 and rDNA (Figure 3G), whereas knockdown of Merm1 enhanced the binding of Dnmt3a to rDNA (Figure 3H), suggesting that Dnmt3a guides Merm1 to rDNA gene bodies. Taken together, these findings demonstrate that Merm1 is recruited by Dnmt3a, thereby establishing hypomethylated gene bodies where Dnmt3a, Pol I, and UBF show co-occupancy. Merm1 co-localizes with Pol I and UBF Given that Merm1, UBF, and Pol I bind to and co-occupy unmethylated gene bodies, we assumed that Merm1 may interact and co-localize with Pol I and UBF. Indeed, we found that Merm1 interacted with RNA Pol I and UBF, but not with productive preinitiation complex components TAFI95 and TIF-IA, which are restricted to rDNA promoters (Figure 4A). To further assess the interactions among Merm1, Pol I, and UBF, we investigated the sub-cellular localization of Merm1 using indirect immunofluorescence microscopy. Immunofluorescent staining of Merm1 with a commercial antibody showed that endogenous Merm1 was observed throughout the nucleus in 293T cells, with a portion co-localized with Pol I and UBF (Supplementary Figure S7A). Furthermore, Merm1 fusion proteins with an HA or GFP tag were generated and transiently transfected into 293T, U2OS, and NIH3T3 cells. The fluorescent signals were captured using a spinning-disk confocal laser scanning microscope or an inverted immunofluorescent microscope, after which the signals were analyzed. HA-tagged Merm1 was most pronounced in the nucleoli of 293T cells, where it partially co-localized with Pol I and UBF (Supplementary Figure S7B). The GFP-tagged Merm1 signal was also enriched in the nucleoli of 293T and NIH3T3 cells, where it partially co-localized with Pol I and UBF (Supplementary Figure S7C and D). Moreover, analysis of the fluorescent signals derived from 3D stacks of GFP-tagged Merm1 at different angles confirmed that Merm1 was associated with Pol I and UBF in U2OS cells (Figure 4B and C; Supplementary Movies S1 and S2). Taken together, these results confirm that Merm1 interacts and co-localizes with Pol I and UBF. Figure 4 View largeDownload slide A fraction of Merm1 interacts and co-localizes with Pol I and UBF within nucleoli. (A) Merm1 interacts with the Pol I transcription machinery. Cell lysates were incubated with an anti-Merm1 antibody, and co-precipitated proteins were monitored by immunoblotting. Ten percent of the input is shown. (B) Merm1 partially co-localizes with Pol I and UBF in nucleoli. Fluorescent images of GFP-tagged Merm1, Pol I, and UBF in human U2OS cells are displayed. Scale bar, 2 μm. (C) 3D isosurface images of Merm1, Pol I, and UBF from two different angles were generated with image stacks using Volocity 6.3. Figure 4 View largeDownload slide A fraction of Merm1 interacts and co-localizes with Pol I and UBF within nucleoli. (A) Merm1 interacts with the Pol I transcription machinery. Cell lysates were incubated with an anti-Merm1 antibody, and co-precipitated proteins were monitored by immunoblotting. Ten percent of the input is shown. (B) Merm1 partially co-localizes with Pol I and UBF in nucleoli. Fluorescent images of GFP-tagged Merm1, Pol I, and UBF in human U2OS cells are displayed. Scale bar, 2 μm. (C) 3D isosurface images of Merm1, Pol I, and UBF from two different angles were generated with image stacks using Volocity 6.3. Interplay between Merm1 and Dnmt3a facilitates Pol I elongation The finding that Merm1 is enriched at gene bodies and interacts with Pol I and UBF, but not with promoter preinitiation complex components, suggests that the interplay between Dnmt3a and Merm1 may regulate rDNA transcription elongation. Therefore, the level of pre-rRNA synthesis in Merm1+/−, Dnmt3a−/−, and DKO cells was assessed by qRT-PCR. DKO cells displayed severely impaired rDNA transcription in comparison with that of Merm1+/− or Dnmt3a−/− cells (Supplementary Figure S8). Short 5′-fluorouridine (FUrd) pulse-labeling of DKO cells revealed drastic decreases in the rates of rRNA synthesis and proliferation (Figure 5A and B), suggesting that interaction between Dnmt3a and Merm1 facilitated rDNA transcription. However, examination of the occupancy of Pol I at rDNA in the wild-type and DKO cells using ChIP assays produced the unexpected result that, in contrast with compromised pre-rRNA synthesis, simultaneous loss of Merm1 and Dnmt3a led to increased occupancy of Pol I at gene bodies (Figure 5C). Whole-genome scale ChIP-seq analysis of Pol I confirmed that concurrent depletion of Merm1 and Dnmt3a increased Pol I enrichment at rDNA gene bodies (Figure 5D). These findings suggested that the severe inhibition of transcription by DKO cells may be caused by retardation of Pol I elongation and accumulation at gene bodies. To test this hypothesis, an actinomycin D (AMD) release assay was performed to compare Pol I elongation in wild-type and DKO cells. Cells were treated with a low concentration of AMD for 1 h to stop rDNA transcription and trap Pol I at rDNA promoters, after which the medium containing AMD was replaced with new medium without AMD to reinitialize movement of Pol I from promoters to gene bodies (Figure 5E). Following the AMD treatment and re-initialization of Pol I movement, Pol I enrichment at rDNA gene bodies was tested. AMD significantly decreased enrichment of Pol I at rDNA gene bodies, but gradual re-occupancy of Pol I at rDNA gene bodies was observed in 293T and DKO cells following the removal of medium containing AMD (Figure 5F). However, DKO cells exhibited a much slower re-occupancy rate of Pol I at rDNA gene bodies in comparison with that of wild-type cells (Figure 5F), suggesting that simultaneous deletion of Merm1 and Dnmt3a led to retardation of Pol I movement along gene bodies. Taken together, these findings demonstrate that Merm1 and Dnmt3a synergistically facilitate rDNA transcription by supporting Pol I elongation at rDNA gene bodies. Figure 5 View largeDownload slide Merm1 and Dnmt3a are required for Pol I transcription elongation. (A) Simultaneous knockout of one allele of Merm1 and two alleles of Dnmt3a impairs rDNA transcription. Merm1+/−/Dnmt3a−/− cells (Merm1+/−/Dnmt3a−/− #1 and Merm1+/−/Dnmt3a−/− #2, denoted as DKO #1 and DKO #2 below) were mixed with 293T cells, labeled for 15 min with 2 mM FUrd, and stained with antibodies against Dnmt3a and BrdU. Merm1+/−/Dnmt3a−/− cells are encircled. The bar diagram shows results obtained by subjecting 300 Merm1+/−/Dnmt3a−/− cells and 293T cells to FUrd staining. Scale bar, 5 μm. (B) Simultaneous depletion of Merm1 and Dnmt3a (DKO #1 and DKO #2) further inhibits cell proliferation in comparison with the inhibitory effect induced by depletion of Merm1 (Merm1+/−#1) or Dnmt3a (Dnmt3a−/− #1) alone. Cells were counted every 24 h. (C) Depletion of Merm1 or Dnmt3a impairs the association of Pol I with rDNA, while simultaneous depletion of these proteins increases Pol I occupancy. ChIP assays were performed using primers amplifying the 28S rRNA coding region to show the relative rDNA occupancy of Pol I in the indicated cells. Error bars represent SD (n = 3). (D) Pol I occupancy at rDNA gene body regions is increased in Merm1+/−/Dnmt3a−/− cells. Upper panel: Pol I ChIP-seq profiles for rDNA. Lower panel: overlay tracks of Pol I ChIP-seq data from the indicated cell lines. The locations of the 18S and 28S rRNA coding regions are shown at the bottom. (E) Scheme of the AMD release assay. 293T and Merm1+/−/Dnmt3a−/− cells (DKO #1 and DKO #2) were treated with 0.05 μg/ml AMD for 1 h, followed by replacement of the medium with AMD-free medium to allow transcription for 8, 16, or 24 h. Groups that were not subjected to AMD treatment (No AMD) or subjected to 1 h of AMD treatment (AMD 1 h) were used as controls. (F) The AMD release assay shows that simultaneous depletion of Merm1 and Dnmt3a reduces the transcription elongation rate of Pol I. ChIP assays were performed using primers that amplify the 28S rRNA coding region to show the relative rDNA occupancy of Pol I in the indicated treated cells. Error bars represent SD (n = 3). *P < 0.1, **P < 0.01, ***P < 0.001, ****P < 0.0001, two-tailed unpaired Student’s t-tests were performed. Figure 5 View largeDownload slide Merm1 and Dnmt3a are required for Pol I transcription elongation. (A) Simultaneous knockout of one allele of Merm1 and two alleles of Dnmt3a impairs rDNA transcription. Merm1+/−/Dnmt3a−/− cells (Merm1+/−/Dnmt3a−/− #1 and Merm1+/−/Dnmt3a−/− #2, denoted as DKO #1 and DKO #2 below) were mixed with 293T cells, labeled for 15 min with 2 mM FUrd, and stained with antibodies against Dnmt3a and BrdU. Merm1+/−/Dnmt3a−/− cells are encircled. The bar diagram shows results obtained by subjecting 300 Merm1+/−/Dnmt3a−/− cells and 293T cells to FUrd staining. Scale bar, 5 μm. (B) Simultaneous depletion of Merm1 and Dnmt3a (DKO #1 and DKO #2) further inhibits cell proliferation in comparison with the inhibitory effect induced by depletion of Merm1 (Merm1+/−#1) or Dnmt3a (Dnmt3a−/− #1) alone. Cells were counted every 24 h. (C) Depletion of Merm1 or Dnmt3a impairs the association of Pol I with rDNA, while simultaneous depletion of these proteins increases Pol I occupancy. ChIP assays were performed using primers amplifying the 28S rRNA coding region to show the relative rDNA occupancy of Pol I in the indicated cells. Error bars represent SD (n = 3). (D) Pol I occupancy at rDNA gene body regions is increased in Merm1+/−/Dnmt3a−/− cells. Upper panel: Pol I ChIP-seq profiles for rDNA. Lower panel: overlay tracks of Pol I ChIP-seq data from the indicated cell lines. The locations of the 18S and 28S rRNA coding regions are shown at the bottom. (E) Scheme of the AMD release assay. 293T and Merm1+/−/Dnmt3a−/− cells (DKO #1 and DKO #2) were treated with 0.05 μg/ml AMD for 1 h, followed by replacement of the medium with AMD-free medium to allow transcription for 8, 16, or 24 h. Groups that were not subjected to AMD treatment (No AMD) or subjected to 1 h of AMD treatment (AMD 1 h) were used as controls. (F) The AMD release assay shows that simultaneous depletion of Merm1 and Dnmt3a reduces the transcription elongation rate of Pol I. ChIP assays were performed using primers that amplify the 28S rRNA coding region to show the relative rDNA occupancy of Pol I in the indicated treated cells. Error bars represent SD (n = 3). *P < 0.1, **P < 0.01, ***P < 0.001, ****P < 0.0001, two-tailed unpaired Student’s t-tests were performed. Discussion Over the past few decades, numerous studies have focused on elucidating the mechanisms underlying DNA methylation and the manner in which it controls gene transcription. With regard to rRNA genes, Dnmt1, Dnmt3a, and Dnmt3b have been shown to localize in the nucleolus and associate with rDNA, implying that they have essential roles in epigenetic control of rDNA transcription (Majumder et al., 2006). It is widely acknowledged that DNA methylation in promoters silences rRNA genes, while active promoters are characterized by DNA hypomethylation (Grummt and Längst, 2013). Dnmt1- and/or Dnmt3b-mediated DNA methylation on rDNA promoters can impair the assembly of the transcription initiation complex, including recruitment of Pol I and UBF (McStay and Grummt, 2008). In contrast, genetic disruption of Dnmt1 and Dnmt3b leads to hypomethylation and activation of rDNA promoters (Majumder et al., 2006; Gagnon-Kugler et al., 2009; Grummt and Längst, 2013). However, neither ectopic Dnmt3a nor decreased expression of Dnmt3a affected rDNA promoter activity (Majumder et al., 2006), indicating that Dnmt3a plays a role distinct from those of Dnmt1 and Dnmt3b. Indeed, unlike Dnmt1 or Dnmt3b, Dnmt3a is not required for the maintenance of methylation at heterochromatic repeat regions (Okano et al., 1999). Deletion of Dnmt1 or Dnmt3b leads to early embryonic lethality, while Dnmt3a-null mice appeared to be grossly normal at birth, but showed postnatal developmental defects and died prematurely, suggesting that Dnmt3a regulates euchromatic expression and postnatal development (Li et al., 1992). More importantly, Dnmt3a has been found to occupy the gene bodies of a large cohort of neurogenic genes, for which it facilitates transcription by functionally antagonizing polycomb repression. These findings reveal that Dnmt3a may maintain active transcription elongation of genes critical for development (Wu et al., 2010). In this study, we revealed that Dnmt3a is involved in the modulation of gene body methylation and transcription elongation of rRNA genes. Thus, our results and others demonstrate Dnmt3a-dependent transcription elongation of Pol I- and Pol II-transcribed genes. In addition, our study identified a repressive regulator of Dnmt3a, nucleolar protein Merm1, which cooperated with Dnmt3a to modulate the methylation status of gene bodies, thereby regulating their capacity for rDNA transcription elongation. Several regulators have been shown to enhance the activity of DNA methyltransferases and play critical roles in maintaining DNA methylation patterns (Chédin et al., 2002; Duymich et al., 2016). Our results reveal that Merm1 acts as a negative regulator of Dnmt3a. Although Merm1 contains a SAM binding motif, it does not possess DNA methyltransferase activity or the conserved catalytic domain common to DNA methyltransferases. Remarkably, Merm1 negatively regulates the DNA methyltransferase activity of Dnmt3a to generate unmethylated rDNA templates, thereby allowing Pol I transcription elongation. Inhibition of methyltransferase activity by Merm1 resulted in association of Dnmt3a with unmethylated gene bodies. Under these circumstances, Dnmt3a could not exert its rDNA methyltransferase activity, and Dnmt3a knockout did not further affect gene body methylation. However, loss of Merm1 restored the DNA methyltransferase activity of Dnmt3a and rendered Dnmt3a able to methylate rDNA gene bodies, thus enhancing rDNA gene body methylation. Our findings provide new insight into the manner in which Dnmt3a and Merm1 interact to modulate the methylation state of gene bodies and control gene transcription elongation. Previously, Merm1 and its yeast homolog Bud23 were identified as 18S rRNA base methyltransferases that play critical roles in pre-rRNA processing (Figaro et al., 2012; Õunap et al., 2013; Tafforeau et al., 2013; Létoquart et al., 2014; Haag et al., 2015; Õunap et al., 2015; Zorbas et al., 2015). It is clear that Merm1 regulates pre-rRNA processing, but it may have additional functions in mammalian cells. First, although mammalian Merm1 can facilitate rRNA methylation, such modification does not seem to be required for pre-processing (Zorbas et al., 2015). Furthermore, human Merm1 and Bud23 have only 47% amino acid similarity. Moreover, yeast does not possess cytosine DNA methyltransferases, so yeast DNA is unmethylated. These findings suggest that human Merm1 and Bud23 have similar, but probably not identical functions (Nakazawa et al., 2011; Jangani et al., 2014; Õunap et al., 2015). Indeed, Merm1 has been shown to drive alterations of epigenetic modification at promoters in mammalian cells (Jangani et al., 2014), and here we verified that Merm1 acts as a repressive regulator of Dnmt3a and activates rDNA transcription. Thus, in addition to its role in pre-rRNA processing, modulation of epigenetic states by Merm1 is required for rDNA transcription. It is not surprising that Merm1 plays dual roles in the processes of rDNA transcription and pre-rRNA processing. Some other proteins, such as Dnmt1, nucleolin, and fibrillarin, have well-characterized roles in multiple steps of rRNA processing, including regulation of rDNA transcription, pre-rRNA splicing, and assembly of ribosomal particles (Egyhazi et al., 1988; Grummt, 1998; Roger et al., 2002). With regard to Dnmt1, the extensive loss of DNA methylation at rDNA loci caused by Dnmt1 knockout reduced rDNA transcription and led to defective rRNA processing (Gagnon-Kugler et al., 2009), suggesting that the ribosomal synthesis rate is controlled by regulating the rates of transcription and subsequent rRNA maturation. In conclusion, we have shown that Merm1 interacts with Pol I and UBF and regulates the activity of Dnmt3a, thereby collaborating to control rDNA transcription elongation. Depletion of Merm1 increased gene body methylation and consequently decreased Pol I binding. Simultaneous knockout of one allele of Merm1 and two alleles of Dnmt3a did not affect gene body methylation, but rather retarded Pol I elongation, resulting in the accumulation of Pol I on rDNA templates. This finding suggests that both Merm1 and Dnmt3a are required for transcriptional elongation of rDNA (Figure 6). In addition, depletion of either Merm1 or Dnmt3a produced similar changes in the expression levels of many Pol II-transcribed genes (Supplementary Figure S9), suggesting that Merm1 and Dnmt3a cooperate to regulate Pol I- and Pol II-mediated gene transcription. Therefore, our results may reveal a general cellular mechanism, in which the generation of unmethylated gene bodies by a repressive regulator of DNA methyltransferases allows transcription elongation. There are alternative pathways that lead to gene bodies demethylation, such as that involving the TET family of demethylases (Santi et al., 1983; He et al., 2011). Therefore, it is essential to explore the functional relationships between TET proteins, Merm1, and Dnmt3a to reveal how they may cooperate to control gene body methylation and transcriptional elongation. Figure 6 View largeDownload slide Proposed model depicting the roles of Merm1 and Dnmt3a in regulating Pol I elongation. (A) Merm1 is recruited by Dnmt3a and inhibits the DNA methyltransferase activity of Dnmt3a, resulting in co-occupancy of Dnmt3a and Pol I at hypomethylated gene bodies, thus allowing Pol I elongation. (B) Upon depletion of Merm1, the methyltransferase activity of Dnmt3a is restored, leading to binding of Dnmt3a to methylated gene bodies and dissociation of Pol I. (C) In the case of simultaneous knockout of one allele of Merm1 and two alleles of Dnmt3a (arrows), loss of these two proteins leads to retardation of elongation and accumulation of Pol I on gene bodies. Figure 6 View largeDownload slide Proposed model depicting the roles of Merm1 and Dnmt3a in regulating Pol I elongation. (A) Merm1 is recruited by Dnmt3a and inhibits the DNA methyltransferase activity of Dnmt3a, resulting in co-occupancy of Dnmt3a and Pol I at hypomethylated gene bodies, thus allowing Pol I elongation. (B) Upon depletion of Merm1, the methyltransferase activity of Dnmt3a is restored, leading to binding of Dnmt3a to methylated gene bodies and dissociation of Pol I. (C) In the case of simultaneous knockout of one allele of Merm1 and two alleles of Dnmt3a (arrows), loss of these two proteins leads to retardation of elongation and accumulation of Pol I on gene bodies. Materials and methods Cell culture U2OS, NIH3T3, 293T, and the derived knockout or knockdown cells were all cultured at 5% CO2, 37°C in Dulbecco’s modified Eagle’s medium (DMEM, Gibco) supplemented with 10% (v/v) fetal bovine serum (FBS). For growth curve, wild-type 293T and knockout cells were in inoculums of 1 × 106, and NIH3T3 and the corresponding knockdown cells were in inoculums of 0.5 × 106 in 100-mm culture dish. Cells were counted with hemocytometer every day and cumulative cell numbers were calculated and presented. Plasmids cDNA encoding human Merm1 (GenBank accession number NM_017528) was PCR amplified with specific primers and the amplicons were ligated into pEGFP-C1, pGEX-4T-1, and pCMV-3 × HA to generate recombinant expression constructs. cDNA encoding mouse Merm1 (GenBank accession number NM_025375.3) was cloned into pEGFP-C1. Transcript variants of Merm1 to produce Merm1-4G4R, Merm1-N, and Merm1-C were cloned and inserted into pCMV-3 × HA to generate recombinant expression vectors. The primers used to generate recombinant constructs are listed in Supplementary Table S1. pcDNA3/Myc-Dnmt1 (Addgene plasmid # 36939), pcDNA3/Myc-Dnmt3a (Addgene plasmid # 35521), and pcDNA3/Myc-Dnmt3b1 (Addgene plasmid # 35522) are gifts from Arthur Riggs (Chen et al., 2005; Li et al., 2006). Antibodies Anti-Dnmt3a (sc-20703), anti-Merm1 (sc-135322), anti-RPA194 (sc-28714), anti-UBF (sc-9131X), and anti-c-Myc (sc-40X) were purchased from Santa Cruz. Anti-α-tubulin (T6199), anti-HA (H9658), and anti-Bromodeoxyuridine (B2531) were obtained from Sigma. Anti-5-methylcytosine (BI-MECY-0100) was from Eurogentec. Anti-H3K4me3 (04-745) and anti-H3K27me3 (07-449) were from Upstate. Anti-Merm1 (ab97911), anti-H3K9me3 (ab8898), anti-H3K36me3 (ab9050), anti-H4K20me3 (ab9053), and anti-H3 (ab1791) were from Abcam. Secondary antibodies IRDye800CW goat anti-mouse IgG (926-32210) and IRDye800CW goat anti-rabbit IgG (926-32211) were purchased from LI-COR. Targeting strategy of knockdown and knockout For knockdown in NIH3T3 cells, the shRNA targeting mouse Merm1 (5′-TCGCAACTCACGGATGATT-3′) was cloned into the retroviral vector pQXCIP. Plat-E cell was transfected with the recombinant pQXCIP constructs to package the retrovirus. The packaged virus was collected and applied to infect NIH3T3 cells. Cells were selected with puromycin for 5 days. For knockout in 293T cells, the CRISPR/Cas9 system was utilized. Recombinant pcDNA3.1-NLS-Cas9-NLS containing human Cas9 and pUC19 containing specific guide RNA were co-transfected into 293T cells. Single clone of 293T was selected and cultured. Cells were harvested to extract genomic DNA and the DNA was followed by Sanger sequencing and DNAMAN alignment analysis. The sequence information of guide RNAs is provided in Supplementary Table S1. Immunofluorescence Cells grown on coverslips were washed with 1× PBS and fixed in 4% paraformaldehyde for 10 min at room temperature. For FUrd labeling, cells were labeled with 2 mM of FUrd for 5 or 15 min before fixation. After washing three times with 1× PBS (5 min each), cells were permeabilized with 0.5% Triton X-100 for 15 min. Then, cells were washed with PBS for three times (5 min each) and were blocked with 1% BSA in 1× PBS (w/v) for 30 min at room temperature. After washing three times with PBS (5 min each), cells were incubated with primary antibodies diluted in PBST (1× PBS + 0.1% Tween-20) at 4°C overnight. After incubation, cells were washed with PBST (1× PBS + 0.1% Triton X-100) for four times. Then cells were incubated with appropriate secondary antibodies (dilution at 1:400, Invitrogen) for 2 h, washed with 0.1% PBST (Triton X-100) and stained with 2 ng/μl of DAPI for 5 min. The coverslips were sealed and the images were recorded with immunofluorescent microscope. The obtained FUrd signals were quantified with Image J software (National Institutes of Health). For immunofluorescence of endogenous Merm1, we used antibodies from Abcam (ab97911) at a concentration of 20 μg/ml. For immunofluorescence of GFP-tagged or HA-tagged Merm1, fluorescent signals were observed under the spinning-disk confocal laser scanning microscope (Nikon) and analyzed with Volocity 6.3 (PerkinElmer). ChIP and real-time quantitative PCR ChIP assays were performed as described (Zhou et al., 2002). Cross-linked chromatin was sonicated to yield fragments of 400−500 base pairs (bp). Then, the chromatin was pre-cleared, 10% input were reserved and the rest were incubated with respective antibodies. Immunoprecipitated proteins were captured with protein A/G Sepharose beads. After reversal of the crosslinking with the existence of RNase A and digestion of proteins with proteinase K, the DNA was purified by phenol–chloroform extraction or chelex-100 resin and amplified by real-time and quantitative PCR (RT-qPCR). The RT-qPCR was run on Light Cycler (Roche) with 2× SYBR Premix Ex Taq (RR420A, Takara). The conditions were set up according to the tutorial manual provided by the manufacturer. The PCR results were acquired and analyzed by normalization to the 10% input. The ChIP primers used in this study are listed in Supplementary Table S2. Co-immunoprecipitation and western blotting assays To prepare nuclear extract, nuclei were lysed in IP buffer (20 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% (v/v) Triton X-100) for 30 min at 4°C and cleared by centrifugation (16000 g, 15 min, 4°C). To prepare whole-cell extract, cells were lysed with RIPA buffer (10 mM Tris-HCl, pH 8.0, 1 mM EDTA, 0.5 mM EGTA, 1% Triton X-100, 0.1% sodium deoxycholate, 0.1% SDS, 140 mM NaCl and add 1 mM PMSF before use) for 1 min and the suspension was diluted 10 times with RIPA buffer without detergent, after incubation for 30 min at 4°C, the lysate was cleared by centrifugation (16000 g, 15 min, 4°C). Then, the lysate was reserved for 10% input and the rest were incubated with respective antibodies at 4°C overnight. Then the protein A/G beads were added to capture the proteins. After washing four times with IP buffer (for nuclear extract) or 0.1× RIPA buffer (for whole-cell extract), the beads carried immunoprecipitated proteins were boiled with 2× SDS loading buffer (80 mM Tris, pH 6.8, 10% glycerol, 2% SDS, 20% β-mercaptoethanol) and centrifuged at 12000 g for 5 min at room temperature. The supernatant and the input were analyzed by western blotting. The protein was subject to SDS-PAGE, transferred to nitrocellulose membrane and incubated with primary antibodies of interest at 4°C overnight. Primary antibodies used were mentioned above. The protein blots were then washed three times in PBST (1× PBS + 0.1% Tween-20) and incubated with diluted secondary antibodies for 2 h at room temperature. And then the blots were washed three times with PBST before visualized under odyssey infrared imaging system (Odyssey, LI-COR). DNA methylation assays To monitor CpG methylation, genomic DNA or ChIP-enriched DNA was treated with sodium bisulfite using EpiTect Bisulfite Kit (Qiagen). Briefly, the genomic DNA was subject to bisulfite conversion reaction. The converted DNA was then purified and amplified by PCR with specific primers. The amplicons were electrophoresed by agarose gel and the purified amplicons were ligated into the pEasy-T5 vector. The recombinant pEasy-T5 constructs were transduced into competent Escherichiacoli, which were plated and incubated at 37°C. At least 15 independent colonies of the E. coli were selected and sequenced. The sequencing results were analyzed by BiQ analyzer (Max Planck Institut Informatik). The primer information is provided in Supplementary Table S2. In vitro DNA methyltransferase activity assays For DNA methyltransferase activity assays, 0.1 μg purified His-tagged Dnmt3a or Dnmt3b was incubated with 0−0.2 μg purified GST-tagged Merm1 or His-tagged Dnmt3L at 37°C for 30 min in 10 μl buffer containing 20 mM Hepes, pH 7.5, 50 mM NaCl, 0.5 mg/ml BSA, 0.5 μCi tritiated SAM (3H-S-adenosyl methionine, PerkinElmer), and 100 ng substrate DNA. After the reaction, DNA was recovered and counted in a scintillation counter. Sequence of the substrate DNA is 5′-GATCGCCGATGCGCGAATCGCGATCGATGCGAT-3′. In addition, quantification of DNA methyltransferase activity was also performed using the EpiQuik DNA Methyltransferase Activity/Inhibition Assay Kit (Epigentek). Methylated DNA immunoprecipitation and next-generation sequencing MeDIP DNA libraries were prepared using the NEBNext DNA Library Prep Master Mix Set for Illumina (E6040, NEB). Briefly, fragmented DNA was end repaired and ligated with adapter. After size selection and purification using AMPure XP Beads (A6380, Beckman), the recovered DNA was denatured and incubated with the specific antibody against 5mC. Then, the immunoprecipitated fragments were captured with protein A/G agarose beads. After digestion of proteins with proteinase K, the DNA fragments were extracted and amplified by adaptor-mediated PCR. Libraries were sequenced using the Illumina HiSeq 2000 sequencing platform. The sequencing reads were aligned to mm9 using bowtie2 under the default parameters, then PCR duplicated reads were removed by Picard (MarkDuplicates). Samtools was used to select properly mapped reads (samtools view -Sb -h -f 2). Finally, peaks were detected by macs (–nomodel –nolambda -w –space = 30). ChIP-Seq Samples were processed for sequencing using NEBNext DNA Library Prep Master Mix Set for Illumina (E6040, NEB). In brief, ChIP-enriched DNA fragments were end repaired and ligated with adapter. After size selection using AMPure XP Beads, the library was amplified by adaptor-mediated PCR. Libraries were sequenced using the Illumina HiSeq 2000 sequencing platform. The alignment step was the same as MeDIP-seq, but the peaks calling were using masc under the default parameters (Landt et al., 2012). Supplementary material Supplementary material is available at Journal of Molecular Cell Biology online. Acknowledgements We thank I. Grummt (German Cancer Research Center, Germany) for the valuable discussions and advice. We thank Ruiqian Li and Ping Zhu (Biodynamic Optical Imaging Center, Peking University, China) for epigenome analysis, Guohong Li (Institute of Biophysics, Chinese Academy of Sciences, China) for assistance with DNA methylation assays. Funding This work was supported by the National Natural Science Foundation of China (31471205, 31171255, 91219101, and 81330009) and the National Basic Research Program of China (2013CB530700). Conflict of interest none declared. Author contributions L.Z. and G.L. designed and performed cellular, molecular, biochemical, and animal experiments; X.H., W.X., Y.G., J.F., and L.Z. contributed to cellular and biochemical experiments; C.Z. contributed to epigenome analysis; T.N. and J.G. provided critical reagents and contributed to data analysis; W.T. and T.N conceived the project and designed experiments; W.T., L.Z., and G.L wrote the manuscript. References Bestor, T., Laudano, A., Mattaliano, R., et al.  . ( 1988). Cloning and sequencing of a cDNA encoding DNA methyltransferase of mouse cells: the carboxyl-terminal domain of the mammalian enzymes is related to bacterial restriction methyltransferases. J. Mol. Biol.  203, 971– 983. Google Scholar CrossRef Search ADS PubMed  Bostick, M., Kim, J.K., Estève, P.-O., et al.  . ( 2007). UHRF1 plays a role in maintaining DNA methylation in mammalian cells. Science  317, 1760– 1764. Google Scholar CrossRef Search ADS PubMed  Bourc’his, D., Xu, G.-L., Lin, C.-S., et al.  . ( 2001). Dnmt3L and the establishment of maternal genomic imprints. Science  294, 2536– 2539. Google Scholar CrossRef Search ADS PubMed  Chen, Z.X., Mann, J.R., Hsieh, C.L., et al.  . ( 2005). Physical and functional interactions between the human DNMT3L protein and members of the de novo methyltransferase family. J. Cell. Biochem.  95, 902– 917. Google Scholar CrossRef Search ADS PubMed  Chédin, F., Lieber, M.R., and Hsieh, C.-L. ( 2002). The DNA methyltransferase-like protein DNMT3L stimulates de novo methylation by Dnmt3a. Proc. Natl Acad. Sci. USA  99, 16916– 16921. Google Scholar CrossRef Search ADS   Doll, A., and Grzeschik, K.H. ( 2001). Characterization of two novel genes, WBSCR20 and WBSCR22, deleted in Williams-Beuren syndrome. Cytogenet. Cell Genet.  95, 20– 27. Google Scholar CrossRef Search ADS PubMed  Duymich, C.E., Charlet, J., Yang, X., et al.  . ( 2016). DNMT3B isoforms without catalytic activity stimulate gene body methylation as accessory proteins in somatic cells. Nat. Commun.  7, 11453. Google Scholar CrossRef Search ADS PubMed  Egyhazi, E., Pigon, A., Chang, J.-H., et al.  . ( 1988). Effects of anti-C23 (nucleolin) antibody on transcription of ribosomal DNA in Chironomus salivary gland cells. Exp. Cell Res.  178, 264– 272. Google Scholar CrossRef Search ADS PubMed  Figaro, S., Wacheul, L., Schillewaert, S., et al.  . ( 2012). Trm112 is required for Bud23-mediated methylation of the 18S rRNA at position G1575. Mol. Cell. Biol.  32, 2254– 2267. Google Scholar CrossRef Search ADS PubMed  Gagnon-Kugler, T., Langlois, F., Stefanovsky, V., et al.  . ( 2009). Loss of human ribosomal gene CpG methylation enhances cryptic RNA polymerase II transcription and disrupts ribosomal RNA processing. Mol. Cell  35, 414– 425. Google Scholar CrossRef Search ADS PubMed  Gruenbaum, Y., Cedar, H., and Razin, A. ( 1982). Substrate and sequence specificity of a eukaryotic DNA methylase. Nature  295, 620– 622. Google Scholar CrossRef Search ADS PubMed  Grummt, I. ( 1998). Regulation of mammalian ribosomal gene transcription by RNA polymerase I. Prog. Nucleic Acid Res. Mol. Biol.  62, 109– 154. Google Scholar CrossRef Search ADS   Grummt, I., and Längst, G. ( 2013). Epigenetic control of RNA polymerase I transcription in mammalian cells. Biochim. Biophys. Acta  1829, 393– 404. Google Scholar CrossRef Search ADS PubMed  Haag, S., Kretschmer, J., and Bohnsack, M.T. ( 2015). WBSCR22/Merm1 is required for late nuclear pre-ribosomal RNA processing and mediates N7-methylation of G1639 in human 18S rRNA. RNA  21, 180– 187. Google Scholar CrossRef Search ADS PubMed  He, Y.-F., Li, B.-Z., Li, Z., et al.  . ( 2011). Tet-mediated formation of 5-carboxylcytosine and its excision by TDG in mammalian DNA. Science  333, 1303– 1307. Google Scholar CrossRef Search ADS PubMed  Hsieh, C.-L. ( 1999). In vivo activity of murine de novo methyltransferases, Dnmt3a and Dnmt3b. Mol. Cell. Biol.  19, 8211– 8218. Google Scholar CrossRef Search ADS PubMed  Jangani, M., Poolman, T.M., Matthews, L., et al.  . ( 2014). The methyltransferase WBSCR22/Merm1 enhances glucocorticoid receptor function and is regulated in lung inflammation and cancer. J. Biol. Chem.  289, 8931– 8946. Google Scholar CrossRef Search ADS PubMed  Landt, S.G., Marinov, G.K., Kundaje, A., et al.  . ( 2012). ChIP-seq guidelines and practices of the ENCODE and modENCODE consortia. Genome Res.  22, 1813– 1831. Google Scholar CrossRef Search ADS PubMed  Leonhardt, H., Page, A.W., Weier, H.-U., et al.  . ( 1992). A targeting sequence directs DNA methyltransferase to sites of DNA replication in mammalian nuclei. Cell  71, 865– 873. Google Scholar CrossRef Search ADS PubMed  Li, E., Bestor, T.H., and Jaenisch, R. ( 1992). Targeted mutation of the DNA methyltransferase gene results in embryonic lethality. Cell  69, 915– 926. Google Scholar CrossRef Search ADS PubMed  Li, H., Rauch, T., Chen, Z.X., et al.  . ( 2006). The histone methyltransferase SETDB1 and the DNA methyltransferase DNMT3A interact directly and localize to promoters silenced in cancer cells. J. Biol. Chem.  281, 19489– 19500. Google Scholar CrossRef Search ADS PubMed  Létoquart, J., Huvelle, E., Wacheul, L., et al.  . ( 2014). Structural and functional studies of Bud23–Trm112 reveal 18S rRNA N7-G1575 methylation occurs on late 40S precursor ribosomes. Proc. Natl Acad. Sci. USA  111, E5518– E5526. Google Scholar CrossRef Search ADS   Majumder, S., Ghoshal, K., Datta, J., et al.  . ( 2006). Role of DNA methyltransferases in regulation of human ribosomal RNA gene transcription. J. Biol. Chem.  281, 22062– 22072. Google Scholar CrossRef Search ADS PubMed  McStay, B., and Grummt, I. ( 2008). The epigenetics of rRNA genes: from molecular to chromosome biology. Annu. Rev. Cell Dev. Biol.  24, 131– 157. Google Scholar CrossRef Search ADS PubMed  Nakazawa, Y., Arai, H., and Fujita, N. ( 2011). The novel metastasis promoter Merm1/Wbscr22 enhances tumor cell survival in the vasculature by suppressing Zac1/p53-dependent apoptosis. Cancer Res.  71, 1146– 1155. Google Scholar CrossRef Search ADS PubMed  Neri, F., Krepelova, A., Incarnato, D., et al.  . ( 2013). Dnmt3L antagonizes DNA methylation at bivalent promoters and favors DNA methylation at gene bodies in ESCs. Cell  155, 121– 134. Google Scholar CrossRef Search ADS PubMed  Okano, M., Bell, D.W., Haber, D.A., et al.  . ( 1999). DNA methyltransferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development. Cell  99, 247– 257. Google Scholar CrossRef Search ADS PubMed  Okano, M., Xie, S., and Li, E. ( 1998). Cloning and characterization of a family of novel mammalian DNA (cytosine-5) methyltransferases. Nat. Genet.  19, 219– 220. Google Scholar CrossRef Search ADS PubMed  Piano, F., Schetter, A.J., Morton, D.G., et al.  . ( 2002). Gene clustering based on RNAi phenotypes of ovary-enriched genes in C. elegans. Curr. Biol.  12, 1959– 1964. Google Scholar CrossRef Search ADS PubMed  Roger, B., Moisand, A., Amalric, F., et al.  . ( 2002). Repression of RNA polymerase I transcription by nucleolin is independent of the RNA sequence that is transcribed. J. Biol. Chem.  277, 10209– 10219. Google Scholar CrossRef Search ADS PubMed  Santi, D.V., Garrett, C.E., and Barr, P.J. ( 1983). On the mechanism of inhibition of DNA-cytosine methyltransferases by cytosine analogs. Cell  33, 9– 10. Google Scholar CrossRef Search ADS PubMed  Sharif, J., Muto, M., Takebayashi, S.-I., et al.  . ( 2007). The SRA protein Np95 mediates epigenetic inheritance by recruiting Dnmt1 to methylated DNA. Nature  450, 908– 912. Google Scholar CrossRef Search ADS PubMed  Suetake, I., Shinozaki, F., Miyagawa, J., et al.  . ( 2004). DNMT3L stimulates the DNA methylation activity of Dnmt3a and Dnmt3b through a direct interaction. J. Biol. Chem.  279, 27816– 27823. Google Scholar CrossRef Search ADS PubMed  Tafforeau, L., Zorbas, C., Langhendries, J.-L., et al.  . ( 2013). The complexity of human ribosome biogenesis revealed by systematic nucleolar screening of pre-rRNA processing factors. Mol. Cell  51, 539– 551. Google Scholar CrossRef Search ADS PubMed  Wu, H., Coskun, V., Tao, J., et al.  . ( 2010). Dnmt3a-dependent nonpromoter DNA methylation facilitates transcription of neurogenic genes. Science  329, 444– 448. Google Scholar CrossRef Search ADS PubMed  Xie, S., Wang, Z., Okano, M., et al.  . ( 1999). Cloning, expression and chromosome locations of the human DNMT3 gene family. Gene  236, 87– 95. Google Scholar CrossRef Search ADS PubMed  Yen, R.-W.C., Vertino, P.M., Nelkin, B.D., et al.  . ( 1992). Isolation and characterization of the cDNA encoding human DNA methyltransferase. Nucleic Acids Res.  20, 2287– 2291. Google Scholar CrossRef Search ADS PubMed  Zhou, Y., Santoro, R., and Grummt, I. ( 2002). The chromatin remodeling complex NoRC targets HDAC1 to the ribosomal gene promoter and represses RNA polymerase I transcription. EMBO J.  21, 4632– 4640. Google Scholar CrossRef Search ADS PubMed  Zorbas, C., Nicolas, E., Wacheul, L., et al.  . ( 2015). The human 18S rRNA base methyltransferases DIMT1L and WBSCR22-TRMT112 but not rRNA modification are required for ribosome biogenesis. Mol. Biol. Cell  26, 2080– 2095. Google Scholar CrossRef Search ADS PubMed  Õunap, K., Käsper, L., Kurg, A., et al.  . ( 2013). The human WBSCR22 protein is involved in the biogenesis of the 40S ribosomal subunits in mammalian cells. PLoS One  8, e75686. Google Scholar CrossRef Search ADS PubMed  Õunap, K., Leetsi, L., Matsoo, M., et al.  . ( 2015). The stability of ribosome biogenesis factor WBSCR22 is regulated by interaction with TRMT112 via ubiquitin-proteasome pathway. PLoS One  10, e0133841. Google Scholar CrossRef Search ADS PubMed  © The Author(s) (2018). Published by Oxford University Press on behalf of Journal of Molecular Cell Biology, IBCB, SIBS, CAS. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Journal of Molecular Cell BiologyOxford University Press

Published: Mar 21, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off