Epigenetic Mechanism of rRNA Gene Silencing: Temporal Order of NoRC-Mediated Histone Modification, Chromatin Remodeling, and DNA Methylation

Plasmids.

pMr600-BH represents a fusion of a 5′-terminal mouse rRNA gene fragment (from positions −328 to +292) with a 3′-terminal BamHI/HinfI fragment (from +334 to +712 with respect to the 3′ end of 28S rRNA) including two terminator elements (T₁ and T₂) separated by 177-bp pUC9 vector sequences. In pMr600₋₁₃₃-BH, the cytosine at −133 has been converted into guanine. pMr1930 contains murine rRNA gene sequences from positions −1930 to +155. pMr-1930ΔT₀ is similar to pMr1930, except that the upstream terminator T₀ (from −174 to −144) has been deleted. pcDNA-FLAG-TIP5 and pMr1930-CBH have been described previously (35). pcDNA-FLAG-Snf2h and -Snf2h_K211R were a gift from Ramin Shlekhattar.

Transfections and RNA analysis.

NIH 3T3 and HEK293T cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. For reporter assays, 5 × 10⁵ NIH 3T3 cells were cotransfected with 1 μg of reporter plasmid and different amounts of pcDNA-FLAG-TIP5 by the calcium phosphate DNA coprecipitation method. Transcripts from reporter plasmids were monitored by hybridization to a ³²P-labeled riboprobe complementary to pUC9 sequences that have been inserted between the 5′- and 3′-terminal rRNA gene fragments of pMr600-BH. Alternatively, transcript levels were analyzed by reverse transcription)-PCR with the SYBR detection system (Roche Applied Science) and primers specific for the rRNA gene reporter and GAPDH (glyceraldehyde-3-phospate dehydrogenase) mRNA. The amounts of transcripts and GAPDH cDNA were quantified by comparing them to a standard curve obtained by logarithmic dilution of the corresponding DNA. Polymerase I (Pol I) transcription was calculated by normalizing the amount of rRNA gene transcripts to GAPDH mRNA.

ChIP and DNA methylation assays.

Chromatin immunoprecipitation (ChIP) assays were performed as described previously (35, 43). Briefly, chromatin from cross-linked cells was sheared by sonication and incubated overnight with specific antibodies, and precipitated proteins were captured with protein G- or protein A-Sepharose saturated with salmon sperm DNA. After elution and reversion of cross-links by heating for 6 h at 65°C, 1 and 3% of precipitated DNA were amplified by PCR with 30 cycles (30 s at 95°C, 40 s at 55°C, 40 s at 72°C) in the presence of 1.5 mM MgCl₂, 200 μM (each) deoxynucleoside triphosphates, and 10 pmol of primers (promoter forward primers, 5′-GACCAGTTGTTCCTTTGAG-3′ [positions −165 to −145] and 5′-GATAGGTACTGACACGCTGTCCTTTCCCTATTA-3′ [−7 to +16]; promoter reverse primer, 5′-AGGACAGCGTGTCAGTACCTATC-3′ [−7 to +16]; 28S rRNA forward primer, 5′-GCGACCTCAGATCAGACGTGG-3′ [+8124 to +8145]; and 28S rRNA reverse primer, 5′-CTTAACGGTTTCACGCCCTC-3′ [+8549 to +8529]). PCR products were separated by electrophoresis on 2% agarose gels and stained with ethidium bromide. Real-time PCR (RT-PCR) was performed in triplicate with a LightCycler (Roche) and the SYBR green detection system. The relative enrichment of the rRNA genes was determined by calculating the ratio of rRNA genes present in the immunoprecipitates to those present in the input chromatin and normalizing the data to those for control reactions containing no antibodies.

To monitor CpG methylation at the rRNA gene reporter, DNA was digested with 20 U of HpaII or MspI before PCR amplification using two different forward primers that map upstream and downstream of the CCGG site at −143 (positions −165 to −145 or −7 to +16) and a reverse primer that is complementary to chloramphenicol acetyltransferase (CAT) gene sequences inserted into the minigene pMr1930-CBH. For quantification, RT-PCR was performed with a LightCycler (Roche). The relative resistance to HpaII digestion was calculated by normalizing the amount of DNA amplified to either of the two forward primers and the reverse primer.

Antibodies.

Antibodies to acetylated histone H4 and the FLAG epitope were from Upstate and Sigma, respectively. Antibodies against di- and trimethylated histone H3-Lys9 were a gift from Thomas Jenuwein.

rRNA gene silencing requires recruitment of NoRC to the rRNA gene promoter by TTF-I.

To study NoRC function in the establishment of heterochromatin, we overexpressed TIP5, the large subunit of NoRC, in HEK293T cells and compared the levels of acetylated histone H4 and di- and trimethylated histone H3-Lys9 for both endogenous rRNA genes and a cotransfected reporter plasmid (Fig. 1A). Consistent with previous results (43), the overexpression of TIP5 markedly decreased the acetylation of histone H4 at the rRNA gene promoter. Moreover, the level of H3-Lys9 dimethylation was significantly increased, whereas the trimethylation of H3-Lys9 was much less affected. Importantly, the same histone modifications were observed both on endogenous rRNA genes (Fig. 1A, left) and on the Pol I reporter plasmid (Fig. 1A, right), underscoring the functional relevance of NoRC-mediated epigenetic changes at the rRNA gene promoter.

TIP5 has been shown to interact with TTF-I, suggesting that NoRC is recruited to the rRNA gene by TTF-I bound to the promoter-proximal terminator T₀ (29, 39). If NoRC was recruited to the rRNA genes by TTF-I, it should not associate with templates lacking the TTF-I binding site. To test this possibility, NIH 3T3 cells were cotransfected with an expression vector encoding FLAG-tagged TIP5 and a reporter plasmid that either contains or lacks the TTF-I binding site T₀ (pMr1930 or pMr1930ΔT₀, respectively). ChIP experiments revealed that TIP5 is associated with the wild-type rRNA gene reporter (Fig. 1B, top), but no binding to the plasmid lacking the TTF-I binding site was observed (Fig. 1B, bottom). Thus, recruitment of NoRC to the rRNA genes depends on the binding of TTF-I to the promoter-proximal terminator T₀.

Given that the association of TIP5 with TTF-I is the first step in the chain of events that lead to the silencing of Pol I transcription, NoRC should not repress transcription on the mutant reporter pMr-1930ΔT₀. Indeed, overexpression of TIP5 strongly reduced transcription from the wild-type but not the mutant template (Fig. 1C). Moreover, neither histone H4 hypoacetylation nor H3-Lys9 dimethylation was observed at the template lacking the terminator T₀ (Fig. 1D). Thus, the silencing of Pol I transcription and the establishment of heterochromatic features requires recruitment of NoRC to the rRNA genes by TTF-I bound to the promoter-proximal target site T₀.

NoRC-mediated DNA methylation depends on histone deacetylation.

To examine whether NoRC-dependent histone modification and de novo DNA methylation may proceed as parallel pathways or are dependent on each other, we assayed the effect of trichostatin A (TSA) and azacytidine (aza-dC), inhibitors of histone deacetylase and DNA methyltransferase, respectively, on rRNA gene silencing. Consistent with the results described above, overexpression of TIP5 reduced the levels of both reporter transcripts and cellular pre-rRNA in a dose-dependent manner (Fig. 2A). The repressive effect of TIP5 on transcriptional activity was relieved by either TSA or aza-dC. No additive or synergistic effect was observed, indicating that histone deacetylation and DNA methylation operate along a common mechanistic pathway. While TIP5-mediated histone H4 deacetylation was impaired by TSA treatment, H3-Lys9 dimethylation was not affected (Fig. 2B). This finding underscores the requirement for histone H4 deacetylation in TIP5-mediated rRNA gene silencing and suggests that histone H3-Lys9 dimethylation occurs independently of histone H4 deacetylation. In contrast, inhibition of DNA methylation by aza-dC treatment affected neither histone H4 deacetylation nor H3-Lys9 dimethylation, indicating that histone deacetylation and methylation occur prior to DNA methylation.

To examine whether histone H4 deacetylation is a prerequisite for DNA methylation, cells overexpressing TIP5 were treated with TSA, and de novo methylation of a cotransfected rRNA gene reporter plasmid (pMr1930-CBH) was analyzed. Cross-linked chromatin was immunoprecipitated with anti-FLAG antibodies (Fig. 2C, middle), DNA was digested with methylation-sensitive isoschizomers HpaII and MspI, and coprecipitated plasmid DNA was amplified with forward primers that map upstream or downstream of the CCGG site (at position −143) in the rRNA gene promoter and a backward primer that is complementary to part of the marker gene (the CAT gene) inserted between the 5′- and 3′-terminal rRNA gene fragments of the reporter construct (Fig. 2C). Consistent with overexpression of TIP5 triggering de novo DNA methylation, a fraction of the reporter plasmid became resistant to HpaII cleavage (Fig. 2C, bottom, lanes 1 and 2). Significantly, de novo DNA methylation did not occur in the presence of TSA (Fig. 2C), indicating that NoRC-mediated DNA methylation depends on prior histone deacetylation. These data demonstrate the mutual interdependence of histone deacetylation and DNA methylation and imply that deacetylation of nucleosomes at the rRNA gene promoter is required for DNA methylation.

Nucleosome remodeling is required for DNA methylation and transcriptional silencing.

NoRC has been shown to use the energy of ATP hydrolysis to alter the topology of nucleosomal DNA (39). To investigate the contribution of NoRC-mediated chromatin remodeling to rRNA gene silencing, we monitored Pol I transcription after coexpressing TIP5 either with SNF2h or with SNF2h_K211R, an ATPase-deficient mutant. Coexpression of SNF2h enhanced the transcriptional repression mediated by TIP5 from 60 to 80% (Fig. 3A). This finding is consistent with coimmunoprecipitation data showing that the majority of ectopic TIP5 associates with endogenous SNF2h (unpublished results) and explains why moderate overexpression of TIP5 is sufficient to silence the rRNA genes. Strikingly, if SNF2h_K211R was coexpressed with TIP5, transcriptional repression was eliminated. This result demonstrates that the ATP-dependent chromatin-remodeling activity of NoRC plays an indispensable role in rRNA gene silencing.

If ATP-dependent nucleosome remodeling is necessary for NoRC-mediated rRNA gene silencing, a step may exist that establishes a link between chromatin remodeling, histone modification, DNA methylation, and initiation complex formation. We therefore sought to examine whether NoRC-dependent chromatin remodeling would facilitate the access of histone-modifying enzymes and DNA methyltransferase to the rRNA genes. For this purpose, we monitored histone H4 acetylation and H3-Lys9 dimethylation at the rRNA gene reporter after the coexpression of FLAG-TIP5 with SNF2h and SNF2h_K211R, respectively. As shown in Fig. 3B, both histone deacetylation and H3-Lys9 dimethylation occurred at the rRNA gene promoter, regardless of whether wild-type or mutant SNF2h was coexpressed with TIP5. This result indicates that the remodeling activity of NoRC is not required for histone deacetylation and H3-Lys9 dimethylation.

Next, we tested whether the remodeling activity of NoRC is required for de novo methylation of the rRNA genes. Consistent with the transcription data, the overexpression of TIP5 increased the amount of HpaII-resistant methylated rRNA genes by about threefold in both the absence and presence of SNF2h (Fig. 3C). In contrast, the overexpression of TIP5 with SNF2h_K211R did not confer HpaII resistance to the rRNA genes. These results indicate that ATP-dependent chromatin remodeling is required for de novo DNA methylation and that DNA methylation occurs subsequent to chromatin remodeling.

NoRC-mediated transcriptional repression requires DNA methylation of CpG at position −133.

Previous work has established that methylation of a single CpG residue at −133 within the UCE of the rRNA gene promoter eliminates binding of the basal transcription factor UBF to nucleosomal DNA, thereby preventing transcription complex formation (34). If methylation of CpG at −133 was the final step in the chain of events that silence the rRNA genes, the overexpression of TIP5 should not repress the transcription of a mutant template in which the cytosine at −133 was replaced by guanosine. Indeed, transcription of the mutant (pMr600₋₁₃₃-BH) was not repressed by increasing amounts of FLAG-TIP5 (Fig. 4A), demonstrating that transcriptional repression by TIP5 requires the methylation of a functionally important CpG residue within the UCE. Moreover, ChIP assays revealed that the overexpression of TIP5 caused the deacetylation and dimethylation of histones at both the wild-type and mutant reporter plasmids (Fig. 4B). This finding implies that histone deacetylation and H3-Lys9 methylation, i.e., modifications that mark heterochromatin, are not sufficient for rRNA gene silencing. In addition, this result reveals that CpG methylation is the final step in the chain of events that silence rRNA gene transcription.

We have previously shown that methylation of CpG at −133 impairs UBF binding to the rRNA genes assembled into chromatin (34). We therefore reasoned that overexpression of TIP5 should not decrease the level of UBF associated with the mutant template that cannot be methylated at CpG at −133. Indeed, overexpression of TIP5 eliminated binding of UBF to the wild-type (pMr600-BH) but not the mutant (pMr600₋₁₃₃-BH) template (Fig. 4C), indicating that NoRC-mediated DNA methylation prevents binding of UBF to the UCE. As a consequence, the assembly of productive transcription initiation complexes is impaired. Thus, NoRC establishes rRNA gene silencing by coordinating a hierarchical order of events that modify nucleosomes, methylate DNA, and prevent the formation of Pol I transcription initiation complexes.