In this study, we applied genome-wide location analysis to uncover genomic features previously not known to associate with Hmt1, the yeast homolog of human PRMT1. To this end, we identified binding of Hmt1 to a number of ncRNA genes such as tRNA and snoRNA genes. Enrichment of Hmt1 binding at these genomic features directly implicates a regulatory role for this protein in biological processes or pathways involving these genomic features. We validated the biological significance of Hmt1 binding at the tRNA genes using RNA hybridization analyses in assaying Hmt1 loss-of-function mutants, which display an increase in the steady-state levels of pre- and mature tRNA when compared to the wild-type cells. This increase is likely attributed to a change in the overall transcriptional output by RNAPIII, as opposed to a defect in the maturation process of tRNAs. The physical association between Hmt1 and TFIIIB component Bdp1 provides a likely explanation for Hmt1’s association at many of the tRNA genes in our ChIP-chip data.
Hmt1 has previously been demonstrated to play a role in the recruitment of mRNP components such as pre-mRNA splicing factors  and mRNA export factors . Furthermore, it regulates transcriptional elongation and termination via methylation of hnRNP and mRNA export factor Npl3 . Thus, it was not surprising for us to find the majority of Hmt1 binding sites fall proximal to or within an ORF. GO-enrichment analysis indicates Hmt1 binding to genes involved in translation, suggesting a potential role for this enzyme in modulating the expression of these genes. Transcriptional profiling experiments carried out previously on both null and catalytically-inactive mutants of Hmt1, however, show no significant changes in the abundance of these transcripts from the wild-type cells . Based on these observations, one possible explanation is that Hmt1’s effects on these genes can only be seen under a specific circumstance, such as a change in the environmental condition or an exposure to certain stress agent. This scenario would support the observed growth phenotype for the Hmt1-null mutants as they display no distinguishable changes in their growth rate from wild-type cells when grown in a rich media, such as YPD . Transcription of HMT1 is repressed in response to heat shock when compared to steady-state growth rate conditions , further suggesting that Hmt1 is needed the most only under a certain growth condition. It is interesting to note that Hmt1 has different occupancy patterns across genes with different transcriptional rates, as this observation demonstrates a potential for different modes of regulation during gene expression by Hmt1. There are two functional implications of Hmt1’s enrichment across the various transcriptional frequency gene classes – first, the highly transcribed gene class contains many of the ribosomal protein genes that encompass many intron-containing genes and Hmt1 has been shown to be more enriched at the 5’-end of these genes. Since Hmt1 affects the co-transcriptional recruitment of pre-mRNA splicing factors , Hmt1’s enrichment across this region may reflect its role in regulating this process. Second, Hmt1’s enrichment across all ORFs shows a “spike” at the 3’-end of the genes, with this trend more pronounced in gene classes with medium and low transcriptional frequencies. This enrichment trend may reflect the association of Hmt1 with known substrates involved in 3’-end processing of a transcript, such as Yra1, Hrp1, and Nab2. Thus, these observations suggest that Hmt1 may methylate its substrate during transcription rather than prior to.
Hmt1 methylates H4R3 in vitro and loss-of-function mutants of Hmt1 display decreased dimethylated H4R3 levels across silent chromatin regions , despite the observation that bulk dimethylated H4R3 is unchanged in hmt1 mutants . Thus, Hmt1-catalyzed H4R3me2 in S. cerevisiae is associated with gene silencing. H4R3 methylation by PRMT1 in vivo is required for many subsequent histone modifications in mammalian cells . It can be inferred then that Hmt1 may also play a role in modulating the levels of other histone modifications. By determining the enrichment or depletion of core histones and other histone modifications for Hmt1-bound ORFs, the data reveal that Hmt1 may coordinate regulation at the promoter with H3-specific acetyltransferases such as SAGA or NuaA3 and the Rpd3(S) complex, which is an H3K36 methyl reader [51–53]. Furthermore, our ORF-body analysis suggests that Hmt1 may coordinate the action of COMPASS (responsible for H3K4 methylation), Set2 (responsible for H3K36 methylation) as well as KDACs [54–57]. Results from our directed ChIP experiments, however, suggests that these histone modification marks may regulate the recruitment of Hmt1 to ORF-containing gene bodies instead, at least for pan-H4 acetylation and the majority of H4K36me2 tested. It is possible that the loss of Hmt1 alone is not sufficient to alter these histone modification states we have identified. Rather, it may be the combination of Hmt1-deletion with the loss of another regulator that effects a significant change in histone modification levels as compared to any single loss-of-function mutant.
Perhaps the most interesting hypothesis we obtained from analyzing our ChIP-chip data is a role for Hmt1 in the biogenesis of tRNAs. In eukaryotic cells, tRNAs are classified on the basis of their anticodon, of which there are 42 different tRNA species in the S. cerevisiae[58, 59]. Furthermore, the actual number of different tRNA species is greater in some cases because there can be species with identical anticodons but sequence differences in the tRNA body . In S. cerevisiae, each tRNA species can be encoded by several genomic copies (tDNAs), of which there are a total of 274 nuclear-encoded tDNAs within the genome . Hmt1 binding is enriched at many of these tDNA alleles and mutants lacking Hmt1 or its catalytic activity display higher levels of the corresponding tRNA abundance. This overall change is likely attributed to a defect in the output rather than the processing during tRNA biogenesis, as both pre- and mature tRNAs increase in a similar fashion in the Hmt1 loss-of-function mutants. If the change in the mature tRNA abundance was attributed to a defect in tRNA processing in the Hmt1 mutants, we would expect the ratio (pre- to mature tRNA) of signals obtained from our RNA hybridization analysis to reflect such.
Transcription of tRNAs is accomplished by RNAPIII and this process requires two gene-internal promoters, boxA and boxB reviewed in [61, 62]. During tRNA gene transcription, these internal promoters are bound by the hetero-heptameric complex TFIIIC. TFIIIC recruits the hetero-trimeric TFIIIB (which consists of Bdp1, Brf1, and TBP) to the upstream region of the tDNA. TFIIIB, in turn, recruits RNAPIII to prompt transcriptional initiation. The physical interaction between Hmt1 and Bdp1 provides a probable mechanism for how Hmt1 may be recruited to these tDNAs. Additionally, this physical interaction gave insights into how Hmt1 may affect the overall tRNA abundance through its ability to regulate RNAPIII transcription of tRNA genes. While the precise molecular mechanisms by which Hmt1 accomplishes this remains to be elucidated, our data provide a valuable clue in that Hmt1 does so via its catalytic activity, based on our data from the Hmt1 catalytically inactive mutant. This points to a yet-to-be discovered substrate of Hmt1 in which its methylation may affect the transcription of tRNA genes by RNAPIII. The biological significance of this transcriptional repression by Hmt1 could reflect a way for coordinating gene expression between RNAPII and RNAPIII, given the known role of Hmt1 in coordinating mRNP biogenesis catalyzed by RNAPII. As it is advantageous for cells to efficiently utilize its resources during growth (especially in sub-optimal conditions), proper coordination would be important for not wasting any valuable resources in executing gene expression programs. Thus, it can be inferred that Hmt1’s function in balancing these gene expression programs may not be as critical when cells are growing under optimal conditions with plenty of resources, as supported by the observation that hmt1-null cells do not display any obvious growth defect in rich media. Further experiments are required to define the specific molecular mechanisms used by Hmt1 in this process as well as other roles Hmt1 may play in the described occupancy relationship for genomic features found within Hmt1 binding sites. In sum, our genome-wide location analysis data do reveal multiple potentially new roles for Hmt1 in the control of eukaryotic gene expression in vivo, especially in the realm of ncRNA biogenesis.