It has been reported that the Paf1 complex associates with RNA pol II and that loss of the Ctr9 or Paf1 components of this complex impairs transcription of a subset of genes. We find that the amount of Paf1 associated with a particular gene is positively correlated with its rate of transcription initiation. The Paf1 complex supports specific histone modifications that are essential for maintenance of telomeric silencing, including monoubiquitylation of histone H2B and methylation of histone H3 at lysine 4. In agreement with previous reports, we detect a loss of telomeric silencing in CTR9 and PAF1 deletion mutants, but demonstrate in addition that the loss of these genes also impairs transcription of telomeric markers.
The Paf1 complex is likely required for efficient transcriptional elongation [15
], but the underlying mechanism has remained unclear. A recent report indicates that Paf1 complex-dependent ubiquitylation of H2B is required for efficient transcription elongation in vitro
, and further suggests that ubiquitylation of H2B permits FACT to efficiently displace H2A/H2B histone dimers [14
]. Here we have shown that the loss of Ctr9 or Paf1 in vivo
is associated with delayed removal of histones from the GAL1
locus upon induction of transcription. This delay in histone displacement may contribute to the decrease in association and altered distribution of RNA pol II that we observed in Ctr9- or Paf1-deficient cells. Our findings may also explain the relative delay in induction of GAL
transcripts in ctr9Δ
strains, relative to wild-type.
We have considered the possibility that a reduction in recruitment of RNA pol II, rather than an impairment of histone eviction, might be responsible for the delay in histone displacement that we observed. In support of our interpretation are two reports documenting recruitment of the Paf1 complex to promoter regions after, but not before, initiation of transcription. In mammalian cells the Paf1 complex is recruited to the RARβ2
promoter after recruitment of RNA pol II and FACT [14
]. Moreover, RNA pol II binds to the yeast ARG1
promoter in the absence of the Paf1 complex; subsequent recruitment of the Paf1 complex is dependent on prior binding of the Spt4 transcription elongation factor and phosphorylation of RNA pol II at Ser5 [36
]. Lastly, association of Rad6 and the SAGA complex with the GAL1
promoter occurs independently of H2B monoubiquitylation [31
]. Taken together, these observations suggest that recruitment of RNA pol II to promoter regions occurs prior to and independent of the actions of Ctr9 or Paf1. Our results are consistent with a direct role for the Paf1 complex in promoting histone eviction but do not exclude the possibility that Paf1 and Ctr9 exert their effects through one or more additional components of the elongating RNA pol II complex.
It has been proposed that newly transcribed genes undergo a so-called "pioneer round" of transcription in which RNA pol II functions in concert with specialized machinery for chromatin modification [37
]. Based on our work and earlier findings we suggest that upon initial induction of GAL1
, Gal4, Rad6/Bre1, SAGA and RNA pol II, among other factors, are recruited and initiate transcription [31
]. Subsequent association of the Paf1 complex with RNA pol II would promote monoubiquitylation of H2B at K123 by Rad6/Bre1 as the polymerase complex traverses the coding sequence [32
]. We imagine that ubiquitylation of K123 could confer upon FACT and Asf1 the ability to evict histones from the DNA in addition to the established roles of FACT and Asf1 in histone displacement and redeposition. Our results suggest that in the absence of the Paf1 complex histones are inefficiently removed and transcriptional elongation is impaired.
While it has been suggested that the Paf1 complex does not affect transcriptional elongation from a GAL1
], this conclusion was based on observations made at 2.5 hr after galactose induction. Because our data indicate that the effects of Ctr9 and Paf1 depletion on histone eviction are most evident early after induction, it is not surprising that differences in elongation rates were not observed at a relatively late time point. Our results suggest that the Paf1 complex exerts its effects on histone displacement during the initial stages of transcriptional activation, and that at later times after induction, histones can be removed even in the absence of the Paf1 complex.
The Paf1 complex and the H2B monoubiquitylation that it promotes are clearly not universally required for gene expression [39
]. Unlike FACT, the genes encoding components of the Paf1 complex and Rad6/Bre1 are inessential [18
]. The available data suggest that the Paf1 complex enhances transcriptional efficiency in those settings in which environmental changes demand rapid transcriptional responses, such as induction of metabolic enzymes or stress response genes. In this regard, it is particularly interesting that one salient phenotype of paf1Δ
strains is sensitivity to a broad range of environmental stressors [42