A new direction for gene looping

Abstract

Upon binding to a promoter, RNA polymerase II can synthesize either a coding mRNA or a divergently transcribed non-coding RNA. In a recent issue of Science, Tan-Wong et al. (2012) find that intragenic looping increases the proper orientation of RNA polymerase II, reducing the production of divergent non-coding transcripts.

Chromatin frequently assumes higher order arrangements that facilitate transcriptional regulation. For example, chromatin loops can bring distal regulatory elements into close proximity to promoters (Krivega and Dean, 2012). Such loops can promote gene expression by allowing distal enhancers to contact a promoter; they can also function to insulate neighboring chromatin domains. Genes themselves can also loop; such intragenic loops occur through interaction of the promoter and the terminator (O’Sullivan et al., 2004). Intragenic looping is transcription-dependent and requires components of the transcription preinitiation complex (TFIIB) and pre-mRNA 3′-end processing complex (Hampsey et al., 2011). Chromosome conformation capture (3C) revealed intragenic looping of many genes, including the yeast genes GAL10 (2.1kb), HEM3 (1.0kb), and FMP27 (7.9kb), as well as the mammalian genes BRCA1 and CD68, and the HIV-1 provirus (Hampsey et al., 2011). Although intragenic looping requires transcription, loss of looping does not strongly affect transcription (Singh and Hampsey, 2007). For a few genes, it has been suggested that intragenic looping might affect their reactivation rate after repression, a phenomenon called transcriptional memory. However, the general functional significance of intragenic looping still remains unclear.

In a recent issue of Science, Proudfoot, Steinmetz and colleagues described work suggesting that intragenic looping plays an important role in regulating divergent transcription, reducing the production of divergently transcribed non-coding RNAs (ncRNAs) (Tan-Wong et al., 2012). The phenomenon of divergent transcription is common to most active promoters in diverse organisms (Seila et al., 2009). Upon assembly of the preinitiation complex, RNA polymerase II (RNAPII) can initiate and transcribe in either direction, one producing an mRNA and the other producing a short, rapidly degraded ncRNA. These cryptic unstable transcripts (CUTs) are widespread but scarce, because they are rapidly degraded by the nuclear exosome (Arigo et al., 2006). They can arise from the nucleosome free regions associated with promoters or the 3′ end of genes (Xu et al., 2009). It is unclear if the production of CUTs has any adaptive value or if it is merely a cost associated with a permissive nucleosome arrangement. However, it is intriguing that CUTs regulate expression of certain mRNAs by recruiting repressive histone modifying factors to the promoter (Camblong et al., 2007).

The authors tested the hypothesis that intragenic looping enhances the directionality of transcription by examining the expression of a divergently transcribed ncRNA at the FMP27 locus in S. cerevisiae (Tan-Wong et al., 2012). A mutation in Ssu72 (ssu72-2), a component of the cleavage/polyadenylation factor that also interacts with the preinitiation factor TFIIB (Hampsey et al., 2011), blocks intragenic looping and leads to increased accumulation of a divergently transcribed ncRNA and increased RNAPII density over FMP27 promoter (Figure 1.). Genome-wide profiling of total RNA in wild type and ssu72-2 mutant strains, using strand-specific microarrays, identified many ncRNAs that were induced. In addition to assessing the effect of Ssu72 loss, the authors also examined the effect of loss of Rrp6, a component of the nuclear exosome(Arigo et al., 2006). When RNA from ssu72-2 cells, rrp6Δ cells, and rrp6Δ ssu72-2 cells was compared with RNA from wild type cells, the authors observed both CUTs (ncRNAs that accumulate in rrp6Δ mutants) as well as additional ncRNAs that accumulated in the ssu72-2 mutants. These additional ncRNA were named SRTs (Ssu72-restricted transcripts). Like CUTs, SRTs frequently arise from promoter regions in a divergent orientation from the gene (Figure 1).

Figure 1. Gene loops enhance transcriptional directionality.

(A) Top: inactive gene, with different portions indicated; bottom: active gene. Actively transcribed genes form an intragenic loop between their promoters and terminators. (B) A mutation in Ssu72 (ssu72-2) results in loss of intragenic looping and divergent transcription of promoter-associated ncRNA. In (A) and (B) PIC, pre-initiation complex; Pol II, RNA polyermase II; CPF, cleavage and polyadenylation factor. Wild type Ssu72 appears dark blue, whereas mutant ssu72-2 appears light blue.

Of the 605 SRTs and 1982 CUTs identified in the array profile, the authors focused on the 135 SRTs and 678 CUTs that were transcribed divergently between tandem ORFs. The ssu72 mutation resulted in additional RNAPII accumulation upstream of TSSs and over SRTs. Mutations in other factors required for intragenic looping, such as TFIIB (Sua7) and Pta1, also increased divergently transcribed SRTs. Additionally, loss of Ssu72 led to increased histone H4 acetylation over SRT-producing promoters. Overall, this suggests that gene looping decreases divergent transcription by a mechanism that involves histone H4 deacetylation. Loss of the histone H4 deacetylase Rco1 led to expression of many ncRNAs. However, the ncRNAs induced by loss of Rco1 are derived from the 3′ end of genes, as opposed to SRTs, which are derived from divergent transcription from promoters. This suggests that intragenic looping has a direct role in regulating transcriptional directionality.

To test if cis mutations that affect intragenic looping would also lead to changes in RNAPII directionality, the authors examined the effects of replacing the polyadenylation signal (PAS) in the 3′ UTR with an Rnt1 cleavage signal (RCS). This results in normal termination but blocks polyadenylation and intragenic looping. Replacement of the PAS with RCS in two yeast genes and in the β-globin transgene in human embryonic kidney cells increased the divergent transcription of ncRNAs by three fold. This suggests that intragenic looping plays a conserved role in regulating transcriptional directionality.

These results suggest that formation of gene loops influence unidirectional transcription. How might this work? Based on the acetylation of histone H4 in promoters of genes that exhibit divergent SRTs, the authors postulate that looping leads to directional histone deacetylation and repression upstream of the promoter. An alternative view is that looping leads to directional acetylation within the loop. Also, because recruitment of RNAPII to the promoter is often rate-limiting, if intragenic looping permits more efficient recycling of RNAPII for reinitiation, it is tempting to speculate that this might bias transcriptional directionality. Many components of the preinitiation complex remain associated with the promoter, potentially serving as a scaffold to allow for such recycling. Consistent with this notion, RNAPII associated with the active hsp70 locus in flies is not readily exchanged with the nuclear pool, suggested that this locus is somehow “compartmentalized” and that RNAPII is recycled (Zobeck et al., 2010). Resolutions of these questions will await a better understanding of how looping affects chromatin structure, histone acetylation and RNAPII function.