SUMMARYThe human genome encodes thousands of non-coding RNAs. Many of these terminate early and are then rapidly degraded, but how their transcription is restricted is poorly understood. In a screen for protein-coding gene termination factors, we identified ZC3H4. However, its depletion causes upregulation and extension of hundreds of unstable transcripts, particularly antisense RNAs and those transcribed from so-called super-enhancers. These loci are occupied by ZC3H4, suggesting that it directly functions in their transcription. Consistently, engineered tethering of ZC3H4 to reporter RNA promotes its degradation by the exosome. ZC3H4 is metazoan-specific - interesting when considering its impact on enhancer RNAs that are less prominent in single-celled organisms. Finally, ZC3H4 loss causes a substantial reduction in cell proliferation, highlighting its overall importance. In summary, we identify ZC3H4 as a factor that plays an important role in restricting non-coding transcription in multi-cellular organisms.

Gene expression is controlled in a dynamic and regulated manner to allow for the consistent and steady expression of some proteins as well as the rapidly changing production of other proteins. Transcription initiation has been a major focus of study because it is highly regulated. However, termination of transcription also plays an important role in controlling gene expression. Transcription termination on protein-coding genes is intimately linked with 3' end cleavage and polyadenylation of transcripts, and it generally results in the production of a mature mRNA that is exported from the nucleus. Termination on many non-coding genes can also result in the production of a mature transcript. Termination is dynamically regulated-premature termination and transcription readthrough occur in response to a number of cellular signals, and these can have varied consequences on gene expression. Here, we review eukaryotic transcription termination by RNA polymerase II (RNAPII), focusing on protein-coding genes.

Significance
.
. Stable endosymbiosis between eukaryotic microbes has driven the evolution of further cellular complexity. Yet the mechanisms that can act to stabilize an emergent eukaryote–eukaryote endosymbiosis are unclear. Using the model facultative endosymbiotic system,
. Paramecium bursaria
. we demonstrate that endosymbiont–host RNA–RNA interactions can drive a cost to host growth upon endosymbiont digestion. These RNA–RNA interactions are facilitated by the host RNA-interference system. For endosymbiont messenger RNA sharing a high level of sequence identity with host transcripts, this process can result in host gene knockdown. We propose that these endosymbiont–host RNA–RNA interactions—“RNA-interference collisions”—represent an emergent mechanism to sanction the host for breakdown of the endosymbiosis, promoting the stability of the facultative endosymbiotic interaction.
.

The human genome encodes thousands of non-coding RNAs. Many of these terminate early and are then rapidly degraded, but how their transcription is restricted is poorly understood. In a screen for protein-coding gene transcriptional termination factors, we identified ZC3H4. Its depletion causes upregulation and extension of hundreds of unstable transcripts, particularly antisense RNAs and those transcribed from so-called super-enhancers. These loci are occupied by ZC3H4, suggesting that it directly functions in their transcription. Consistently, engineered tethering of ZC3H4 to reporter RNA promotes its degradation by the exosome. ZC3H4 is predominantly metazoan -interesting when considering its impact on enhancer RNAs that are less prominent in single-celled organisms. Finally, ZC3H4 loss causes a substantial reduction in cell proliferation, highlighting its overall importance. In summary, we identify ZC3H4 as playing an important role in restricting non-coding transcription in multicellular organisms.

The allosteric and torpedo models have been used for 30 yr to explain how transcription terminates on protein-coding genes. The former invokes termination via conformational changes in the transcription complex and the latter proposes that degradation of the downstream product of poly(A) signal (PAS) processing is important. Here, we describe a single mechanism incorporating features of both models. We show that termination is completely abolished by rapid elimination of CPSF73, which causes very extensive transcriptional readthrough genome-wide. This is because CPSF73 functions upstream of modifications to the elongation complex and provides an entry site for the XRN2 torpedo. Rapid depletion of XRN2 enriches these events that we show are underpinned by protein phosphatase 1 (PP1) activity, the inhibition of which extends readthrough in the absence of XRN2. Our results suggest a combined allosteric/torpedo mechanism, in which PP1-dependent slowing down of polymerases over termination regions facilitates their pursuit/capture by XRN2 following PAS processing.

Many RNA polymerases terminate transcription using allosteric/intrinsic mechanisms, whereby protein alterations or nucleotide sequences promote their release from DNA. RNA polymerase II (Pol II) is somewhat different based on its behavior at protein-coding genes where termination additionally requires endoribonucleolytic cleavage and subsequent 5'→3' exoribonuclease activity. The Pol-II-transcribed small nuclear RNAs (snRNAs) also undergo endoribonucleolytic cleavage by the Integrator complex, which promotes their transcriptional termination. Here, we confirm the involvement of Integrator but show that Integrator-independent processes can terminate snRNA transcription both in its absence and naturally. This is often associated with exosome degradation of snRNA precursors that long-read sequencing analysis reveals as frequently terminating at T-runs located downstream of some snRNAs. This finding suggests a unifying vulnerability of RNA polymerases to such sequences given their well-known roles in terminating Pol III and bacterial RNA polymerase.

RNA polymerase II (Pol II) transcribes hundreds of thousands of transcription units – a reaction always brought to a close by its termination. Because Pol II transcribes multiple gene types, its termination occurs in a variety of ways, with the polymerase being responsive to different inputs. Moreover, it is not just a default process occurring at the end of genes. Promoter-proximal and premature termination is common and might in turn regulate gene expression levels. Although some transcription termination mechanisms have been debated for decades, research is only just underway on emergent processes. We provide an updated view of transcription termination in human cells, highlighting common themes and some interesting differences between the contexts in which it occurs.

Cell-based studies of human ribonucleases traditionally rely on methods that deplete proteins slowly. We engineered cells in which the 3′→5′ exoribonucleases of the exosome complex, DIS3 and EXOSC10, can be rapidly eliminated to assess their immediate roles in nuclear RNA biology. The loss of DIS3 has the greatest impact, causing the substantial accumulation of thousands of transcripts within 60 min. These transcripts include enhancer RNAs, promoter upstream transcripts (PROMPTs), and products of premature cleavage and polyadenylation (PCPA). These transcripts are unaffected by the rapid loss of EXOSC10, suggesting that they are rarely targeted to it. More direct detection of EXOSC10-bound transcripts revealed its substrates to prominently include short 3′ extended ribosomal and small nucleolar RNAs. Finally, the 5′→3′ exoribonuclease, XRN2, has little activity on exosome substrates, but its elimination uncovers different mechanisms for the early termination of transcription from protein-coding gene promoters.

Every transcription cycle ends in termination when RNA polymerase dissociates from the DNA. Although conceptually simple, the mechanism has proven somewhat elusive in eukaryotic systems. Gene-editing and high resolution polymerase mapping now offer clarification of important steps preceding transcriptional termination by RNA polymerase II in human cells.

Termination is a ubiquitous phase in every transcription cycle but is incompletely understood and a subject of debate. We used gene editing as a new approach to address its mechanism through engineered conditional depletion of the 5' → 3' exonuclease Xrn2 or the polyadenylation signal (PAS) endonuclease CPSF73 (cleavage and polyadenylation specificity factor 73). The ability to rapidly control Xrn2 reveals a clear and general role for it in cotranscriptional degradation of 3' flanking region RNA and transcriptional termination. This defect is characterized genome-wide at high resolution using mammalian native elongating transcript sequencing (mNET-seq). An Xrn2 effect on termination requires prior RNA cleavage, and we provide evidence for this by showing that catalytically inactive CPSF73 cannot restore termination to cells lacking functional CPSF73. Notably, Xrn2 plays no significant role in either Histone or small nuclear RNA (snRNA) gene termination even though both RNA classes undergo 3' end cleavage. In sum, efficient termination on most protein-coding genes involves CPSF73-mediated RNA cleavage and cotranscriptional degradation of polymerase-associated RNA by Xrn2. However, as CPSF73 loss caused more extensive readthrough transcription than Xrn2 elimination, it likely plays a more underpinning role in termination.

Poly(A) signals are required for efficient 3' end formation and transcriptional termination of most protein-encoding genes transcribed by RNA polymerase II. However, transcription can extend far beyond the poly(A) site before termination occurs. This implies the existence of further downstream termination signals. In mammals, a variety of sequence elements, in addition to the poly(A) site, have been implicated in the termination process. For example, termination of the human beta- and epsilon-globin genes is mediated by a sequence downstream of the poly(A) site that promotes an RNA cotranscriptional cleavage (CoTC). Here we report the identification of multiple termination sequences in the mouse serum albumin (MSA) 3' flanking region. Many transcripts from this region are cleaved cotranscriptionally, implying that such cleavage of pre-mRNA may be a more general feature of transcriptional termination.

In the past few years, especially since the discovery of RNA interference (RNAi), our understanding of the role of RNA in gene expression has undergone a significant transformation. This change has been brought about by growing evidence that RNA is more complex and transcription more promiscuous than has previously been thought. Many of the new transcripts are of so-called noncoding RNA (ncRNA); i.e. RNA that does not code for proteins such as mRNA, or intrinsic parts of the cellular machinery such as the highly structured RNA components of ribosomes (rRNA) and the small nuclear RNA (snRNA) components of the splicing machinery. It is becoming increasingly apparent that ncRNAs have very important roles in gene expression. This paper focuses on work from our laboratory in which we have investigated the roles and turnover of ncRNA located within the gene pre-mRNA, which we refer to as intragenic ncRNA. Also discussed are some investigations of intergenic ncRNA transcription and how these two classes of ncRNA may interrelate.

New evidence indicates that termination of transcription is an important regulatory step, closely related to transcriptional interference and even transcriptional initiation. However, how this occurs is poorly understood. Recently, in vivo analysis of transcriptional termination for the human beta-globin gene revealed a new phenomenon--co-transcriptional cleavage (CoTC). This primary cleavage event within beta-globin pre-messenger RNA, downstream of the poly(A) site, is critical for efficient transcriptional termination by RNA polymerase II. Here we show that the CoTC process in the human beta-globin gene involves an RNA self-cleaving activity. We characterize the autocatalytic core of the CoTC ribozyme and show its functional role in efficient termination in vivo. The identified core CoTC is highly conserved in the 3' flanking regions of other primate beta-globin genes. Functionally, it resembles the 3' processive, self-cleaving ribozymes described for the protein-encoding genes from the myxomycetes Didymium iridis and Physarum polycephalum, indicating evolutionary conservation of this molecular process. We predict that regulated autocatalytic cleavage elements within pre-mRNAs may be a general phenomenon and that functionally it may provide the entry point for exonucleases involved in mRNA maturation, turnover and, in particular, transcriptional termination.