The complex eukaryotic transcriptome: unexpected pervasive transcription and novel small RNAs

Top Cited Papers
Publisher Website
Google Scholar
Abstract
Initial transcriptome analyses were limited to the quantification of the transcripts that correspond to known features, such as mRNA genes or stable non-coding RNAs. Projects aimed at the exhaustive identification of RNAs, unbiased by genome annotations, have uncovered unexpected complexity of the transcriptome in eukaryotes. Several transcripts are generated from genomic regions that were previously thought to be silent or antisense to genes. This widespread genomic transcription is often called 'pervasive transcription'. Advances in understanding the complexity of eukaryotic transcriptomes have been driven by technological breakthroughs. Several distinct approaches have recently evolved to allow rapid, unbiased, genome-wide analyses of transcriptomes. First, DNA microarrays have been developed to the point at which tiled oligonucleotides span whole genomes (genomic tiling microarrays). Second, the development of next-generation sequencing techniques has allowed the efficient and quantitative analysis of sequence tags that either cover whole transcripts or are enriched for the 5′ or 3′ extremities of the RNAs. The new genomic approaches for transcriptome analysis have uncovered a variety of non-coding transcripts. Recently, several classes of small non-coding RNAs have been found to be associated with gene promoters in animals. Although these different classes of RNA differ in their characteristics — such as their modal length — their distribution with respect to gene transcription start sites (TSSs) is remarkably similar. Some of these small RNAs are transcribed in the same orientation as the mRNAs and are usually located a short distance downstream of the gene TSS, and others are transcribed in the opposite direction to the mRNA from upstream of the gene TSS. The origin of these promoter-associated RNAs is unknown. However, several observations point to a possible relationship with the so-called 'paused' polymerases, that is, polymerases that engage in transcription but that pause a few dozens of nucleotides downstream of the TSS. In particular, the distribution of the promoter-associated small RNAs is similar to that of 'engaged' RNA polymerase II, as determined by a novel genome-wide run-on technique. However, how the small RNAs and paused polymerases might be related remains puzzling and is far from established. In yeast, an important part of pervasive transcription gives rise to highly unstable transcripts called cryptic unstable transcripts (CUTs). These transcripts are heterogeneous at their 3′ ends and range in size from ∼200 to ∼600 nucleotides. Another class of more stable transcripts has been distinguished and named stable unannotated transcripts (SUTs), although there is not a clear demarcation between the two classes. Like the promoter-associated small RNAs found in animals, these RNAs are mostly transcribed from nucleosome-free regions, in particular those associated with gene promoters. In addition, they show a divergent distribution profile. However, this profile is not completely equivalent to that observed in animals, as their TSSs are almost exclusively located upstream of the gene TSSs, whether or not they are transcribed in the sense orientation or in divergent orientation relative to the gene. The majority of CUTs and SUTs are divergent from their associated mRNAs. One model for their origin, which is supported by mutational analysis of one example, is that the assembly of pre-initiation complexes (PICs) during transcription initiation is poorly polarized, and so cryptic PICs are often assembled in the wrong orientation relative to the gene. The transcripts they generate are efficiently degraded by an efficient quality control mechanism. The nature of the quality control mechanism that targets CUTs for rapid degradation is well understood. This mechanism is coupled to the peculiar mode of termination of transcription for these RNAs, which resembles the transcription termination of small nucleolar RNAs. This mode of termination is coupled to exonucleolytic degradation by the exosome, assisted by a novel poly(A) polymerase-containing complex called the Trf4–Air2–Mtr4p polyadenylation (TRAMP) complex. Although the role of divergent CUTs in regulation is unknown, several different specific regulation mechanisms have been described that use antisense SUTs or generate sense CUTs. How widespread the use of these unconventional regulation mechanisms is remains to be determined. Likewise, in animals, the general role of promoter-associated transcription remains enigmatic, although several different mechanisms have been described that make use of such transcripts as effectors of gene regulation. Importantly, whatever the precise mechanism is that generates promoter-associated small non-coding RNAs in yeast and animals, these studies indicate that transcription initiation is a poorly polarized process and many, if not most, promoter regions therefore seem to be intrinsically bidirectional.