NEW YORK (GenomeWeb News) – A trio of papers appearing online yesterday in Nature Genetics is providing new insights into the networks regulating transcription in animal cells. The papers, by three different research groups, were all done as part of the fourth stage of the Functional Annotation of Mouse, or FANTOM4, project.
In the first study, researchers from Australia, Japan, Italy, and Scotland examined the role of repetitive elements in regulating transcriptional activity. The team found that between six and 30 percent of the mouse and human RNA transcripts they evaluated initiate inside repetitive parts of the genome — usually in a tissue and/or species specific way.
"With the use of powerful, high-throughput methodologies, we were able to elucidate in depth the extent and character of repetitive element transcription in mammalian cells," lead author Geoffrey Faulkner, a graduate student at the University of Queensland's Institute for Molecular Bioscience, and his co-authors wrote. "[R]epetitive elements are a pervasive source of transcription and transcriptional regulation and therefore must be considered in future studies of the genome as a 'transcription machine.'"
Between about a third and half of mammalian genomes consist of repetitive elements, which can arise from retrotranspositions of short or long interspersed nuclear elements, long terminal repeats, or genomic DNA. Retrotransposons can serve as sources of alternative promoters, terminators, exons, splice junctions, and, potentially, non-coding RNA.
And because retrotransposons contribute to everything from genome evolution to gene expression, Faulkner and his colleagues decided to characterize the retrotransposon transcriptome in mammalian cells.
The researchers used CAGE, a method for sequencing the 5' ends of messenger RNAs, in conjunction with deep sequencing. This so-called deepCAGE method provided the opportunity to more extensively explore transcription start sites and related regions.
Mapping these CAGE tags to a reference genome provides large-scale information about transcription start sites. For example, when the researchers mapped 65 million human tags and 18.5 million mouse tags from 176 CAGE libraries, they found that about 18 percent of mouse transcription start sites and more than 31 percent of human transcription start sites fell within repetitive elements.
Even so, the researchers found that transcripts with transcription start sites within retrotransposons were generally not as highly expressed as other transcripts.
The transcripts initiated in retrotransposons were also different from one tissue to the next. For instance, nearly a third of CAGE tags from human embryonic tissue mapped to repeat regions, compared with roughly 15 percent of tags from human brain tissue. The patterns varied over time and between species.
In their subsequent experiments, Faulkner and his colleagues investigated the type of transcription initiation, promoters, and sequences associated with retrotransposons. They also looked at the consequences of retrotransposon start sites on nearby genes and detected 23,000 potential regulatory regions that are linked to retrotransposons.
"Our findings highlight the global impact of retrotransposon transcription on the evolution and functional output of the mammalian transcriptome," Faulkner and his co-authors wrote. "[W]e suggest that retrotransposons are multifaceted regulators of the functional output of the mammalian transcriptome."
In the second paper, lead author Ryan Taft, a researcher at the University of Queensland's Institute for Molecular Bioscience and his colleagues used deep sequencing of small RNAs to identify and characterize a new class of small RNAs — located near transcription start sites in human, chicken, and Drosophila cells — that they dubbed transcription initiation RNAs.
These tiRNAs are usually around 18 nucleotides long and are usually derived from sequence near transcription start sites. Because they occur in several animal species and are often associated with highly expressed transcripts, the researchers concluded that tiRNAs "may be a general feature of transcription in metazoan and possibly all eukaryotes."
"Although we do not yet know whether tiRNAs are simply signatures of transcription or have a particular function, their robust and evolutionarily conserved non-random distributions and distinct size characteristics indicate that these RNAs are not random noise but a new and distinctive size class of RNAs reflective of an ancient mechanism centrally associated with transcription in metazoans, and possibly with eukaryotic transcription in general," the authors wrote.
The team relied on information about transcription start sites from both the RefSeq database and systematic deepCAGE tag data in a human acute monocytic leukemia cell line. Taft and his colleagues found 2,312 human tiRNAs located at more than 20 percent of promoters identified by either deepCAGE sequencing or expressed RefSeq genes. Even so, most tiRNAs identified in human cells were slightly downstream of transcription start sites.
The researchers also found nearly 2,000 tiRNAs in chicken embryo libraries and between 3,750 and 29,722 tiRNAs in Drosophila libraries, depending on sequencing depth used to generate the library data.
The team's deepCAGE experiments indicated tiRNAs were often linked to higher expression levels. And in both human and Drosophila cells, tiRNAs were most often found along with genes associated with high expression or specific tissue or life cycle stages. The tiRNAs also showed a slight association with RNA Pol II binding. Although the researchers did not nail down a function for tiRNAs, they speculated that they contribute to transcriptional control, potentially by regulating chromatin.
"Our findings also suggest that deep-sequencing efforts aimed at a comprehensive compendium of small RNAs in eukaryotes is far from complete," Taft and his co-authors concluded.
For the third study, members of the FANTOM Consortium and Japan's RIKEN Omics Science Center used an approach called motif activity response analysis, or MARA, to investigate changes in transcriptional networks and target genes in the human myeloid leukemia cell line THP-1 as it differentiated.
THP-1 cells stop actively growing and differentiate into cells resembling mature monocytes and macrophages following treatment with a compound called phorbol myristate acetate. The researchers exploited this property, using deepCAGE, cDNA microarrays, computational methods, and other techniques to characterize transcriptional regulators involved in this PMA-stimulated transition.
The approach allowed the team to integrate data on different aspects of regulatory dynamics during cellular differentiation, including interactions between transcription start sites, transcript expression, and regulatory motifs. The researchers used an siRNA knockdown screen to verify the regulatory roles of 52 transcription factors.
Overall, their research suggests that the transcriptional networks behind cellular growth arrest and differentiation are complicated and interconnected. "Our results indicate that cellular states are constrained by complex networks involving both positive and negative regulatory interactions among substantial numbers of transcription factors, and that no single transcription factor is both necessary and sufficient to drive the differentiation process," the team wrote.
The authors noted that they have provided an online resource called EdgeExpressDB, through the FANTOM4 website, which "allows users to explore our annotations of the structure, expression, and regulation of promoters genome-wide."
The team is also making human promoters, transcription factor motifs, transcription factor sites genome-wide, and predicted expression effects from the study available through the Swiss Regulon site and providing a web interface through Swiss Regulon for researchers who want to perform MARA on their own expression data.