05 Dec 2008

In its textbook version, gene transcription is a one-way street. Under the influence of transcription promoter regions, RNA polymerase jumps on DNA at the start sites and gets off at the end of genes. However tidy that view seems, a quartet of papers in today’s online edition of Science draws a different and more complex picture of traffic flow on genes. The studies, each of which uses different methodologies, find RNA polymerase II boarding both upstream and downstream of start sites, and generating antisense RNA at many places across the genome. Together, the reports reveal that antisense transcription is more common and more regulated than previously thought.

The results offer insight into a fundamental process in cells, and it is too early to tell what, if any, connections researchers will eventually draw to Alzheimer’s or other diseases. Nonetheless, interest will be high in the possibilities that the studies open up for new insights into the complex orchestration of gene expression. AD researchers will also be interested in a recent study on protein-RNA interactions from Robert Darnell and colleagues that provides a new way to look at RNA alternative splicing, with important applications to neurological diseases.

In the first paper, researchers from the lab of Phillip Sharp at MIT were sequencing short RNAs from embryonic stem cells when they found a group of RNAs some 20 nucleotides small that mapped close to transcription start sites. First authors Amy Seila, Mauro Calabrese, and colleagues found sense transcripts that clustered to sites 50 nucleotides downstream of start sites, as well as antisense RNAs that matched sequences 250 nucleotides upstream. The snippets were not microRNAs, as they occurred in cells that lacked the Dicer enzyme required to produce microRNAs. The short transcripts were found in all the cell types the investigators examined, and on half of all genes. The more active the transcription of a gene was, the more short transcripts were produced. The initiation points of the short transcripts corresponded to sites where RNA polymerase is known to sit on DNA both upstream and downstream of transcription start sites, and the new results suggest that the enzyme is actively transcribing from both locales.

Similar results come from Leighton Core, Joshua Waterfall, and John Lis at Cornell University in Ithaca, New York. This group started out by mapping the position and orientation of transcriptionally active RNA polymerase II over the entire genome using a transcription run-on method coupled with high-throughput sequencing. After elongating nascent transcripts in human lung cancer cells in the presence of an inhibitor of initiation, the investigators sequenced 25 million transcripts and mapped them back to the genome.

In all, more than half of genes gave evidence of antisense transcription and most promoters had polymerase bound upstream facing in an antisense direction. Core and colleagues found a similar distribution of sense and antisense transcripts spanning the transcription start site as did the Sharp group. Both groups also found that the upstream transcripts did not elongate past the promoter sequence, in contrast to downstream, sense-facing polymerases that tended to go on to produce elongated transcripts. It is not clear how the polymerase knows which direction to go to produce coding RNAs, or what the regulatory ramifications of upstream polymerase binding and transcription will be. Both studies suggest that divergent transcription is a marker for active promoters, and may play a role in rapid regulation of gene expression.

On a hunt for more elusive RNAs, Torben Jensen and colleagues at Aarhus University, Denmark, took an approach of blocking RNA degradation to discover unstable RNAs that would normally escape detection. First author Pascal Preker engineered human cells to lack the exosome machinery that degrades some kinds of RNA. In the cells, the researchers found a buildup of upstream transcripts driven from gene promoters. The transcripts were both sense and antisense, were polyadenylated, and were produced at levels comparable to the sense, downstream transcripts. The authors speculate that the previously unrecognized transcripts might affect promoter structure or in some other way regulate transcription, and they conclude that the generality of promoter upstream transcripts “hints at a more complex regulator chromatin structure around the TSS than was previously anticipated.”

The last report looks specifically at the antisense transcriptome, and finds these RNAs to be widespread and cell-type specific. The work comes from the lab of Nickolas Papadopoulos and colleagues at Johns Hopkins Kimmel Cancer Center in Baltimore, Maryland, who developed a method to determine unambiguously which RNAs were transcribed from each strand of DNA using bisulfite modification of cytodine residues in RNA and subsequent sequencing. First author Yiping He and colleagues generated four million sequence tags from cellular RNA that could be assigned a specific position in the genome. They found antisense transcripts from close to half of genes (in agreement with Lis and colleagues). In addition, they determined that antisense transcripts were concentrated in certain parts of genes, namely in exons and upstream of promoters. This indicated that antisense transcription was not a random event, but was under some form of regulation. “Antisense transcripts thus appear to be a pervasive feature of human cells, suggesting that they are a fundamental component of gene regulation,” the authors write. However, just what that function might be remains to be seen. One possibility is that antisense participates in cell-type specific gene regulation. In agreement with this idea, He and colleagues found that in five different cell types, different genes showed antisense transcription.

After transcription comes processing of RNA, and the alternative splicing of sense mRNA is thought to play an important role, especially in the brain, to create the diversity of proteins that subserve complex neuronal functions. A recent paper in Nature from Robert Darnell and colleagues at Rockefeller University in New York provides a handle on understanding RNA splicing in vivo on a genomic scale. First author Donny Licatalosi and colleagues have developed a modification of the chromatin-immunoprecipitation (CHIP) technique used to study DNA-protein complexes that allows them to probe RNA-protein interactions in living tissues. Putting protein-RNA crosslinking and immunoprecipitation together with high-throughput sequence analysis, they did a genomewide survey to map the binding of the neuron-specific splicing factor Nova to RNAs in mouse brain. The results revealed that the site of Nova binding affects the outcome of splicing events, and identified novel sites of action for Nova, including an unexpected role in the alternative polyadenylation of mRNA. Interestingly, Nova binding seemed to promote long 3’ untranslated regions, a hallmark of brain mRNAs and potential targets for regulation by microRNAs.

“Understanding RNA regulation is an area of great interest in neurologic disease, including AD," Darnell wrote in an e-mail to ARF. “We think in fact that this platform may prove directly relevant to AD—just look at the RNA defect in FTDP-17 as a prime example.” Darnell, who as a professor of cancer biology heads the Laboratory of Molecular Neuro-Oncology at Rockefeller, is looking to follow up that thought and make a foray into a new field. He says he has just sent off his first AD grant application.—Pat McCaffrey

Comments

1. • Robert Darnell
     
   • Posted: 05 Dec 2008

   Comment on four Science papers

   Taken together, it seems that the findings propose that active genes remain "poised" for transcription by making short RNAs in the regions of their promoter, as if keeping the engine running. This appears to be a general phenomenon, not specific to AD-related genes per se. On the other hand, these short RNAs have the potential themselves to be regulated (e.g., by miRNAs). Although there is no evidence to support such a speculation, any new means of regulating gene regulation opens a new crack in the door of manipulating gene regulation for the treatment of disease. It will be quite interesting to see how this story plays out.

   Wherever new gene regulation is discovered, there is the potential that disease-associated mutations may exist that could have clinical consequences. These papers emphasize that relying solely on identification of coding region mutations in disease may be too narrow an approach—the entire structure of a gene, from upstream (promoter-proximal) regions, through alternative exons and 3' ends, will need to be considered.

   View all comments by Robert Darnell

Make a Comment

To make a comment you must login or register.

References

No Available References

Further Reading

No Available Further Reading