More than meets the eye as Affymetrix chases transcription

With the increasing recognition of the importance of non-coding RNA in regulation and control of gene expression, it now appears there is much more to the genome than meets the eye.

In a project that makes the most of their cutting-edge array technologies, US company Affymetrix is preparing to map the entire human genome for all potential transcription sites. The company's vice-president of biological sciences, Thomas Gingeras, was at the Lorne Genome Conference in Victoria last week, talking up the company's Transcriptome project, a collaboration with the NIH's National Cancer Institute.

"A couple of years ago, we entered into a funded relationship with the National Cancer Institute at the NIH to build an array that in an unbiased fashion would interrogate across chromosome 21 and 22 for sites where transcription might be happening," Gingeras told Australian Biotechnology News.

Starting from the basis that every nucleotide could be considered a potential site for transcription, the non-repetitive sequences of human chromosomes 21 and 22 were interrogated every 35 nucleotides using oligonucleotide arrays, Gingeras explained. Only poly-adenylated RNA isolated from the nucleus and the cytoplasm was examined, to avoid picking up RNAs destined for degradation.

The results gave a tantalising hint that the genome contained far more information that it was being credited with.

"There was about an order of magnitude more transcription going on than could be correlated with annotation maps," Gingeras says. With cDNAs making up about 10 per cent of the genome, and ESTs (including cDNAs) about 30 per cent, that left about two thirds of the genome located away from annotated areas. Of that two thirds, about half was found to be embedded in genes, while the other half was found to be tens of kilobases away from the nearest annotation.

Much greater detail

Using cloning and sequencing to further analyse the RNA, the array results were confirmed, showing that the majority of the sequences were non-coding RNAs. Localisation experiments showed that the RNA was processed (splicing and poly-adenylation) and mostly located in the nucleus, perhaps awaiting transport. The data was published last year in Science.

Since then, Affymetrix researchers, in collaboration with Kevin Struhl at Harvard University, have been looking at transcription in much greater detail. Using the binding sites for three transcription factors, c-Myc, Sp-1 and p53, they have mapped across chromosome 21 and 22, and what they have found is quite striking, says Gingeras.

Around 60 per cent of the sites are located either within or 3' to genes, or well away from any real annotations. Another 27 per cent are located 5' to the genes and the remainder are ambiguous, although Gingeras believes they will also be found to be embedded in genes.

The discovery begs the question whether the transcription factor binding sites are marking the start positions of non-coding RNAs, such as antisense transcripts. According to Gingeras, they are, but further details will have to wait until an upcoming paper has been published.

"The take-home message is that we believe there is a global architecture including antisense transcripts associated with many well-characterised genes," he said. "Antisense implies negative regulatory action, but really there can be more than one kind of function."

The next step, Gingeras said, will be to make maps of the entire genome at a higher resolution. Affymetrix plans to do this in two stages, with 10 chromosomes to be mapped at a five-nucleotide resolution by November this year, and the remaining chromosomes to be completed by about November 2004, subject to approval of the funding required.

The company is also committed to building other maps associated with regulation of gene expression, such as other transcription factor binding sites, DNA methylation sites, and chromatin structures. And the data is being made publicly available by Affymetrix, said Gingeras, as a condition of the funding.

Reference: 'Large-Scale Transcriptional Activity in Chromosomes 21 and 22', Kapranov P, Cawley SE, Drenkow J, Bekiranov S, Strausberg RL, Fodor SPA and Gingeras TR, Science 2002 May 3; 296: 916-919.