Biological Chemistry: Study Finds Regulated Transcription of Novel RNAs Neuroscience: Old-line Antibiotic Saves Neurons After Spinal Cord Trauma Cell Biology: Structure Turns Iron Entry into Cells on Its Head Political Action Students Get Face to Face with Mass. Voters Letter to the Editor   Monkey Protein Blocks Infection by HIV Iron Maidens May Be at Higher Risk for Diabetes Protein Interferes with RNAi   The HMS Faculty Council Academy Honors 75 New Members MD-PhD Program Celebrates 30 Years Finalists Named for Teaching Award, Votes Invited Honors and Advances   Patient-Doctor Gets Dinged on Wards   House Fire Exposes Gaps in Care   Front Page

BIOLOGICAL CHEMISTRY Study Finds Regulated Transcription of Novel RNAs Noncoding Transcripts May Be Widespread in Genome A tour of almost any neighborhood on trash day reveals how loosely people define the word junk. The same may be said for a tour of the annotated human genome, because even though many scientists abhor the term junk DNA, it is widely used to describe the millions of base pairs that--as yet--have no ascribed function. But in the genome, as in the trash, what seems of little use to some often turns out to be surprisingly valuable. In the Feb. 20 Cell, Kevin Struhl, the David Wesley Gaiser professor of biological chemistry and molecular pharmacology at HMS, together with Thomas Gingeras and colleagues at the genomics technology company Affymetrix, reports that thousands of such gems may be hiding in the "junk" of chromosomes 21 and 22; by inference, they may be abundant throughout the entire human genome. Distribution of transcription factor binding sites. Mapping studies show that binding sites for p53, cMyc, and Sp1 are not only at the 5' end of known genes. Just as many sites are upstream of novel transcripts, and even more are found within or slightly downstream of known genes. The location of others cannot be clearly defined. Bases Loaded The work builds on the 2002 discovery by Gingeras that the number of RNAs transcribed from chromosomes 21 and 22 may be an order of magnitude greater than predicted by the annotated human genome. Philipp Kapranov, Simon Cawley, and colleagues at Affymetrix had probed cDNAs prepared from 11 different human cell lines with "tiled arrays," pools of small DNAs, each 25 bases long, that can bind RNA transcribed from any location on either chromosome. Containing more than a million oligonucleotides, these chips were designed so that probes bind at 35 base pair intervals, on average, along the chromosomes. "The ability to query transcription in such an 'unbiased' manner and at such resolution is a vast improvement over traditional approaches, where one depends on probes that recognize only known genes," Struhl explained. "The tiled array spans the whole chromosome and allows us to probe the parts that we know nothing about." Though Gingeras's earlier data suggested that there may be a lot more transcription going on in mammalian cells than previously thought, questions remained--most notably, whether such transcripts have any cellular function and if their synthesis is regulated. Struhl recognized that his lab's expertise with another kind of "CHIP," chromatin immunoprecipitation, could help them tackle these questions. "The ability to query transcription in such an 'unbiased' manner and at such resolution is a vast improvement over traditional approaches, where one depends on probes that recognize only known genes." --Kevin Struhl In the genome, the presence of a transcription factor binding site is a good indication that the transcription machinery is recruited near that location, and that RNA synthesis actually occurs there. Chromatin immunoprecipitation uses antibodies to isolate fragments of DNA that contain such transcription factor sites. For example, an antibody to the well-known tumor suppressor p53 might precipitate hundreds of different DNA fragments, each with its own p53-binding site. Struhl realized that probing his CHIPs with the tiled-array chips from Affymetrix could reveal new binding sites that might be important for regulating synthesis of both known and unknown transcripts, like the ones Kapranov had found previously. With this in mind, postdoctoral fellows Huck Ng and Edward Sekinger prepared CHIPs from mammalian cell lines using antibodies to three well-studied transcription factors: p53; cMyc, which plays an important role in the transformation of many cells to tumors; and Sp1, one of the most common transcription factors found in mammalian cells. They passed these CHIPs to Affymetrix, where Cawley and Stefan Bekiranov tested the precipitates with the tiled array chips to chromosomes 21 and 22. When the Chips Are Down The experiment revealed more than 1,100 sites, which could be narrowed to 866 kilobase-long stretches of DNA, where one or more of the transcription factors bound. Significantly, only a small percentage of these regions were discovered in the most likely place, just upstream of a known gene (see chart). The remainder were ambiguous, identified either within or at the distal end of a known gene, or, as in the case of about one quarter of all regions, found upstream of potential novel transcripts. To see if some of the more novel transcripts actually exist, the Affymetrix team used the polymerase chain reaction, amplifying just over 80 percent of the potential transcripts from mammalian-cell RNA, indicating that the majority of them really are made in the cell. The function of noncoding RNAs is unclear. Kevin Struhl believes that in some cases the transcripts per se may not be important, but that the act of transcription might open up the chromatin structure and affect regulation of nearby genes. (Photo by Graham Ramsay) But what might be the function of these novel transcripts? Not unexpectedly, they have little coding capacity--if they had, they would most likely have been discovered by now. "Many of those within the last exon or downstream of known genes could be involved in regulation," said Struhl, "because they could form a sense/antisense pair with the upstream gene." In fact, about 30 percent of these transcripts share a transcription factor binding site with a proximal upstream gene, suggesting that there could even be some kind of coordinated regulation of transcription. Experimental evidence backs up this suggestion. When Struhl and colleagues treated human cell lines with retinoic acid, then tested changes in expression using the tiled arrays, they found that similar numbers of coding and noncoding RNAs were either up- or down-regulated. Because these novel transcripts now need to be individually characterized, it may be some time before a picture develops of what they do. "Perhaps the transcripts per se are not even important," suggested Struhl, venturing that the act of transcribing a small noncoding RNA may help to open up the chromatin structure and facilitate transcription of the "real" gene that lies just upstream. --Tom Fagan