Acknowledgements We thank M. Kobayashi, D. Engel, M. Muramatsu and T. Honjo for discussion and comments on the manuscript, M. Busslinger for information on RT–PCR analysis, N. Kaneko for the preparation of tissue sections, and M. Kanno for the use of FACS machinery. This work was supported by Grants-in-aid from the Ministry of Education, Culture, Sport, Science and Technology of Japan and Hiroshima University 21st century COE programme. A.M. was initially supported by the Research Fellowships for Young Scientists from the Japanese Society for Promotion of Science. Competing interests statement The authors declare that they have no competing financial interests. Correspondence and requests for materials should be addressed to K.I. (igarak@hiroshima-u.ac.jp). .............................................................. Intergenic transcription is required to repress the Saccharomyces cerevisiae SER3 gene Joseph A. Martens, Lisa Laprade & Fred Winston Department of Genetics, Harvard Medical School, 77 Avenue Louis Pasteur, Boston, Massachusetts 02115, USA ............................................................................................................................................................................. Transcription by RNA polymerase II in Saccharomyces cerevisiae and in humans is widespread, even in genomic regions that do not encode proteins 1–6 . The purpose of such intergenic transcrip- tion is largely unknown, although it can be regulatory 7,8 . We have discovered a role for one case of intergenic transcription by studying the S. cerevisiae SER3 gene. Our previous results demonstrated that transcription of SER3 is tightly repressed during growth in rich medium 9 . We now show that the regulatory region of this gene is highly transcribed under these conditions and produces a non-protein-coding RNA (SRG1). Expression of the SRG1 RNA is required for repression of SER3. Additional experiments have demonstrated that repression occurs by a transcription-interference mechanism in which SRG1 transcrip- tion across the SER3 promoter interferes with the binding of activators. This work identifies a previously unknown class of transcriptional regulatory genes. The S. cerevisiae SER3 gene encodes a phosphoglycerate dehy- drogenase that catalyses a step in serine biosynthesis 10 . To investi- gate SER3 repression in greater detail, we performed chromatin immunoprecipitation (ChIP) experiments to test for the association of general transcription factors with the SER3 regulatory region in vivo. Surprisingly, our results showed that under repressing conditions, a significant level of both TATA-binding protein (TBP) and RNA polymerase II (Pol II) bind 5 0 of the SER3 TATA element (Fig. 1a). Additional ChIP experiments showed that other factors, including phosphorylated forms of Pol II and mRNA capping factors, are also associated with this region (unpublished data). Thus, when SER3 is repressed, its regulatory region is associated with factors required for active transcription. As described below, this region contains a second TATA element (shown in Fig. 1a). To determine whether transcription occurs in the SER3 regula- tory region, we performed transcription run-on experiments. These results showed that there is a high level of active transcription (Fig. 1b, probes 3–5). This transcription occurs on the same strand as SER3 as determined by its hybridization specificity for single- stranded probes. We gained additional insight into the transcription occurring 5 0 of SER3 by comparing the DNA sequence of this region to that of four yeasts closely related to S. cerevisiae 11,12 . This analysis (Supplementary Fig. 1) revealed that, in addition to the TATA element proximal to the SER3 coding region (2103 relative to the SER3 ATG), there is a second TATA element at 2558 that is perfectly conserved in all five yeasts. Short conserved sequences 5 0 of this TATAelement suggest the presence of regulatory sites. Within the transcribed region, there are three conserved sequence motifs between 2262 and 2156. Deletion analysis strongly suggests that this region functions as a SER3 upstream activating sequence (UAS) (Supplementary Fig. 2). Outside of these sequence elements there is no significant conservation or open reading frame in this region, suggesting that it does not encode a protein. Because this region is transcribed and, as will be shown, has an identified function, we have designated it as the SRG1 gene (SER3 regulatory gene 1). We will refer to the TATA element at 2558 as the SRG1 TATA element. To test for a role for SRG1 in SER3 repression, we constructed a mutation in the SRG1 TATA sequence, from TATAAA to CCTAGG (srg1-1). By northern analysis, the SRG1 transcript in a wild-type strain appears primarily as a diffuse band of approximately 550 bases (Fig. 2a). A longer RNA transcript is also present at a lower level whose length is consistent with initiation at the SRG1 initiation site and readthrough across SER3. In the srg1-1 TATA mutant, both RNAs are undetectable. Significantly, SER3 transcrip- tion is derepressed to a high level in this mutant, demonstrating an Figure 1 Evidence for active transcription 5 0 of SER3. a, ChIP analysis at SER3. ChIP analysis of TBP and Pol II (Rpb3) was performed on wild-type strain FY2245 grown in repressing conditions (YPD medium). A representative set of PCR reactions from two-fold dilutions of chromatin is shown. The regions amplified by PCR (A–D) are marked by blue bars in the diagram of the SER3 promoter on the right. The control primer set amplifies a region of chromosome V that lacks open reading frames 25 . The graphs summarize the results of three independent experiments, with each value representing the average and standard error of the fold enrichment (ratio of the percentage of SER3 immunoprecipitated to the percentage of the control region immunoprecipitated). b, Transcription run-on analysis at SER3. Transcription run-on analysis was performed on wild-type strain FY2097 grown in repressing conditions. A schematic of SER3 is shown, with the green bars representing antisense oligonucleotides that detect transcription from the Watson strand. The two yellow bars (N4, N5) represent sense oligonucleotides used as negative controls. An oligonucleotide that detects transcription of ACT1 was included as a positive control. letters to nature NATURE | VOL 429 | 3 JUNE 2004 | www.nature.com/nature 571 ©2004 Nature Publishing Group