Genome-wide occupancy profile of the RNA polymerase III machinery in Saccharomyces cerevisiae reveals loci with incomplete transcription complexes

Genes were originally defined more than a century ago by variants or mutations that alter the phenotype of an organism, and this method of identification remains useful today. Genes have also been defined by virtue of the RNAs or proteins they encode, and in many such cases, the phenotypes of mutations have not been assessed. With the advent of complete genomic sequences, genes are defined primarily by computational analysis. Although powerful, computational methods are necessarily biased by the specific properties used to define the genes. As a consequence, computational analyses often fail to identify real genes or, conversely, identify genes that are biologically meaningless. For example, as defined primarily by open reading frames (ORFs) greater than 99 amino acids, the yeast Saccharomyces cerevisiae was originally found to have 6,275 possible ORFs, 5,885 of which were predicted to be translated into proteins (9). Subsequent identification of transcribed regions of the yeast genome revealed 160 ORFs that had escaped annotation based on the original sequence-derived criteria (39). Recent computational studies based on sequence homologies among closely related yeast species suggest that approximately 500 of the originally defined genes are invalid, and they identify more than 40 smaller genes that were initially unrecognized (2, 19). Thus, in order to define the complete collection of genes in an organism, it is essential to use large-scale experimental approaches to complement the computational analysis. In eukaryotic organisms, RNA polymerase (Pol) III is responsible for the transcription of tRNAs and a few other nontranslated RNAs. Although greatly outnumbered by Pol II-transcribed, mRNA-encoding genes, Pol III genes are transcribed at very high frequencies and in fact constitute a much larger fraction of the total cellular RNA. In the yeast Saccharomyces cerevisiae, the non-tRNA genes transcribed by Pol III include the RNA component of RNase P (RPR1), the U6 small nuclear RNA (SNR6), the cytoplasmic RNA of the signal recognition particle (SCR1), and the 5S rRNA. In mammalian cells, the RNA component of RNase MRP is also transcribed by Pol III; its yeast counterpart NME1 has not been identified as a Pol III gene. Several human genes, including the 7SK gene, the RNase MRP RNA gene, and the U6 RNA gene, have extragenic promoter elements and use a specialized Pol recruitment complex known as SNAPc (36), which is not found in yeast. tRNAs and other Pol III genes share specific sequence and structural properties, including conserved A and B block sequences typically found within the coding region (7), thereby making it possible to search an entire genome for Pol III genes by computer-based methods. Such approaches revealed a total of 275 tRNA genes in the complete S. cerevisiae genome (10, 25, 32), as well as Pol III-transcribed RNA170 (31). Northern analysis of long intergenic regions (31) and recent computational analysis based on conservation of secondary structure across species (26) have revealed a number of previously unpredicted noncoding RNAs. In addition, at least one as-yet-unidentified Pol III gene has been implicated in the processing of the major ribosomal (Pol I-dependent) transcript (13). Thus, it is unclear how many additional Pol III genes exist in the yeast genome. The Pol III transcription apparatus consists of the three multisubunit general transcription factors TFIIIB, TFIIIC, and RNA Pol III, with an additional factor, TFIIIA, playing a role in the transcription of the 5S rRNA (8). The assembly factor TFIIIC consists of six subunits, and it recognizes the conserved A and B blocks. These DNA sequence elements are usually found embedded within the RNA-coding sequence, though occasionally, as with SNR6, the B block may be located outside the transcribed region. TFIIIB, which is recruited by TFIIIC to a region upstream of the transcriptional start site, comprises the TATA-binding protein, TBP, the TFIIB-related factor Brf1, and the SANT domain protein Bdp1. RNA Pol III consists of 17 subunits, 10 of which are unique to Pol III, the others being common to two or all three of the yeast RNA Pols (8, 36). We use chromatin immunoprecipitation, followed by microarray hybridization, to determine the genome-wide distribution of the Pol III transcription apparatus. The Pol III transcriptome in yeast includes essentially all tRNA genes, the previously identified non-tRNA genes, and SNR52, which encodes a small nucleolar RNA. These results are in agreement with recent reports that appeared after the work here was completed (11, 34). Unexpectedly, we identify ZOD1, a locus of unknown function associated with all components of the Pol III machinery, as well as eight ETC loci that are occupied by TFIIIC but not by TFIIIB or Pol III. The B-block sequences and association of the Pol III factors at these loci are conserved across Saccharomyces species, raising the possibilities of regulated Pol III recruitment and novel functions of the Pol III machinery.