Transfer RNA (tRNA) genes and other RNA polymerase III transcription units are dispersed in high copy throughout nuclear genomes, and can antagonize RNA polymerase II transcription in their immediate chromosomal locus. Previous work in Saccharomyces cerevisiae found that this local silencing required subnuclear clustering of the tRNA genes near the nucleolus. Here we show that the silencing also requires nucleosome participation, though the nature of the nucleosome interaction appears distinct from other forms of transcriptional silencing. Analysis of an extensive library of histone amino acid substitutions finds a large number of residues that affect the silencing, both in the histone N-terminal tails and on the nucleosome disk surface. The residues on the disk surfaces involved are largely distinct from those affecting other regulatory phenomena. Consistent with the large number of histone residues affecting tgm silencing, survey of chromatin modification mutations shows that several enzymes known to affect nucleosome modification and positioning are also required. The enzymes include an Rpd3 deacetylase complex, Hos1 deacetylase, Glc7 phosphatase, and the RSC nucleosome remodeling activity, but not multiple other activities required for other silencing forms or boundary element function at tRNA gene loci. Models for communication between the tRNA gene transcription complexes and local chromatin are discussed.

Maf1 is a negative regulator of RNA polymerase III (Pol III) in yeast. Maf1-depleted cells manifest elevated tRNA transcription and inability to grow on non-fermentable carbon source, such as glycerol. Using genomic microarray approach, we examined the effect of Maf1 deletion on expression of Pol II-transcribed genes in yeast grown in medium containing glycerol. We found that transcription of FBP1 and PCK1, two major genes controlling gluconeogenesis, was decreased in maf1Δ cells. FBP1 is located on chromosome XII in close proximity to a tRNA-Lys gene. Accordingly we hypothesized that decreased FBP1 mRNA level could be due to the effect of Maf1 on tgm silencing (tRNA gene mediated silencing). Two approaches were used to verify this hypothesis. First, we inactivated tRNA-Lys gene on chromosome XII by inserting a deletion cassette in a control wild type strain and in maf1Δ mutant. Second, we introduced a point mutation in the promoter of the tRNA-Lys gene cloned with the adjacent FBP1 in a plasmid and expressed in fbp1Δ or fbp1Δ maf1Δ cells. The levels of FBP1 mRNA were determined by RT-qPCR in each strain. Although the inactivation of the chromosomal tRNA-Lys gene increased expression of the neighboring FBP1, the mutation preventing transcription of the plasmid-born tRNA-Lys gene had no significant effect on FBP1 transcription. Taken together, those results do not support the concept of tgm silencing of FBP1. Other possible mechanisms are discussed.