Characterisation of the DNA binding domain of the yeast RAP1 protein.

The 827 amino acid yeast RAP1 protein interacts with DNA to regulate gene expression at numerous unrelated loci in the yeast genome. By a combination of amino, carboxy and internal deletions, we have defined an internal 235 amino acid fragment of the yeast RAP1 protein that can bind efficiently to the RAP1 binding site of the PGK Upstream Activation Sequence (UAS). This domain spans residues 361 to 596 of the full length protein and lacks any homology to the DNA binding 'zinc finger' or 'helix-turn-helix' structural motifs. All the RAP1 binding sites we have tested bind domain 361-596, arguing that RAP1 binds all its chromosomal sites via this domain. The domain could not be further reduced in size suggesting that it represents the minimal functional DNA binding domain. The relevance of potential regions of secondary structure within the minimal binding domain is discussed.     

Characterization of the DNA-binding activity of GCR1: in vivo evidence for two GCR1-binding sites in the upstream activating sequence of TPI of Saccharomyces cerevisiae.

GCR1 gene function is required for high-level glycolytic gene expression in Saccharomyces cerevisiae. Recently, we suggested that the CTTCC sequence motif found in front of many genes encoding glycolytic enzymes lay at the core of the GCR1-binding site. Here we mapped the DNA-binding domain of GCR1 to the carboxy-terminal 154 amino acids of the polypeptide. DNase I protection studies showed that a hybrid MBP-GCR1 fusion protein protected a region of the upstream activating sequence of TPI (UASTPI), which harbored the CTTCC sequence motif, and suggested that the fusion protein might also interact with a region of the UAS that contained the related sequence CATCC. A series of in vivo G methylation protection experiments of the native TPI promoter were carried out with wild-type and gcr1 deletion mutant strains. The G doublets that correspond to the C doublets in each site were protected in the wild-type strain but not in the gcr1 mutant strain. These data demonstrate that the UAS of TPI contains two GCR1-binding sites which are occupied in vivo. Furthermore, adjacent RAP1/GRF1/TUF- and REB1/GRF2/QBP/Y-binding sites in UASTPI were occupied in the backgrounds of both strains. In addition, DNA band-shift assays were used to show that the MBP-GCR1 fusion protein was able to form nucleoprotein complexes with oligonucleotides that contained CTTCC sequence elements found in front of other glycolytic genes, namely, PGK, ENO1, PYK, and ADH1, all of which are dependent on GCR1 gene function for full expression. However, we were unable to detect specific interactions with CTTCC sequence elements found in front of the translational component genes TEF1, TEF2, and CRY1. Taken together, these experiments have allowed us to propose a consensus GCR1-binding site which is 5'-(T/A)N(T/C)N(G/A)NC(T/A)TCC(T/A)N(T/A)(T/A)(T/G)-3'.     

Concerted action of the transcriptional activators REB1, RAP1, and GCR1 in the high-level expression of the glycolytic gene TPI.

In Saccharomyces cerevisiae, the TPI gene product, triosephosphate isomerase, makes up about 2% of the soluble cellular protein. Using in vitro and in vivo footprinting techniques, we have identified four binding sites for three factors in the 5' noncoding region of TPI: a REB1-binding site located at positions -401 to -392, two GCR1-binding sites located at positions -381 to -366 and -341 to -326, and a RAP1-binding site located at positions -358 to -346. We tested the effects of mutations at each of these binding sites on the expression of a TPI::lacZ gene fusion which carried 853 bp of the TPI 5' noncoding region integrated at the URA3 locus. The REB1-binding site is dispensable when material 5' to it is deleted; however, if the sequence 5' to the REB1-binding site is from the TPI locus, expression is reduced fivefold when the site is mutated. Because REB1 blocks nucleosome formation, the most likely function of its binding site in the TPI controlling region is to prevent the formation of nucleosomes over the TPI upstream activation sequence. Mutations in the RAP1-binding site resulted in a 10-fold reduction in expression of the reporter gene. Mutating either GCR1-binding site alone had a modest effect on expression of the fusion. However, mutating both GCR1-binding sites resulted in a 68-fold reduction in the level of expression of the reporter gene. A LexA-GCR1 fusion protein containing the DNA-binding domain of LexA fused to the amino terminus of GCR1 was able to activate expression of a lex operator::GAL1::lacZ reporter gene 116-fold over background levels. From this experiment, we conclude that GCR1 is able to activate gene expression in the absence of REB1 or RAP1 bound at adjacent binding sites. On the basis of these results, we suggest that GCR1 binding is required for activation of TPI and other GCR1-dependent genes and that the primary role of other factors which bind adjacent to GCR1-binding sites is to facilitate of modulate GCR1 binding in vivo.     

Genetic Evidence That the Meiotic Recombination Hotspot at the His4 Locus of Saccharomyces Cerevisiae Does Not Represent a Site for a Symmetrically Processed Double-Strand Break

In the yeast Saccharomyces cerevisiae, the binding of the Rap1 protein to a site located between the 5' end of the HIS4 gene and the 3' end of BIK1 stimulates meiotic recombination at both flanking loci. By using strains that contain mutations located in HIS4 and BIK1, we found that most recombination events stimulated by the binding of Rap1 involve HIS4 or BIK1, rather than bidirectional events including both loci. The patterns of aberrant segregation indicate that most of the Rap1-stimulated recombination events do not represent the symmetric processing of a double-strand DNA break located at the Rap1-binding site.     

Saccharomyces cerevisiae RAP1 binds to telomeric sequences with spatial flexibility

A wide divergence has been detected in the telomeric sequences among budding yeast species. Despite their length and homogeneity differences, all these yeast telomeric sequences show a conserved core which closely matches the consensus RAP1-binding sequence. We demonstrate that the RAP1 protein binds this sequence core, without involving the diverged sequences outside the core. In Saccharomyces castellii and Saccharomyces dairensis specific classes of interspersed variant repeats are present. We show here that a RAP1-binding site is formed in these species by connecting two consecutive 8 bp telomeric repeats. DNase I footprint analyses specify the binding site as the 13 bp sequence CTGGGTGTCTGGG. The RAP1 protein also binds the variant repeats, although with a lowered affinity. However, a split footprint is produced when RAP1 binds a variant repeat where the two half-sites of the binding site are separated by an additional 6 nt. This is probably caused by the intervening sequence looping out of the RAP1-DNA complex. We suggest that the bipartite subdomain structure of the RAP1 protein allows it to remodel telomeric chromatin, a feature which may be of great relevance for telomeric chromatin assembly and structure in vivo.     

Nonidentical DNA-binding sites of endonuclease NaeI recognize different families of sequences flanking the recognition site.

NaeI endonuclease uses a two-site binding mechanism to cleave substrate DNA: reaction-rate studies imply that occupancy of the second DNA site causes an allosteric change in the protein that enables DNA cleavage at the first site [Conrad, M., & Topal, M. D. (1989) Proc. Natl. Acad. Sci. U.S.A. 86, 9707-9711]. Measurements of relative binding affinities for 14-base-pair DNA fragments containing the NaeI recognition sequence GCCGGC and various flanking sequences showed that the two DNA-binding sites are not identical. G.C-rich flanking sequences were preferred by the activator binding site, whereas A.T-rich flanking sequences were preferred by the substrate binding site: GGGTGCCGGCAGGG was preferred 8-fold more by the activator site but 14-fold less by the substrate site than TTTCGCCGGCGTTT. Substitution of pyrimidine or 7-deazapurine for purine immediately 3' to GCCGGC reduced DNA affinity for only the activator site by up to 26-fold, implying that the activator DNA-binding site requires N-7 base contacts immediately flanking GCCGGC. The implications of nonidentical DNA-binding sites, one of which binds a specific DNA site to allosterically activate the other, are discussed.