We examined the stability of microsatellites of different repeat unit lengths in Saccharomyces cerevisiae strains deficient in DNA mismatch repair. The msh2 and msh3 mutations destabilized microsatellites with repeat units of 1, 2, 4, 5, and 8 bp; a poly(G) tract of 18 bp was destabilized several thousand-fold by the msh2 mutation and about 100-fold by msh3. The msh6 mutations destabilized microsatellites with repeat units of 1 and 2 bp but had no effect on microsatellites with larger repeats. These results argue that coding sequences containing repetitive DNA tracts will be preferred target sites for mutations in human tumors with mismatch repair defects. We find that the DNA mismatch repair genes destabilize microsatellites with repeat units from 1 to 13 bp but have no effect on the stability of minisatellites with repeat units of 16 or 20 bp. Our data also suggest that displaced loops on the nascent strand, resulting from DNA polymerase slippage, are repaired differently than loops on the template strand.Eukaryotic genomes often contain regions of DNA (called microsatellites or minisatellites) in which a single base or a small number of bases is tandemly repeated. In this paper, repetitive tracts with repeats of 1 to 13 bp will be considered microsatellites and tracts with repeats of more than 16 bp will be considered minisatellites. Both microsatellites and minisatellites are unstable, frequently undergoing deletions and additions (10,13,19). In vitro replication experiments demonstrate that DNA polymerase frameshift errors occur in repetitive sequences (17), and most of the available in vivo data suggest that alterations in microsatellite length reflect DNA polymerase slippage events (19,27). This mechanism predicts a transient dissociation of the template and the nascent strand during replication of the microsatellite (28). Due to the repetitive nature of the tract, the two DNA strands can reassociate out of register, leaving one or more unpaired repeats on either the template or nascent strand (see Fig. 1). If the distortion caused by these unpaired bases is not removed from the newly synthesized strand by the DNA mismatch repair system, the result will be a loss (if the unpaired bases are on the template strand) or a gain (if the unpaired bases are on the nascent strand) of one or more repeats. As expected from this model, mutations in the genes required for DNA mismatch repair greatly increase the rate of instability of repetitive DNA sequences in Escherichia coli, the yeast Saccharomyces cerevisiae, and human cells (19,22,27).In E. coli, two of the proteins involved in DNA mismatch repair are MutS (involved in recognition of the DNA mismatch) and MutL (involved in interactions between MutS and other proteins) (22). Homologs of these proteins have been identified in yeast and mammals. In yeast, the effects of mutations in the mutL homologs MLH1 and PMS1 and in the mutS homologs MSH2, MSH3, and MSH6 on the stability of a 33-bp poly(GT) repeat have been examined previously (14,26,27). Mutations in MLH1, P...

Yeast chromosomes terminate in tracts of simple repetitive DNA (poly[G1-3T]). Mutations in the gene TEL1 result in shortened telomeres. Sequence analysis of TEL1 indicates that it encodes a very large (322 kDa) protein with amino acid motifs found in phosphatidylinositol/protein kinases. The closest homolog to TEL1 is the human ataxia telangiectasia gene.

Homologous recombination is an important mechanism for the repair of DNA damage in mitotically dividing cells. Mitotic crossovers between homologues with heterozygous alleles can produce two homozygous daughter cells (loss of heterozygosity), whereas crossovers between repeated genes on non-homologous chromosomes can result in translocations. Using a genetic system that allows selection of daughter cells that contain the reciprocal products of mitotic crossing over, we mapped crossovers and gene conversion events at a resolution of about 4 kb in a 120-kb region of chromosome V of Saccharomyces cerevisiae. The gene conversion tracts associated with mitotic crossovers are much longer (averaging about 12 kb) than the conversion tracts associated with meiotic recombination and are non-randomly distributed along the chromosome. In addition, about 40% of the conversion events have patterns of marker segregation that are most simply explained as reflecting the repair of a chromosome that was broken in G1 of the cell cycle.

Activation of phospholipase C-dependent inositol polyphosphate signaling pathways generates distinct messengers derived from inositol 1,4,5-trisphosphate that control gene expression and mRNA export. Here we report the regulation of telomere length by production of a diphosphorylinositol tetrakisphosphate, PP-IP 4 , synthesized by the KCS1 gene product. Loss of PP-IP 4 production results in lengthening of telomeres, whereas overproduction leads to their shortening. This effect requires the presence of Tel1, the yeast homologue of ATM, the protein mutated in the human disease ataxia telangiectasia. Our data provide in vivo evidence of a regulatory link between inositol polyphosphate signaling and the checkpoint kinase family and describe a third nuclear process modulated by phospholipase C activation.Appropriate cellular responses to environmental changes involve intracellular second messenger systems that transduce the signals from cytoplasm to nucleus, thereby initiating adaptive genetic programs. One well described intracellular messenger system works through receptor-coupled activation of phospholipase C (PLC), 1 which hydrolyzes phosphatidylinositol 4,5-bisphosphate to yield inositol 1,4,5-trisphosphate (IP 3 ) (reviewed in Refs. 1 and 2). In metazoans, cellular production of IP 3 functions as a signal for calcium release from intracellular stores through allosteric activation of an IP 3 receptor channel. Recent studies indicate that IP 3 also plays an important role as precursor to multiple inositol polyphosphates (IPs), each with potentially unique signaling capability (reviewed in Ref. 3). This signaling potential is not restricted to the cytoplasm, because IPs are known to function in nuclear processes in budding yeast (reviewed in Ref. 4). Activation of yeast phospholipase C (Plc1) produces IP 3 , which is rapidly phosphorylated by the dual specificity 6-/3-kinase Ipk2 (initially cloned as ArgRIII/Arg82 (5)) to yield IP 4 and IP 5 (6 -8). This phosphorylation step is coupled to transcriptional regulation, possibly through chromatin remodeling (7, 9 -11). IP 5 is then phosphorylated by the 2-kinase Ipk1, generating IP 6 , which is required for efficient mRNA export from the nucleus (12-14). Both IP 5 and IP 6 can be converted to the diphosphoryl inositols PP-IP 4 and PP-IP 5 through the action of Kcs1, a kinase required for normal vacuolar morphology (15-18). PP-IP 5 has recently been suggested to play a role in chemotaxis in Dictyostelium (19). A nuclear role for Kcs1 is implied by its initial cloning as a regulator of mitotic DNA recombination, and inositol kinase activity is required for this regulation, but further understanding of its nuclear activity is lacking (20,21).Recent work has provided further linking of IP production and nuclear function through an important family of protein serine/threonine kinases known as phosphatidylinositol 3-kinase related kinases (PIKKs). IP 6 stimulates DNA repair by non-homologous end joining with mammalian proteins in vitro (22). Non-homologous end joining can b...