Saccharomyces cerevisiae - Semantic Scholar

© 2000 Oxford University Press

Human Molecular Genetics, 2000, Vol. 9, No. 1

93–100

CGG/CCG repeats exhibit orientation-dependent instability and orientation-independent fragility in Saccharomyces cerevisiae Bala S. Balakumaran, Catherine H. Freudenreich+ and Virginia A. Zakian§ Department of Molecular Biology, Princeton University, Princeton, NJ 08544, USA Received 3 September 1999; Revised and Accepted 15 October 1999

An expansion to >200 CGG/CCG repeats (hereafter called CGG) in the 5′′ region of the FMR1 gene causes fragile X syndrome, and this locus becomes a folatesensitive fragile site. We used Saccharomyces cerevisiae as a model system to study the stability and fragility of CGG repeats. Tracts of (CGG)81 and (CGG)160 were integrated onto a yeast chromosome in both orientations relative to the nearest replication origin. Tracts of this length are pre-mutation alleles in humans, with a high probability of expansion in future generations. The CGG tracts in yeast colonies showed a length-dependent instability with longer tracts being more prone to contraction than shorter tracts. In addition, there was an orientation bias for tract stability with tracts having fewer contractions when the CCG strand was the template for lagging strand synthesis. Expansions of the CGG tracts also occurred in an orientation-dependent manner, although at a lower frequency than contractions. To determine whether CGG tracts are fragile sites in yeast, the CGG tracts were flanked by direct repeats, and the rate of recombination between the repeats determined. Strains carrying the (CGG)160 tract in either orientation had a large increase in their rate of recombination compared with a no-tract control strain. Because this increase was dependent on genes involved in doublestrand break repair, recombination was likely to be initiated by CGG tract-induced breakage between the direct repeats. The observation of orientationdependent instability and orientation-independent fragility suggests that at least some aspects of their underlying mechanisms are different. INTRODUCTION Expansion of a CGG/CCG (hereafter called CGG) trinucleotide repeat tract in the 5 promoter region of the FMR1 gene at the FRAXA site is the cause of fragile X syndrome (1,2). In normal individuals, there are 6–54 copies of CGG repeats at FRAXA. An expansion to 60–200 repeats is termed a +Present §To

pre-mutation because the disease is not yet manifest in individuals carrying these alleles, but these pre-mutations increase the probability of expansion in future generations. In fragile-X patients, there are >200 copies of the repeat, and in some cases as many as 2000 repeats (reviewed in ref. 3). The CGG tracts and a CpG island adjacent to the FMR1 promoter are methylated in expanded alleles, resulting in the absence of FMR1 transcription and the disease phenotype (4–6). An expanded CGG tract at the FRAXE locus causes mental retardation (7). There are three additional loci where CGG repeats expand, but in these cases expansion is not known to be associated with disease (8). Another feature of expanded CGG tracts is their ability to act as fragile sites, which are regions on human chromosomes defined cytologically by their tendency to break non-randomly when exposed to specific chemicals or conditions (8). Fragile sites are classified into categories based on their frequency of expression and the conditions that induce them (9). All five of the known rare folate-sensitive, fragile sites in humans are loci with expanded CGG tracts (2,8). Several other inherited trinucleotide diseases are caused by the expansion of CTG/CAG or GAA/TTC repeats (3,10). In vitro studies have shown that all three trinucleotide repeat sequences can form unusual secondary structures under specific ionic and pH conditions, and that the stability of these secondary structures is dependent on the nature and number of the repeats (11–14). For example, oligonucleotides containing either CGG or CCG repeats form stable hairpins in vitro, with the CGG repeats having a higher melting temperature (Tm) than CCG repeats for oligonucleotides of 15–25 repeats (11,15). In addition, a CGG strand of 11–20 repeats also forms intrastrand quadruplex structures (16,17). Model organisms have been used to study trinucleotide repeats. CGG tracts cloned into bacterial plasmids show a length- and orientation-dependent stability, with longer tracts generating more contractions and the tracts being less stable in the orientation in which the CGG strand is the lagging strand template (18,19). CTG repeats show a length- and orientationdependent stability in both bacteria (20,21) and yeast (22–25). Although DNA replication in eukaryotic and prokaryotic cells has many similar features, unlike prokaryotic DNA, eukaryotic DNA is organized into nucleosomes. CGG repeats assemble into nucleosomes more poorly than any other tested sequence, a property that is further decreased by DNA

address: Department of Biology, Tufts University, Medford, MA 02155, USA whom correspondence should be addressed. Tel: +1 609 258 6770; Fax: +1 609 258 1701; Email: [email protected]

94


Figure 1. Orientation of the CGG tracts at the LYS2 locus on chromosome II. A tract of (CGG)81 or (CGG)160 was inserted such that the tracts and the adjacent URA3 gene were flanked by direct repeats of 708 bp of 3LYS2 sequence. In orientation I, the top strand is CGG, and in orientation II, the top strand is CCG. Replication proceeds from left to right through this region of the chromosome (22).

methylation (26–28). The unusual chromatin-forming properties of CGG repeats might influence both expansion and fragility. Here we use Saccharomyces cerevisiae as a model organism to study the stability and fragility of CGG repeats on a eukaryotic chromosome. In yeast, CGG tracts displayed a length- and orientation-dependent stability and acted as lengthdependent fragile sites. In contrast to stability, fragility of CGG repeats was orientation independent. RESULTS Integration of CGG tracts onto a yeast chromosome To investigate the properties of the trinucleotide repeat CGG in a eukaryotic chromosomal context, a (CGG)81 or (CGG)160 repeat tract was integrated at the LYS2 locus on chromosome II in the yeast S.cerevisiae (Fig. 1). Most of the CGG tracts at the FMR1 locus in humans contain two interruptions, one after 10 repeats and the other after another 15 repeats (29,30). The (CGG)81 sequence used here was isolated originally from a fragile-X patient (18). This tract, which has two interruptions, had the following sequence: (CGG)11AGG(CGG)60CAG(CGG)8. The (CGG)160 repeat tract is a dimer of the (CGG)81 tract and contains a 5 bp non-repeat sequence at the junction of the two tracts. (CGG)81 and (CGG)160 mimic pre-mutation alleles, which at FRAXA range in size from 60 to 200 repeats (3). The CGG repeat tracts were cloned into a yeast integration vector containing both a URA3 marker and 708 bp of 3LYS2 DNA. This plasmid was linearized and targeted to the LYS2 locus in such a way as to create a duplication of the 3LYS2 DNA with the CGG tract and the URA3 selectable marker between the flanking 3LYS2 cassettes (Fig. 1). Previous experiments using two-dimensional gel electrophoresis established that replication proceeds through the LYS2 locus in such a way that the top strand is the template for lagging strand synthesis (22), a result confirmed for strains containing CGG tracts (B. Lenzmeier, B.S. Balakumaran and V.A. Zakian, unpublished data). Strains were made such that the tracts were integrated in both orientations with respect to the direction of replication, i.e. with the top strand being either C- or G-rich. In addition, a control strain was made that

Figure 2. Stability of (CGG)81 and (CGG)160 tracts is dependent on length and orientation. Yeast genomic DNA was digested with SacI and ClaI, which flank the CGG tract, and separated by electrophoresis on a 1.5% agarose gel. Representative Southern blots probed with a 32P-labeled CGG fragment are shown. The arrow (right) indicates the expected size of the starting tract length. (A) Tract lengths of (CGG)81 colonies in both orientations. In orientation II, lanes 2 and 9 contain DNA from colonies that had contractions, whereas lane 5 has DNA from a colony that had both a contraction and starting length tract. In orientation I, lanes 4, 5 and 10 had contractions and lane 3 had both a contraction and starting length tract. (B) Tract lengths of (CGG)160 colonies in both orientations. In orientation II, lanes 2, 3, 7, 11, 12, 14, 18 and 19 were mostly full length. In orientation I, lanes 1, 4, 7, 8 and 18 were mostly full length.

contained only the URA3 gene flanked by the same 3LYS2 direct repeats. Ura+ yeast transformants were selected, and the structure of the integrants was confirmed by Southern blot analysis (data not shown). CGG tract length was determined in independent transformants, and transformants with full-length CGG tracts were chosen for further analysis. For the (CGG)160 strain in which the top strand was G-rich, there was no transformant that had a pure population of cells with a full-length tract. In this case, the transformants were re-streaked and daughter colonies were analyzed to identify ones with mostly full-length tracts for further analysis. CGG/CCG repeats exhibit a length- and orientationdependent stability In order to determine the stability of the CGG repeats on a yeast chromosome during mitotic growth, yeast colonies of known tract lengths were streaked for single colonies on plates lacking uracil (–ura) to select for the presence of the repeat tracts. The resulting daughter colonies, which represent 20 generations from a single cell to a medium sized colony, were picked and grown in –ura medium for an additional 10 generations, and genomic DNA was prepared for analysis of tract length. As PCR amplification did not give consistent results, Southern blot analysis was used to obtain an accurate determination of CGG tract length. For each experiment, tract lengths were determined on 8–12 colonies from a re-streak. The experiment was repeated a minimum of eight times such that DNA from >75 colonies was examined for each tract length and orientation. Tract lengths from colonies that contained >50% cells with shorter or longer tract sizes were classified as contractions or expansions, respectively. In most cases of changes in tract length, >90% of tracts from the colony had a single discrete size (Fig. 2).


95

Table 1. Summary of the stability of tract length analysis Total

Contraction Expansion (%) (%)

No change (%)

Orientation I (CGG)81

85

29.4

1.2

69.4

Orientation II (CGG)81

79

12.6

3.8

83.6

Orientation I (CGG)160

248

85.1

0.0

14.9

Orientation II (CGG)160

213

59.2

1.9

37.0

Results of the (CGG)81 and (CGG)160 tracts in both orientations are classified into three categories. Lanes that contained >50% of cells with tract lengths shorter or longer than starting length tracts were classified as contractions or expansions, respectively, and those with >50% of cells at the starting length as no change.

Figure 3. CGG tracts expand in yeast. Southern blots were carried out as described in Figure 2. Lanes with tract expansions in (CGG) 81 and (CGG)160 are indicated by an asterisk (*). The starting tract length is indicated on the left (arrow) and molecular weight (MW) markers on the right.

Although CGG tracts were unstable in both orientations, a pronounced orientation-dependent difference in repeat stability was observed. For both (CGG)81 and (CGG)160 repeat tracts, the orientation in which the top strand was C-rich (orientation II) was more stable than the orientation in which the top strand was G-rich (orientation I). The (CGG) 81 tract had 13% contractions in orientation II and 29% contractions in orientation I (Fig. 2A; Table 1), a statistically significant difference (P < 0.02) (31). This difference was even more pronounced with the (CGG)160 tract. In orientation II, 59% of the colonies had contractions, but 85% had contractions in orientation I (Fig. 2B; Table 1; P < 0.01). Thus, for both the (CGG)81 and (CGG)160 tracts when the CGG strand was the template for lagging strand synthesis, the tracts had more contractions. CGG repeats in yeast also showed a length-dependent stability. (CGG)81 repeats were more stable than (CGG)160 repeats (Fig. 2; Table 1). Even strains with (CGG)81 in the more unstable orientation I had fewer contractions (29%) than strains with (CGG)160 in the more stable orientation II (59%; Table 1). When the (CGG)81 tract size in either orientation was measured by Southern analysis, the bands were discrete sizes with little evidence of smaller products, suggesting that the tracts were quite stable during colony growth (Fig. 2A). However, the (CGG)160 tract in either orientation usually had a smear of CGG hybridization below it, indicating that the long

Figure 4. CGG tracts are length-dependent fragile sites. (A) Schematic diagram of the fragility assay. The starting cells contained a CGG tract and the URA3 gene flanked by LYS2 direct repeats (Fig. 1), and were FOAS owing to the presence of URA3. Breakage at the tract is expected to be repaired by recombination between the direct repeats, generating cells that have lost the URA3 gene and are FOAR. Rates of FOAR were also determined in a no-tract control strain. The distance between the direct repeats is 4.3 kb plus the size of the CGG tract. The constructs shown in (A) are in orientation II, but in (B) rates of recombination were determined for the no-tract control and (CGG)81 and (CGG)160 strains in both orientations. (B) The rate of recombination depends on CGG tract length and is orientation independent. The average rate of generating FOAR cells 10–7 (y-axis) is shown for strains containing no tract, (CGG)81 and (CGG)160 in both orientations. Each experiment was done two to six times. For each strain, 10 colonies that had full-length tracts were used in the fluctuation analysis. The standard error for each experiment is indicated. Note that the y-axis is in logarithmic scale. The numbers above each column represent the fold increase compared with the no-tract control strain.

tracts were subject to continuous contractions during colony growth (Fig. 2B). When expansions were analyzed, no striking difference was observed between the two tract sizes (Table 1). There were 4/164 (2.4%) colonies in the (CGG)81 strains with expansions of 5, 5, 5 and 21 repeats, respectively, and 4/461 (0.9%) colonies in the (CGG)160 strains with expansions of 5, 5, 22 and 33 repeats, respectively (Table 1; Fig. 3). In addition there was

96


Figure 5. The rate of CGG-mediated recombination depends on genes involved in double-strand break repair. The experiment was done in the wild-type strain, the wild-type strain grown on 0.05 M hydroxyurea, and strains with a deletion in RAD50, RAD52 or RAD1 genes. For each strain, 10 colonies that had full-length tracts were used in the fluctuation analysis. The average from at least two different experiments is plotted, except for (CGG)81 in hydroxyurea, which is a single experiment. The standard error for each experiment is indicated. (A) The rate of CGG-mediated recombination as measured by the rate of generating FOAR cells 10–5. The rate and the standard error for each experiment is noted. (B) The average rate of generating FOAR cells 10–7 (y-axis) is shown for no-tract control, (CGG)81 and (CGG)160 strains in the more stable orientation II. Note that the y-axis is in logarithmic scale. The fold difference compared with wild type for no-tract, 81 and 160 repeats is indicated above each column.

one case for each tract length, where a subpopulation of the cells in a colony had undergone an expansion of 50 repeats. In all but one case, expansions occurred in colonies with the tract in the more stable orientation. When all the expansions including the ones that occurred in only a subpopulation of a colony were considered, there were nine expansions in orientation II [four in (CGG)81 and five in (CGG)160], but only one in orientation I [from (CGG)81 (Table 1)]. Thus, orientations I and II had a statistically significant difference in expansion rates (P < 0.02), with expansions more likely to be recovered in orientation II, the relatively more stable orientation. CGG/CCG repeats act as length-dependent fragile sites All the known rare folate-sensitive, fragile sites on human chromosomes are expanded CGG trinucleotide tracts (8). To determine whether CGG tracts act as fragile sites on yeast chromosomes, we used a genetic assay that was previously used to detect CTG-tract-mediated chromosome breakage (23). This assay has the benefit that it can detect even low

levels of CGG fragility. As described above, a (CGG)81 or a (CGG)160 repeat tract and the URA3 gene were integrated at the LYS2 locus on chromosome II in such a way that they were flanked by 708 bp of 3LYS2 DNA (Fig. 4A). Yeast cells with a functional URA3 gene are sensitive to the drug 5-fluro-orotic acid (FOA) and hence will die on plates containing FOA (32). If recombination occurs between the 3 LYS2 repeats, the URA3 gene will be lost, yielding an FOAR cell. A double-strand break between two direct repeats increases the rate of recombination between the repeats (33). Therefore, if the CGG tract acts as a break site in vivo, the rate of recombination between the flanking LYS2 sequences will increase, resulting in an increase in the number of FOAR colonies. Fluctuation analysis was used to measure the rate of recombination in cells with a (CGG)81 or a (CGG)160 tract in both orientations as well as in a corresponding no-tract control strain. Cells were grown on plates non-selectively, allowing recombination to occur in a subset of cells in each colony (34). Individual colonies were picked as agar plugs, resuspended in water, and a portion of each was spread on plates containing FOA to select for those cells that had lost the URA3 gene. A second portion of the resuspended cells was diluted and plated on non-selective plates to determine the total number of cells. The rate of generating FOAR cells was calculated by the method of the median (35,36). One complication with this approach is that, even in the stable orientation, (CGG)160 tracts underwent frequent contractions (Fig. 2B). Thus, to determine the effects of tract length on recombination rate, it was important to confirm that the colonies used in the fluctuation analysis maintained the original tract length. Tract length was determined by Southern blot analysis for all colonies tested, and only those that maintained full-length tracts were used to calculate rates of recombination (data not shown). The structure of the LYS2 locus in FOAR colonies was also checked by Southern analysis. All colonies tested (20/20) showed a structure consistent with recombinational loss of URA3 (data not shown). Results of the fluctuation analysis showed that strains with the (CGG)81 tract in either orientation had a modest 2(orientation II) to 4-fold (orientation I) increase in the rate of generating FOAR cells compared with the no-tract control strain (Fig. 4B). In contrast, strains with the (CGG)160 tract in either orientation had a large increase in recombination rate, 236-fold for orientation II and 100-fold for orientation I (Fig. 4B), compared with the no-tract control. The modest 2-fold effect of orientation on the recombination rates of both (CGG)81 and (CGG)160 tracts was not statistically significant (P = 0.4) (Fig. 4B). Thus, CGG tracts increased recombination between the LYS2 direct repeats in a length-dependent, orientation-independent manner. The CGG-mediated increase in FOAR cells depends on genes involved in double-strand break repair A genetic approach was used to determine whether the CGGtract-mediated increase in FOAR cells was due to an increase in the occurrence of double-strand breaks near the CGG tract (Fig. 5). If recombination proceeded via a double-strand break, mutations in genes involved in double-strand break repair should lead to reduced or no recombination. We constructed rad50, rad1 and rad52 versions of the control strain and the


(CGG)81 and (CGG)160 strains in the stable orientation, and determined the rate of generating FOAR cells in each. Rad50p is involved in the processing of double-strand breaks such that double-strand break-initiated recombination is slowed but not eliminated in its absence (33). Rad1p along with Rad10p is part of an endonuclease complex that plays a critical role in the single-strand annealing mechanism for double-strand break repair by removing the overhanging non-homologous 5 strands to create the repair substrate (37–39). Rad52p is a central player in recombination, and its absence eliminates most pathways of double-strand break repair in yeast (40,41). In the absence of either Rad1p or Rad52p, the rate of generating FOAR cells was greatly reduced, as expected if recombination were initiated by a double-strand break (Fig. 5). The rate of generating FOAR cells was 200-fold lower in both the rad52 and rad1 (CGG)160 strains compared with a wild-type (CGG)160 strain. Even though the rate of generating FOAR colonies for the (CGG)160 rad1 and rad52 strains was only 0.5% of wild-type, these strains still exhibited a 10- to 15-fold higher rate of recombination than the rad1 or rad52 no-tract control strain (Fig. 5). This result suggests that in the absence of gene products critical for double-strand break repair, breakage still occurred near the CGG tracts, and that the double-strand break can be repaired, albeit inefficiently, in the absence of either Rad52p or Rad1p. Recombination in both the (CGG)81 and no-tract control rad52 and rad1 strains was reduced up to 33-fold. The absence of Rad50p also decreased recombination rates for all strains 10-fold compared with wild-type versions of these strains, with the difference in the ratio between the test and control strains still maintained (Fig. 5). The large increase in the rate of generating FOAR cells in the presence of the (CGG)160 tract and the dependency of this increase on genes required for processing double-strand breaks suggests that the (CGG)160 tract induces a double-strand break between the direct repeats. The presence of hydroxyurea in yeast slows DNA replication by decreasing nucleotide pools (42). This reduction is thought to result in slowing of DNA polymerase, which in turn causes replication fork pausing and increased DNA breakage. When cells were grown in 0.05 M hydroxyurea, a modest increase in the rate of generating FOAR cells was observed for both the notract and the CGG tract strains (Fig. 5). This result is reminiscent of the increase in breakage at expanded CGG loci in human cells grown in folate- or thymidine-deprived media, conditions that also slow DNA replication (43). DISCUSSION The data in this paper show that (CGG)81 and (CGG)160 tracts were both unstable in yeast (Figs 2 and 3), with changes in tract length occurring in 16% [(CGG)81; orientation II] to 85% [(CGG)160; orientation I] of colonies (Table 1). Although contractions and expansions were both detected, contractions were up to 30 times more common than expansions (Table 1). Instability was both length and orientation dependent. Contractions were 3-fold more frequent for (CGG)160 than (CGG)81 (Table 1). Although (CGG)81 and (CGG)160 tracts were prone to contractions in both orientations, orientation I, where the CGG strand was the template for lagging strand synthesis (Fig. 1), had more contractions than orientation II (Table 1). In contrast, CGG repeats in orientation II were more

97

likely to expand. Another paper published whilst this manuscript was in preparation showed a similar orientation bias in yeast for CGG tracts ranging from 30 repeats with two interruptions to 74 uninterrupted repeats (44). This study also documented that short CGG tracts are stable in yeast as no contractions or expansions were detected in 48 colonies with the (CGG)30 tract in either orientation. CGG and CCG oligonucleotides both form hairpins in vitro, but the CGG repeats have a higher Tm compared with a CCG tract of similar length (11,15). Also, only the CGG strand forms intra-strand quadruplex structures (16,17). A replication slippage model proposes that hairpins are more likely to form on the template strand for lagging strand synthesis since this strand has more single-strand character than its complement (45). The orientation-dependent stability of CGG repeats in yeast is consistent with a model in which the CGG strand also forms more stable secondary structures in vivo, since the tracts were more prone to contraction when the CGG strand was the template for the lagging strand synthesis and to expansions when CGG was the sequence of the Okazaki fragments. As CGG repeats show the same orientation preference in yeast as they do in bacteria (18,19,46), the ability to form secondary structures is probably determined primarily by DNA sequence, rather than reflecting an unusual nucleosomal structure for CGG repeats (26–28). The data in this paper are most comparable with our earlier work on CTG repeats (22,23) as the repeat tracts were integrated at the same site on chromosome II in both studies. Like CGG repeats, CTG repeats show length- and orientationdependent stability in yeast (22,23), a conclusion also reached by others (24,47). However, CTG tracts expand less often than CGG tracts; using a similar Southern analysis of individual colonies, expansions were not detected in the (CTG)130 strain (22, and unpublished data). In addition, CTG repeats in the unstable orientation were more prone to contraction than CGG repeats in the unstable orientation. Full-length tracts of (CTG)130 in the unstable orientation were not recovered (22), whereas 69% of (CGG)81 and 15% of (CGG)160 colonies in the less stable orientation had mostly full-length tracts after 30 cell divisions (Table 1). Finally, CTG (23) and CGG (Fig. 4B) tracts are both length-dependent fragile sites in yeast. As the rate of recombination for the (CGG)160 tract (3.4 10–3) was similar to that for (CTG)250 (2.4 10–3), CGG tracts appear to break more often than CTG tracts of equal length. The orientation dependence of CTG and CGG repeats in both yeast and Escherichia coli is consistent with models in which the strand that forms the more stable hairpin is the template for lagging strand synthesis in the unstable orientation (and vice versa). However, contrary to the data presented here, this model predicts that CGG repeats will have a stronger orientationdependent stability than CTG repeats because the difference in Tm for CGG versus CCG oligonucleotides is much greater than for CTG versus CAG (15). Thus, aspects of DNA or chromatin structure in addition to the ability to form stable hairpins probably contribute to orientation-dependent stability in yeast, a conclusion also reached in a study on repair rates of different trinucleotide sequences during yeast meiosis (48). It would be interesting to study the effect of both replication timing and proximity to heterochromatin-like DNA on the stability of CGG tracts, as these conditions might mimic the fully methylated environment at the FMR1 locus in humans.

98


How does the stability of CGG repeats in yeast compare with their behavior in human cells? Although there is no information on the direction of replication through expandable CGG tracts in humans, the same orientation bias for expansion in both yeast and bacteria makes a strong prediction that expandable tracts are present in the stable orientation in humans. As in yeast and bacteria, long CGG tracts in humans are increasingly unstable, but, unlike yeast where contractions predominate (Table 1), pre-mutation sized CGG tracts in maternal human carriers almost always expand in future generations (reviewed in ref. 49). However, this difference in expandability may be more apparent than real. Expansion of CGG repeats in humans is probably restricted to the germline; indeed in early development, when DNA is largely unmethylated, large CGG tracts contract in somatic cells (50). Genome-wide DNA methylation that occurs early in human development appears to protect long CGG tracts against further deletions (50). As yeast lacks CpG methylation, the milieu of a mitotic yeast cell may mimic that of the early human embryo in terms of stability of CGG repeats. This possibility can be tested by expressing a CpG-specific methylase in yeast (51) and determining its effects on CGG stability. A major finding of this paper is that CGG tracts are preferential sites of breakage on yeast chromosomes (Fig. 4B). In the strain with the (CGG)81 tract, recombination was only slightly increased over a no-tract control. However, in a strain with a (CGG)160 tract, recombination was increased 100- to 200-fold. As this increase was dependent on genes that mediate double-strand break repair (Fig. 5), the presence of the (CGG)160 tract likely increased recombination by increasing the frequency of double-strand breaks on the yeast chromosome between the direct repeats. We did not detect discrete breakage products by electrophoresis and Southern hybridization, although we often detected CGG-dependent smears that migrated in the expected region (data not shown). These data suggest that the CGG tracts induced breaks at multiple sites within the region separating the direct repeats, rather than at a discrete site near or within the CGG tract. Alternatively, the frequency and/or half-life of CGG-induced breakage products may be too low to allow their easy detection. Given the orientation-dependent stability of CGG tracts in yeast (Table 1), the orientation independence (Fig. 4B) of fragility was unexpected. Two-dimensional gel analysis of CGG tracts of up to 70 repeats in bacterial plasmids show that CGG tracts cause replication fork pausing in an orientation-dependent manner: more pausing is detected when CGG repeats are on the template for lagging strand synthesis (46), a result also seen in yeast with (CGG)81 and (CGG)160 tracts integrated onto chromosome II (B. Lenzmeier, B.S. Balakumaran and V.A. Zakian, unpublished data). The results presented here indicate that the fragility of (CGG)160 repeats is unlikely to be determined exclusively by the extent of pausing during DNA replication but rather is influenced by an orientation-independent aspect of CGG repeats, such as chromatin structure. In human cells, loci containing CGG tracts do not express the fragile site when the tracts are short, consistent with the length-dependent fragility seen in yeast (52). However, expanded CGG tracts in human cells are not thought to break

in cultured cells unless there is a strong selection for the expression of fragile sites by perturbing DNA replication (53,54). Indeed, the failure to recover the expected products of CGG-mediated chromosome breakage in individuals with fragile X syndrome has led to the interpretation that breakage at fragile sites is rare or non-existent in people. However, expanded CGG-tract-induced breakage might occur in humans but escape detection, either because the broken chromosome induces apoptosis (55) or because these breaks are poor substrates for telomere addition (56) or recombinational repair. The demonstration of CGG-tract-induced breakage in unperturbed, wild-type yeast cells suggests that this issue should be re-investigated in humans. MATERIALS AND METHODS A 708 bp EcoRV–PvuII fragment of LYS2 was isolated from pTD27 (a gift from T. Davis, University of Washington, Seattle, WA). After its ends were blunted by treatment with T4 DNA polymerase, this fragment was cloned in both orientations at a blunted XhoI site of the yeast integration vector, pRS306, which contains the URA3 gene (57). (CGG)81 and (CGG)160 tracts were isolated from XbaI–HindIII-digested pRW3316 and pRW3308 [a gift from Dr R.D. Wells (18)]. The CGG tracts were cloned into the XbaI–HindIII sites of pRS306 that contained a 708 bp LYS2 fragment inserted at the XhoI site. The plasmids with CGG tracts were transformed into E.coli SURE cells (Stratagene, La Jolla, CA) and grown at 30C on Luria–Bertani medium with ampicillin selection. For transformation of yeast, integrating plasmids were linearized with AflII, which cuts within the LYS2 fragment, and transformed into YPH500 (MAT ura3-52 ade2-101ochre trp1-63 his3200 leu2-1 lys2-801am) (57) using the lithium acetate protocol and dimethylsulfoxide (58,59). The open reading frames of the rad50 and rad1 genes were precisely deleted by transformation of a fragment generated by PCR using HIS3 as the template and primers containing flanking sequences of the genes to be deleted (60). The rad52 gene was deleted using the rad52:LEU2 fragment from the plasmid pSM20 (a gift from D. Schild, Lawrence Livermore National Laboratory, Berkeley, CA). The structure of all transformants was determined by restriction digestion and Southern blot analysis. To determine the stability of CGG tracts, transformants with full-length tracts were streaked for single cells, and genomic DNA was prepared from the daughter colonies using a glass-bead protocol (61). Genomic DNA was digested with SacI and ClaI, separated on a 1.5% agarose gel and blotted onto a nylon membrane. A 32P-labeled CGG fragment isolated from XbaI– HindIII-digested pRW3308 was used for hybridization overnight in Superstarks solution (50% formamide, 20 SSC, 1 Denhardt’s solution, 2% SDS, 0.5% milk, 1 mg of herring sperm DNA per ml). The blots were washed twice for 10 min each time in 2 SSC, 0.1% SDS at 25C, followed by two 20 min washes in 0.1 SSC, 0.1% SDS at 65C. The blots were then exposed either to a Kodak X-omat film or to a phosphor imager. Contingency table statistical tests (2 and Fisher’s tests) were used to analyze the results of orientation-dependent stability (31). For fluctuation analysis, cells were streaked on non-selective yeast extract/peptone/dextrose (YEPD) (62) plates for single colonies except for experiments involving hydroxyurea, in which case plates contained 0.05 M hydroxy-


urea, and colonies were picked as agar plugs and resuspended in water. A portion was plated on FOA plates and an appropriate dilution was plated on YEPD plates for estimating the total number of colonies. The rate of FOAR was calculated using the method of the median (35,36). ACKNOWLEDGEMENTS We thank R. Wells for the generous gift of plasmids containing CGG repeats, S. Warren for helpful discussions on the behavior of CGG repeats in humans, K. Petren for help with statistical analysis, and B. Lenzmeier and L. Vega for their comments on the manuscript and discussion of data throughout the course of these experiments. We also thank B. Lenzmeier for permission to cite his unpublished data. This work was supported by NIH grant CA79441. REFERENCES 1. Caskey, C.T., Pizzuti, A., Fu, Y.H., Fenwick Jr, R.G. and Nelson, D.L. (1992) Triplet repeat mutations in human disease. Science, 256, 784–789. 2. Oostra, B.A. and Verkerk, A.J. (1992) The fragile X syndrome: isolation of the FMR-1 gene and characterization of the fragile X mutation. Chromosoma, 101, 381–387. 3. Ashley, C.T. and Warren, S.T. (1995) Trinucleotide repeat expansion and human disease. Annu. Rev. Genet., 29, 703–728. 4. Pieretti, M., Zhang, F., Fu, Y.H., Warren, S.T., Oostra, B.A., Caskey, C.T. and Nelson, D.L. (1991) Absence of expression of the FMR-1 gene in fragile X syndrome. Cell, 66, 817–822. 5. Sutcliffe, J.S., Nelson, D.L., Zhang, F., Pieretti, M., Caskey, C.T., Saxe, D. and Warren, S.T. (1992) DNA methylation represses FMR-1 transcription in fragile X syndrome. Hum. Mol. Genet., 1, 397–400. 6. Tapscott, S.J., Klesert, T.R., Widrow, R.J., Stoger, R. and Laird, C.D. (1998) Fragile-X syndrome and myotonic dystrophy: parallels and paradoxes. Curr. Opin. Genet. Dev., 8, 245–253. 7. Knight, S.J., Flannery, A.V., Hirst, M.C., Campbell, L., Christodoulou, Z., Phelps, S.R., Pointon, J., Middleton-Price, H.R., Barnicoat, A., Pembrey, M.E. et al. (1993) Trinucleotide repeat amplification and hypermethylation of a CpG island in FRAXE mental retardation. Cell, 74, 127–134. 8. Sutherland, G.R. and Richards, R.I. (1999) Fragile sites—cytogenetic similarity with molecular diversity. Am. J. Hum. Genet., 64, 354–359. 9. Berger, R., Bloomfield, C.D. and Sutherland, G.R. (1985) Report of the committee on chromosome rearrangements in neoplasia and on fragile sites. Cytogenet. Cell Genet., 40, 490–535. 10. Reddy, P.S. and Housman, D.E. (1997) The complex pathology of trinucleotide repeats. Curr. Opin. Cell Biol., 9, 364–372. 11. Mitas, M. (1997) Trinucleotide repeats associated with human disease. Nucleic Acids Res., 25, 2245–2254. 12. Ohshima, K. and Wells, R.D. (1997) Hairpin formation during DNA synthesis primer realignment in vitro in triplet repeat sequences from human hereditary disease genes. J. Biol. Chem., 272, 16798–16806. 13. Darlow, J.M. and Leach, D.R.F. (1998) Secondary structures in d(CGG) and d(CCG) repeat tracts. J. Mol. Biol., 275, 3–16. 14. Sakamoto, N., Chastain, P.D., Parniewski, P., Ohshima, K., Pandolfo, M., Griffith, J.D. and Wells, R.D. (1999) Sticky DNA: self-association properties of long GAA.TTC repeats in R.R.Y triplex structures from Friedreich’s ataxia. Mol. Cell, 3, 465–475. 15. Gacy, A.M. and McMurray, C.T. (1998) Influence of hairpins on template reannealing at trinucleotide repeat duplexes: a model for slipped DNA. Biochemistry, 37, 9426–9434. 16. Nadel, Y., Weisman-Shomer, P. and Fry, M. (1995) The fragile X syndrome single strand d(CGG)n nucleotide repeats readily fold back to form unimolecular hairpin structures. J. Biol. Chem., 270, 28970–28977. 17. Usdin, K. (1998) NGG-triplet repeats form similar intrastrand structures: implications for the triplet expansion diseases. Nucleic Acids Res., 26, 4078–4085. 18. Shimizu, M., Gellibolian, R., Oostra, B.A. and Wells, R.D. (1996) Cloning, characterization and properties of plasmids containing CGG triplet repeats from the FMR-1 gene. J. Mol. Biol., 258, 614–626.

99

19. Hirst, M.C. and White, P.J. (1998) Cloned human FMR1 trinucleotide repeats exhibit a length- and orientation-dependent instability suggestive of in vivo lagging strand secondary structure. Nucleic Acids Res., 26, 2353– 2358. 20. Kang, S., Jaworski, A., Ohshima, K. and Wells, R.D. (1995) Expansion and deletion of CTG repeats from human disease genes are determined by the direction of replication in E.coli. Nature Genet., 10, 213–218. 21. Bowater, R.P., Rosche, W.A., Jaworski, A., Sinden, R.R. and Wells, R.D. (1996) Relationship between Escherichia coli growth and deletions of CTG.CAG triplet repeats in plasmids. J. Mol. Biol., 264, 82–96. 22. Freudenreich, C.H., Stavenhagen, J.B. and Zakian, V.A. (1997) Stability of a CTG/CAG trinucleotide repeat in yeast is dependent on its orientation in the genome. Mol. Cell. Biol., 17, 2090–2098. 23. Freudenreich, C.H., Kantrow, S.M. and Zakian, V.A. (1998) Expansion and length-dependent fragility of CTG repeats in yeast. Science, 279, 853–856. 24. Maurer, D.J., O’Callagan, B.L. and Livingston, D.M. (1996) Orientation dependence of trinucleotide CAG repeat instability in Saccharomyces cerevisiae. Mol. Cell. Biol., 16, 6617–6622. 25. Miret, J.J., Pessoa-Brandao, L. and Lahue, R.S. (1997) Instability of CAG and CTG trinucleotide repeats in Saccharomyces cerevisiae. Mol. Cell. Biol., 17, 3382–3387. 26. Wang, Y.H. and Griffith, J. (1996) Methylation of expanded CCG triplet repeat DNA from fragile X syndrome patients enhances nucleosome exclusion. J. Biol. Chem., 271, 22937–22940. 27. Wang, Y.H., Gellibolian, R., Shimizu, M., Wells, R.D. and Griffith, J. (1996) Long CCG triplet repeat blocks exclude nucleosomes: a possible mechansim for the nature of fragile sites in chromosomes. J. Mol. Biol., 263, 511–516. 28. Godde, J.S., Kass, S.U., Hirst, M.C. and Wolffe, A.P. (1996) Nucleosome assembly on methylated CGG triplet repeats in the fragile X mental retardation gene 1 promoter. J. Biol. Chem., 271, 24325–24328. 29. Verkerk, A.J.M.H., Pieretti, M., Sutcliffe, J.S., Fu, Y.-H., Kuhl, D.P.A., Pizzuti, A., Reiner, O., Richards, S., Victoria, M.F., Zhang, F. et al. (1991) Identification of a gene (FMR-1) containing a CGG repeat coincident with a breakpoint cluster region exhibiting length variation in fragile X syndrome. Cell, 65, 905–914. 30. Kunst, C.B. and Warren, S.T. (1994) Cryptic and polar variation of the fragile X repeat could result in predisposing normal alleles. Cell, 77, 853– 861. 31. Zar, J. (1984) Biostatistical Analysis. 2nd edn. Prentice Hall, NJ. 32. Boeke, J.D., LaCroute, F. and Fink, G.R. (1984) A positive selection for mutants lacking orotidine-5-phosphate decarboxylase activity in yeast: 5fluoro-orotic acid resistance. Mol. Gen. Genet., 197, 345–346. 33. Sugawara, N. and Haber, J.E. (1992) Characterization of double-strand break-induced recombination: homology requirements and single-stranded DNA formation. Mol. Cell. Biol., 12, 563–575. 34. Hartwell, L. and Smith, D. (1985) Altered fidelity of mitotic chromosome transmission in cell cycle mutants of S.cerevisiae. Genetics, 110, 381–395. 35. Lea, D.E. and Coulson, C.A. (1949) The distribution of the numbers of mutants in bacterial populations. J. Genet., 49, 264–285. 36. Jones, M.E., Thomas, S.M. and Rogers, A. (1994) Luria–Delbruck fluctuation experiments: design and analysis. Genetics, 136, 1209–1216. 37. Sung, P., Reynolds, P., Prakash, L. and Prakash, S. (1993) Purification and characterization of the Saccharomyces cerevisiae RAD1/RAD10 endonuclease. J. Biol. Chem., 268, 26391–26399. 38. Tomkinson, A.E., Bardwell, A.J., Bardwell, L., Tappe, N.J. and Friedberg, E.C. (1993) Yeast DNA repair and recombination proteins Rad1 and Rad10 constitute a single-stranded-DNA endonuclease. Nature, 362, 860–862. 39. Bardwell, A.J., Bardwell, L., Tomkinson, A.E. and Friedberg, E.C. (1994) Specific cleavage of model recombination and repair intermediates by the yeast Rad1–Rad10 DNA endonuclease. Science, 265, 2082–2085. 40. Prakash, S., Prakash, L., Burke, W. and Montelone, B. (1980) Effects of the Rad52 gene on recombination in Saccharomyces cerevisiae. Genetics, 94, 31–50. 41. Malone, R.E. and Esposito, R.E. (1980) The RAD52 gene is required for homothallic interconversion of mating types and spontaneous mitotic recombination in yeast. Proc. Natl Acad. Sci. USA, 77, 503–507. 42. Hartwell, L.H. (1976) Sequential function of gene products relative to DNA synthesis in the yeast cell cycle. J. Mol. Biol., 104, 803–814. 43. Sutherland, G.R. and Richards, R.I. (1995) The molecular basis of fragile sites in human chromosomes. Curr. Opin. Genet. Dev., 5, 323–327. 44. White, P.J., Borts, R.H. and Hirst, M.C. (1999) Stability of the human fragile X (CGG)(n) triplet repeat array in Saccharomyces cerevisiae deficient in aspects of DNA metabolism. Mol. Cell. Biol., 19, 5675–5684.

100


45. McMurray, C.T. (1995) Mechanisms of DNA expansion. Chromosoma, 104, 2–13. 46. Samadashwily, G.M., Raca, G. and Mirkin, S.M. (1997) Trinucleotide repeats affect DNA replication in vivo. Nature Genet., 17, 298–304. 47. Miret, J.J., Pessoa-Brandao, L. and Lahue, R.S. (1998) Orientationdependent and sequence-specific expansions of CTG/CAG trinucleotide repeats in Saccharomyces cerevisiae. Proc. Natl Acad. Sci. USA, 95, 12438–12443. 48. Moore, H., Greenwell, P.W., Liu, C.P., Arnheim, N. and Petes, T.D. (1999) Triplet repeats form secondary structures that escape DNA repair in yeast. Proc. Natl Acad. Sci. USA, 96, 1504–1509. 49. Nelson, D. (1998) Fragile-X syndrome. In Jameson, J. (ed.), Principles of Molecular Medicine. Humana Press, Totowa, NJ, pp. 1063–1067. 50. Malter, H.E., Iber, J.C., Willemsen, R., de Graaff, E., Tarleton, J.C., Leisti, J., Warren, S.T. and Oostra, B.A. (1997) Characterization of the full fragile X syndrome mutation in fetal gametes. Nature Genet., 15, 165–169. 51. Kladde, M.P., Xu, M. and Simpson, R.T. (1996) Direct study of DNA– protein interactions in repressed and active chromatin in living cells. EMBO J., 15, 6290–6300. 52. Yu, S., Mulley, J., Loesch, D., Turner, G., Donnelly, A., Gedeon, A., Hillen, D., Kremer, E., Lynch, M., Pritchard, M. et al. (1992) Fragile-X syndrome: unique genetics of the heritable unstable element. Am. J. Hum. Genet., 50, 968–980. 53. Warren, S.T., Zhang, F., Licameli, G.R. and Peters, J.F. (1987) The fragile X site in somatic cell hybrids: an approach for molecular cloning of fragile sites. Science, 237, 420–423.

54. Glover, T.W. and Stein, C.K. (1988) Chromosome breakage and recombination at fragile sites. Am. J. Hum. Genet., 43, 265–273. 55. Evan, G. and Littlewood, T. (1998) A matter of life and cell death. Science, 281, 1317–1322. 56. Hanish, J.P., Yanowitz, J.L. and De Lange, T. (1994) Stringent sequence requirements for the formation of human telomeres. Proc. Natl Acad. Sci. USA, 91, 8861–8865. 57. Sikorski, R.S. and Hieter, P. (1989) A system of shuttle vectors and yeast host strains designed for efficient manipulation of DNA in Saccharomyces cerevisiae. Genetics, 122, 19–27. 58. Schiestl, R.H. and Gietz, R.D. (1989) High efficiency transformation of intact yeast cells using single stranded nucleic acids as a carrier. Curr. Genet., 16, 339–346. 59. Hill, J., Ian, K.A., Donald, G. and Griffiths, D.E. (1991) DMSO-enhanced whole-cell yeast transformation. Nucleic Acids Res., 19, 5791. 60. Lorenz, M.C., Muir, R.S., Lim, E., McElver, J., Weber, S.C. and Heitman, J. (1995) Gene disruption with PCR products in Saccharomyces cerevisiae. Gene, 158, 113–117. 61. Rose, M.D., Winston, F. and Hieter, P. (1990) Methods in Yeast Genetics: A Laboratory Course Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 62. Zakian, V.A. and Scott, J.F. (1982) Construction, replication and chromatin structure of TRP1 RI circle, a multiple-copy synthetic plasmid derived from Saccharomyces cerevisiae chromosomal DNA. Mol. Cell. Biol., 2, 221– 232.