Involvement of Very Short DNA Tandem Repeats and the Influence of the RAD52 Gene on the Occurrence of Deletions in Saccharomyces cerevisiae

Corresponding author: Jean-Luc Souciet, Laboratoire de Microbiologie et de Génétique, UPRES-A 7010, Institut de Botanique, 28, rue Goethe, 67083 Strasbourg Cedex, France., souciet{at}gem.u-strasbg.fr (E-mail)

Communicating editor: M. JOHNSTON

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Chromosomal rearrangements, such as deletions, duplications, or Ty transposition, are rare events. We devised a method to select for such events as Ura⁺ revertants of a particular ura2 mutant. Among 133 Ura⁺ revertants, 14 were identified as the result of a deletion in URA2. Of seven classes of deletions, six had very short regions of identity at their junctions (from 7 to 13 bp long). This strongly suggests a nonhomologous recombination mechanism for the formation of these deletions. The total Ura⁺ reversion rate was increased 4.2-fold in a rad52 strain compared to the wild type, and the deletion rate was significantly increased. All the deletions selected in the rad52 context had microhomologies at their junctions. We propose two mechanisms to explain the occurrence of these deletions and discuss the role of microhomology stretches in the formation of fusion proteins.

CHROMOSOMAL double-strand breaks (DSBs) are potentially lethal events that cells have to repair. DSBs can occur spontaneously, during recombination, or can be induced by damage to DNA. Their repair can lead to genomic rearrangements, such as deletions, duplications, translocations, and other chromosomal abnormalities, which may lead to cell death or other malfunctions, including carcinogenesis in mammalian cells (MEUTH 1989 ; TLSTY et al. 1995 ). There are two major pathways for the repair of DSBs in eukaryotic cells (for a review see PAQUES and HABER 1999 ). In Saccharomyces cerevisiae, the first repair pathway is based upon homologous recombination, which depends on the RAD52 epistasis group of genes. The homologous recombination pathway requires similar sequences at least 50–100 bp long (SUGAWARA and HABER 1992 ). Repair can be by homologous recombination sensu stricto, or single-strand annealing (SSA). Homologous recombination sensu stricto consists of copying a homologous DNA region located on a sister chromatid or on a homologous chromosome to repair a DNA lesion. Events such as conversion can achieve this (RESNICK 1976 ). The SSA pathway processes a DSB surrounded by tandem repeats, at least 50–100 bp long, to form a chromosomal deletion (LIN et al. 1984 ; HABER 1992 ).

The second pathway for DSB repair is the nonhomologous pathway, also called the nonhomologous end joining (NHEJ) pathway. This repair pathway requires little (2–20 bp) or no sequence homology between the ends of the DSBs (MEZARD et al. 1992 ; KRAMER et al. 1994 ; MEZARD and NICOLAS 1994 ). The molecular mechanisms of this pathway are being investigated by many groups (for a review see TSUKAMOTO and IKEDA 1998 ). There are at least four types of DNA end-joining mechanism. The first one requires no complementary sequence at either end: two protruded single-stranded ends are ligated and the resulting gap is filled in by DNA synthesis (ROTH et al. 1985 ; PFEIFFER and VIELMETTER 1988 ). The second involves the joining of DNA ends through the pairing of short complementary sequences (ROTH and WILSON 1986 ; PFEIFFER and VIELMETTER 1988 ), as happens with complementary overhanging DNA ends produced by endonucleases. These short complementary sequences can be considered to be microhomology regions. The third mechanism is based upon adding untemplated nucleotides to one blunt end, which enables it to base pair with the other end (THODE et al. 1990 ). The last mechanism, which is a rare event, involves inserting an unrelated DNA fragment, called filler DNA, into the junction during the joining reaction (ROTH et al. 1985 ; ROTH and WILSON 1986 ). In S. cerevisiae, this filler DNA can be Ty sequences (MOORE and HABER 1996A ; TENG et al. 1996 ).

NHEJ is more frequent in mammalian cells than is homologous recombination. In yeast, the predominant pathway is homologous recombination. While some DSBs are repaired via the NHEJ pathway, the pairing of short complementary sequences does most of these repair events. The RAD52 gene, along with the RAD51 and RPA genes, is implicated in the homologous recombination pathway. For instance, Rad52p is required for SSA between long direct repeats. Several different studies have shown that the NHEJ DSB repair pathway is RAD52 independent (MEZARD and NICOLAS 1994 ; MOORE and HABER 1996B ).

Most studies on NHEJ or homologous recombination in yeast have used in vivo artificial systems, such as plasmid constructs or engineered genomic loci (KRAMER et al. 1994 ; MOORE and HABER 1996B ). These systems allow the fairly rapid detection and characterization of the recombination events leading to cell survival, but may not faithfully reproduce the chromosomal environment. Our lab has previously developed a genetic screen allowing the in vivo selection of spontaneous genomic rearrangements (EXINGER and LACROUTE 1979 ; ROELANTS et al. 1995 ). This genetic screen is based on the use of a particular allele of the S. cerevisiae URA2 gene. The URA2 gene encodes a 240-kD multifunctional protein composed of glutamine amido transferase (GATase), carbamoyl phosphate synthetase (CPSase) and aspartate transcarbamoylase (ATCase) catalytic domains, and one inactive dihydroorotase (DHOase)-like domain (Fig 1). Ura2p catalyzes the first two steps in the pyrimidine biosynthesis pathway by its CPSase-ATCase activities. The ura2 15-30-72 mutant allele used in our genetic screen bears three mutations: two nonsense mutations and a frameshift, all of them located in the GATase-CPSase encoding region. The ATCase domain is inactive in the ura2 15-30-72 mutant strain, mainly because of the polarity of nonsense mutations, so the strain is a uracil auxotroph. The functional reactivation of the ATCase domain is sufficient to generate Ura⁺ prototrophs (as described in EXINGER and LACROUTE 1979 ). The frequency of the ATCase reactivation events is very low, 10^-10. Three types of reactivation events have been previously described: insertion of a Ty1 retrotransposon in the URA2 coding sequence upstream of the last mutation (BACH 1984 ; ROELANTS et al. 1995 ), deletion of the sequence bearing the three mutations (EXINGER and LACROUTE 1979 ; ROELANTS et al. 1995 ), and duplication of the ATCase coding sequence and its insertion elsewhere in the genome under the dependence of a resident promoter (BACH et al. 1995 ; ROELANTS et al. 1995 ). The usual distribution observed for these reactivation events is 2/3 Ty1 insertion, 1/6 deletion, and 1/6 duplication (ROELANTS et al. 1995 ).

View larger version (11K): In this window In a new window Download PPT slide   Figure 1. The general organization of the URA2 locus, with the catalytic domains, and the locations of the three point mutations in the ura2 15-30-72 allele. The thick line indicates the URA2 coding sequence, which is 6645 nucleotides long. ns 15 and ns 30 represent the sites of the nonsense mutations, fs 72 represents the site of the frameshift. GATase, glutamine amidotransferase; CPSase, carbamoylphosphate synthetase; DHOase-like, dihydroorotase-like; ATCase, aspartate transcarbamoylase. The diagrams are not drawn to scale.

In this work, we have investigated the ATCase reactivation events due to deletion of URA2 sequences. To understand the way spontaneous recombination affects the reshaping of chromosomes, we selected independent revertants that were characterized at the molecular level and sequenced. The presence of microhomologies at the ends of the deletions suggests that a nonhomologous recombination mechanism is involved in the formation of these deletions. We have also examined how the recombination gene RAD52 affected this mechanism. The possibility of the recombination of direct DNA tandem repeats located in different genes, leading to new genes encoding a fusion protein, is discussed.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Media and yeast strains:
Yeast cells were grown at 30° in YPG or YNB, liquid or solid (2% agar) medium, appropriately supplemented. For the genetic techniques, methods described by MORTIMER and HAWTHORNE 1966 were followed. Yeast transformation was done as previously described (BECKER and GUARENTE 1991 ).

All strains used in this study are isogenic derivatives of FL100 (ATCC 28583). The three independent point-mutated ura2-15, ura2-30, and ura2-72 strains were crossed to construct the mutated ura2 15-30-72 strain, according to EXINGER and LACROUTE 1979 . Strain AW1 was obtained by crossing a ura2 15-30-72 trp1-4 Gal⁺ strain with a his3200 strain followed by selection of a ura2 15-30-72 trp1-4 his3200 Gal⁺ haploid strain. rad52 strain AW3 was constructed by a single-step gene replacement of the AW1 strain with a PCR-generated DNA fragment, according to WACH et al. 1994 . DNA amplification was performed on matrix plasmid pFA6a, with 65-mer oligonucleotides containing, respectively, 18 and 19 bases of 5' and 3' kanMX4 multiple cloning site sequences as described in WACH et al. 1994 and, respectively, 47 and 46 bases complementary to specific regions 5' and 3' to the RAD52 gene. The S1 primer was 5'-GACGAAAAATATAGCGGCGGGCGGGT TACGCGACCGGTATCGAATGGCGTACGCTGCAGGTCGAC-3', and the S2 primer was 5'-ATAAATAATGATGCAAATTTTTTATTTGTTTCGGCCAGGAAGCGTTATCGATGAATTCGAGCTCG -3'; bold letters denote pFA6a sequences flanking the kanMX4 module. The kanMX4-rad52 PCR fragment was then used to transform AW1 haploid strain and G418^R transformants were selected on YPG medium with G418 (Sigma, St. Louis) at a final concentration of 200 mg/liter. The DNA replacement pattern at the RAD52 locus of G418^R transformants was confirmed by PCR amplification.

Selection of Ura⁺ revertants:
One isolated colony of the ura2 15-30-72 strain was used to inoculate one YPD plate and incubated at 30° for 5–6 days. Then, the cells were harvested, resuspended in water, spread on YNB appropriately supplemented without uracil, and incubated at 30° to allow growth of Ura⁺ revertants. Simultaneously, a 100-µl aliquot of each culture was diluted 10^-6 times, plated on YPG medium, and grown at 30° to estimate the total number of viable cells that were plated in each experiment. Spontaneous revertants were generally recovered after 7 days at 30° and were then subjected to molecular analysis. The frequency of ATCase reactivation was extremely low, ranging from 10^-11 to 10^-9. Three types of ATCase reactivation events were obtained. The Ura⁺ revertants were considered independent when isolated from independent Ura^- cultures or when isolated from the same culture but were clearly shown to be the results of different molecular events.

Mutation rate determination:
The rates of reversion to Ura⁺ of two to three sets of 5–10 independent cultures were calculated by the maximum-likelihood method as described in LEA and COULSON 1949 . As there are several types of events that can lead to ATCase reactivation, the total reversion rates were obtained by summing the mutation rates associated with each of the different mutations. The 95% confidence limits were calculated using Student's t-test.

Southern blot analysis:
Total DNA from S. cerevisiae was prepared following HOFFMAN and WINSTON 1987 . Restriction endonuclease digestions were carried out as described by the manufacturers. DNA blots were prepared from pulsed-field gel electrophoresis (PFGE) and conventional agarose gels by vacuum transfer of DNA to Hybond-N⁺ membrane (Amersham Pharmacia Biotech, Arlington Heights, IL). Digoxygenin (DIG)-labeled DNA probe was prepared using the DIG DNA labeling and detection kit, and signal detection was performed using the luminescent substrate CDP-Star (Roche Diagnostics).

PCR amplification:
The names, locations, sequences, and temperatures of melting (T_m) of the primers described in this work are listed in Table 1. Amplification was performed using as DNA matrix 1 µl of a 1/100 dilution of yeast total DNA preparation. Amplification with Taq DNA polymerase from Sigma was carried out in an Eppendorf (Madison, WI) thermocycler or in a MJ Research (Watertown, MA) thermocycler, using the following procedure: 94° for 2 min, 33 cycles at 94° for 30 sec, x° for 30 sec (according to T_m determination rules), and 72° for 2 min followed by 72° for 10 min. This procedure amplified DNA fragments up to 2.5 kb long. DNA products were analyzed and sized on 1% agarose gels.

To characterize the three point mutations specific to the ura2 15-30-72 allele, oligonucleotides 3207-3213R and 3216-577 were used on the AW1 and FL100 strains. Then, the resulting PCR fragments were sequenced on both strands with, respectively, the oligonucleotide pairs 3207/3212R and 3212/3213R for the 3207-3213R fragment and 3438/598 for the 3216-577 fragment.

DNA sequencing and sequence analysis:
Double-strand DNA fragments obtained after PCR amplification were purified using the Gene Clean II kit from Bio-Rad (Richmond, CA). Double-strand DNA sequencing was then performed on these PCR-purified fragments using the method described by SANGER et al. 1977 and the Thermo Sequenase radiolabeled terminator cycle sequencing kit from Amersham Pharmacia Biotech. Nucleotide localization in the alleles is referred to the A + 1 position of the first ATG of the coding sequence. The URA2 gene accession no. is M27174 (SOUCIET et al. 1989 ). We performed FASTA analyses after sequencing of the PCR products to determine the precise boundaries of the deletion events. This was done with UWGCG programs (DEVEREUX et al. 1984 ) on a VAX II/750 minicomputer. Other general analyses of the nucleotide sequences were done using DNA Strider 1.3.

PFGE:
Chromosomal DNA was prepared as described by CARLE and OLSON 1985 . Chromosomes were separated on a 1% agarose gel (Pharmacia, Piscataway, NJ) in 0.5x TBE buffer at 7 V/cm for 24 hr with a pulse time of 45 sec and an angle of 120°, using a Bio-Rad CHEF-DRIII mapper apparatus. The gel was stained with ethidium bromide to identify the chromosomal pattern specific for each strain.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Molecular characterization of the three point mutations specific to the ura2 15-30-72 allele:
The ura2 15-30-72 allele was constructed using the genetic approach depicted in EXINGER and LACROUTE 1979 , but was never characterized at the molecular level. The position of each point mutation was detected by comparing the nucleotide sequences of the ura2 15-30-72 Ura^- AW1 strain to that of the wild-type FL100 strain, as described in MATERIALS AND METHODS. Comparison of the mutated and wild-type sequences indicates that nonsense mutation ura2-15 (opal) corresponds to an A to T transversion (codon 88 AGA becomes TGA) at position +262 and nonsense mutation ura2-30 (ochre) corresponds to a C to T transition at nucleotide 916 of the gene (codon 306 CAA becomes TAA). Both of these nonsense mutations are located in the GATase coding sequence. Mutation ura2-72 is a frameshift at position +2972 (G nucleotide missing in codon 991 CGT), in the middle of the CPSase coding sequence (Fig 1).

Molecular characterization of the Ura⁺ revertants:
Revertants from the ura2 15-30-72 strain were isolated as described in MATERIALS AND METHODS. Southern blots were analyzed to discriminate between duplication events and Ty insertions or deletion events as described in ROELANTS et al. 1995 . The Ty insertions in the URA2 coding sequence were then characterized by PCR using an oligonucleotide specific to the LTR of Ty1 and a second oligonucleotide of the coding sequence of ATCase (ROELANTS et al. 1995 ). The deletion events were characterized by PCR as reported in Fig 2. No band was detectable for the ura2 15-30-72 allele using the oligonucleotide pair 38,3–AW1, distant from 5307 bp, under our standard PCR conditions. But large deletions between primers 38,3 and AW1 could be detected with these PCR conditions and the lengths of the amplified DNA fragments obtained helped to size the deletions (Fig 2A). No amplification between oligonucleotides 38,3 and AW1 after all these successive screening steps meant that larger deletion events had occurred in some strains. These deletions were further characterized by primer-walking PCR (Fig 2B). Oligonucleotide pairs located upstream of the AW1 primer (or downstream of the 38,3 primer) were used to locate the borders of these deletions, with the AW1 strain as a positive control for the PCR experiment. The lack of amplification between a pair of close oligonucleotides indicated a deletion of the region containing one of the two primers.

View larger version (29K): In this window In a new window Download PPT slide   Figure 2. Identification by PCR amplification of Ura+ strains due to deletion of URA2 sequences. (A) First screening step for deleted strains. The strains identified as potential deletion events after Southern blot analyses and Ty-specific PCR amplifications were subjected to another round of PCR, using primer 38,3 localized in the 5' UTR of the URA2 gene, and primer AW1, specific to the URA2 coding phase. (B) Primer-walking PCR. For large deletions that cannot be detected with the oligonucleotide pair 38,3-AW1, a second PCR screening step was performed. The primer-walking technique uses the presence or absence of a PCR amplification signal between close primer pairs to define the boundaries of a sequence. X represents chromosome X. The diagrams are not drawn to scale.

The above screening steps allowed us to identify 14 independent deletion strains among the Ura⁺ revertants obtained after selections. After roughly mapping the 5' and 3' borders of all the deletion events (Fig 2B), we determined the exact borders and sizes of the deletions in the Ura⁺ revertants by DNA sequencing the products of the PCR amplifications (see MATERIALS AND METHODS). The 14 independent deletions selected in our system fell into seven different categories (classes A–G, Table 2).

Analysis of the deletion events:
The overall localization of the deletions obtained in our genetic system revealed three general types of deletion events (Fig 3): (i) a URA2 internal deletion (classes C–F) retained the URA2 wild-type promoter and ATG initiator codon; (ii) a promoter deletion (classes A and B) deleted part of the URA2 wild-type promoter and the ATG initiator codon without fusing the remainder of the ura2 gene to another gene; (iii) a deletion-fusion event (class G) was an in-frame fusion between part of the ura2 allele and another gene located on chromosome X, the PBS2 gene, leading to the deletion of the intervening TRK1 gene. The region encompassing the two nonsense mutations and the frameshift was always deleted (Table 2). Moreover, the 3' ends of the deletions were located in the CPSase coding region (except for the class C events), so the new truncated ura2 alleles contained the DHOase-like domain.

View larger version (21K): In this window In a new window Download PPT slide   Figure 3. Three kinds of general deletions that could explain ATCase activity recovery. At the top is the chromosomal organization of the near surroundings of the URA2 locus. The numbers under the drawings are the chromosomal coordinates of the corresponding genes, as given by the complete genome sequence of the S288C S. cerevisiae strain. The arrows indicate the direction of translation for the genes. The diagrams are not drawn to scale.

The class C deletion events are the largest URA2 internal deletions producing a functional ATCase domain ever characterized: 5315 bp of ura2 internal sequence was deleted, removing the URA2 DHOase-like domain. The protein synthesized from this particular ura2 allele is predicted to be 438 amino acids (aa) long and consists almost exclusively of the ATCase coding domain. Thus, these class C alleles encode a minimal ATCase.

Presence of direct repeats at the ends of the deletions:
There are identical sequences at both ends of all the deletion classes, with one exception. These sequences are organized as perfect direct repeats that differ for each class of deletions and are 7–13 bp long. The sequences of these direct repeats with their surrounding sequences on each side of the deletions are given in Fig 4. The class A deletion is the only one without direct repeats at its junction.

View larger version (43K): In this window In a new window Download PPT slide   Figure 4. The repeated sequences at the boundaries of the deletions. The boxed sequences correspond to the sequences of the deletant strains as read after sequencing of the PCR products. The repeated sequences are represented by boldface letters in the boxes (except for class A, which has no repeated sequence). For each class of deletion, the upper line represents the URA2 or PBS2 sequences located on the 5' side of the deletion and the lower line represents the URA2 sequences on the 3' side of the deletion. For each class of deletion, at the beginning and at the end of the sequences, the position of these sequences with respect to the URA2 or PBS2 gene is given. The sequences are given 5' to 3' from left to right.

Deletion of the RAD52 gene increases the frequency of the deletions:
The repeated sequences at the ends of the deletions can be seen as the microhomology sequences used for single-strand pairing in one of the NHEJ DSB repair pathways. As the NHEJ DSB repair pathway is RAD52 independent, we investigated the RAD52 dependence of the deletions obtained in our genetic system. We also wanted to know, in case deletion events could be selected in a rad52 background, if these deletions could take place between direct repeats. We constructed a ura2 15-30-72 rad52 strain (see MATERIALS AND METHODS) and selected Ura⁺ revertants, following the selection scheme described above, in parallel with the wild-type control strain.

The results of these selections in the RAD52 and rad52 backgrounds are summarized in Table 3. The ATCase reversion rate in the RAD52 control strain was in the range of that obtained for previous selections (10^-10). The overall rate of ATCase reactivation events in the rad52 background was 4.2 times greater than in the RAD52 background. Thus the inactivation of the homologous recombination pathway by deletion of the RAD52 gene favored the appearance of Ura⁺ revertants.

The molecular characteristics of the Ura⁺ revertants and the corresponding mutation rates are shown in Table 3. The four independent reactivation events selected in the RAD52 background included one Ty1 insertion, one deletion, and two duplications. In the rad52 background, 68% of the 19 independent reactivation events were deletions and 32% were duplications. At the number of cells analyzed, no Ty insertion was obtained in the rad52 strain. There was a significant increase (12-fold) of the deletion rate from the RAD52 to the rad52 background and 2.7 times more (from 25 to 68%) deletions among the reactivation events in the rad52 background than in the RAD52 background. There was no significant increase in the duplication rate from the RAD52 to the rad52 background.

The molecular characterization of the ends of the deletions selected in the RAD52 and rad52 backgrounds reveals two new classes of deletion events:
The only Ura⁺-deleted strain obtained in these selections in a RAD52 background did not belong to any of the seven deletion classes described above. This strain was therefore assigned to a new deletion class, H (Table 2). The sequencing of the deletion ends showed an 11-bp-long imperfect direct repeat at the junction (Fig 4). The junction sequence of the deleted strain was identical to the repeat located on the 5' side of the deletion.

Five of the 13 independent deletions obtained in the rad52 background belonged to deletion class B, 4 to class C, and 2 remain to be characterized at the nucleotide level. The last 2 independent deletions had the same location in the URA2 gene, which differed from the 8 already characterized. They therefore form a new deletion class, I (Table 2). These new deleted strains had an 11-bp-long direct repeat at their junctions (Fig 4).

The newly discovered deletion classes H and I resembled the other deletion events already characterized in this work. They were internal URA2 sequence deletions and they both deleted the region between the first nonsense mutation and the frameshift mutation (Table 2). The 3' end of these deletions was located in the CPSase coding region as were the majority of deletions selected.

The chromosomal profile of the deleted strains is not modified:
We checked the chromosomal profiles of the deletion strains to see whether deletion events on chromosome X were associated with larger rearrangements. To do so, we compared the PFGE profiles of one Ura⁺ representative of each class of deletion events with the profile of the AW1 Ura^- progenitor cell. The chromosomal profiles of the Ura⁺ strains showed no detectable differences from the Ura^- progenitor cell, even for chromosome X, where the deletion events took place (data not shown). We therefore assume that the deletions events found in our genetic screen are not associated with other large structural rearrangements.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The aim of this work was to characterize chromosomal rearrangements in S. cerevisiae and to elucidate the mechanisms of their formation. The mechanisms should to be similar to those generally used by the cell, since neither additional sequences nor plasmids are needed. We used a specific allele of the URA2 gene to select for three types of rearrangements: insertion of a Ty1 retrotransposon, deletion of chromosomal sequences, and duplications. For this study, we concentrated on the deletion events producing a functional ATCase-coding allele.

The deletion events have nonhomologous recombination characteristics:
Rad52p is essential for the homologous recombination pathway and prevents NHEJ. VAN DYCK et al. 1999 suggested that the presence or the absence of Rad52p directs a DSB in a Rad52-dependent homologous recombination pathway or in a Ku-dependent nonhomologous end-joining pathway. We observed that the global ATCase reactivation rate in a ura2 15-30-72 rad52 strain was 4.4 times greater and that the deletion rate was significantly increased (12 times) compared to the ura2 15-30-72 RAD52 control strain. Thus, the deletion events we observed are RAD52 independent. These deletion events could thus be due to nonhomologous recombination. We even suggest that inactivation of the RAD52-dependent homologous recombination pathway favors the deletion mechanism in our system. The deletion events we obtained in our gain-of-function assay are RAD52 independent, whereas CHEN et al. 1998 concluded that interaction between short direct repeats requires Rad52p. The authors used a loss-of-function assay, selecting Can^r mutants by the loss of the CAN1 gene activity. To explain this difference, we suggest that the number of mutants analyzed at the level of mutation rate reached in their study could probably not allow the recovery of deletion events.

DSB repair by nonhomologous recombination can be by blunt end ligation, ligation joining through pairing of short complementary sequences, addition of untemplated nucleotides, and eventually insertion of filler DNA. The deletions we selected were formed only by blunt end ligation (for deletion class A) or joining through pairing of short complementary sequences (for the other deletion classes). All the deletion events we characterized in the rad52 background, and most of those obtained in the RAD52 background, show perfect microhomologies at their junctions.

The commonly proposed model to explain the presence of such short direct repeats at the ends of the deletions is a SSA-like nonhomologous recombination (for a scheme and a review see PAQUES and HABER 1999 ). Some authors have proposed another mechanism, replication fork slippage. This has been mostly described for deletions in Escherichia coli (ALBERTINI et al. 1982 ; SAVESON and LOVETT 1999 ), but also in humans by EFSTRATIADIS et al. 1980 and in yeast by GORDENIN et al. 1993 and TRAN et al. 1995 . To elucidate the deletion formation mechanism(s), it will be necessary to study the occurrence of deletion events in backgrounds known to affect NHEJ [such as hdf1 (yku70) or mre11] or replication fork slippage (pol3) and in mutants able to suppress microhomology-mediated and nonhomology-mediated gross chromosomal rearrangements, such as, respectively, rfa1 or rad27 and rad50, mre11, or xrs2 (CHEN and KOLODNER 1999 ).

Repeated sequences and chromosomal reorganization:
The repeated sequences involved here can allow for chromosomal rearrangements, such as deletion of DNA sequences up to 10 kb long. Even longer deletions implicating direct repeats could perhaps be obtained, but since our strains are haploid, the deletion event cannot encompass any essential gene nor destabilize the chromosomal structure. Actually, in our haploid strain, the proximity of the essential GCD14 gene to the URA2 gene limits the size of the largest deletion event (JAQUET and JAUNIAUX 1999 ). There should be a greater tolerance for genomic rearrangements in a diploid strain than in a haploid strain. Repeated sequences might then allow larger chromosomal deletions than in a haploid strain, as long as the deleted chromosome is stable.

Most of the deletions characterized in this work result in fusion proteins. The deleted alleles were mainly internal URA2 fusion between the GATase and CPSase (classes D–F, H, and I) or ATCase (class C) domains, but deletion class G is an in-frame fusion between the PBS2 and URA2 genes. This type of fusion gene could be a motor for evolution because of the change in the regulation of one of the partners of the fusion. The newly created gene may even encode a new function. Thus direct repeats can contribute greatly to genome reorganization.

	ACKNOWLEDGMENTS

We thank Alain Nicolas and Pierre Netter for helpful discussions of this work. We thank Jacques Belliard and Nicolas Pech for valuable help with statistical analyses. A.J.W. is supported by a grant from the French Ministère de l'Education Nationale, de la Recherche et de la Technologie. This work was supported in part by a CNRS grant (Programme Génome 97-0693). The yeast genetic department of the UPRES-A 7010 is a member of the Strasbourg Génopole.

Manuscript received November 2, 1999; Accepted for publication June 15, 2000.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ALBERTINI, A. M., M. HOFER, M. P. CALOS, and J. H. MILLER, 1982  On the formation of spontaneous deletions: the importance of short sequence homologies in the generation of large deletions. Cell 29:319-328[Medline].

BACH, M. L., 1984  Ty1-promoted expression of aspartate transcarbamylase in the yeast Saccharomyces cerevisiae.. Mol. Gen. Genet. 194:395-401[Medline].

BACH, M. L., F. ROELANTS, J. DE MONTIGNY, M. HUANG, and S. POTIER et al., 1995  Recovery of gene function by gene duplication in Saccharomyces cerevisiae.. Yeast 11:169-177[Medline].

BECKER, D. M. and L. GUARENTE, 1991  High-efficiency transformation of yeast by electroporation. Methods Enzymol. 194:182-187[Medline].

CARLE, G. F. and M. V. OLSON, 1985  An electrophoretic karyotype for yeast. Proc. Natl. Acad. Sci. USA 82:3756-3760[Abstract/Free Full Text].

CHEN, C. and R. D. KOLODNER, 1999  Gross chromosomal rearrangements in Saccharomyces cerevisiae replication and recombination defective mutants. Nat. Genet. 23:81-85[Medline].

CHEN, C., K. UMEZU, and R. D. KOLODNER, 1998  Chromosomal rearrangements occur in S. cerevisiae rfa1 mutator mutants due to mutagenic lesions processed by double-strand-break repair. Mol. Cell 2:9-22[Medline].

DEVEREUX, J., P. HAEBERLI, and O. SMITHIES, 1984  A comprehensive set of sequence analysis programs for the VAX. Nucleic Acids Res. 12:387-395.

EFSTRATIADIS, A., J. W. POSAKONY, T. MANIATIS, R. M. LAWN, and C. O'CONNELL et al., 1980  The structure and evolution of the human beta-globin gene family. Cell 21:653-668[Medline].

EXINGER, F. and F. LACROUTE, 1979  Genetic evidence for the creation of a reinitiation site by mutation inside the yeast ura2 gene. Mol. Gen. Genet. 173:109-113[Medline].

GORDENIN, D. A., K. S. LOBACHEV, N. P. DEGTYAREVA, A. L. MALKOVA, and E. PERKINS et al., 1993  Inverted DNA repeats: a source of eukaryotic genomic instability. Mol. Cell. Biol. 13:5315-5322[Abstract/Free Full Text].

HABER, J. E., 1992  Exploring the pathways of homologous recombination. Curr. Opin. Cell Biol. 4:401-412[Medline].

HOFFMAN, C. S. and F. WINSTON, 1987  A ten-minute DNA preparation from yeast efficiently releases autonomous plasmids for transformation of Escherichia coli.. Gene 2–3:267-272.

JAQUET, L. and J. C. JAUNIAUX, 1999  Disruption and basic functional analysis of five chromosome X novel ORFs of Saccharomyces cerevisiae reveals YJL125c as an essential gene for vegetative growth. Yeast 15:51-61[Medline].

KRAMER, K. M., J. A. BROCK, K. BLOOM, J. K. MOORE, and J. E. HABER, 1994  Two different types of double-strand breaks in Saccharomyces cerevisiae are repaired by similar RAD52-independent, nonhomologous recombination events. Mol. Cell. Biol. 14:1293-1301[Abstract/Free Full Text].

LEA, D. E. and C. A. COULSON, 1949  The distribution of numbers of mutants in bacterial populations. J. Genet. 49:264-285.

LIN, F. L., K. SPERLE, and N. STERNBERG, 1984  Model for homologous recombination during transfer of DNA into mouse L cells: role for DNA ends in the recombination process. Mol. Cell. Biol. 4:1020-1034[Abstract/Free Full Text].

MEUTH, M., 1989 Illegitimate recombination in mammalian cells, pp. 833–860 in Mobile DNA, edited by D. E. BERG and M. M. HOWE. American Society for Microbiology, Washington, DC.

MEZARD, C. and A. NICOLAS, 1994  Homologous, homeologous, and illegitimate repair of double-strand breaks during transformation of a wild-type strain and a rad52 mutant strain of Saccharomyces cerevisiae.. Mol. Cell. Biol. 4:1278-1292.

MEZARD, C., D. POMPON, and A. NICOLAS, 1992  Recombination between similar but not identical DNA sequences during yeast transformation occurs within short stretches of identity. Cell 70:659-670[Medline].

MOORE, J. K. and J. E. HABER, 1996a  Capture of retrotransposon DNA at the sites of chromosomal double-strand breaks. Nature 383:644-646[Medline].

MOORE, J. K. and J. E. HABER, 1996b  Cell cycle and genetic requirements of two pathways of nonhomologous end-joining repair of double-strand breaks in Saccharomyces cerevisiae.. Mol. Cell. Biol. 16:2164-2173[Abstract].

MORTIMER, R. K. and D. C. HAWTHORNE, 1966  Genetic mapping in Saccharomyces cerevisiae.. Genetics 53:165-173[Free Full Text].

PÂQUES, F. and J. E. HABER, 1999  Multiple pathways of recombination induced by double-strand breaks in Saccharomyces cerevisiae.. Microbiol. Mol. Biol. Rev. 2:349-404.

PFEIFFER, P. and W. VIELMETTER, 1988  Joining of nonhomologous DNA double strand breaks in vitro.. Nucleic Acids Res. 16:907-924[Abstract/Free Full Text].

RESNICK, M. A., 1976  The repair of double-strand breaks in DNA: a model involving recombination. J. Theor. Biol. 59:97-106[Medline].

ROELANTS, F., S. POTIER, J. L. SOUCIET, and J. DE MONTIGNY, 1995  Reactivation of the ATCase domain of the URA2 gene complex: a positive selection method for Ty insertions and chromosomal rearrangements in Saccharomyces cerevisiae.. Mol. Gen. Genet. 246:767-773[Medline].

ROTH, D. B. and J. H. WILSON, 1986  Nonhomologous recombination in mammalian cells: role for short sequence homologies in the joining reaction. Mol. Cell. Biol. 6:4295-4304[Abstract/Free Full Text].

ROTH, D. B., T. N. PORTER, and J. H. WILSON, 1985  Mechanisms of nonhomologous recombination in mammalian cells. Mol. Cell. Biol. 5:2599-2607[Abstract/Free Full Text].

SANGER, F., S. NICKLEN, and A. R. COULSON, 1977  DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74:5463-5467[Abstract/Free Full Text].

SAVESON, C. J. and S. T. LOVETT, 1999  Tandem repeat recombination induced by replication fork defects in Escherichia coli requires a novel factor, RadC. Genetics 152:5-13[Abstract/Free Full Text].

SOUCIET, J. L., M. NAGY, M. LE GOUAR, F. LACROUTE, and S. POTIER, 1989  Organization of the yeast URA2 gene: identification of a defective dihydroorotase-like domain in the multifunctional carbamoylphosphate synthetase-aspartate transcarbamylase complex. Gene 79:59-70[Medline].

SUGAWARA, N. and J. E. HABER, 1992  Characterization of double-strand break-induced recombination: homology requirements and single-stranded DNA formation. Mol. Cell. Biol. 12:563-575[Abstract/Free Full Text].

TENG, S. C., B. KIM, and A. GABRIEL, 1996  Retrotransposon reverse-transcriptase-mediated repair of chromosomal breaks. Nature 383:641-644[Medline].

THODE, S., A. SCHAFER, P. PFEIFFER, and W. VIELMETTER, 1990  A novel pathway of DNA end-to-end joining. Cell 60:921-928[Medline].

TLSTY, T. D., A. BRIOT, A. GUALBERTO, I. HALL, and S. HESS et al., 1995  Genomic instability and cancer. Mutat. Res. 337:1-7[Medline].

TRAN, H. T., N. P. DEGTYAREVA, N. N. KOLOTEVA, A. SUGINO, and H. MASUMOTO et al., 1995  Replication slippage between distant short repeats in Saccharomyces cerevisiae depends on the direction of replication and the RAD50 and RAD52 genes. Mol. Cell. Biol. 15:5607-5617[Abstract].

TSUKAMOTO, Y. and H. IKEDA, 1998  Double-strand break repair mediated by DNA end-joining. Genes Cells 3:135-144[Abstract].

VAN DYCK, E., A. Z. STASIAK, A. STASIAK, and S. C. WEST, 1999  Binding of double-strand breaks in DNA by human Rad52 protein. Nature 398:728-731[Medline].

WACH, A., A. BRACHAT, R. POHLMANN, and P. PHILIPPSEN, 1994  New heterologous modules for classical or PCR-based gene disruptions in Saccharomyces cerevisiae.. Yeast 10:1793-1808[Medline].

This article has been cited by other articles:

				M. Fidalgo, R. R. Barrales, J. I. Ibeas, and J. Jimenez Adaptive evolution by mutations in the FLO11 gene PNAS, July 25, 2006; 103(30): 11228 - 11233. [Abstract] [Full Text] [PDF]

				J. Schacherer, J. de Montigny, A. Welcker, J.-L. Souciet, and S. Potier Duplication processes in Saccharomyces cerevisiae haploid strains Nucleic Acids Res., November 3, 2005; 33(19): 6319 - 6326. [Abstract] [Full Text] [PDF]

				J. H. Carlson, S. Hughes, D. Hogan, G. Cieplak, D. E. Sturdevant, G. McClarty, H. D. Caldwell, and R. J. Belland Polymorphisms in the Chlamydia trachomatis Cytotoxin Locus Associated with Ocular and Genital Isolates Infect. Immun., December 1, 2004; 72(12): 7063 - 7072. [Abstract] [Full Text] [PDF]

				V. Leh-Louis, B. Wirth, S. Potier, J.-L. Souciet, and L. Despons Expansion and Contraction of the DUP240 Multigene Family in Saccharomyces cerevisiae Populations Genetics, August 1, 2004; 167(4): 1611 - 1619. [Abstract] [Full Text] [PDF]

				J. Schacherer, Y. Tourrette, J.-L. Souciet, S. Potier, and J. de Montigny Recovery of a Function Involving Gene Duplication by Retroposition in Saccharomyces cerevisiae Genome Res., July 1, 2004; 14(7): 1291 - 1297. [Abstract] [Full Text] [PDF]

				X. Yu and A. Gabriel Ku-Dependent and Ku-Independent End-Joining Pathways Lead to Chromosomal Rearrangements During Double-Strand Break Repair in Saccharomyces cerevisiae Genetics, March 1, 2003; 163(3): 843 - 856. [Abstract] [Full Text] [PDF]