Virulence and Karyotype Analyses of rad52 Mutants of Candida albicans: Regeneration of a Truncated Chromosome of a Reintegrant Strain (rad52
http://www.100md.com
感染与免疫杂志 2005年第12期
Georgetown University Medical Center, Department of Microbiology & Immunology, Washington, D.C.
Department of Microbiology, Universdad de Extremadura, Badajoz, Spain
Department of Immunology, Microbiology & Parasitology, Universidad del Pais Vasco, Leioia, Vizcaya, Spain
ABSTRACT
The virulence of Candida albicans mutants lacking one or both copies of RAD52, a gene involved in homologous recombination (HR), was evaluated in a murine model of hematogenously disseminated candidiasis. In this study, the virulence of the rad52 mutant was dependent upon the inoculum concentration. Mice survived at a cell inoculum of 1 x 106, but there was a decrease in survival time at dosages of 1.5 x 106 and especially at 3 x 106 cells per animal. The heterozygote RAD52/rad52 behaved like wild type, whereas a reintegrant strain was intermediate in its ability to cause death compared to these strains and to the avirulent rad52/rad52 null at inocula of 1 x 106 and 1.5 x 106 cells. A double mutant, lig4/lig4/rad52/rad52, was avirulent at all inocula used. PCR analysis of the RAD52 and/or LIG4 loci showed that all strains recovered from animals matched the genotype of the inoculated strains. Analysis of the electrophoretical karyotypes indicated that the inoculated, reintegrant strain carried a large deletion in one copy of chromosome 6 (the shortest homologue, or Chr6b). Interestingly, truncated Chr6b was regenerated in all the strains recovered from moribund animals using the homologue as a template. Further, regeneration of Chr6b was paralleled by an increase in virulence that was still lower than that of wild type, likely because of the persistent loss of heterozygosity in the regenerated region. Overall, our results indicate that systemic candidiasis can develop in the absence of HR, but simultaneous elimination of both recombination pathways, HR and nonhomologous end-joining, suppresses virulence even at very high inocula.
INTRODUCTION
Candida albicans exhibits considerable phenotypic variation both in vitro and in the patient. While the belief was that C. albicans lacked a sexual cycle and therefore was unable to create variability through genome recombination, there is ample evidence now that mating type-like (MTL) loci are present in the genome and that compatible strains can mate (13, 14, 25). However, the low frequency at which mating seems to occur in nature (26), the failure to identify a meiotic process, and evidence that the tetraploid progeny of mating are generally thought to return to a near-diploid state through chromosome loss (4) suggest that mating is not the usual way through which C. albicans creates variability (26). In the absence of a meiosis-driven recombination of new alleles, two other mechanisms may account for the variability in this organism. The first mechanism includes gene regulation, generally accomplished through signal transduction pathways following perception by cells of an environmental signal, as has been observed in morphological transition and adaptation to stress growth conditions (6, 34). A second mechanism that accounts for variability is spontaneous or induced genetic changes, including chromosomal alterations, mutations, and loss of heterozygosity (LOH), as has been reported by a number of investigators in laboratory strains as well as from clinical isolates (for recent reviews see references 19, 24, and 32); all of these genetic events are facilitated by the diploid genome of C. albicans (28). Chromosomal changes, as identified by electrokaryotyping methods, occur as translocations as well as through aneuploidy (19, 32, 34). An example of a chromosomal change as a result of aneuploidy which results in an altered phenotype is the loss of chromosome 5 and the growth of that strain in the presence of sorbose as the only carbon source (15, 16). Alterations in chromosome copy number have also been associated with resistance to fluconazole in vitro; these strains are either monosomic for chromosome 4 or trisomic for chromosome 3 (30). Karyotypic rearrangements, including those that result in homozygosity at the MTL locus, are also well documented in clinical strains (19). The mechanism(s) by which strains of C. albicans develop aneuploidy is not understood. However, it has been suggested that for Saccharomyces cerevisiae, an increase in DNA double-strand breaks (DSB) occurs but that normal cell surveillance and DNA repair are bypassed (12, 17). A second possibility is nondisjunction at mitosis. In addition to chromosomal rearrangements, LOH occurs at measurable rates in vitro as well as in vivo (9, 10). LOH may result from chromosome loss, and in fact this is the primary mechanism through which MTL homozygosis occurs in C. albicans (37). However, LOH may also result from mutations or mitotic recombination. It has been reported that in vivo, LOH may occur either by itself or in combination with chromosomal alterations, resulting in the appearance of new phenotypic traits (10).
In S. cerevisiae and apparently in C. albicans, DSB are preferentially repaired by homologous recombination (HR), in part a function of the RAD52 gene product as well as a complex of other proteins (7, 29). In the absence of Rad52, both organisms may still repair DSB using a second pathway that uses a nonhomologous (illegitimate) end-joining (NHEJ) mechanism (1, 7, 23, 29). NHEJ requires a completely different set of proteins, including the homologues of the mammalian Ku proteins (yKu70 and yKu80) as well as Lig4 and its associated proteins Lif1 and Lif2. We previously identified a LIG4 homologue of C. albicans and have constructed mutants in this gene. The LIG4 of C. albicans is involved in NHEJ and, further, lig4 mutants are partially defective in morphogenesis and are avirulent in a murine model of hematogenously disseminated candidiasis (1-3). HR also drives mitotic recombination, including local gene conversion (no crossing over) as well as mitotic crossing over; these events may result in LOH at a single (or a few) locus or along long tracts of the chromosome, respectively. In order to determine the contribution of the HR pathway in DNA repair and mitotic recombination, we have recently reported several phenotypes of the C. albicans rad52 mutants and demonstrated that Rad52p is (i) critical to the repair of DNA damage caused by either UV light or the radiomimetic compound methylmethane sulfonate; (ii) absolutely required for the integration of linear DNA with long flanking sequence homology, i.e., critical for HR; and (iii) critical in maintaining the length of telomeres (7). Recent results have also indicated a role for Rad52p in the maintenance of the genomic stability in C. albicans (G. Larriba et al., ASM Conf. Candida Candidiasis, Austin, Tex., abstr. 83B, 2004; unpublished results).
Because of the role of Rad52p in repairing DSB, mitotic recombination, and in maintaining genome stability, we have examined the virulence of rad52 mutants in a murine model of hematogenously disseminated candidiasis. Our reasons for this study include not only an analysis of the role of homologous recombination and chromosome stability in virulence but also categorization of the changes that occur in the C. albicans chromosomes during disease development in both the presence and the absence of HR and NHEJ.
MATERIALS AND METHODS
Strains. The strains used in this study are listed in Table 1; their construction was described previously (1, 7).
Growth rates. Liquid cultures were started by inoculation of a 250-ml flask containing 50 ml of YPD (1% yeast extract, 2% glucose, 2% peptone) or SC (synthetic complete) with a suspension of cells previously grown in the same medium. For each medium, cultures were adjusted to a final optical density at 600 nm of 0.05 using a spectrophotometer and incubated at 30°C. Samples were taken at the indicated times, and the optical density at 600 nm was determined in a spectrophotometer. Appropriate dilutions of each sample were plated in duplicate on YPD plates to determine the number of CFU.
OMP decarboxylase assays. We followed previously described procedures for orotidine 5'-monophosphate (OMP) decarboxylase (OMPase) assays (5, 20). All strains were grown overnight at 30°C, washed, and then subcultured in 50 ml of YPD and grown at 30°C until the cultures reached mid-log phase (approximately 6 h). Cultures were centrifuged, washed with 1 mM Tris-HCl buffer (pH 7.5) containing 1 mM dithiothreitol (DTT), and suspended in 2 ml of phosphate buffer (pH 7.5). Cells were transferred to microcentrifuge tubes and broken with glass beads by vortexing for 2 min. The homogenates were transferred to other tubes and centrifuged for 10 min at 13,000 rpm to remove particulate matter. Supernatants were collected and stored in an ice bath, and the protein content was determined by the Bradford method using bovine serum albumin as a standard. To perform OMP decarboxylase assays, reaction mixtures contained 500 μl phosphate buffer (pH 6.0), 10 μl 0.1 M -mercaptoethanol, 50 μl 1 mM OMP, 470 μl cell extract in 1-ml quartz cuvettes. Enzyme activity was measured as a decrease in absorbance per min at 285 nm for 25 min using a water blank as a control. The OMP decarboxylase activity was calculated by Beer's law using a molar extinction coefficient of 1.65 x 103 cm–1 M–1. One unit of enzyme activity was defined as the amount of enzyme required for conversion of 1 μM of OMP to UMP per min (μM/mg of protein/min).
Nucleic acid extraction and analysis. Standard techniques were routinely used for DNA manipulation. Genotyping was carried out by PCR analysis of the RAD52 and/or LIG4 locus and, when indicated, by Southern hybridization. For PCR analysis of the RAD52 locus, we used the previously described pair of oligonucleotides RV1 and RV2, which flank the disrupted region of the RAD52 open reading frame (ORF) (7), as well as another primer set, URA3.1 (5'-GGTATAGAAATGCTGCTTGG-3') and URA3.2 (5'-CGAATCGGCACTACAGC-3'), which are complementary to segments near the 5' and 3' ends of the URA3 ORF, respectively. Primer sets RV1 and RV2 amplify 1.5- and 1.4-kb fragments from wild-type RAD52 and RAD52::hisG constructs, respectively, but not the RAD52::hisG-URA3-hisG construct. Identification of the latter was carried out in two PCRs; the first one used RV2-URA3.1 to yield a 1.8-kb band and the second one used RV1-URA3.2 to yield a 1.3-kb fragment. PCR was performed as described elsewhere (31) except that the annealing temperature was always 62°C and, for the amplification RV1-RV2, the concentration of deoxynucleoside triphosphate was 2 mM. For PCR analysis of the LIG4 locus, we used the oligonucleotides LIG4.3 (5'-GTATACCAGAAGTAAGATGGC-3') and LIG4.4 (5'-CAGGGTGCCTGCTCGAGTGTC-3'), which flank the LIG4 ORF (1), and an annealing temperature of 60°C.
Karyotype analysis. Chromosomal analysis of the same strains inoculated and recovered from mice was accomplished as described previously (1, 7). A 0.1-ml sample of exponentially growing cultures of C. albicans was used to inoculate 10 ml of YPD. Cultures were shaken for 48 h at 30°C, collected by centrifugation, washed twice with 50 mM EDTA, pH 8, and suspended in 1 ml of CPES buffer (40 mM citric acid, 120 mM sodium phosphate, 20 mM EDTA, pH 8, 1.2 M sorbitol, and 5 mM DTT, with 0.2 mg Zymolyase 20,000A). To this suspension, 1 ml of CPE buffer (minus sorbitol and DTT) containing 1% low-melting agarose was added, mixed gently, transferred to a sample mold, and stored at –20°C. The solid agar plugs were then transferred to test tubes containing 6 ml of CPE buffer and stored at 30°C for 4 h. The CPE buffer was replaced with TESP buffer (1 M Tris-HCl, 0.5 M EDTA, 2% sodium dodecyl sulfate [SDS], 1 mg ml–1 proteinase K), and the gels were incubated overnight at 50°C and then washed with Tris-EDTA three times at 50°C and six times at room temperature. Gels were stored at 4°C in 50 mM EDTA, pH 8. Two different protocols were then used to separate chromosomes. In the first protocol, all chromosomes were separated except for both homologues of chromosomes 6 and 7. The gel samples were electrophoresed in 0.6% agarose for 24 h at 80 V with a 120- to 300-s linear ramp and then for 48 h at 80 V with a 420- to 900-s linear ramp in a rotating gel electrophoresis apparatus (Rotaphor; Biometra). A second protocol was also used to separate both homologues of the smaller chromosomes, 6 and 7 (21, 22). In this case, the gel samples were run in 1% agarose at 180 V with a linear ramp of 60 to 120 s, 120° included angle for 48 h, and then at 120 V with a 300- to 420-s linear ramp, 120°C included angle, for 48 h in the same apparatus.
Southern analysis of karyotypes was carried out as described previously (7), using a marker of the SfiI 6C fragment (COX12) as probe. Hybridization bands were visualized using a Molecular Imager (Bio-Rad Laboratories).
Animal experiments. C. albicans strains were grown in YPD medium at 30°C to stationary phase. Cells were harvested by centrifugation, washed twice in calcium- and magnesium-free phosphate-buffered saline (PBS; BioSource International), and suspended to a density of 5 x 106 CFU per ml on the basis of hemocytometer counts prior to use. For the animal experiments, we followed previously published methods (1). Groups of 10 male BALB/c mice (18 to 20 g each; Harlan Laboratories) were injected intravenously via the lateral tail vein with 0.2 ml (106 CFU) of each strain listed in Table 1. In addition, with strain TCR2.2 (rad52/rad52), we also infected mice similarly but with 0.3 ml of a cell suspension (1.5 x 106 CFU). We also determined the virulence of the rad52 (TCR2.2), lig4-deleted mutant (CEA2), and lig4/rad52 double mutant (EAT2) strains, each used at an inoculum concentration of 3 x 106 cells per mouse. All mice were observed twice daily for signs of morbidity and, if moribund, animals were euthanized by CO2 inhalation. Concomitantly, each C. albicans strain was used to inoculate 15 additional mice. Five mice from each group were sacrificed by CO2 inhalation at 24, 48, and 72 h postinfection, and the kidneys from mice infected with each strain were removed, weighed, and homogenized in 5.0 ml of PBS. Homogenates were diluted in PBS, and aliquots were plated on YPD agar supplemented with 50 μg of streptomycin per ml to prevent bacterial growth. Plates were incubated at 30°C for 48 h, and the numbers of CFU per gram of tissue were then quantitated. Also at 48 h, kidneys from mice infected with CAF2-1, TCR2.2, CEA2, or EAT2 at a dose of 3 x 106 CFU were removed, fixed in 10% formalin, and prepared for histological examination using the periodic acid-Schiff stain. To verified genetic constructs, strains recovered from animals were frozen and subjected to PCR and karyotype analysis as described elsewhere.
Statistical analysis. In order to determine strain differences in virulence and tissue counts by statistical analysis, we performed a Kaplan-Meier test of significance by a log-rank test.
RESULTS
Growth rates and OMPase activities of strains. As shown in Fig. 1, when incubated in YPD at 30°C, the rad52 strain grew significantly slower than the heterozygote TCR1, whereas the growth rate of the reintegrant strain TCR3.2.1 was similar to that of other Rad52+ strains. Disruption of lig4 in the rad52 background did not modify the growth rate of the latter. On the basis of the respective growth curves, we calculated generation times of 140 min for both rad52 null strains (TCR2.2 [rad52/rad52] and EAT2 [lig4/lig4 rad52/rad52]) and 75 to 80 min for each of the Rad52+ strains, including CAF2-1 (not shown), heterozygote TCR1, revertant TCR3.2.1, and CEA2 (lig4), which was reported by us previously (1). Additionally, because of the URA3 positional effect that can occur in mutants constructed using the Urablaster method, we measured OMPase activity in all strains. Using C. albicans SC5314 as a reference, we found the OMPase activities for all strains were as follows: CAF2-1, 63.3%; TCR1, 53.3%; TCR2.2, 61.3%; TCR3.2.1; 49.3%; EAT2, 49.6%. These data suggest that, when located at the RAD52 locus, URA3 is expressed at nearly the same level as in its original location; furthermore, the OMPase activity of TCR2.2 (61.3%) was higher than that of TCR3.2.1 (49.3%), yet TCR2.2 was less virulent (see below). Interestingly, the OMPase activity of CEA2 was only 21.4% of wild-type cells, suggesting that this value probably influenced our previous results on the lack of virulence of the null strain lig4/lig4 (see below).
Virulence of rad52 strains. We compared the virulence of the wild type (CAF2-1), the homozygous (TCR2.2 [rad52/rad52]) and heterozygous (TCR1 [RAD52/rad52]) mutants, and a reintegrant strain (TCR3.2.1 [rad52/rad52::RAD52]), as well as a null lig4 strain (CEA2 [lig4/lig4]) and a double null mutant lig4 rad52 (EAT2 [lig4/lig4 rad52/rad52]) (Table 1) according to methods previously described (1).
The survival of mice infected with 1 x 106 to 1.5 x 106 CFU of each strain of C. albicans described above is represented in Fig. 2A. Wild-type C. albicans (CAF2-1) at a dosage of 1 x 106 CFU killed all animals within 3 days postinfection, similar to the heterozygous strain (RAD52/rad52 [TCR1]; P = 0.45), whereas the gene-rescued strain (rad52/rad52::RAD52 [TCR3.2.1]) displayed a significantly lower killing ability than the heterozygote (P < 0.004) in spite of the fact that each strain exhibited a similar growth rate and OMPase activity. Several reasons may account for this behavior (see below). The homozygous null strain (rad52/rad52 [TCR2.2]) was avirulent at a dose of 1 x 106, similar to strains CEA2 and EAT2 (Fig. 2A). However, if the concentration of strain TCR2.2 used to infect mice was increased to 1.5 x 106 cells, then 30% of the mice died by 28 days postinfection (Fig. 2A). We also compared the survival of mice infected with 3 x 106 cells of strains TCR2.2, CEA2 (lig4/lig4), and EAT2 (lig4/lig4/rad52/rad52) (Fig. 2B). We found that the virulence of the rad52/rad52 and lig4/lig4 mutants is similar (increased) at this cell concentration, while the double mutant (EAT2) remains avirulent. Thus, virulence of the rad52/rad52 mutant is dependent upon the inoculum concentration, while the double mutant remains avirulent even though the growth rates of both strains are similar. The CFU/g of kidney was determined for each strain at 24 to 72 h postinfection (Fig. 3), except for animals infected with CAF2-1, since all were moribund by 72 h. The colony counts of strains CAF2-1, TCR1 (RAD52/rad52), and TCR3.2.1 (rad52/rad52::RAD52) remained high (5.5 x 106 to 6.23 x 106) during the first 48 to 72 h postinfection. However, for strain TCR2.2 (rad52/rad52), at the same inoculum dose (1 x 106 CFU), the highest CFU/g kidney was always lower (4.65 x 106) (and even dropped by almost 1 log by 72 h), and the same was true for the double mutant EAT2 (lig4/lig4 rad52/rad52) (P < 0.0001, TCR2.2 and EAT2 versus all strains). In order to obtain and maintain a CFU/g in kidney in the range of 5.7 x 106 (i.e., similar to that reached with the Rad52+ strains inoculated with the lower dose of 1 x 106 CFU), it was necessary to raise the inoculum dose of both rad52/rad52 null strains to 3 x 106 CFU. Interestingly, only then were the single mutants (CEA2 and TCR2.2) virulent, whereas the double mutant remained avirulent (Fig. 2B). As described before (1), the lig4 null strain, which grew as fast as the parental CAF2-1, was avirulent at an inoculum dose of 106. We show here that the lig4/lig4 null strain was virulent when the dosage was increased to 3 x 106, although, in support of the role of Lig4p in virulence, even at this higher dose, the lig4/lig4 strain was less virulent than the parental CAF2-1 at 1 x 106. The fact that the double mutant rad52/rad52/lig4/lig4, which exhibits a growth rate similar to the rad52/rad52 mutant, is avirulent even at the highest dose (3 x 106) not only confirms a role for Lig4 in virulence but also indicates that killing by rad52/rad52 mutant cells is not a direct effect of the large amount of cells inoculated. The histopathology of kidneys from mice infected for 48 h is shown for selected strains in Fig. 4. The presence of filamentous growth in the kidney was observed, but the amount of growth varied among strains. Interestingly, animals infected with the double mutant lig4/lig4/rad52/rad52 survived the course of the experiment, but the organism could still be recovered from tissue in high levels and could be seen in the histological sections of infected tissue (Fig. 3 and 4).
Genotyping of strains recovered from mice. To verify that the strains recovered from the infected animals were derived from those used to infect mice, we first analyzed the RAD52 and/or LIG4 loci using PCR. For that purpose, we selected strains recovered from the kidneys of animals infected for 48 h that had been inoculated with 1 x 106 cells of CAF2-1, TCR1(RAD52/rad52), TCR2.2 (rad52/rad52), TCR3.2.1 (rad52/rad52::RAD52), CEA2 (lig4/lig4), and EAT2 (lig4/lig4 rad52/rad52). The hisG-URA3-hisG-interrupted allele was found in the TCR1-inoculated mice (Fig. 5A, lanes 1 and 6) and all recovered strains of TCR1 (one of three isolates is shown) (Fig. 5A, lanes 2 and 7). TCR1 strains also carried the wild-type RAD52 allele (not shown).
PCR analysis of 10 isolates recovered from mice that were inoculated with TCR2.2 detected the hisG-URA3-hisG-disrupted RAD52 allele, indicating that the recovered strains (two examples are shown in Fig. 5A, lanes 4 and 5) were derived from the inoculation strain (Fig. 5A, lane 3; compare lane 8 with lanes 9 and 10). As expected, all TCR2.2 strains carried the hisG-disrupted allele (Fig. 5B, lanes 3, 4, and 5 for the inoculated and same two recovered strains, respectively). Parental strain CAI4 (Fig. 5B, lane 1) and strain TCR1.1 (Fig. 5B, lane 6) showed the expected PCR results.
A previous analysis of the RAD52 locus in the revertant TCR3.2.1 suggested that this strain had lost the RAD52::hisG-disrupted allele (7). These analyses were hindered by the presence of a nonspecific band in the PCR products (7), as well as some genetic instability detected in this strain (see below). However, under the new PCR conditions that eliminate the contaminant, we found that TCR3.2.1 carries both a wild-type (1.5-kb band) and a hisG-disrupted (1.4-kb band) RAD52 allele (Fig. 5B, lane 7). Both alleles were also present in five strains recovered from the animals inoculated with TCR3.2.1 (two examples are shown in Fig. 5B, lanes 8 and 9).
As shown in Fig. 5C for the inoculated and two recovered strains of CEA2 (lig4::hisG/lig4::hisG-URA3-hisG), each carries both the lig4::hisG-URA3-hisG allele (indicated by the presence of a 1.8-kb band with oligonucleotides LIG4.3-URA3.1 and a 1.3kb band with the oligonucleotides LIG.4.4-URA3.2) and the lig4::hisG allele(indicated by the presence of the 1.4-kb band with oligonucleotides LIG4.3-LIG4.4) (1). Figure 5D shows that three recovered EAT2 strains yielded the same PCR products as the inoculated EAT2 (lig4rad52) (lig4::hisG/lig4::hisG rad52::hisG/rad52::hisG-URA3-hisG), which correspond to the rad52::hisG-URA3-hisG allele (1.8-kb and 1.3-kb bands with oligonucleotides RV2-URA3.1 and RV1-URA3.2, respectively), as well as the rad52::hisG (1.4-kb band with oligonucleotides RV1-RV2) and the lig4::hisG (1.4-kb band with oligonucleotides LIG4.3-LIG4.4) alleles. Overall, our results indicate that the recovered strains are derived from the inoculated strains and not from an exogenous C. albicans contaminant.
Karyotypic analysis of the recovered strains. In order to investigate the occurrence of genomic alterations that could have occurred during the infection (in vivo), we compared chromosomal karyotypes of all strains. Figure 6A shows the standard chromosomal karyotypes of inoculated (one from each strain; Fig. 6A, lanes 1, 3, 7, and 12) and representative recovered strains of CAF2-1 (Fig. 6, lane 2), TCR1 (Fig. 6, lanes 4 to 6), TCR2.2 (3 of 10 isolates analyzed; Fig. 6, lanes 13 to 15), and TCR3.2.1 (4 of 5 analyzed; Fig. 6, lanes 8 to 11). Analysis of the shorter chromosomes of the same strains is shown in Fig. 6B. A shift in the mobility of chromosome R (ChrR) was observed in all Rad52+ strains and, in particular, strains CAF2-1 and one TCR1 strain recovered from infected animals (compare lanes 1 and 3 with lanes 2 and 4, respectively). This observation suggests the existence of frequent recombination events between the rRNA gene cistrons during the course of the infection. This is not surprising, since chromosome R rearrangements are well documented in wild-type (Rad52+) strains of C. albicans (references 20, 26, and 33 and references therein). Interestingly, changes in the mobility of ChrR were not as obvious in the case of the rad52 null strains (TCR2.2 and EAT2) (Fig. 6A, lanes 12 to 15 and lanes 16 to 19, respectively), suggesting that in wild-type (Rad52+) strains this event likely involves HR. In the absence of Rad52, other genetic events that do not require the Rad52 protein, such as deletions mediated by single strand annealing (29), could result in the loss of some rRNA cistrons with the concomitant shortening of ChrR, as observed in one TCR2.2 strain recovered from the animals (Fig. 6A, lane 15). Both homologues of ChrR were observed with the EAT2-inoculated strain (Fig. 6A, lane 16). Two recovered strains yielded the same pattern (lanes 17 and 18), but in the third one, either both copies of ChrR comigrated with the shorter homologue or the largest one was lost (lane 19). In agreement with our observations, recent results indicate that rRNA repeats cannot expand in the absence of Rad52 (18). Further analysis of the sizes of the rRNA cistrons in inoculated and recovered rad52 mutants appears necessary before establishing definitive conclusions. Apart from the changes involving ChrR, we did not observe gross chromosomal rearrangements in the rad52 null strains (TCR2.2 and EAT2) (Fig. 6A). This observation is in strong contrast with the genomic instability that characterizes these mutants in S. cerevisiae (17, 29, 38). Our explanation is that strains carrying gross chromosomal rearrangements derived from deficiencies in HR are less fit to survive in the animal and are eliminated. Still, it seems that the presence of the rad52 mutation equalizes the size of both homologues of chromosome 6, and this characteristic was maintained in the recovered strains (Fig. 6B, lanes 12 and 13 to 15). By contrast, the same mutation exacerbated the difference in the size of both homologues of chromosome 7, and this feature was also maintained in the recovered strains. Whether these variations in the size of homologues are specific for the rad52 mutation remain to be investigated.
As expected from previous results (7), the reintegrant TCR3.2.1 showed a supernumerary chromosome migrating above Chr5 (SN1) (Fig. 6A and B, lane 7). This extra band was conserved in all the rescued strains. In addition, the improved resolution of our electrophoresis karyotypes allowed us to detect two additional features of the revertant strain TCR3.2.1, the apparent presence of a single copy of Chr6 and an additional extra band running ahead of Chr7 (SN2) (Fig. 6A and B, lane 7). Interestingly, the four recovered strains from animals inoculated with 1 x 106 cells of TCR 3.2.1 had regained both copies of Chr6 and concomitantly had lost the extra band (Fig. 6A and B, lanes 8 to 11). Furthermore, the electrokaryotypes obtained under conditions that separated both homologues of chromosomes 6 and 7 (Fig. 6B) allowed us to identify the single copy of Chr6 present in the inoculated revertant with the largest homologue of that chromosome (from now on, Chr6a) (lane 7), whereas four recovered strains carried both homologues (lanes 8 to 11). Other reports have described the differences in size between the C. albicans homologues that are due to the number of repeats in the major repeat sequence (references 19, 22, and 24 and references therein). Southern blot analysis using a probe of chromosome 6 (COX12), which is located in the SfiI fragment 6C, labeled Chr6a and SN2 in the inoculated TCR3.2.1 and both homologues of Chr6 in the recovered strains (Fig. 6C). These results unambiguously demonstrate that SN2 was a truncated Chr6b that was regenerated in the host to its original size, most probably using the homologue Chr6a as a template. We favor this possibility, since if cells were able to use randomly any chromosome as a template, we should expect to find a full spectrum of sizes for the regenerated Chr6b. The molecular mechanisms responsible for the truncation/deletion as well as for the regeneration of this chromosome are being investigated.
In view of these results and taking into account that the heterozygote TCR1 was significantly more virulent than the reintegrant TCR3.2.1 (Fig. 2A), we were interested in comparing the virulence of the reintegrant strain recovered from mice to that of both the original reintegrant and the heterozygote used to infect animals. Therefore, the same murine model was used with strains CAF2-1, TCR1, TCR3.2.1 (original), and TCR3.2.1 (recovered) at an inoculum dosage of 1 x 106 yeast cells per mouse (Fig. 7). We show that the recovered, reintegrant strain is statistically more virulent than the original reintegrant (P < 0.0001, TCR3.2.1r1 versus TCR3.2.1). However, the recovered reintegrant still did not reach the virulence of the heterozygote.
DISCUSSION
We report that virulence of DNA repair mutants is dependent upon inoculum dosage and that the double mutant (lig4/lig4/rad52/rad52) is avirulent even at concentrations of 3 x106 yeast cells per mouse. We also observed that the gene-reconstituted mutant (rad52/RAD52) was not as virulent as a heterozygote; several possibilities could account for this observation. First, truncated Chr6 may be lost in some cells and, as a consequence, these cells likely become less viable or even die after a few generations. This would not be surprising, since in S. cerevisiae many chromosomes that lose a telomere are themselves lost (11, 27, 33). Accordingly, the real dose injected into the animals could be lower than the theoretical one. However, the CFU/g of kidney for each strain did not differ significantly during the first 72 h (Fig. 3), in spite of the fact that by 96 h, almost 100% of mice infected with TCR3.2.1 were alive while almost all mice infected with TCR1 were dead (Fig. 2A). This observation does not weaken the first possibility, since perhaps many cells of the original TCR3.2.1 grow well on agar plates but not in the animal because of the lack of Chr6b. The lag period in killing observed with the original TCR3.2.1 could be required for the regeneration of Chr6b in some cells and the concomitant increase in virulence. Loss of Chr6b and duplication of Chr6a was not detected in any of the reintegrant strains recovered from the animals. If this happened, it is likely that the resultant strains would not be selected in the animal. A second explanation that could also account for the lower virulence of TCR3.2.1 is the decrease in gene dosage that affects all the ORFs included in the deleted region, which also implies LOH in that fragment. Our results have demonstrated unambiguously that those cells that repair the truncated Chr6b are selected in the animal. This is not surprising either. In S. cerevisiae, a centric chromosomal fragment may be repaired after being segregated through several generations (11, 27, 33). Assuming that TCR3.2.1 cells use the Chr6a as a template during the repair process, they should regain the original gene dosage but still maintain the LOH in the deleted region. In order to investigate how these alterations of Chr6b may affect virulence, we compared the virulence of one strain recovered from animals (TCR3.2.1r1) with both the original reintegrant TCR.3.2.1 and the heterozygote TCR1. Interestingly, the recovered reintegrant strain was statistically more virulent than the original reintegrant, a phenotype that should be associated with repair of chromosome 6b (Fig. 7). However, it was still less virulent than the heterozygote, likely because repair caused LOH in all markers located in the regenerated region.
The conclusions from our study include the following. First, whereas there is no doubt that C. albicans undergoes genetic alterations in the host (10, 11, 22, 31, 36), the importance of these events in disease has not been determined. Our results indicate that the virulence of rad52 strains is dose dependent, being avirulent for mice at a dose of 106 cells. However, this avirulence could be attributed to its longer generation time. The same null strain was still able to kill at doses of 1.5 x 106 and especially 3 x 106, indicating that at higher inoculum densities, the generation time is less critical to the development of candidiasis, at least in this system. Second, our results indicate that, under these conditions, severe C. albicans infection may occur in the absence of mitotic recombination (HR). Thus, at the same inoculum density (3 x 106), the double mutant lig4rad52 remained avirulent even though it displayed the same doubling time as the rad52 single mutant. Since both mutants yielded a similar OMPase activity, this observation agrees with our previous report (confirmed in the present work) on the relevance of NHEJ to C. albicans virulence, in spite of the fact that the lig4 null strain used (CEA2) displayed a significantly lower OMPase activity than CAF2. Third, repair of a truncated chromosome in vitro has been shown to occur in the derivatives of strain RM100#13, but this occurred only occasionally, since most of its genetically manipulated derivatives, including BWP17, maintained the truncation (35). Here, we have shown for the first time the ability of C. albicans to repair a truncated chromosome in the animal, giving rise to a clone of cells that was selected during the course of the infection. Virulence of this strain was enhanced following the repair of Chr6b but did not reach that of wild type or the heterozygous strains, probably because of LOH in the truncated region. These results and the fact that we did not detect any recovered strain with a duplicated Chr6a emphasize the importance of the heterozygosity of some markers in the virulence of C. albicans.
ACKNOWLEDGMENTS
This study was supported by a Public Health Service grant, NIH-NIAID 1 R01 AI51949 to G.L. and R.C., and grant 2PR03A044 from Junta de Extremadura to E.A.
We thank Bebe Magee for providing the marker of the SfiI fragments 6C. We also thank Belen Hermosa and Leocadia Franco for their technical support.
REFERENCES
1. Andaluz, E., R. Calderone, G. Reyes, and G. Larriba. 2001. Phenotypic analysis and virulence of Candida albicans LIG4 mutants. Infect. Immun. 69:137-147.
2. Andaluz, E., T. Ciudad, J. Rubio-Coque, R. Calderone, and G. Larriba.1999 . Cell cycle regulation of a DNA ligase-encoding gene (CaLIG4) from Candida albicans. Yeast 15:1199-1210.
3. Andaluz, E., G. Larriba, and R. Calderone. 1996. A Candida albicans gene encoding a DNA ligase. Yeast 12:893-898.
4. Bennett, R. J., and A. D. Johnson. 2003. Completion of a parasexual cycle in Candida albicans by induced chromosome loss in tetraploid strains. EMBO J. 22:2505-2515.
5. Brand, A., D. M. MacCallum, A. J. P. Brown, N. A. R. Gow, and F. Odds. 2004. Ectopic expression of URA3 can influence the virulence phenotypes and proteome of Candida albicans but can be overcome by targeted reintegration of URA3 at the RPS10 locus. Eukaryot. Cell 3:900-909.
6. Brown, A. J. P. 2002. Morphogenetic signaling pathways in Candida albicans, p.95 -106. In R. Calderone (ed.), Candida and Candidiasis. ASM Press, Inc., Washington, D.C.
7. Ciudad, T., E. Andaluz, E., O. Steinberg-Neifach, N. Lue, N. Gow, R. Calderone, and G. Larriba. 2004. Homologous recombination in Candida albicans: role of CaRad52p in DNA repair, integration of linear DNA fragments and telomere length. Mol. Microbiol. 53:1177-1194.
8. Fonzi, W. A., and M. Y. Irwin. 1993. Isogenic strain construction and gene mapping in Candida albicans. Genetics 134:717-728.
9. Forche, A., P. T. Magee, B. B. Magee, and G. May.2004 . Genome-wide single-nucleotide polymorphism map of Candida albicans. Eukaryot. Cell 3:705-714.
10. Forche, A., G. May, and P. T. Magee. 2005. Demonstration of loss of heterozygosity by single-nucleotide polymorphism microarray analysis and alterations in strain morphology in Candida albicans strains during infection. Eukaryot. Cell 4:156-165.
11. Galgoczy, D. J., and D. P. Toczyski. 2001. Checkpoint adaptation precedes spontaneous and damage-induced genomic instability in yeast. Mol. Cell. Biol. 21:1710-1718.
12. Hiraoka, M., K. Watanabe, K. Umezu, and H. Maki. 2000. Spontaneous loss of heterozygosity in diploid Saccharomyces cerevisiae. Genetics 156:1531-1548.
13. Hull, C. M., and A. D. Johnson. 1999. Identification of a mating type-like locus in the asexual pathogenic yeast Candida albicans. Science 285:1271-1275.
14. Hull, C. M., R. M. Raisner, and A. D. Johnson. 2000. Evidence for mating of the asexual yeast Candida albicans. Science 289:307-310.
15. Janbon, G., F. Sherman, and E. Rustchenko. 1998. Monosomy of a specific chromosome determines L-sorbose utilization: a novel regulatory mechanism in Candida albicans. Proc. Natl. Acad. Sci. USA 95:5150-5155.
16. Janbon, G., F. Sherman, and E. Rustchenko. 1999. Appearance and properties of L-sorbose-utilizing mutants of Candida albicans obtained on a selective plate. Genetics 153:653-664.
17. Klein, H. L. 2001. Spontaneous chromosome loss in Saccharomyces cerevisiae is suppressed by DNA damage checkpoint functions. Genetics 159:1501-1509.
18. Kobayashi, T., T. Horiuchi, P. Tongaonkar, L. Vu, and M. Nomura.2004 . SIR2 regulates recombination between different rDNA repeats, but not recombination within individual rRNA genes in yeast. Cell 117:441-453.
19. Larriba, G. 2004. Genome instability, recombination, and adaptation, p. 285-334. In G. San Blas and R. Calderone (ed.), Pathogenic fungi: host interactions and emerging strategies for control. Horizon Press, Norwich, United Kingdom.
20. Lay, J., L. K. Henry, J. Clifford, Y. Koltin, C. E. Bulawa, and J. M. Becker.1998 . Altered expression of selectable marker URA3 in gene-disrupted Candida albicans strains complicates interpretation of virulence studies. Infect. Immun. 66:5301-5306.
21. Legrand, M., P. Lephart, A. Forche, P. T. Magee, and B. B. Magee. 2004. Homozygosity at the MTL locus in clinical strains of Candida albicans is correlated with karyotypic rearrangements. Mol. Microbiol. 52:1451-1462.
22. Lephart, P. R., H. Chibana, and P. T. Magee.2005 . Effect of the major repeat sequence on chromosome loss in Candida albicans. Eukaryot. Cell 4:733-741.
23. Lewis, L. K., and M. A. Resnick. 2000. Tying up loose ends: non-homologous end-joining in Saccharomyces cerevisiae. Mutat. Res. 451:71-89.
24. Magee, P. T., and H. Chibana. 2002. The genomes of Candida albicans and other Candida species, p.293 -304. In R. A. Calderone (ed.), Candida and candidiasis. ASM Press, Inc., Washington, D.C.
25. Magee, B. B., and P. T. Magee. 2000. Induction of mating in Candida albicans by construction of MTLa and MTL strains.Science 289:310-313.
26. Magee, B. B., and P. T. Magee. 2004. Through a glass opaquely: the biological significance of mating in Candida albicans. Curr. Opin. Microbiol. 7:661-665.
27. Malkova, A., E. L. Ivanov, and J. E. Haber.1996 . Double-strand break repair in the absence of RAD51 in yeast: a possible role for break-induced replication.Proc. Natl. Acad. Sci. USA 93:7131-7136.
28. Olaiya, A. F., and S. J. Sogin. 1979. Ploidy determination of Candida albicans. J. Bacteriol. 140:1043-1049.
29. Paques, F., and J. E. Haber. 1999. Multiple pathways of recombination induced by double-strand breaks in Saccharomyces cerevisiae. Microbiol. Mol. Biol. Rev. 63:349-404.
30. Perepnikhatka, V., F. J. Fischer, M. Niimi, R. A. Baker, R. D. Cannon, Y. K. Wang, F. Sherman, and E. Rustchenko. 1999. Specific chromosome alterations in fluconazole-resistant mutants of Candida albicans. J. Bacteriol. 181:4041-4049.
31. Rustad, T. R., D. A. Stevens, M. A. Pfaller, and T. C. White. 2002. Homozygosity at the Candida albicans MTL locus associated with azole resistance.Microbiology 148:1061-1072.
32. Rustchenko, E., and F. Sherman. 2003. Genetic instability in Candida albicans, p. 732-776. In D. H. Howard (ed.), Pathogenic fungi in humans and animals. Marcel Dekker, New York, N.Y.
33. Sandell, L. L., and V. A. Zakian. 1993. Loss of a yeast telomere: arrest, recovery, and chromosome loss.Cell 75:729-739.
34. Soll, D. R. 2002. Phenotypic switching, p.123 -142. In R. A Calderone (ed.), Candida and candidiasis. ASM Press, Inc., Washington, D.C.
35. Selmecki, A., S. Bergmann, and J. Berman. 2005. Comparative genome hybridization reveals widespread aneuploidy in Candida albicans laboratory strains. Mol. Microbiol. 55:1553-1565.
36. Tavanti, A., N. A. R. Gow, M. C. J. Maiden, F. C. Odds, and D. J. Shaw.2004 . Genetic evidence for recombination in Candida albicans based in haplotype analysis. Fungal Genet. Biol. 41:553-562.
37. Wu, W., C. Pujol, S. R. Lockhart, and D. Soll.2005 . Chromosome loss followed by duplication is the major mechanism of spontaneous mating-type homozygosis in Candida albicans. Genetics 169:1311-1327.
38. Yoshida, J., K. Umezu, and H. Maki. 2003. Positive and negative roles of homologous recombination in the maintenance of genome stability in Saccharomyces cervisiae. Genetics 164:31-46.(Neeraj Chauhan, Toni Ciud)
Department of Microbiology, Universdad de Extremadura, Badajoz, Spain
Department of Immunology, Microbiology & Parasitology, Universidad del Pais Vasco, Leioia, Vizcaya, Spain
ABSTRACT
The virulence of Candida albicans mutants lacking one or both copies of RAD52, a gene involved in homologous recombination (HR), was evaluated in a murine model of hematogenously disseminated candidiasis. In this study, the virulence of the rad52 mutant was dependent upon the inoculum concentration. Mice survived at a cell inoculum of 1 x 106, but there was a decrease in survival time at dosages of 1.5 x 106 and especially at 3 x 106 cells per animal. The heterozygote RAD52/rad52 behaved like wild type, whereas a reintegrant strain was intermediate in its ability to cause death compared to these strains and to the avirulent rad52/rad52 null at inocula of 1 x 106 and 1.5 x 106 cells. A double mutant, lig4/lig4/rad52/rad52, was avirulent at all inocula used. PCR analysis of the RAD52 and/or LIG4 loci showed that all strains recovered from animals matched the genotype of the inoculated strains. Analysis of the electrophoretical karyotypes indicated that the inoculated, reintegrant strain carried a large deletion in one copy of chromosome 6 (the shortest homologue, or Chr6b). Interestingly, truncated Chr6b was regenerated in all the strains recovered from moribund animals using the homologue as a template. Further, regeneration of Chr6b was paralleled by an increase in virulence that was still lower than that of wild type, likely because of the persistent loss of heterozygosity in the regenerated region. Overall, our results indicate that systemic candidiasis can develop in the absence of HR, but simultaneous elimination of both recombination pathways, HR and nonhomologous end-joining, suppresses virulence even at very high inocula.
INTRODUCTION
Candida albicans exhibits considerable phenotypic variation both in vitro and in the patient. While the belief was that C. albicans lacked a sexual cycle and therefore was unable to create variability through genome recombination, there is ample evidence now that mating type-like (MTL) loci are present in the genome and that compatible strains can mate (13, 14, 25). However, the low frequency at which mating seems to occur in nature (26), the failure to identify a meiotic process, and evidence that the tetraploid progeny of mating are generally thought to return to a near-diploid state through chromosome loss (4) suggest that mating is not the usual way through which C. albicans creates variability (26). In the absence of a meiosis-driven recombination of new alleles, two other mechanisms may account for the variability in this organism. The first mechanism includes gene regulation, generally accomplished through signal transduction pathways following perception by cells of an environmental signal, as has been observed in morphological transition and adaptation to stress growth conditions (6, 34). A second mechanism that accounts for variability is spontaneous or induced genetic changes, including chromosomal alterations, mutations, and loss of heterozygosity (LOH), as has been reported by a number of investigators in laboratory strains as well as from clinical isolates (for recent reviews see references 19, 24, and 32); all of these genetic events are facilitated by the diploid genome of C. albicans (28). Chromosomal changes, as identified by electrokaryotyping methods, occur as translocations as well as through aneuploidy (19, 32, 34). An example of a chromosomal change as a result of aneuploidy which results in an altered phenotype is the loss of chromosome 5 and the growth of that strain in the presence of sorbose as the only carbon source (15, 16). Alterations in chromosome copy number have also been associated with resistance to fluconazole in vitro; these strains are either monosomic for chromosome 4 or trisomic for chromosome 3 (30). Karyotypic rearrangements, including those that result in homozygosity at the MTL locus, are also well documented in clinical strains (19). The mechanism(s) by which strains of C. albicans develop aneuploidy is not understood. However, it has been suggested that for Saccharomyces cerevisiae, an increase in DNA double-strand breaks (DSB) occurs but that normal cell surveillance and DNA repair are bypassed (12, 17). A second possibility is nondisjunction at mitosis. In addition to chromosomal rearrangements, LOH occurs at measurable rates in vitro as well as in vivo (9, 10). LOH may result from chromosome loss, and in fact this is the primary mechanism through which MTL homozygosis occurs in C. albicans (37). However, LOH may also result from mutations or mitotic recombination. It has been reported that in vivo, LOH may occur either by itself or in combination with chromosomal alterations, resulting in the appearance of new phenotypic traits (10).
In S. cerevisiae and apparently in C. albicans, DSB are preferentially repaired by homologous recombination (HR), in part a function of the RAD52 gene product as well as a complex of other proteins (7, 29). In the absence of Rad52, both organisms may still repair DSB using a second pathway that uses a nonhomologous (illegitimate) end-joining (NHEJ) mechanism (1, 7, 23, 29). NHEJ requires a completely different set of proteins, including the homologues of the mammalian Ku proteins (yKu70 and yKu80) as well as Lig4 and its associated proteins Lif1 and Lif2. We previously identified a LIG4 homologue of C. albicans and have constructed mutants in this gene. The LIG4 of C. albicans is involved in NHEJ and, further, lig4 mutants are partially defective in morphogenesis and are avirulent in a murine model of hematogenously disseminated candidiasis (1-3). HR also drives mitotic recombination, including local gene conversion (no crossing over) as well as mitotic crossing over; these events may result in LOH at a single (or a few) locus or along long tracts of the chromosome, respectively. In order to determine the contribution of the HR pathway in DNA repair and mitotic recombination, we have recently reported several phenotypes of the C. albicans rad52 mutants and demonstrated that Rad52p is (i) critical to the repair of DNA damage caused by either UV light or the radiomimetic compound methylmethane sulfonate; (ii) absolutely required for the integration of linear DNA with long flanking sequence homology, i.e., critical for HR; and (iii) critical in maintaining the length of telomeres (7). Recent results have also indicated a role for Rad52p in the maintenance of the genomic stability in C. albicans (G. Larriba et al., ASM Conf. Candida Candidiasis, Austin, Tex., abstr. 83B, 2004; unpublished results).
Because of the role of Rad52p in repairing DSB, mitotic recombination, and in maintaining genome stability, we have examined the virulence of rad52 mutants in a murine model of hematogenously disseminated candidiasis. Our reasons for this study include not only an analysis of the role of homologous recombination and chromosome stability in virulence but also categorization of the changes that occur in the C. albicans chromosomes during disease development in both the presence and the absence of HR and NHEJ.
MATERIALS AND METHODS
Strains. The strains used in this study are listed in Table 1; their construction was described previously (1, 7).
Growth rates. Liquid cultures were started by inoculation of a 250-ml flask containing 50 ml of YPD (1% yeast extract, 2% glucose, 2% peptone) or SC (synthetic complete) with a suspension of cells previously grown in the same medium. For each medium, cultures were adjusted to a final optical density at 600 nm of 0.05 using a spectrophotometer and incubated at 30°C. Samples were taken at the indicated times, and the optical density at 600 nm was determined in a spectrophotometer. Appropriate dilutions of each sample were plated in duplicate on YPD plates to determine the number of CFU.
OMP decarboxylase assays. We followed previously described procedures for orotidine 5'-monophosphate (OMP) decarboxylase (OMPase) assays (5, 20). All strains were grown overnight at 30°C, washed, and then subcultured in 50 ml of YPD and grown at 30°C until the cultures reached mid-log phase (approximately 6 h). Cultures were centrifuged, washed with 1 mM Tris-HCl buffer (pH 7.5) containing 1 mM dithiothreitol (DTT), and suspended in 2 ml of phosphate buffer (pH 7.5). Cells were transferred to microcentrifuge tubes and broken with glass beads by vortexing for 2 min. The homogenates were transferred to other tubes and centrifuged for 10 min at 13,000 rpm to remove particulate matter. Supernatants were collected and stored in an ice bath, and the protein content was determined by the Bradford method using bovine serum albumin as a standard. To perform OMP decarboxylase assays, reaction mixtures contained 500 μl phosphate buffer (pH 6.0), 10 μl 0.1 M -mercaptoethanol, 50 μl 1 mM OMP, 470 μl cell extract in 1-ml quartz cuvettes. Enzyme activity was measured as a decrease in absorbance per min at 285 nm for 25 min using a water blank as a control. The OMP decarboxylase activity was calculated by Beer's law using a molar extinction coefficient of 1.65 x 103 cm–1 M–1. One unit of enzyme activity was defined as the amount of enzyme required for conversion of 1 μM of OMP to UMP per min (μM/mg of protein/min).
Nucleic acid extraction and analysis. Standard techniques were routinely used for DNA manipulation. Genotyping was carried out by PCR analysis of the RAD52 and/or LIG4 locus and, when indicated, by Southern hybridization. For PCR analysis of the RAD52 locus, we used the previously described pair of oligonucleotides RV1 and RV2, which flank the disrupted region of the RAD52 open reading frame (ORF) (7), as well as another primer set, URA3.1 (5'-GGTATAGAAATGCTGCTTGG-3') and URA3.2 (5'-CGAATCGGCACTACAGC-3'), which are complementary to segments near the 5' and 3' ends of the URA3 ORF, respectively. Primer sets RV1 and RV2 amplify 1.5- and 1.4-kb fragments from wild-type RAD52 and RAD52::hisG constructs, respectively, but not the RAD52::hisG-URA3-hisG construct. Identification of the latter was carried out in two PCRs; the first one used RV2-URA3.1 to yield a 1.8-kb band and the second one used RV1-URA3.2 to yield a 1.3-kb fragment. PCR was performed as described elsewhere (31) except that the annealing temperature was always 62°C and, for the amplification RV1-RV2, the concentration of deoxynucleoside triphosphate was 2 mM. For PCR analysis of the LIG4 locus, we used the oligonucleotides LIG4.3 (5'-GTATACCAGAAGTAAGATGGC-3') and LIG4.4 (5'-CAGGGTGCCTGCTCGAGTGTC-3'), which flank the LIG4 ORF (1), and an annealing temperature of 60°C.
Karyotype analysis. Chromosomal analysis of the same strains inoculated and recovered from mice was accomplished as described previously (1, 7). A 0.1-ml sample of exponentially growing cultures of C. albicans was used to inoculate 10 ml of YPD. Cultures were shaken for 48 h at 30°C, collected by centrifugation, washed twice with 50 mM EDTA, pH 8, and suspended in 1 ml of CPES buffer (40 mM citric acid, 120 mM sodium phosphate, 20 mM EDTA, pH 8, 1.2 M sorbitol, and 5 mM DTT, with 0.2 mg Zymolyase 20,000A). To this suspension, 1 ml of CPE buffer (minus sorbitol and DTT) containing 1% low-melting agarose was added, mixed gently, transferred to a sample mold, and stored at –20°C. The solid agar plugs were then transferred to test tubes containing 6 ml of CPE buffer and stored at 30°C for 4 h. The CPE buffer was replaced with TESP buffer (1 M Tris-HCl, 0.5 M EDTA, 2% sodium dodecyl sulfate [SDS], 1 mg ml–1 proteinase K), and the gels were incubated overnight at 50°C and then washed with Tris-EDTA three times at 50°C and six times at room temperature. Gels were stored at 4°C in 50 mM EDTA, pH 8. Two different protocols were then used to separate chromosomes. In the first protocol, all chromosomes were separated except for both homologues of chromosomes 6 and 7. The gel samples were electrophoresed in 0.6% agarose for 24 h at 80 V with a 120- to 300-s linear ramp and then for 48 h at 80 V with a 420- to 900-s linear ramp in a rotating gel electrophoresis apparatus (Rotaphor; Biometra). A second protocol was also used to separate both homologues of the smaller chromosomes, 6 and 7 (21, 22). In this case, the gel samples were run in 1% agarose at 180 V with a linear ramp of 60 to 120 s, 120° included angle for 48 h, and then at 120 V with a 300- to 420-s linear ramp, 120°C included angle, for 48 h in the same apparatus.
Southern analysis of karyotypes was carried out as described previously (7), using a marker of the SfiI 6C fragment (COX12) as probe. Hybridization bands were visualized using a Molecular Imager (Bio-Rad Laboratories).
Animal experiments. C. albicans strains were grown in YPD medium at 30°C to stationary phase. Cells were harvested by centrifugation, washed twice in calcium- and magnesium-free phosphate-buffered saline (PBS; BioSource International), and suspended to a density of 5 x 106 CFU per ml on the basis of hemocytometer counts prior to use. For the animal experiments, we followed previously published methods (1). Groups of 10 male BALB/c mice (18 to 20 g each; Harlan Laboratories) were injected intravenously via the lateral tail vein with 0.2 ml (106 CFU) of each strain listed in Table 1. In addition, with strain TCR2.2 (rad52/rad52), we also infected mice similarly but with 0.3 ml of a cell suspension (1.5 x 106 CFU). We also determined the virulence of the rad52 (TCR2.2), lig4-deleted mutant (CEA2), and lig4/rad52 double mutant (EAT2) strains, each used at an inoculum concentration of 3 x 106 cells per mouse. All mice were observed twice daily for signs of morbidity and, if moribund, animals were euthanized by CO2 inhalation. Concomitantly, each C. albicans strain was used to inoculate 15 additional mice. Five mice from each group were sacrificed by CO2 inhalation at 24, 48, and 72 h postinfection, and the kidneys from mice infected with each strain were removed, weighed, and homogenized in 5.0 ml of PBS. Homogenates were diluted in PBS, and aliquots were plated on YPD agar supplemented with 50 μg of streptomycin per ml to prevent bacterial growth. Plates were incubated at 30°C for 48 h, and the numbers of CFU per gram of tissue were then quantitated. Also at 48 h, kidneys from mice infected with CAF2-1, TCR2.2, CEA2, or EAT2 at a dose of 3 x 106 CFU were removed, fixed in 10% formalin, and prepared for histological examination using the periodic acid-Schiff stain. To verified genetic constructs, strains recovered from animals were frozen and subjected to PCR and karyotype analysis as described elsewhere.
Statistical analysis. In order to determine strain differences in virulence and tissue counts by statistical analysis, we performed a Kaplan-Meier test of significance by a log-rank test.
RESULTS
Growth rates and OMPase activities of strains. As shown in Fig. 1, when incubated in YPD at 30°C, the rad52 strain grew significantly slower than the heterozygote TCR1, whereas the growth rate of the reintegrant strain TCR3.2.1 was similar to that of other Rad52+ strains. Disruption of lig4 in the rad52 background did not modify the growth rate of the latter. On the basis of the respective growth curves, we calculated generation times of 140 min for both rad52 null strains (TCR2.2 [rad52/rad52] and EAT2 [lig4/lig4 rad52/rad52]) and 75 to 80 min for each of the Rad52+ strains, including CAF2-1 (not shown), heterozygote TCR1, revertant TCR3.2.1, and CEA2 (lig4), which was reported by us previously (1). Additionally, because of the URA3 positional effect that can occur in mutants constructed using the Urablaster method, we measured OMPase activity in all strains. Using C. albicans SC5314 as a reference, we found the OMPase activities for all strains were as follows: CAF2-1, 63.3%; TCR1, 53.3%; TCR2.2, 61.3%; TCR3.2.1; 49.3%; EAT2, 49.6%. These data suggest that, when located at the RAD52 locus, URA3 is expressed at nearly the same level as in its original location; furthermore, the OMPase activity of TCR2.2 (61.3%) was higher than that of TCR3.2.1 (49.3%), yet TCR2.2 was less virulent (see below). Interestingly, the OMPase activity of CEA2 was only 21.4% of wild-type cells, suggesting that this value probably influenced our previous results on the lack of virulence of the null strain lig4/lig4 (see below).
Virulence of rad52 strains. We compared the virulence of the wild type (CAF2-1), the homozygous (TCR2.2 [rad52/rad52]) and heterozygous (TCR1 [RAD52/rad52]) mutants, and a reintegrant strain (TCR3.2.1 [rad52/rad52::RAD52]), as well as a null lig4 strain (CEA2 [lig4/lig4]) and a double null mutant lig4 rad52 (EAT2 [lig4/lig4 rad52/rad52]) (Table 1) according to methods previously described (1).
The survival of mice infected with 1 x 106 to 1.5 x 106 CFU of each strain of C. albicans described above is represented in Fig. 2A. Wild-type C. albicans (CAF2-1) at a dosage of 1 x 106 CFU killed all animals within 3 days postinfection, similar to the heterozygous strain (RAD52/rad52 [TCR1]; P = 0.45), whereas the gene-rescued strain (rad52/rad52::RAD52 [TCR3.2.1]) displayed a significantly lower killing ability than the heterozygote (P < 0.004) in spite of the fact that each strain exhibited a similar growth rate and OMPase activity. Several reasons may account for this behavior (see below). The homozygous null strain (rad52/rad52 [TCR2.2]) was avirulent at a dose of 1 x 106, similar to strains CEA2 and EAT2 (Fig. 2A). However, if the concentration of strain TCR2.2 used to infect mice was increased to 1.5 x 106 cells, then 30% of the mice died by 28 days postinfection (Fig. 2A). We also compared the survival of mice infected with 3 x 106 cells of strains TCR2.2, CEA2 (lig4/lig4), and EAT2 (lig4/lig4/rad52/rad52) (Fig. 2B). We found that the virulence of the rad52/rad52 and lig4/lig4 mutants is similar (increased) at this cell concentration, while the double mutant (EAT2) remains avirulent. Thus, virulence of the rad52/rad52 mutant is dependent upon the inoculum concentration, while the double mutant remains avirulent even though the growth rates of both strains are similar. The CFU/g of kidney was determined for each strain at 24 to 72 h postinfection (Fig. 3), except for animals infected with CAF2-1, since all were moribund by 72 h. The colony counts of strains CAF2-1, TCR1 (RAD52/rad52), and TCR3.2.1 (rad52/rad52::RAD52) remained high (5.5 x 106 to 6.23 x 106) during the first 48 to 72 h postinfection. However, for strain TCR2.2 (rad52/rad52), at the same inoculum dose (1 x 106 CFU), the highest CFU/g kidney was always lower (4.65 x 106) (and even dropped by almost 1 log by 72 h), and the same was true for the double mutant EAT2 (lig4/lig4 rad52/rad52) (P < 0.0001, TCR2.2 and EAT2 versus all strains). In order to obtain and maintain a CFU/g in kidney in the range of 5.7 x 106 (i.e., similar to that reached with the Rad52+ strains inoculated with the lower dose of 1 x 106 CFU), it was necessary to raise the inoculum dose of both rad52/rad52 null strains to 3 x 106 CFU. Interestingly, only then were the single mutants (CEA2 and TCR2.2) virulent, whereas the double mutant remained avirulent (Fig. 2B). As described before (1), the lig4 null strain, which grew as fast as the parental CAF2-1, was avirulent at an inoculum dose of 106. We show here that the lig4/lig4 null strain was virulent when the dosage was increased to 3 x 106, although, in support of the role of Lig4p in virulence, even at this higher dose, the lig4/lig4 strain was less virulent than the parental CAF2-1 at 1 x 106. The fact that the double mutant rad52/rad52/lig4/lig4, which exhibits a growth rate similar to the rad52/rad52 mutant, is avirulent even at the highest dose (3 x 106) not only confirms a role for Lig4 in virulence but also indicates that killing by rad52/rad52 mutant cells is not a direct effect of the large amount of cells inoculated. The histopathology of kidneys from mice infected for 48 h is shown for selected strains in Fig. 4. The presence of filamentous growth in the kidney was observed, but the amount of growth varied among strains. Interestingly, animals infected with the double mutant lig4/lig4/rad52/rad52 survived the course of the experiment, but the organism could still be recovered from tissue in high levels and could be seen in the histological sections of infected tissue (Fig. 3 and 4).
Genotyping of strains recovered from mice. To verify that the strains recovered from the infected animals were derived from those used to infect mice, we first analyzed the RAD52 and/or LIG4 loci using PCR. For that purpose, we selected strains recovered from the kidneys of animals infected for 48 h that had been inoculated with 1 x 106 cells of CAF2-1, TCR1(RAD52/rad52), TCR2.2 (rad52/rad52), TCR3.2.1 (rad52/rad52::RAD52), CEA2 (lig4/lig4), and EAT2 (lig4/lig4 rad52/rad52). The hisG-URA3-hisG-interrupted allele was found in the TCR1-inoculated mice (Fig. 5A, lanes 1 and 6) and all recovered strains of TCR1 (one of three isolates is shown) (Fig. 5A, lanes 2 and 7). TCR1 strains also carried the wild-type RAD52 allele (not shown).
PCR analysis of 10 isolates recovered from mice that were inoculated with TCR2.2 detected the hisG-URA3-hisG-disrupted RAD52 allele, indicating that the recovered strains (two examples are shown in Fig. 5A, lanes 4 and 5) were derived from the inoculation strain (Fig. 5A, lane 3; compare lane 8 with lanes 9 and 10). As expected, all TCR2.2 strains carried the hisG-disrupted allele (Fig. 5B, lanes 3, 4, and 5 for the inoculated and same two recovered strains, respectively). Parental strain CAI4 (Fig. 5B, lane 1) and strain TCR1.1 (Fig. 5B, lane 6) showed the expected PCR results.
A previous analysis of the RAD52 locus in the revertant TCR3.2.1 suggested that this strain had lost the RAD52::hisG-disrupted allele (7). These analyses were hindered by the presence of a nonspecific band in the PCR products (7), as well as some genetic instability detected in this strain (see below). However, under the new PCR conditions that eliminate the contaminant, we found that TCR3.2.1 carries both a wild-type (1.5-kb band) and a hisG-disrupted (1.4-kb band) RAD52 allele (Fig. 5B, lane 7). Both alleles were also present in five strains recovered from the animals inoculated with TCR3.2.1 (two examples are shown in Fig. 5B, lanes 8 and 9).
As shown in Fig. 5C for the inoculated and two recovered strains of CEA2 (lig4::hisG/lig4::hisG-URA3-hisG), each carries both the lig4::hisG-URA3-hisG allele (indicated by the presence of a 1.8-kb band with oligonucleotides LIG4.3-URA3.1 and a 1.3kb band with the oligonucleotides LIG.4.4-URA3.2) and the lig4::hisG allele(indicated by the presence of the 1.4-kb band with oligonucleotides LIG4.3-LIG4.4) (1). Figure 5D shows that three recovered EAT2 strains yielded the same PCR products as the inoculated EAT2 (lig4rad52) (lig4::hisG/lig4::hisG rad52::hisG/rad52::hisG-URA3-hisG), which correspond to the rad52::hisG-URA3-hisG allele (1.8-kb and 1.3-kb bands with oligonucleotides RV2-URA3.1 and RV1-URA3.2, respectively), as well as the rad52::hisG (1.4-kb band with oligonucleotides RV1-RV2) and the lig4::hisG (1.4-kb band with oligonucleotides LIG4.3-LIG4.4) alleles. Overall, our results indicate that the recovered strains are derived from the inoculated strains and not from an exogenous C. albicans contaminant.
Karyotypic analysis of the recovered strains. In order to investigate the occurrence of genomic alterations that could have occurred during the infection (in vivo), we compared chromosomal karyotypes of all strains. Figure 6A shows the standard chromosomal karyotypes of inoculated (one from each strain; Fig. 6A, lanes 1, 3, 7, and 12) and representative recovered strains of CAF2-1 (Fig. 6, lane 2), TCR1 (Fig. 6, lanes 4 to 6), TCR2.2 (3 of 10 isolates analyzed; Fig. 6, lanes 13 to 15), and TCR3.2.1 (4 of 5 analyzed; Fig. 6, lanes 8 to 11). Analysis of the shorter chromosomes of the same strains is shown in Fig. 6B. A shift in the mobility of chromosome R (ChrR) was observed in all Rad52+ strains and, in particular, strains CAF2-1 and one TCR1 strain recovered from infected animals (compare lanes 1 and 3 with lanes 2 and 4, respectively). This observation suggests the existence of frequent recombination events between the rRNA gene cistrons during the course of the infection. This is not surprising, since chromosome R rearrangements are well documented in wild-type (Rad52+) strains of C. albicans (references 20, 26, and 33 and references therein). Interestingly, changes in the mobility of ChrR were not as obvious in the case of the rad52 null strains (TCR2.2 and EAT2) (Fig. 6A, lanes 12 to 15 and lanes 16 to 19, respectively), suggesting that in wild-type (Rad52+) strains this event likely involves HR. In the absence of Rad52, other genetic events that do not require the Rad52 protein, such as deletions mediated by single strand annealing (29), could result in the loss of some rRNA cistrons with the concomitant shortening of ChrR, as observed in one TCR2.2 strain recovered from the animals (Fig. 6A, lane 15). Both homologues of ChrR were observed with the EAT2-inoculated strain (Fig. 6A, lane 16). Two recovered strains yielded the same pattern (lanes 17 and 18), but in the third one, either both copies of ChrR comigrated with the shorter homologue or the largest one was lost (lane 19). In agreement with our observations, recent results indicate that rRNA repeats cannot expand in the absence of Rad52 (18). Further analysis of the sizes of the rRNA cistrons in inoculated and recovered rad52 mutants appears necessary before establishing definitive conclusions. Apart from the changes involving ChrR, we did not observe gross chromosomal rearrangements in the rad52 null strains (TCR2.2 and EAT2) (Fig. 6A). This observation is in strong contrast with the genomic instability that characterizes these mutants in S. cerevisiae (17, 29, 38). Our explanation is that strains carrying gross chromosomal rearrangements derived from deficiencies in HR are less fit to survive in the animal and are eliminated. Still, it seems that the presence of the rad52 mutation equalizes the size of both homologues of chromosome 6, and this characteristic was maintained in the recovered strains (Fig. 6B, lanes 12 and 13 to 15). By contrast, the same mutation exacerbated the difference in the size of both homologues of chromosome 7, and this feature was also maintained in the recovered strains. Whether these variations in the size of homologues are specific for the rad52 mutation remain to be investigated.
As expected from previous results (7), the reintegrant TCR3.2.1 showed a supernumerary chromosome migrating above Chr5 (SN1) (Fig. 6A and B, lane 7). This extra band was conserved in all the rescued strains. In addition, the improved resolution of our electrophoresis karyotypes allowed us to detect two additional features of the revertant strain TCR3.2.1, the apparent presence of a single copy of Chr6 and an additional extra band running ahead of Chr7 (SN2) (Fig. 6A and B, lane 7). Interestingly, the four recovered strains from animals inoculated with 1 x 106 cells of TCR 3.2.1 had regained both copies of Chr6 and concomitantly had lost the extra band (Fig. 6A and B, lanes 8 to 11). Furthermore, the electrokaryotypes obtained under conditions that separated both homologues of chromosomes 6 and 7 (Fig. 6B) allowed us to identify the single copy of Chr6 present in the inoculated revertant with the largest homologue of that chromosome (from now on, Chr6a) (lane 7), whereas four recovered strains carried both homologues (lanes 8 to 11). Other reports have described the differences in size between the C. albicans homologues that are due to the number of repeats in the major repeat sequence (references 19, 22, and 24 and references therein). Southern blot analysis using a probe of chromosome 6 (COX12), which is located in the SfiI fragment 6C, labeled Chr6a and SN2 in the inoculated TCR3.2.1 and both homologues of Chr6 in the recovered strains (Fig. 6C). These results unambiguously demonstrate that SN2 was a truncated Chr6b that was regenerated in the host to its original size, most probably using the homologue Chr6a as a template. We favor this possibility, since if cells were able to use randomly any chromosome as a template, we should expect to find a full spectrum of sizes for the regenerated Chr6b. The molecular mechanisms responsible for the truncation/deletion as well as for the regeneration of this chromosome are being investigated.
In view of these results and taking into account that the heterozygote TCR1 was significantly more virulent than the reintegrant TCR3.2.1 (Fig. 2A), we were interested in comparing the virulence of the reintegrant strain recovered from mice to that of both the original reintegrant and the heterozygote used to infect animals. Therefore, the same murine model was used with strains CAF2-1, TCR1, TCR3.2.1 (original), and TCR3.2.1 (recovered) at an inoculum dosage of 1 x 106 yeast cells per mouse (Fig. 7). We show that the recovered, reintegrant strain is statistically more virulent than the original reintegrant (P < 0.0001, TCR3.2.1r1 versus TCR3.2.1). However, the recovered reintegrant still did not reach the virulence of the heterozygote.
DISCUSSION
We report that virulence of DNA repair mutants is dependent upon inoculum dosage and that the double mutant (lig4/lig4/rad52/rad52) is avirulent even at concentrations of 3 x106 yeast cells per mouse. We also observed that the gene-reconstituted mutant (rad52/RAD52) was not as virulent as a heterozygote; several possibilities could account for this observation. First, truncated Chr6 may be lost in some cells and, as a consequence, these cells likely become less viable or even die after a few generations. This would not be surprising, since in S. cerevisiae many chromosomes that lose a telomere are themselves lost (11, 27, 33). Accordingly, the real dose injected into the animals could be lower than the theoretical one. However, the CFU/g of kidney for each strain did not differ significantly during the first 72 h (Fig. 3), in spite of the fact that by 96 h, almost 100% of mice infected with TCR3.2.1 were alive while almost all mice infected with TCR1 were dead (Fig. 2A). This observation does not weaken the first possibility, since perhaps many cells of the original TCR3.2.1 grow well on agar plates but not in the animal because of the lack of Chr6b. The lag period in killing observed with the original TCR3.2.1 could be required for the regeneration of Chr6b in some cells and the concomitant increase in virulence. Loss of Chr6b and duplication of Chr6a was not detected in any of the reintegrant strains recovered from the animals. If this happened, it is likely that the resultant strains would not be selected in the animal. A second explanation that could also account for the lower virulence of TCR3.2.1 is the decrease in gene dosage that affects all the ORFs included in the deleted region, which also implies LOH in that fragment. Our results have demonstrated unambiguously that those cells that repair the truncated Chr6b are selected in the animal. This is not surprising either. In S. cerevisiae, a centric chromosomal fragment may be repaired after being segregated through several generations (11, 27, 33). Assuming that TCR3.2.1 cells use the Chr6a as a template during the repair process, they should regain the original gene dosage but still maintain the LOH in the deleted region. In order to investigate how these alterations of Chr6b may affect virulence, we compared the virulence of one strain recovered from animals (TCR3.2.1r1) with both the original reintegrant TCR.3.2.1 and the heterozygote TCR1. Interestingly, the recovered reintegrant strain was statistically more virulent than the original reintegrant, a phenotype that should be associated with repair of chromosome 6b (Fig. 7). However, it was still less virulent than the heterozygote, likely because repair caused LOH in all markers located in the regenerated region.
The conclusions from our study include the following. First, whereas there is no doubt that C. albicans undergoes genetic alterations in the host (10, 11, 22, 31, 36), the importance of these events in disease has not been determined. Our results indicate that the virulence of rad52 strains is dose dependent, being avirulent for mice at a dose of 106 cells. However, this avirulence could be attributed to its longer generation time. The same null strain was still able to kill at doses of 1.5 x 106 and especially 3 x 106, indicating that at higher inoculum densities, the generation time is less critical to the development of candidiasis, at least in this system. Second, our results indicate that, under these conditions, severe C. albicans infection may occur in the absence of mitotic recombination (HR). Thus, at the same inoculum density (3 x 106), the double mutant lig4rad52 remained avirulent even though it displayed the same doubling time as the rad52 single mutant. Since both mutants yielded a similar OMPase activity, this observation agrees with our previous report (confirmed in the present work) on the relevance of NHEJ to C. albicans virulence, in spite of the fact that the lig4 null strain used (CEA2) displayed a significantly lower OMPase activity than CAF2. Third, repair of a truncated chromosome in vitro has been shown to occur in the derivatives of strain RM100#13, but this occurred only occasionally, since most of its genetically manipulated derivatives, including BWP17, maintained the truncation (35). Here, we have shown for the first time the ability of C. albicans to repair a truncated chromosome in the animal, giving rise to a clone of cells that was selected during the course of the infection. Virulence of this strain was enhanced following the repair of Chr6b but did not reach that of wild type or the heterozygous strains, probably because of LOH in the truncated region. These results and the fact that we did not detect any recovered strain with a duplicated Chr6a emphasize the importance of the heterozygosity of some markers in the virulence of C. albicans.
ACKNOWLEDGMENTS
This study was supported by a Public Health Service grant, NIH-NIAID 1 R01 AI51949 to G.L. and R.C., and grant 2PR03A044 from Junta de Extremadura to E.A.
We thank Bebe Magee for providing the marker of the SfiI fragments 6C. We also thank Belen Hermosa and Leocadia Franco for their technical support.
REFERENCES
1. Andaluz, E., R. Calderone, G. Reyes, and G. Larriba. 2001. Phenotypic analysis and virulence of Candida albicans LIG4 mutants. Infect. Immun. 69:137-147.
2. Andaluz, E., T. Ciudad, J. Rubio-Coque, R. Calderone, and G. Larriba.1999 . Cell cycle regulation of a DNA ligase-encoding gene (CaLIG4) from Candida albicans. Yeast 15:1199-1210.
3. Andaluz, E., G. Larriba, and R. Calderone. 1996. A Candida albicans gene encoding a DNA ligase. Yeast 12:893-898.
4. Bennett, R. J., and A. D. Johnson. 2003. Completion of a parasexual cycle in Candida albicans by induced chromosome loss in tetraploid strains. EMBO J. 22:2505-2515.
5. Brand, A., D. M. MacCallum, A. J. P. Brown, N. A. R. Gow, and F. Odds. 2004. Ectopic expression of URA3 can influence the virulence phenotypes and proteome of Candida albicans but can be overcome by targeted reintegration of URA3 at the RPS10 locus. Eukaryot. Cell 3:900-909.
6. Brown, A. J. P. 2002. Morphogenetic signaling pathways in Candida albicans, p.95 -106. In R. Calderone (ed.), Candida and Candidiasis. ASM Press, Inc., Washington, D.C.
7. Ciudad, T., E. Andaluz, E., O. Steinberg-Neifach, N. Lue, N. Gow, R. Calderone, and G. Larriba. 2004. Homologous recombination in Candida albicans: role of CaRad52p in DNA repair, integration of linear DNA fragments and telomere length. Mol. Microbiol. 53:1177-1194.
8. Fonzi, W. A., and M. Y. Irwin. 1993. Isogenic strain construction and gene mapping in Candida albicans. Genetics 134:717-728.
9. Forche, A., P. T. Magee, B. B. Magee, and G. May.2004 . Genome-wide single-nucleotide polymorphism map of Candida albicans. Eukaryot. Cell 3:705-714.
10. Forche, A., G. May, and P. T. Magee. 2005. Demonstration of loss of heterozygosity by single-nucleotide polymorphism microarray analysis and alterations in strain morphology in Candida albicans strains during infection. Eukaryot. Cell 4:156-165.
11. Galgoczy, D. J., and D. P. Toczyski. 2001. Checkpoint adaptation precedes spontaneous and damage-induced genomic instability in yeast. Mol. Cell. Biol. 21:1710-1718.
12. Hiraoka, M., K. Watanabe, K. Umezu, and H. Maki. 2000. Spontaneous loss of heterozygosity in diploid Saccharomyces cerevisiae. Genetics 156:1531-1548.
13. Hull, C. M., and A. D. Johnson. 1999. Identification of a mating type-like locus in the asexual pathogenic yeast Candida albicans. Science 285:1271-1275.
14. Hull, C. M., R. M. Raisner, and A. D. Johnson. 2000. Evidence for mating of the asexual yeast Candida albicans. Science 289:307-310.
15. Janbon, G., F. Sherman, and E. Rustchenko. 1998. Monosomy of a specific chromosome determines L-sorbose utilization: a novel regulatory mechanism in Candida albicans. Proc. Natl. Acad. Sci. USA 95:5150-5155.
16. Janbon, G., F. Sherman, and E. Rustchenko. 1999. Appearance and properties of L-sorbose-utilizing mutants of Candida albicans obtained on a selective plate. Genetics 153:653-664.
17. Klein, H. L. 2001. Spontaneous chromosome loss in Saccharomyces cerevisiae is suppressed by DNA damage checkpoint functions. Genetics 159:1501-1509.
18. Kobayashi, T., T. Horiuchi, P. Tongaonkar, L. Vu, and M. Nomura.2004 . SIR2 regulates recombination between different rDNA repeats, but not recombination within individual rRNA genes in yeast. Cell 117:441-453.
19. Larriba, G. 2004. Genome instability, recombination, and adaptation, p. 285-334. In G. San Blas and R. Calderone (ed.), Pathogenic fungi: host interactions and emerging strategies for control. Horizon Press, Norwich, United Kingdom.
20. Lay, J., L. K. Henry, J. Clifford, Y. Koltin, C. E. Bulawa, and J. M. Becker.1998 . Altered expression of selectable marker URA3 in gene-disrupted Candida albicans strains complicates interpretation of virulence studies. Infect. Immun. 66:5301-5306.
21. Legrand, M., P. Lephart, A. Forche, P. T. Magee, and B. B. Magee. 2004. Homozygosity at the MTL locus in clinical strains of Candida albicans is correlated with karyotypic rearrangements. Mol. Microbiol. 52:1451-1462.
22. Lephart, P. R., H. Chibana, and P. T. Magee.2005 . Effect of the major repeat sequence on chromosome loss in Candida albicans. Eukaryot. Cell 4:733-741.
23. Lewis, L. K., and M. A. Resnick. 2000. Tying up loose ends: non-homologous end-joining in Saccharomyces cerevisiae. Mutat. Res. 451:71-89.
24. Magee, P. T., and H. Chibana. 2002. The genomes of Candida albicans and other Candida species, p.293 -304. In R. A. Calderone (ed.), Candida and candidiasis. ASM Press, Inc., Washington, D.C.
25. Magee, B. B., and P. T. Magee. 2000. Induction of mating in Candida albicans by construction of MTLa and MTL strains.Science 289:310-313.
26. Magee, B. B., and P. T. Magee. 2004. Through a glass opaquely: the biological significance of mating in Candida albicans. Curr. Opin. Microbiol. 7:661-665.
27. Malkova, A., E. L. Ivanov, and J. E. Haber.1996 . Double-strand break repair in the absence of RAD51 in yeast: a possible role for break-induced replication.Proc. Natl. Acad. Sci. USA 93:7131-7136.
28. Olaiya, A. F., and S. J. Sogin. 1979. Ploidy determination of Candida albicans. J. Bacteriol. 140:1043-1049.
29. Paques, F., and J. E. Haber. 1999. Multiple pathways of recombination induced by double-strand breaks in Saccharomyces cerevisiae. Microbiol. Mol. Biol. Rev. 63:349-404.
30. Perepnikhatka, V., F. J. Fischer, M. Niimi, R. A. Baker, R. D. Cannon, Y. K. Wang, F. Sherman, and E. Rustchenko. 1999. Specific chromosome alterations in fluconazole-resistant mutants of Candida albicans. J. Bacteriol. 181:4041-4049.
31. Rustad, T. R., D. A. Stevens, M. A. Pfaller, and T. C. White. 2002. Homozygosity at the Candida albicans MTL locus associated with azole resistance.Microbiology 148:1061-1072.
32. Rustchenko, E., and F. Sherman. 2003. Genetic instability in Candida albicans, p. 732-776. In D. H. Howard (ed.), Pathogenic fungi in humans and animals. Marcel Dekker, New York, N.Y.
33. Sandell, L. L., and V. A. Zakian. 1993. Loss of a yeast telomere: arrest, recovery, and chromosome loss.Cell 75:729-739.
34. Soll, D. R. 2002. Phenotypic switching, p.123 -142. In R. A Calderone (ed.), Candida and candidiasis. ASM Press, Inc., Washington, D.C.
35. Selmecki, A., S. Bergmann, and J. Berman. 2005. Comparative genome hybridization reveals widespread aneuploidy in Candida albicans laboratory strains. Mol. Microbiol. 55:1553-1565.
36. Tavanti, A., N. A. R. Gow, M. C. J. Maiden, F. C. Odds, and D. J. Shaw.2004 . Genetic evidence for recombination in Candida albicans based in haplotype analysis. Fungal Genet. Biol. 41:553-562.
37. Wu, W., C. Pujol, S. R. Lockhart, and D. Soll.2005 . Chromosome loss followed by duplication is the major mechanism of spontaneous mating-type homozygosis in Candida albicans. Genetics 169:1311-1327.
38. Yoshida, J., K. Umezu, and H. Maki. 2003. Positive and negative roles of homologous recombination in the maintenance of genome stability in Saccharomyces cervisiae. Genetics 164:31-46.(Neeraj Chauhan, Toni Ciud)