Allele-specific gene targeting in Candida albicans results from heterology between alleles

Kyle Yesland1 and William A. Fonzi1

Department of Microbiology and Immunology, Georgetown University, 3900 Reservoir Road NW, Washington, DC 20007-2197, USA1

Author for correspondence: William A. Fonzi. Tel: +1 202 687 1135. Fax: +1 202 687 1800. e-mail: fonziw{at}gusun.georgetown.edu


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
The opportunistic fungal pathogen Candida albicans is asexual and diploid. Thus, introduction of recessive mutations requires targeted gene replacement of two alleles to effect expression of a recessive phenotype. This is often performed by recycling of a URA3 marker gene that is flanked by direct repeats of hisG. After targeting to a locus, recombination between the repeats excises URA3 leaving a single copy of hisG in the disrupted allele. The remaining functional allele is targeted in a second transformation with the same URA3 marked construct. Replacement can be highly biased toward one allele. At the PHR1 locus, there was an approximately 50-fold preference for replacement of the disrupted versus the functional allele in a heterozygous mutant. This preference was reduced six- to eightfold when the transforming DNA lacked the hisG repeats. Nonetheless, there remained a sixfold preference for targeting a particular allele of PHR1 and this was evident even in transformations of the parental strain containing two wild-type alleles of PHR1. Both wild-type alleles were cloned and nucleotide sequence comparison revealed 24 heterologies over a 2 kb region. Using restriction site polymorphisms to distinguish alleles, it was observed that transformation with the cloned DNA of allele PHR1-1 preferentially targeted allele 1 of the genome. Transformations with PHR1-2 exhibited the reciprocal specificity. In both these instances, heterology was present in the flanking regions of the transforming DNA. When the transforming DNA was chosen from a region 100% identical in both alleles, alleles 1 and 2 were targeted with equal frequency. It is concluded that sequence heterology between alleles results in an inherent allele specificity in targeted recombination events.

Keywords: Candida albicans, recombination, gene targeting, heterozygosity

The GenBank accession numbers for the sequences reported in this paper are AF247189 and AF247190 for PHR1-1 and PHR1-2, respectively.


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Candida albicans is an opportunistic fungal pathogen of humans. Because of its clinical significance, the biology of this fungus has been the subject of numerous studies (Odds, 1988 ). Despite this extensive investigation, relatively little is known of its basic genetic behaviour. This is due, in part, to its genetic intractability, an asexual life cycle and a diploid genome (Odds, 1988 ). The use of parasexual techniques has provided some information and suggests that C. albicans behaves much like other fungi with, perhaps, a few distinctions (Poulter, 1990 ).

Currently, genetic manipulation of C. albicans relies almost exclusively on reverse genetic methods. The demonstration of transformation (Kelly et al., 1988 , 1987 ) coupled with workable methods of gene replacement (Fonzi & Irwin, 1993 ; Gorman et al., 1991 ) has fostered numerous molecular genetic studies of C. albicans biology. Although improved methods of gene replacement have been developed (Morschhauser et al., 1999 ; Wilson et al., 1999 , 2000 ), most studies to date have relied on recycling of a URA3 marker using positive and negative selection (Fonzi & Irwin, 1993 ). This method, first developed for Saccharomyces cerevisiae, employs a cassette consisting of direct repeats of Salmonella typhimurium hisG flanking a URA3 marker gene (Alani et al., 1987 ; Fonzi & Irwin, 1993 ). The cassette is inserted within the cloned gene of interest, which provides homologous flanking sequences to direct targeted gene replacement. Integration of transforming DNA in C. albicans occurs largely via homologous recombination (Kelly et al., 1987 ). The integrated DNA restores uridine prototrophy to the ura3 recipient cells. Spontaneous recombination between the hisG sequences results in loss of the URA3 marker and retention of a single copy of hisG at the disrupted locus (Fonzi & Irwin, 1993 ). The Uri- uridine auxotrophs resulting from this excision event are selected on 5-fluoroorotic-acid-containing medium (Boeke et al., 1984 ).

Because C. albicans is diploid, expression of recessive mutations requires disruption of two alleles. The heterozygous Uri- mutants are transformed with the same cassette construction, selecting again for uridine prototrophy. Statistically, the cassette is expected to integrate with equal probability into either the previously disrupted allele or the remaining wild-type allele. In practice this is not the case. The ratio of targeting to the wild-type versus the disrupted allele can vary from 1 in 3 to 1 in 35, depending on the locus under investigation (Fonzi & Irwin, 1993 ; Sentandreu et al., 1998 ). The factors responsible for this allele specificity and locus-specific variation are unknown.

In this study we examined allele-specific integration at the PHR1 locus (Saporito-Irwin et al., 1995 ). In a heterozygous mutant with the genotype phr1{Delta}::hisG/PHR1, transforming DNA recombined with the phr1{Delta}::hisG allele approximately 50-times more frequently than with the wild-type allele. This selectivity was only partially due to the additional homology between the disrupted allele and the transforming DNA provided by the hisG sequences. Restriction site polymorphisms permitted identification and cloning of both wild-type alleles of PHR1, designated PHR1-1 and PHR1-2. Transformation with the cloned alleles demonstrated that PHR1-1 preferentially recombined with the resident copy of allele 1 and that PHR1-2 preferentially recombined with the resident copy of allele 2. The nucleotide sequences of the two alleles were approximately 1% divergent overall and these heterologies were concentrated in the 5' and 3' flanking regions of the transforming DNAs. Transformation with PHR1 sequences that were identical in both alleles eliminated the recombinational bias. The results suggest that in C. albicans, as in other organisms, homeologous recombination is less efficient than homologous recombination and that this imparts allele specificity to targeted integration events.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Strains and growth conditions.
Candida albicans strains CAF3-1 (ura3{Delta}/ura3{Delta}) and CAS-6 (ura3{Delta}/ura3{Delta} phr1{Delta}::hisG/PHR1) were described previously (Fonzi & Irwin, 1993 ; Saporito-Irwin et al., 1995 ). Strains YZP1-3 and YZP1-9 are independent mutants of the same genotype as CAS-6 and were constructed as described for CAS-6 (Saporito-Irwin et al., 1995 ). The strains were routinely cultured on YPD medium (Sherman, 1991 ) supplemented with 25 µg uridine ml-1. YNB medium, consisting of 2% glucose and 0·67% Difco yeast nitrogen base, was used in selecting Uri+ transformants. Medium 199 buffered at pH 8 was used to assess growth and morphological phenotypes as described previously (Saporito-Irwin et al., 1995 ). Media were solidified with 1·5% agar. Cultures were incubated at 30 °C.

Gene cloning and plasmid constructions.
Both alleles of PHR1 were cloned by hybridization screening of a {lambda}GEM-12 genomic library derived from strain SC5314 (Gillum et al., 1984 ). The library and screening conditions have been described by Birse et al. (1993) . The 2·1 kb insert from plasmid pSMS-24 (Saporito-Irwin et al., 1995 ) was used as the hybridization probe to detect PHR1 sequences. The two alleles were distinguished by a BamHI restriction site polymorphism, which resulted in hybridization bands of 8·4 and 7·4 kb on Southern blots of genomic DNA. {lambda} clones containing the corresponding fragments were isolated and the inserts were subcloned into the BamHI site of plasmid pBSK+ (Stratagene). The alleles contained on the 7·4 and 8·4 kb fragments were designated PHR1-1 and PHR1-2, respectively.

Plasmids p{Delta}PHR1-1::URA3 and p{Delta}PHR1-2::URA3 contained analogous deletions in alleles 1 and 2, respectively. These were constructed by first subcloning a 2·5 kb EcoRI–EcoRV fragment containing PHR1 into a modified pBSK+ vector (Stratagene). The vector lacked the ClaI recognition site in the polylinker region, which was destroyed by Klenow polymerase fill-in of the restricted site followed by ligation. The subclones were digested with ClaI to remove nt +397 to +1235 of the PHR1 coding region and ligated with a 1·1 kb ClaI fragment containing URA3. The URA3 fragment was generated by PCR amplification using the forward primer USP-1 (5'-CCATCGATTGCTGTAGTGCCATTGAT-3') and the reverse primer RP-2 (5'-CCATCGATAAAGTGAAAATCTCCCCCT-3'). The amplification product encompassed nt –283 to +840, relative to the start codon of URA3 and incorporated flanking ClaI restriction sites. URA3 was cloned in the same transcriptional orientation as PHR1 in both plasmids.

Plasmid pSMS-23 contains a deletion mutation of PHR1 similar to that of p{Delta}PHR1-1::URA3 and p{Delta}PHR1-2::URA3. It differs in that nt +397 to +1235 of the PHR1 coding region were substituted with hisG-URA3-hisG sequences (Fonzi & Irwin, 1993 ). Construction of this plasmid was described previously (Saporito-Irwin et al., 1995 ).

Plasmid pSMS24-URA contained an insertional disruption of PHR1-2. This plasmid was constructed by blunt-end ligation of a 1·3 kb ScaI–XbaI fragment containing URA3 into the BspEI site located at nt +889 of the PHR1 coding region in plasmid pSMS-24 (Saporito-Irwin et al., 1995 ). The BspEI and XbaI ends were made blunt by Klenow polymerase fill-in.

Transformation and screening.
Strains were transformed as described previously (Mühlschlegel & Fonzi, 1997 ) and Uri+ transformants were selected on YNB plates. For phenotypic screening of strains CAS-6, YZP1-3 and YZP1-9, transformants were patched to Medium 199 plates buffered to pH 8. These were incubated at 37 °C for 48 h. The resulting colonies were examined macroscopically for the presence of hyphae and microscopically for cell morphology. Colonies lacking filamentation and containing enlarged, highly vacuolated cells were scored as PHR1 null mutants (Saporito-Irwin et al., 1995 ). Between 100 and 500 colonies were scored for each transformation. The results are presented as the fraction of transformants exhibiting the null mutant phenotype.

Genotypic screening of CAF3-1 transformants was performed by Southern blot analysis. The methods for Southern blot preparation and hybridization were described previously (Mühlschlegel & Fonzi, 1997 ). The 2·1 kb EcoRI fragment of PHR1 (Saporito-Irwin et al., 1995 ) was used as the hybridization probe. Genomic DNA from cells transformed with p{Delta}PHR1-1::URA3 or p{Delta}PHR1-2::URA3 was digested sequentially with EcoRI and MscI. The MscI site is located at position +23 of allele 2, but is absent from allele 1. Replacement of PHR1-1 was distinguished by the loss of the native 9 kb EcoRI–EcoRI hybridizing band and acquisition of a 1·21 kb EcoRI–EcoRI band due to the proximal EcoRI introduced by integration of URA3. The loss of the native 8·7 kb MscI–EcoRI band and acquisition of a 0·9 kb MscI–EcoRI band distinguished replacement of PHR1-2.

Genomic DNA from cells transformed with pSMS24-URA was digested with MscI and PvuII. Replacement of PHR1-1 was distinguished by loss of a native 3·2 kb MscI–PvuII hybridization band and acquisition of a 2·8 kb MscI–MscI band due to insertion of a proximal MscI site present in URA3. Loss of a native 1·4 kb MscI–PvuII band and acquisition of a 0·99 kb MscI–MscI band distinguished replacement of PHR1-2.


   RESULTS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Selective integration at the PHR1 locus
In previous genetic studies the PHR1 locus was mutated by gene replacement with a HindIII–PvuII fragment derived from plasmid pSMS-23 (Saporito-Irwin et al., 1995 ). This fragment contained 660 bp of PHR1 sequence 5' and 185 bp 3' of a hisG-URA3-hisG insertion. The hisG-URA3-hisG cassette replaced 839 bp of the PHR1 ORF between nt +397 and +1236. Heterozygous mutants resulting from the replacement of one allele of PHR1 were readily obtained (Saporito-Irwin et al., 1995 ). Limited attempts to replace the remaining functional allele of the heterozygous mutant with this same HindIII–PvuII fragment were unsuccessful due to preferential recombination with the disrupted allele (unpublished data). The magnitude of this preference was examined by transforming heterozygous mutants and using a phenotypic assay to discriminate between recombination with the functional or the disrupted allele. If the transforming DNA recombined with the disrupted allele, the strain would retain a functional copy of PHR1 and exhibit a wild-type phenotype. Recombination with the functional allele would result in a homozygous null mutant, which exhibits pH-dependent morphological defects. PHR1 null mutants are normal when cultured at pH<6·0, but fail to form hyphae and produce enlarged, rounded and highly vacuolated cells when cultured at alkaline pH (Saporito-Irwin et al., 1995 ). Strain CAS-6 (phr1{Delta}::hisG/PHR1) was transformed with the HindIII–PvuII fragment from pSMS-23 and the Uri+ transformants were patched to Medium 199, pH 8·0. After growth at the restrictive pH, the phenotype of the cells was examined. Of the 147 transformants scored, only three exhibited the null mutant phenotype, a frequency of 2·0x10-2. A very similar allelic preference was observed in transformations of two independent heterozygous strains YZP1-3 and YZP1-9, which yielded mutant frequencies of 3·3x10-2 and 1·4x10-2, respectively. These frequencies are minimal estimates since there is no phenotypic distinction between replacement of the disrupted allele and non-homologous integration at other loci, both of which would behave as wild-type. However, non-homologous integration of this DNA fragment is rare (Saporito-Irwin et al., 1995 ), so failure to account for these events introduces only a minor error in the calculated frequencies. This highly biased targeting also cannot be attributed to selective recovery of wild-type cells, since the transformants were selected at acidic pH, which is permissive for growth of PHR1 null mutants (Saporito-Irwin et al., 1995 ). Thus we conclude that the exogenous transforming DNA preferentially recombines with the disrupted allele and only rarely replaces the wild-type allele.

Effect of hisG sequences on recombination
Preferential targeting of the disrupted allele might reflect the simultaneous presence of hisG sequences in the disrupted allele and the exogenous transforming DNA. As a consequence, the disrupted allele has homology with the exogenous DNA that the wild-type allele does not. By transforming with DNA lacking these sequences the effect of hisG on recombinational preference was assessed. Plasmid p{Delta}PHR1-2::URA3 was analogous to pSMS-23 used in the previous experiment in that both had deletions of the same PHR1 sequences and, as discussed subsequently, both were derived from the same allele of PHR1. However, instead of a hisG-URA3-hisG insert, p{Delta}PHR1-2::URA3 contained an insert of URA3 alone. Digestion of p{Delta}PHR1-2::URA3 with PstI and PvuII released a DNA fragment in which URA3 was flanked by PHR1 sequences essentially identical to those of the HindIII–PvuII digest of pSMS-23. The fragments differed in only a few nucleotides at the 5' end resulting from cleavage in the polylinker of the plasmid. Despite the absence of hisG sequences, preferential recombination with the disrupted allele was still evident (Table 1). The median frequency of mutant transformants was 15x10-2, a roughly sixfold increase in comparison with the previous transformations using hisG-containing DNA. However, the disrupted allele was still targeted at a sixfold higher frequency than the wild-type allele. This indicates that the simultaneous presence of hisG sequences in the disrupted allele and on the transforming DNA does have a selective influence on recombination. However, it does not fully account for the allelic preference of recombination at the PHR1 locus.


View this table:
[in this window]
[in a new window]
 
Table 1. Effect of allelic source, fragment length and sequence divergence on recombinational preference

 
Inherent allele-specific recombination
Unlike the parental strain in which both alleles of PHR1 are wild-type, the alleles in the heterozygous mutant are distinguished by the presence of the hisG insert within the disrupted allele. This creates an inherent structural asymmetry at the locus and, perhaps, in recombination events. To assess whether the insert per se was affecting recombination, allele targeting was examined in the parental strain. The alleles of PHR1 can be distinguished on Southern blots by virtue of a BamHI restriction site polymorphism (Fig. 1). Strain CAF3-1 was transformed with the HindIII–PvuII fragment from pSMS-23 and the genotype of 11 Uri+ transformants was determined by Southern blot analysis. In all 11 isolates the 8·4 kb band, corresponding to allele 2, was absent and replaced with a fragment of approximately 12 kb (Fig. 1). Thus, the transforming DNA was preferentially recombining with PHR1-2. Furthermore, Southern blot analysis of the heterozygous strains CAS-6, YZP1-3 and YZP1-6 demonstrated that PHR1-2 was the allele disrupted in these strains (data not shown). Therefore, the preferential targeting of the disrupted locus in the transformation of these strains was also due to preferential recombination with PHR1-2. Together these results indicated that PHR1-2 recombined at a higher frequency than PHR1-1 and that this was independent of the presence or absence of an insert within the allele.



View larger version (25K):
[in this window]
[in a new window]
 
Fig. 1. Allele specific recombination of transforming DNA with PHR1-2. Genomic DNA from CAF3-1 (lane 1) and 11 independent transformants (lanes 2–12) was digested with BamHI and examined by Southern blot hybridization with PHR1. The sizes of the wild-type bands are indicated on the left; the 7·4 and 8·4 kb bands correspond to alleles PHR1-1 and PHR1-2, respectively.

 
The cause of this asymmetry could lie with either or both reactants in the recombination event, the resident genomic DNA or the exogenous transforming DNA. PHR1-1 and PHR1-2 may have intrinsic differences in their ability to recombine, specificity may be imparted by some attribute of the transforming DNA or specificity may be an interdependent effect. Examination of restriction site polymorphisms demonstrated that the PHR1 sequences in plasmid pSMS-23 were derived from allele 2 (data not shown). Thus, in the preceding experiments the transforming DNA derived from allele 2 was preferentially recombining with the resident allele 2. This suggested that specificity was interdependent. If so, then transforming DNA derived from PHR1-1 should preferentially recombine with allele 1 in the genome.

By transforming with identical DNA fragments derived from either PHR1-1 or PHR1-2 the relationship between the allelic source of the transforming DNA and the allele specificity of recombination was tested. Plasmids p{Delta}PHR1-1::URA3 and p{Delta}PHR1-2::URA3 were identically constructed, except that the PHR1 sequences were derived from the cloned allele 1 or 2, respectively. In the preceding experiment, transformation of the heterozygous disruptants with a PstI–PvuII digest of p{Delta}PHR1-2::URA3 resulted in preferential recombination with the disrupted allele as indicated by the low frequency of null mutants (Table 1). The analogous transformation using p{Delta}PHR1-1::URA3 DNA yielded the reciprocal results; the mutant frequency increased to 81x10-2, indicating preferential replacement of the functional allele. Since the functional allele was PHR1-1 in these strains, p{Delta}PHR1-1::URA3 DNA selectively recombined with the corresponding allele in the genome. The magnitude of this bias was similar to that of p{Delta}PHR1-2::URA3 DNA for allele 2. Several other restriction enzyme digests, which altered the length of the flanking sequences, were tested in the transformations and in each case the allele specificity of integration was maintained (Table 1).

To ensure that these results were not an artefact of the locus asymmetry in the heterozygous mutants, DNA from allele 1 or 2 was transformed into the parental strain CAF3-1. Plasmid p{Delta}PHR1-1::URA3 was digested with BsaBI and AhdI. The recognition sites for these enzymes are located at positions +313 and +1306, respectively. Plasmid p{Delta}PHR1-2::URA3 was digested with BsaBI alone since two recognition sites are present, one at +313 and the other +1300. The 3' ends differ due to a restriction site polymorphism. The transformants were screened by Southern blot hybridization using an MscI restriction site polymorphism to distinguish which allele had recombined. Of 24 isolates recovered from transformation with p{Delta}PHR1-1::URA3 DNA, 22 were disrupted in allele 1. Conversely, 15 of 16 transformants recovered from transformation with p{Delta}PHR1-2::URA3 DNA were disrupted in allele 2. Therefore, exogenous DNA derived from allele 1 preferentially recombines with allele 1 of the genome and the inverse preference is observed for allele 2, irrespective of the presence or absence of a non-homologous insert within one of the alleles.

Allele-specific recombination due to sequence heterology
One obvious explanation for the self-preference of allele targeting lies in the diverged nucleotide sequence of the two alleles as reflected in their multiple restriction site polymorphisms (Fig. 2). As a consequence, recombination between the exogenous DNA and its genomic counterpart is a homologous event, but recombination with the alternate allele would be a homeologous event. Homeologous recombination is less efficient than homologous recombination (Chen & Jinks-Robertson, 1998 ; Elliott et al., 1998 ; Nassif & Engels, 1993 ; Stambuk & Radman, 1998 ) and this could account for the bias in gene replacement.



View larger version (10K):
[in this window]
[in a new window]
 
Fig. 2. Distribution of sequence heterologies at the PHR1 locus. A restriction map of the region sequenced from both alleles is shown. The vertical bars below the map indicate the location of nucleotide differences between PHR1-1 and PHR1-2. The rectangle indicates the coding region of PHR1 and flanking regions are indicated by the solid line. The dashed line indicates vector sequence. The hatched region between the ClaI sites indicates the region replaced with either hisG-URA3-hisG or URA3 in the transforming DNAs. The asterisk indicates the site of URA3 insertion in pSMS24-URA. Polymorphic restriction sites are indicated by the particular allele containing the site.

 
The nucleotide sequence of both alleles was determined for a region extending from -263 bp upstream to 119 bp downstream of the PHR1 coding region. Comparison of alleles 1 and 2 over this 2028 bp region identified 24 nt differences, including 18 transitions, 5 transversions and a 1 bp deletion (Table 2). Seven of the 20 nt differences within the coding region were associated with amino acid substitutions (Table 3). Heterologies were unevenly distributed and concentrated in the 5' and 3' ends of the gene (Fig. 2). In relation to the transforming DNA fragments, the number of diverged nucleotides in the 5'PHR1 sequences ranged from 16 in the PstI–PvuII digests to none in the BsaBI digests (Table 1). The 3' flanking sequences contained from as many as 5 to as few as 1 diverged nucleotides (Table 1).


View this table:
[in this window]
[in a new window]
 
Table 2. Sequence heterology of PHR1 alleles

 

View this table:
[in this window]
[in a new window]
 
Table 3. Substitution mutations in PHR1

 
If this heterology was responsible for the observed allele-specific gene replacement, then transforming DNA derived from a region 100% identical between the two alleles should recombine with either allele at equal frequency. The previous plasmids could not be used to test this prediction since one of the heterologous sites was only 6 bp from the marker insertion site. Consequently, plasmid pSMS24-URA was constructed. Digestion of this plasmid with BsaBI and NheI released a fragment in which the 153 bp 5' and 202 bp 3' of the marker gene were identical in both alleles. Strain CAF3-1 was transformed with the pSMS24-URA DNA and the genotype of the Uri+ transformants were determined by Southern blot analysis. Of 26 transformants examined, PHR1-1 was replaced in 14 and PHR1-2 was replaced in the other 12. Therefore, allele-specific targeting is lost in the absence of sequence heterology.


   DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Targeted gene replacements at the PHR1 locus exhibited a highly biased allele preference. In transformations of heterozygous mutants, there was an approximately 50-fold preference for replacement of the disrupted allele over the wild-type allele. This specificity was a consequence of two factors, the parallel presence of hisG sequences in the disrupted allele and the exogenous DNA and the sequence divergence of PHR1-1 and PHR1-2. Increasing the length of the complementary sequences has a positive affect on the frequency of targeted integration in a number of organisms (Hasty et al., 1991 ; Papadopoulou & Dumas, 1997 ; Rothstein, 1991 ) and this may account for the selectivity imparted by hisG. However, the frequency of transformation in yeast is also proportional to the number of target sequences available for recombination (Wilson et al., 1994 ). The presence of the hisG sequences in the disrupted allele, in a sense, provides an additional target in this allele since recombination could occur between the hisG-URA3-hisG cassette of the exogenous DNA even in the absence of flanking PHR1 sequences. Furthermore, because the cassette has two copies of hisG this may also increase the opportunity for such recombination events.

The hisG insert in the disrupted allele could potentially affect recombination by other mechanisms as well. For instance, the insert increased by >500 bp, the distance between PHR1 homologous sequences, and may alter the chromatin structure, nucleosome positioning or rate of transcription initiation, any of which might influence recombination frequency (Grigoriev & Hsieh, 1997 ; Kirkpatrick et al., 1999 ; Wu & Lichten, 1994 ). However, the significance of these effects appear minimal since they would be predicted to have a similar effect on recombination with transforming DNA from either PHR1-1 or PHR1-2 and this was not the case. Exogenous PHR1-1 DNA replaced the endogenous PHR1-1 at a similar frequency in both parental and heterozygous strains.

The effect of allelic sequence divergence, which was approximately 1·2% overall for PHR1, presumably reflects the greater efficiency of homologous recombination in comparison with homeologous transformation. Recombination between like alleles was a homologous event while recombination between the diverged alleles was a homeologous event. It is well established that sequence heterology interferes with recombination in prokaryotic and eukaryotic species (Chen & Jinks-Robertson, 1998 ; Elliott et al., 1998 ; Nassif & Engels, 1993 ; Stambuk & Radman, 1998 ) and this includes targeted gene replacements (Leung et al., 1997 ; Negritto et al., 1996 ). In S. cerevisiae a single nucleotide difference within a 350 bp region can reduce recombination fourfold (Datta et al., 1997 ). Reduced recombination is due to interference by the mismatch repair system with heteroduplex formation between diverged sequences (Chen & Jinks-Robertson, 1998 ; Leung et al., 1997 ).

The results of this study imply that the mismatch repair system is similarly effective in Candida albicans. As shown by the data in Table 1, a single nucleotide difference was sufficient to bias recombination frequencies. In all of these transformations, recombination with the homologous allele was favoured four- to fivefold over recombination with the heterologous allele. The only exception occurred in transformations using the BsaBI fragment of p{Delta}PHR1-2::URA3, which contained only a single nucleotide difference from PHR1-1. This fragment unexpectedly recombined 30-times more frequently with the homologous allele. Preliminary experiments suggest that this atypical behaviour is a peculiarity of this particular DNA fragment (unpublished results). In these experiments, designed to test reciprocity, a heterozygous mutant was constructed in which PHR1-1 was disrupted instead of PHR1-2 as in the present experiments. Transformations of this heterozygote analogous to those in Table 1 yielded quantitatively inverted mutant frequencies as expected. The only exception was again with the BsaBI fragment of p{Delta}PHR1-2::URA3. The frequency of mutants in these experiments was 14% rather than the expected 97%. However, it should be noted that the mutant frequency increased about fivefold.

Allele-specific gene targeting has been reported for other loci in C. albicans (Sentandreu et al., 1998 ) and this is presumably related to the same factors influencing gene replacement of PHR1. Although few loci in C. albicans have been examined for heteroallelic differences at the nucleotide level, heterology has been evident at all loci examined. Both alleles of SAP4, ALS2 and ALS4 have been completely or partially sequenced (Hoyer et al., 1998 ; Miyasaki et al., 1994 ) and roughly 0·5% divergence was found for each locus. This compares to the 1·2% divergence of PHR1 alleles. Variations in the degree and distribution of these heterologies could account for the observed locus to locus variations in allele-specific gene targeting. Of course this does not rule out additional factors. In C. albicans the frequency of recombination with exogenous DNA correlates with the transcriptional activity of the targeted locus (Srikantha et al., 1995 ). Given that C. albicans is naturally heterozygous for many recessive mutations (Whelan & Magee, 1981 ; Whelan et al., 1980 ; Whelan & Soll, 1982 ), it would not be surprising to find allelic variations in transcription or perhaps DNA topology, which could influence recombination (Paques & Haber, 1999 ).

There are both practical and theoretical implications to the results reported here. From a pragmatic perspective, the inability to inactivate both alleles of a locus cannot be taken as prima facie evidence that the gene is essential. Since a single cloned allele is typically employed in such experiments, sequence divergence between alleles may greatly reduce recombination and prevent production of the homozygous replacement. Of more biological significance is the implied effect of the mismatch repair system in generating genomic diversity in C. albicans. As discussed by Vuli’c et al. (1997 ), the mismatch repair system serves dual roles in the evolution of genetic diversity, it prevents the accumulation of mutations and inhibits genetic exchange between diverged sequences. In an asexual diploid like C. albicans this may have the interesting consequence of fostering allelic divergence. Although the recombination and repair system would limit the rate of accumulation of mutations, once a mutation did occur, mismatch repair would limit interchromosomal recombination and act to maintain heterozygosity. This would promote the accumulation of additional heterozygous mutations, which would further inhibit interchromosomal recombination. The net effect would be divergence of alleles. Environmental and physiological influences on expression of the mismatch repair system would periodically relax the barrier on chromosomal exchange and allow improved traits to be brought to homozygosity and deleterious mutations to be lost (Vuli’c et al., 1997 ). This periodic recombination between diverged alleles may serve a genetic role analogous to sexual recombination in other organisms.


   ACKNOWLEDGEMENTS
 
We wish to thank Yonghong Zhang and Terrie Nghiem for their excellent technical assistance in conducting these studies. This work was supported by Public Health Services grant GM47727 from the National Institutes of Health and the Burroughs Wellcome Fund Scholar Award in Molecular Pathogenic Mycology.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Alani, E., Cao, L. & Kleckner, N. (1987). A method for gene disruption that allows repeated use of URA3 selection in the construction of multiply disrupted yeast strains. Genetics 116, 541-545.[Abstract/Free Full Text]

Birse, C. E., Irwin, M. Y., Fonzi, W. A. & Sypherd, P. S. (1993). Cloning and characterization of ECE1, a gene expressed in association with cell elongation of the dimorphic pathogen Candida albicans. Infect Immun 61, 3648-3655.[Abstract]

Boeke, J. D., LaCroute, F. & Fink, G. R. (1984). A positive selection for mutants lacking orotidine-5'-phosphate decarboxylase activity in yeast: 5-fluoro-orotic acid resistance. Mol Gen Genet 197, 345-346.[Medline]

Chen, W. & Jinks-Robertson, S. (1998). Mismatch repair proteins regulate heteroduplex formation during mitotic recombination in yeast. Mol Cell Biol 18, 6225-6237.

Datta, A., Hendrix, M., Lipsitch, M. & Jinks-Robertson, S. (1997). Dual roles for DNA sequence identity and the mismatch repair system in the regulation of mitotic crossing-over in yeast. Proc Natl Acad Sci USA 94, 9757-9762.[Abstract/Free Full Text]

Elliott, B., Richardson, C., Winderbaum, J., Nickoloff, J. A. & Jasin, M. (1998). Gene conversion tracts from double-strand break repair in mammalian cells. Mol Cell Biol 18, 93-101.[Abstract/Free Full Text]

Fonzi, W. A. & Irwin, M. Y. (1993). Isogenic strain construction and gene mapping in Candida albicans. Genetics 134, 717-728.[Abstract/Free Full Text]

Gillum, A. M., Tsay, E. Y. H. & Kirsch, D. R. (1984). Isolation of the Candida albicans gene for orotidine-5'-phosphate decarboxylase by complementation of S. cerevisiae ura3 and E. coli pyrF mutations. Mol Gen Genet 198, 179-182.[Medline]

Gorman, J. A., Chan, W. & Gorman, J. W. (1991). Repeated use of GAL1 for gene disruption in Candida albicans. Genetics 129, 19-24.[Abstract/Free Full Text]

Grigoriev, M. & Hsieh, P. (1997). A histone octamer blocks branch migration of a Holliday junction. Mol Cell Biol 17, 7139-7150.[Abstract]

Hasty, P., Rivera-Perez, J. & Bradley, A. (1991). The length of homology required for gene targeting in embryonic stem cells. Mol Cell Biol 11, 5586-5591.[Medline]

Hoyer, L. L., Payne, T. L. & Hecht, J. E. (1998). Identification of Candida albicans ALS2 and ALS4 and localization of Als proteins to the fungal cell surface. J Bacteriol 180, 5334-5343.[Abstract/Free Full Text]

Kelly, R., Miller, S. M., Kurtz, M. B. & Kirsch, D. R. (1987). Directed mutagenesis in Candida albicans: One-step gene disruption to isolate ura3 mutants. Mol Cell Biol 7, 199-207.[Medline]

Kelly, R., Miller, S. M. & Kurtz, M. B. (1988). One-step gene disruption by cotransformation to isolate double auxotrophs in Candida albicans. Mol Gen Genet 214, 24-31.[Medline]

Kirkpatrick, D. T., Wang, Y. H., Dominska, M., Griffith, J. D. & Petes, T. D. (1999). Control of meiotic recombination and gene expression in yeast by a simple repetitive DNA sequence that excludes nucleosomes. Mol Cell Biol 19, 7661-7671.[Abstract/Free Full Text]

Leung, W.-Y., Malkova, A. & Haber, J. E. (1997). Gene targeting by linear duplex DNA frequently occurs by assimilation of a single strand that is subject to preferential mismatch correction. Proc Natl Acad Sci USA 94, 6851-6856.[Abstract/Free Full Text]

Miyasaki, S. H., White, T. C. & Agabian, N. (1994). A fourth secreted aspartyl proteinase gene (SAP4) and a CARE2 repetitive element are located upstream of the SAP1 gene in Candida albicans. J Bacteriol 176, 1702-1710.[Abstract]

Morschhauser, J., Michel, S. & Staib, P. (1999). Sequential gene disruption in Candida albicans by FLP-mediated site-specific recombination. Mol Microbiol 32, 547-556.[Medline]

Mühlschlegel, F. A. & Fonzi, W. A. (1997). PHR2 of Candida albicans encodes a functional homolog of the pH-regulated gene PHR1 with an inverted pattern of expression. Mol Cell Biol 17, 5960-5967.[Abstract]

Nassif, N. & Engels, W. (1993). DNA homology requirements for mitotic gap repair in Drosophila. Proc Natl Acad Sci USA 90, 1262-1266.[Abstract]

Negritto, M. T., Wu, X., Kuo, T., Chu, S. & Bailis, A. M. (1996). Influence of DNA sequence identity on efficiency of targeted gene replacement. Mol Cell Biol 17, 278-286.[Abstract]

Odds, F. C. (1988). Candida and Candidosis. A Review and Bibliography, 2nd edn. London: Baillière Tindall.

Papadopoulou, B. & Dumas, C. (1997). Parameters controlling the rate of gene targeting frequency in the protozoan parasite Leishmania. Nucleic Acids Res 25, 4278-4286.[Abstract/Free Full Text]

Paques, F. & Haber, J. E. (1999). Multiple pathways of recombination induced by double-stranded breaks in Saccharomyces cerevisiae. Microbiol Mol Biol Rev 63, 349-404.[Abstract/Free Full Text]

Poulter, R. (1990). Classical methods for the genetic analysis of Candida albicans. In The Genetics of Candida , pp. 75-123. Edited by D. R. Kirsch, R. Kelly & M. B. Kurtz. Boca Raton, FL: CRC Press.

Rothstein, R. (1991). Targeting, disruption, replacement, and allele rescue: Integrative transformation in yeast. Methods Enzymol 194, 281-301.[Medline]

Saporito-Irwin, S. M., Birse, C. E., Sypherd, P. S. & Fonzi, W. A. (1995). PHR1, a pH-regulated gene of Candida albicans, is required for morphogenesis. Mol Cell Biol 15, 601-613.[Abstract]

Sentandreu, M., Elorza, M. V., Sentandreu, R. & Fonzi, W. A. (1998). Cloning and characterization of PRA1, a gene encoding a novel pH-regulated antigen of Candida albicans. J Bacteriol 180, 282-289.[Abstract/Free Full Text]

Sherman, F. (1991). Getting started with yeast. Methods Enzymol 194, 3-21.[Medline]

Srikantha, T., Morrow, B., Schröppel, K. & Soll, D. R. (1995). The frequency of integrative transformation at phase-specific genes of Candida albicans correlates with their transcriptional state. Mol Gen Genet 246, 342-352.[Medline]

Stambuk, S. & Radman, M. (1998). Mechanism and control of interspecies recombination in Escherichia coli. I. Mismatch repair, methylation, recombination and replication functions. Genetics 150, 533-542.[Abstract/Free Full Text]

Vuli’c, M., Dionisio, F., Taddei, F. & Radman, M. (1997). Molecular keys to speciation: DNA polymorphisms and the control of genetic exchange in enterobacteria. Proc Natl Acad Sci USA 94, 9763-9767.[Abstract/Free Full Text]

Whelan, W. L. & Magee, P. T. (1981). Natural heterozygosity in Candida albicans. J Bacteriol 145, 896-903.[Medline]

Whelan, W. L. & Soll, D. R. (1982). Mitotic recombination in Candida albicans: recessive lethal alleles linked to a gene required for methionine biosynthesis. Mol Gen Genet 187, 477-485.[Medline]

Whelan, W. L., Partridge, R. M. & Magee, P. T. (1980). Heterozygosity and segregation in Candida albicans. Mol Gen Genet 180, 107-113.[Medline]

Wilson, J. H., Leung, W., Bosco, G., Dieu, D. & Haber, J. E. (1994). The frequency of gene targeting in yeast depends on the number of target copies. Proc Natl Acad Sci USA 91, 177-181.[Abstract]

Wilson, R. B., Davis, D. & Mitchell, A. P. (1999). Rapid hypothesis testing with Candida albicans through gene disruption with short homology regions. J Bacteriol 181, 1868-1874.[Abstract/Free Full Text]

Wilson, R. B., Davis, D., Enloe, B. M. & Mitchell, A. P. (2000). A recyclable Candida albicans URA3 cassette for PCR product-directed gene disruptions. Yeast 16, 65-70.[Medline]

Wu, T. C. & Lichten, M. (1994). Meiosis-induced double-strand break sites determined by yeast chromatin structure. Science 263, 515-518.[Medline]

Received 30 March 2000; revised 30 June 2000; accepted 4 July 2000.