Genetics of Candida albicans - mmbr.asm.org · GENETICS OF C. ALBICANS 227 NATURALHETEROZYGOSITYAND...

MICROBIOLOGICAL REVIEWS, Sept. 1990, p. 226-241 Vol. 54, No. 30146-0749/90/030226-16$02.00/0Copyright © 1990, American Society for Microbiology

Genetics of Candida albicansSTEWART SCHERER' AND PAUL T. MAGEE2*

Department of Microbiology, University of Minnesota School of Medicine, Minneapolis, Minnesota 55455,1 andDepartment of Genetics and Cell Biology, University of Minnesota, St. Paul, Minnesota 551082

INTRODUCTION ................................................. 226NATURAL HETEROZYGOSITY AND GENETIC DIVERSITY.............................................. ...227GENOME OF C. ALBICANS ................................................. 227

Electrophoretic Karyotype ................................................. 227Cloned Genes ................................................. 228Repeated Sequences .................................................. 230Molecular Epidemiology and Taxonomy ................................................. 230

DNA TRANSFORMATION .................................................. 231Transformation Protocols................................................. 231Replication of the Transforming DNA ................................................. 231Directed Chromosomal Mutations ................................................. 231

PARASEXUAL GENETICS ................................................. 232Spheroplast Fusion .................................................. 232Genetic Mapping.................................................. 233

GENETIC INSTABILITY AND PHENOTYPIC VARIATION................................................. 233VIRULENCE DETERMINANTS.................................................. 235DRUG RESISTANCE ................................................. 236

Fluoropyrimidines ................................................. 236Drugs That Target Sterol Pathways ................................................. 237

CANDIDA SPECIES ................................................. 237C. stellatoidea .......................................................... 237Other Medically Important Candida species .................................................. 238

CONCLUSIONS AND PROSPECTS ................................................. 238ACKNOWLEDGMENTS ................................................. 238LITERATURE CITED................................................. 238

INTRODUCTION

Candida albicans is an imperfect yeast which exists as acommensal with a large number of animals (93, 109). Inindividuals with impaired immunity or in patients who havesuffered an insult to their normal microflora, infections withC. albicans or related species such as Candida tropicalis canbe a serious medical problem (21, 65, 79). Life-threateningCandida infections are frequently seen in transplant recipi-ents. In some studies, more than one third of cancer patientsat autopsy had evidence of systemic Candida infections.Candida esophagitis is among the most frequent of theopportunistic infections seen in acquired immunodeficiencysyndrome patients. Vaginal Candida infections are a signif-icant source of morbidity among women of childbearing age.C. albicans also causes oral thrush in infants. The limitednumber of effective and safe antifungal antibiotics exacer-bates these problems. Given the serious nature of diseasescaused by the organism and the role of genetics in ourunderstanding of bacterial pathogenesis, development of thegenetics of C. albicans is an important endeavor.

C. albicans is capable of a yeast-to-hyphal-phase transi-tion (dimorphic transition) and a variety of high-frequencyphenotypic transitions, ranging from differences in colony

* Corresponding author.

morphology to differences in cell shape, surface, and perme-ability. These two properties may be important in pathogen-esis (119) and are of great intrinsic interest as biologicalprocesses (109).

In the past 10 years, work on the genetics and molecularbiology of C. albicans has increased greatly. Beginning withthe demonstration that the organism as usually isolated isdiploid and naturally heterozygous (80, 141), the modern eraof Candida genetics has moved rapidly from the isolation ofmutants and the initial use of parasexual genetics to molec-ular genetics and electrophoretic karyotyping. The result ofthese advances is that a significant amount is becomingknown about the genome and the genetic map of the organ-ism. More importantly, the techniques necessary to gain newknowledge even more rapidly are becoming available. This isin large part a result of the similarity of the organism withSaccharomyces cerevisiae at the molecular level. The tech-niques used for the molecular and genetic manipulation of C.albicans have been the subject of recent reviews (56, 72).At the same time that the genetics of the organism was

becoming more approachable, careful studies of its biologywere leading to new information about some of its charac-teristics and to the reinterpretation of old information. Ittherefore seems appropriate at this time to review thegenetics of C. albicans, both classical and molecular, and theresulting insights into its virulence, its epidemiology, and itsrelationships with other yeasts.

226

on May 25, 2019 by guest

http://mm

br.asm.org/

Dow

nloaded from

http://mmbr.asm.org/

GENETICS OF C. ALBICANS 227

NATURAL HETEROZYGOSITY ANDGENETIC DIVERSITY

The early studies on the genetics of C. albicans involvedmainly the isolation of auxotrophic variants (6, 58). UVirradiation was the mutagen used, and a variety of differentphenotypes was isolated. The major rationale for theseexperiments was to determine whether various phenotypeswere correlated with a change in virulence. From a geneti-cist's point of view, the interest of these studies in that avariety of phenotypes was isolated with some ease and thatthere was a strong similarity to many of the phenotypes of S.cerevisiae mutants such as red adenine auxotrophs. Theseresults rendered most plausible the hypothesis that theorganism was haploid, like the common laboratory strains ofS. cerevisiae, and stimulated efforts to find its sexual phase.The question of the ploidy of C. albicans was raised anew

by Olaiya and Sogin (80), who carried out DNA measure-ments on several strains and found that they containedapproximately the same amount of DNA as did diploid S.cerevisiae. Furthermore, the survival curves after treatmentwith UV light, ethyl methanesulfonate, or nitrosoguanidinefor C. albicans and diploid S. cerevisiae were very similar.The problems with these experiments are, of course, thatone cannot assume similar genome size or DNA repairsystems in making interspecies comparisons. In fact, C.albicans apparently lacks a photoreactivation repair system(98). The question was soon resolved by the genetic demon-stration that C. albicans is heterozygous at several loci andtherefore must be largely diploid.

Strong evidence for a diploid genome was put forward byWhelan et al. (141), who showed that for several clinicalisolates UV irradiation gave a very biased spectrum ofauxotrophs, with methionine requirements predominating.One possible explanation for this was that the strains wereheterozygous and that the treatment induced mitotic cross-ing over, rendering the recessive auxotrophic mutationshomozygous. If both progeny of the crossover event sur-vive, one should be able to find a sectored colony with onepart homozygous auxotroph and the other homozygousprototroph. This prediction was tested by isolating both theauxotrophic and the prototrophic progeny of the putativecrossing-over event and asking whether the prototroph stillgave a biased auxotrophic spectrum. In fact, the spectrumwas random, as would be expected if the original heterozy-gosity had been eliminated by mitotic recombination. Fur-thermore, revertants of the auxotrophic progeny behavedlike the original parent, in that they gave the same auxotrophalmost exclusively. The simplest explanation for these ex-periments is that the organism is diploid. Extending thisobservation, Poulter et al. (87) found that multiply auxo-trophic strains could be serially reverted to give prototrophswhich, when irradiated, gave back double or triple auxo-trophs, indicating that the auxotrophs were produced bymitotic crossing over and that the markers were linked.The extent of natural heterozygosity has been demon-

strated in several ways. As mentioned above, Whelan et al.(138, 141) showed that a large number of clinical isolates areheterozygous for recessive mutations which lead to auxotro-phies. Poulter (84) estimated that approximately 5% of C.albicans isolates are heterozygous for one of the suf muta-tions (leading to inability to reduce sulfite) and that anadditional 5% are heterozygous for a variety of other muta-tions. Since these studies used the observation that a limitedspectrum of auxotrophs was generated by UV irradiation asthe criterion for heterozygosity, the numbers must be re-

garded as a lower limit. Heterozygosity near the centromerewould rarely be revealed in these experiments, and leakymutations might escape detection. So too, of course, wouldtemperature-sensitive mutations (most of the experimentswere done at 34°C) and those which affect a parameter suchas growth on a carbon source other than glucose. Among theother types of heterozygosity detectable by growth pheno-type are resistance to 5-fluorocytosine (5-FC) (17, 135),failure to elaborate an extracellular protease (15), and reces-sive lethal mutations (142). One can therefore assume thatmore than 10% of C. albicans isolates are heterozygous forrecessive mutations. The extent of DNA polymorphism andassociated natural heterozygosity is described below.The existence of recessive lethal mutations was demon-

strated by Whelan and Soll with strain Ca526 (142). Using asuf gene as marker, they demonstrated that the presence orabsence of colonies sectored for prototrophy and auxotro-phy, expected from a mitotic crossover, could be explainedby postulating the existence of recessive lethal mutationslinked to the auxotrophy. No other such mutations havebeen reported, but that may be because it requires a fairlyelaborate genetic analysis to show their existence. If they arecommon, it makes the existence of an independent haploidphase problematic.An important implication of the elevated ploidy of Can-

dida strains is that it provides a significantly greater potentialfor genetic diversity than would be found in a haploid strain.Presently we know of few organisms which totally lack thepossibility of genetic exchange, and our view of C. albicansas an imperfect fungus may well be altered as we find outmore about its biology. However, if it has lost its sexualcycle, then diploidy, heterozygosity, mitotic recombination,and phenotypic instability may be its mode of generating andmaintaining genetic diversity.

GENOME OF C. ALBICANS

Many of the earliest studies of the C. albicans genomefocused on determination of the ploidy of clinical isolates(90). Direct measurements of DNA content by diphenyl-amine assays (90) or flow cytometry (19) gave values near 37fg. These numbers are very close to the value obtained withdiploid S. cerevisiae. C. albicans DNA contains about 0.1%5-methylcytosine during growth in the yeast phase andsignificantly less during mycelial growth (95). Genome sizescalculated by reassociation kinetics (19.6 and 14.8 fg bydifferent methods) agree well with the DNA-per-cell values ifone assumes a diploid genome (90). The range of valuesgiven above corresponds to a haploid genome size (includingmitochondrial DNA) of 14 to 18 megabases (Mb). Examina-tion of the electrophoretic karyotype revealed on pulsed-field gels has provided graphic evidence of the diploid natureof C. albicans.

Electrophoretic Karyotype

C. albicans was one of the first species examined bypulsed-field electrophoresis (66, 71, 78, 116, 117, 132). Itschromosomes were found to be, on average, somewhatlarger than those of S. cerevisiae and were useful for testingparameters for the resolution of molecules in the 1- to 3-Mbrange (82). Aside from the mitochondrial DNA, no small,circular plasmids have been found in C. albicans. Thenatural diversity of the species and its diploid nature com-plicated the determination of the number of distinct chromo-somes. The existing genetic data indicated that several

VOL. 54, 1990


http://mm

br.asm.org/

Dow

nloaded from


228 SCHERER AND MAGEE

FC18 WO-1 .24

ACT1 RDNI

CDC21 - A CT SNC'. L YS 2

ADE2

CHR4

LYS2

L YS? SNCJ DFRI1 CHR?4CHR6DFR1

CHIR6 SNC:;

FIG. 1. C. albicans karyotypes. The electrophoretic karyotypesand genetic maps of strains WO-1.24 (an auxotrophic derivative ofstrain WO-1) and reference strain FC18 are compared. WO-1.24 hasan additional chromosome with a mobility intermediate betweenstandards 4 and 5 plus a much smaller chromosome. The positions ofthe bands are relative order on contour-clamped homogeneouselectric field gels (132) and do not represent absolute mobility.Cloned genes in the central column label bands of similar size inboth strains. Probes listed on the sides highlight differences betweenthe strains. Sources of data are listed in Table 2. Additional data forWO-1.24 are from B. B. Magee (personal communication). For adiscussion of the largest band in the karyotype, see the text. Thesequence labeled "LYS2" does not complement all S. cerevisiaelys2 mutants (B. B. Magee, personal communication).

linkage groups would be found, but these data were notadequate to provide an estimate for the total number ofchromosomes.There is substantial chromosome length polymorphism

when different C. albicans strains are compared (66, 70, 71,78, 116). These polymorphisms may be homozygous orheterozygous. It is the rule, rather than the exception, that inany strain at least one pair of homologous chromosomes canbe resolved from each other by pulsed-field gel electropho-resis. Since molecules must differ in size by at least tens ofkilobases to be resolved by this method, this result wouldsuggest that extensive regions of material were present onone homolog and absent on the other. Superimposed uponthe chromosome length polymorphism is some instability inthe karyotype of strains maintained in the laboratory, vari-ation in ploidy, and the presence of additional, much smallerchromosomes in some strains. Suzuki et al. (127) havereported changes in ploidy and chromosome rearrangementsassociated with changes in morphology. Strain WO-1, thewhite-opaque switching strain, has an additional, muchsmaller chromosome consisting at least in part of sequencesfrom the conventional set and also contains a second rear-

ranged chromosome (Fig. 1). Kelly et al. (49) found that onestrain, SGY129, appeared to be tetrasomic for the URA3-containing chromosome (chromosome 3), with each of twohomologs being present twice.

Without an extensive genetic map to serve. as a guide, amolecular approach was taken to analyze the electrophoretickaryotype and determine which bands represent distinctlinkage groups. An extensive array of cloned single-copysequences of both known and unknown functions was usedto probe Southern blots of separated chromosome-sizedDNA molecules (70). Analysis of these data led to theconclusion that several C. albicans strains have seven pairsof chromosomes. These seven linkage groups would containmore than 13 Mb of DNA. With the addition of the multicopy41-kb mitochondrial genome, there is reasonably good agree-ment with the genome size calculated above.A complication in this analysis is the possibility that the

chromosomes of two distinct linkage groups would have thesame electrophoretic mobility. In strain 1012A, there isclearly an additional large DNA molecule in the electropho-retic karyotype (63). Its mobility is very similar to that ofchromosome 1 in strain FC18 (the largest), and it containsthe rDNA but lacks ACT] (Table 1). Evidence is accumu-lating that the largest band in the electrophoretic karyotypeof reference strain FC18 may consist of two physical linkagegroups more easily resolved in other strains (E. Rustchenko-Bulgac, personal communication). The sizes of the largestchromosomes are not known with sufficient accuracy to usecomparison with the genome size to resolve the issue. Thesedata and the extensive variability in the smaller chromo-somes described above lead one to the view that there is nostandard karyotype for the species.The physical map derived from a small number of refer-

ence strains may, in many instances, be difficult to reconcilewith data obtained from a particular clinical isolate ofinterest. The extensive natural variation in the karyotypemakes it very difficult to imagine an efficient meiotic cycle

for the species. For the purposes of this review, we will usean existing seven-chromosome numbering system as a ref-erence (numbered 1 through 7, largest to smallest [70]) withthe former chromosome 1 designated L for the large chro-mosomes. At present, there are insufficient data to assignunambiguously the many cloned genes derived from thelarge chromosomes into two distinct physical linkagegroups. These data are summarized in Table 1.

Cloned GenesMost of the cloned C. albicans genes have been isolated

through their counterparts in S. cerevisiae. Several ap-proaches have been taken including complementation (27,42, 52, 54, 57, 94), sequence homology (67, 75, 114), and theability of certain C. albicans sequences to confer newphenotypes on Saccharomyces strains (70). Genes cloned inthis fashion range from the highly conserved actin (75) andtubulin (114) genes to genes involved in amino acid biosyn-thesis (94) and sugar utilization (70). The first gene isolated

TABLE 1. Assignment of cloned genes to the karyotype of C. albicans FC18'

Group Size (kb)b DNA probe(s)

LC 3,000 ACTI ADEI CDC3 CDCio GAL] MGLI RDN1 SOR9 TRPI TUB22 2,500 CDC21 ERG7 HEM3 HIS3 HPTI PRAI RAS]3 2,000 ADE2 ILV2 SOR2 URA34 1,800 L YS15 1,500 CAGI ERGIJ6 1,300 ARSI BEN]7 1,200 ARG57 DFRI LEU2a For sources of data, see Tables 2, 3, and 4.b The chromosome sizes are estimates and will vary considerably when different strains are examined.c Group L may consist of two physical linkage groups (see text).

MICROBIOL. REV.


http://mm

br.asm.org/

Dow

nloaded from



Gene (protein)"

TABLE 2. Genes isolated through S. cerevisiae function

Chromosomeb Isolation method'

ACT] (actin)ADEI (CAIR:aspartate ligase)ADE2 (AIR carboxylase)ARG4 (arginosuccinate lyase)BEN]CAGI (G protein, ot subunit)CDC3CDC1OCDC21 (thymidylate synthase)eCHSI (chitin synthase I)DFRJ (dihydrofolate reductase)ERG7 (2,3-oxidosqualene cyclase)ERGll (lanosterol-14a-demethylase)fGAL] (galactokinase)HEM3 (uroporphyrin-I synthase)HIS3 (imidazole glycerol phosphate-dehydratase)HPTJ (hypoxanthine:guanine phosphoribosyl transferase)ILV2 (acetolactate pyruvate lyase)LEU2 (isopropylmalate dehydrogenase)MGLIPRAIRAS]SOR2SOR9STA1 (glucoamylase)TRPJ (PRA isomerase)TUB2 (,3-tubulin)URA3 (orotidine monophosphate decarboxylase)

L Homology SequenceL Complementation Compleme3 Complementation Compleme

Complementation6 New function5 Homology SequenceL Complementation SequenceL Complementation Sequence2 Complementation Sequence

Complementation7 Homology Sequence2 Complementation5 Complementation SequenceL Complementation2 Complementation Phenotype2 Complementation2 Complementation3 Homology Function i7 Complementation PhenotypeL New function2 Homology2 Homology Function i3 New functionL New function

New functionL ComplementationL Homology Sequence3 Complementation Phenotype

Verification method

onts C. albicans adelents C. albicans ade2

e after gene disruption

in S. cerevisiae


in S. cerevisiae


"Gene-enzyme relationships are from Jones and Fink (44). Nomenclature conforms to that used with S. cerevisiae.b Chromosome assignments are as in Table 1.c Isolation methods are sequence homology with a Saccharomyces gene, complementation of existing S. cerevisiae mutants, or the ability of the Candida

sequence to confer a new selectable phenotype on S. cerevisiae.d A, S. Scherer, unpublished data; B, P. Russel and S. Wagner, Abstr. Annu. Meet. Am. Soc. Microbiol. 1989, F80, p. 471; C, B. DiDomenico, personal

communication; D, J. Hicks, personal communication; E, C. Thrash-Bingham and J. Gorman, unpublished data.e CDC21 = TMPI.f ERGII ERG16.

by complementation was URA3, which, like its S. cerevisiaecognate, also functions in Escherichia coli (27). Genescorresponding to the markers in standard Saccharomycestransformation hosts, such as LEU2, HIS3, and TRPJ, were

readily isolated. In some instances it has been possible toverify the identity of these genes by their ability to comple-ment characterized C. albicans mutants; however, the num-

ber of C. albicans mutants with known biochemical defectsremains relatively small. Two genes in this category are theADEJ and ADE2 loci, whose mutants are red. Some clonedgenes have been used to create Candida mutants throughgene disruption methods, and the identity of these sequencesis confirmed on the basis that these mutants have a predict-able phenotype. Examples in this group include LEU2 andURA3 (48, 49). The nucleotide sequences of several clonedgenes, such as those encoding actin and tubulin, have beenobtained and confirm their identities. Table 2 summarizesexisting data on these cloned single-copy DNA sequences.The ability to confer new phenotypes on S. cerevisiae with

Candida genes extends the range of those that might becloned in Saccharomyces strains. Magee et al. used theinability of S. cerevisiae to utilize certain sugars as the basisfor gene isolation based on expression (70). Two clonesisolated through their ability to confer resistance in S.cerevisiae to the antibiotics benomyl and methotrexate were

found to be identical (M. Fling, personal communication).The sequences encoded neither of the targets of those drugs,and their nature remains unknown. Most studies of expres-sion of Candida genes in S. cerevisiae have used high-copy-

number YEp vectors, but it has been demonstrated forURA3 that complementation can be obtained at one copy perhaploid genome (27).Genes such as ADEJ and ADE2, which have played

important roles in the development of the parasexual geneticsystem, are the tools for alignment of the physical andgenetic maps of C. albicans. Essential to the completion ofthis task is the isolation of genes corresponding to Candidaauxotrophs whose biochemical defect is unknown. This willdecrease reliance on anonymous DNA segments and asso-

ciated restriction fragment length polymorphisms (RFLPs)for chromosome assignment of linked genes. Recently, Gos-horn and Scherer (unpublished data) have been able toisolate several Candida biosynthetic genes by complemen-tation in C. albicans. This approach will provide many new

selectable markers for DNA transformation of Candidaspecies and makes possible the isolation of genes affectingfunctions not found in S. cerevisiae. The frequency at whichCandida genes complement S. cerevisiae mutants and thenearly complete set of defined biochemical Saccharomycesmutants (44) should, in many cases, permit rapid identifica-tion of the encoded enzyme. These additional cloned single-copy sequences are listed in Table 3.The sequenced Candida genes have conformed with what

has been learned about the structure of S. cerevisiae genes.In cases when intervening sequences exist, they contain the

highly conserved S. cerevisiae splicing recognition se-

quences (114). There is substantial codon bias in the abun-dantly expressed genes, as is seen with other fungi. Work to

Referencesd

75A54B70DCC111564, 5550, E52, 62, E7057, E70, 94AD42, 48, E7067D7070569411427, 49, 70

VOL. 54, 1990


http://mm

br.asm.org/

Dow

nloaded from



TABLE 3. Additional cloned single-copy sequencesa

Name Chromosome Function Referencesb

ARG57C 7 Complements Candida mutant AARS1 6 ARS 55, BCHR4 4 Unknown 70CHR6 6 Unknown BCHR7 7 Unknown 70LYSId 4 Complements Candida mutant ASNC3 5 Unknown BSNC9 5 Unknown BTEFI Translation factor lot 125TEF2 Translation factor la 125

a Sequences of unknown function were isolated to provide specific DNAprobes for chromosomes 4, 6, and 7 in strain FC18 and the small chromosomein strain WO-1 (SNC3 and SNC9).

b A, A. Goshorn, personal communication; B, B. B. Magee, personalcommunication.

c The arginine gene complements the auxotrophy in strain STN57 (128).d The LYS] gene corresponds to the S. cerevisiae function (44) as well as

the previous Candida mutant designation (36).

date has not provided information on the precise nature of aCandida promoter.Examination of single-copy genes provides an additional

measure of heterozygosity in C. albicans. There has beenone published report of such RFLPs, those found for theURA3 locus by Kelly et al. (48). In examining two strainswhich they had used as recipients for a gene disruptionexperiment, these workers found that the pattern of EcoRIsites was different on the two homologs in each strain.Interestingly, one pattern of sites appeared in both strains,indicating that at least one homolog was similar in the twoCandida isolates.

Repeated SequencesMost of the repeated DNA in C. albicans consists of the

mitochondrial DNA and the rDNA tandem array (Table 4).The number of copies of the rDNA has been estimated bydigestion of chromosomal DNA with a 6-base recognitionenzyme that does not cleave the rDNA (e.g., BamHI) andelectrophoresis on pulsed-field gels. Two bands are obtainedin the 0.5- to 1-Mb range, which most probably derive fromthe rDNA genes on each homolog (B. B. Magee, personalcommunication). This would imply about 40 to 80 copies ofthe 12- to 14-kb repeat unit per haploid genome. Themitochondrial genome has been analyzed in some detail byWills et al. (143, 144). It is a 41-kb circle containing a largeinverted repeat. It is present in lower copy number than isthe rDNA.Candida species contain several dispersed, repeated fam-

ilies of DNA sequences (Table 4). These were sought in

TABLE 4. Repeated sequences

Familya No. of copies Repeat size References(distribution) (kb)

rDNA 40-80 (tandem array) 12-14 29, 69Mitochondrial DNA ca. 30 (mitochondria) 41 142, 14327A ca. 10 (dispersed) ca. 15 105Ca3 ca. 10 (dispersed) 120-122Ca7 (telomeres) 120-122

a The 27A and Ca3 gene families are homologous, but their preciserelationship is not yet established. An additional repeated sequence has beenidentified in C. albicans (16); however, it also may be related to one of thefamilies listed above (see text).

several laboratories as a route to sensitive probes for typingC. albicans strains. Two approaches were taken for theidentification of such sequences. Both involved hybridiza-tion of total genomic DNA probes to clones of C. albicansDNA. In such an experiment, increasing copy number in thegenome leads to increased signals in the final hybridizationexperiment. Scherer and Stevens (106) sought species-spe-cific sequences and probed the same clones in parallel withC. tropicalis DNA to eliminate conserved regions fromstudy. Soll et al. (120-122) used mitochondrial DNA andrDNA sequences as probes to specifically exclude thosehighly repeated' elements.

Several classes of sequences emerged from these efforts(Table 3). One of these families appears to be telomeric (Ca7)and produces new DNA polymorphisms at extremely highrates in some strains. The Ca3 and 27A gene families arespecific to C. albicans and Candida stellatoidea. The 27Afamily produces new DNA polymorphisms in the laboratoryat low but detectable rates. New polymorphisms are oftenseen when drug-resistant mutants are selected. These cloneswere isolated from different C. albicans strains and used toprobe genomic DNA from unrelated strains, so th'eir preciserelationship remains unclear; however, the 27A family ishomologous to the Ca3 family (S. Scherer, unpublishedobservations). Ca3 and Ca7 were isolated in the samelaboratory and are known to be distinct. Because of thechoice of restriction enzymes used, the relationship of anadditional gene family isolated by Cutler et al. (16) to all ofthe repeated sequences described above is unknown. Asimilar analysis of repeated sequences has proven effectivewith C. tropicalis (122).

Molecular Epidemiology and Taxonomy

The extensive DNA polymorphism found in C. albicansand other Candida species has led to the rapid developmentof a variety of DNA-based typing systems for epidemiologicapplications. These methods differ greatly in their resolutionand are therefore best suited to somewhat different tasks.Unlike biochemical tests (133), DNA-based methods canreveal many distinct types through common methodologyrather than a single plus-or-minus result. However, detec-tion of too many types makes the construction of usefulsubgroupings with potential predictive value more difficult.'One method makes use of the high level of chromosome

length polymorphisms and examines strains for karyotypicdifferences. As the karyotypes of C. albicans and C. tropi-calis are reasonably well conserved, this approach also haspotential for species typing (71). This method has provenextremely valuable in discriminating between C. albicansand its closest relative, C. stellatoidea (61).A second approach is simple examination of the fluores-

cence pattern seen after restriction endonuclease cleavage(usually with EcoRI) followed by electrophoresis on a con-ventional agarose gel (105). Even a modest degree of se-quence divergence will result in a strikingly different diges-tion pattern. From the background pattern of bands one caneasily determine which Candida species is present, and therepeated rDNA and mitochondrial DNA add sufficient poly-morphism to distinguish dozens of subtypes within C. albi-cans (76, 124). EcoRI digestion of C. albicans DNA yieldsthree prominent bands derived from the rDNA (69). Themiddle of these three bands has been found in only two sizes,3.7 and 4.2 kb. Strains with the 3.7-kb band predominate;those with the the 4.2-kb band include about 10% of recentclinical isolates (124). The other two fragments are highly

MICROBIOL. REV.


http://mm

br.asm.org/

Dow

nloaded from



polymorphic. This technique has been used to examine therelationship to one another of C. albicans isolates found ininfections of an individual, to study venereal and nosocomialtransmission, and to correlate DNA types with the previ-ously described groupings based on biotyping studies (76,124). Southern blots derived from these gels can be used forsubsequent detection of RFLPs. Infections of an individualare generally clonal in origin, and indistinguishable isolatescan be recovered from sexual partners and from patients inthe same hospital. There is neither strong concordance norrandom association of groups derived by molecular methodsand the earlier phenotypic groups (124).

Several DNA probes have proven useful for RFLP detec-tion in C. albicans. Species can be verified by using therelatively conserved gene encoding actin (1, 75). Strains canbe typed by using the repeated but homogeneous mitochon-drial DNA (81). The most sensitive typing methods make useof the dispersed, repeated sequences described above (23,106, 120-122). These probes will easily distinguish all C.albicans strains that are not epidemiologically related andare capable of detecting polymorphism that arises during thecourse of a single infection. In principle, these probes couldbe used to determine the lineages of cells in an infection inmuch the same way that mosaics are used to analyzedevelopment of multicellular organisms.The lack of a sexual cycle precludes the use of mating and

the production of viable, fertile meiotic products for thedefinition of Candida species. As a result, traditional typingschemes have relied on arrays of biochemical and growthtests to distinguish Candida species (133). The developmentof DNA-based typing schemes has led to a reevaluation ofthe traditional species designations; in most cases thesetaxonomic distinctions have held up well. However, Can-dida parapsilosis appears to include three rather distinctgroupings, one of which contains at least one strain typed asCandida mauritiana (105). C. albicans and C. stellatoideaare the only Candida species which produce germ tubes inserum. For these two species, the only reliable differencewas that C. albicans could utilize sucrose (60). Molecularmethods have led to a reexamination of the relationship of C.albicans to C. stellatoidea, which will be described in detailbelow.

DNA TRANSFORMATION

The DNA transformation system developed for C. albi-cans by Kurtz et al. (54) shares many similarities with thoseused with other fungi, particularly S. cerevisiae. The keys toits development were the availability of genes clonedthrough their S. cerevisiae cognates and mutants of thecorresponding C. albicans genes. Among the few C. albi-cans mutants with well-defined deficiencies are the twoadenine auxotrophs which accumulate a red pigment. Theinitial experiments, using an ADE2 gene isolated by comple-mentation in S. cerevisiae, achieved a transformation fre-quency similar to that seen in the first experiments with S.cerevisiae. As the development of the Candida DNA trans-formation system has, in many ways, paralleled that of the S.cerevisiae system (8), here we will point out the similaritiesand differences in the behavior of these yeasts.

Transformation Protocols

Several very diverse procedures exist for transformationof S. cerevisiae. One closely resembles spheroplast fusionprocedures (37). Osmotically stabilized spheroplasts are

incubated with DNA in the presence of calcium chloride,polyethylene glycol is added, and the cells are plated in aregeneration agar overlay or onto osmotically stabilizedplates. Although a somewhat different protocol was orig-inally described (54), a procedure essentially identical to thatused with S. cerevisiae will work well with C. albicans.

Spheroplast transformation has proven effective withstrains of disparate genetic backgrounds. Although somedifferences in transformation efficiency have been noted,they are not of the magnitude seen with unrelated S. cere-visiae hosts (8). The set of strains available for testingremains limited to those with mutations in certain clonedgenes, but is rapidly increasing. Development of dominantmarkers for the transformation system will permit the trans-formation of the more highly virulent prototrophic clinicalisolates.

In contrast, the lithium procedure (41), which works wellwith a variety of fungi, is largely ineffective with C. albicans.Several other approaches are available for transformation offungal cells; these include electroporation (134) and directtransfer during conjugation with bacteria (33). The efficacyof these techniques with Candida species remains to bedetermined.

Replication of the Transforming DNAThe Saccharomyces transformation system has thus far

proven an excellent model for the modes of replication of thetransforming DNA. The initial experiments by Kurtz et al.(54) involved integrating vectors. A modest degree of non-homologous recombination is seen, but it has not interferedwith the targeting of linear molecules for gene disruptionexperiments (49).

Sequences were soon identified that were capable ofpromoting autonomous replication of the transforming DNA(ARS elements) accompanied by greatly enhanced transfor-mation frequencies. Although their behavior is similar to thatseen in S. cerevisiae, the ARS-containing plasmids arelargely replicated in Candida species as multimers (55). Thissubtle distinction may complicate determination of the struc-ture of the transforming DNA and its recovery into E. coli.Multimeric replication of autonomously replicating plasmidsis seen in other fungal transformation systems (24).Centromeres are genetically defined by segregation pat-

terns in meiosis and cytologically defined through the ap-pearance of metaphase chromosomes. Candida centromeresequences have not yet been described. Whether they havethe simple structure seen in S. cerevisiae (35) or the morecomplex features of Schizosaccharomyces pombe cen-tromeres (31), the lack of a meiotic cycle will complicate thediscrimination of centromeres from other sequences that canenhance mitotic stability of ARS plasmids. CentromericDNA could be distinguished by the fact that dicentricplasmids constructed in vitro are not tolerated in vivo (73).

Directed Chromosomal MutationsHomologous integration of transforming DNA has permit-

ted the development of methods for the construction in vitroof previously unidentified mutants of C. albicans. Thesemolecular approaches make possible the construction ofcertain multiply marked strains that are not readily producedthrough parasexual genetics. Gene disruption proceduresalso avoid additional deleterious mutations that often accom-pany conventional mutagenesis. These techniques, althoughgenerally similar to those used with S. cerevisiae, arecomplicated by the diploid nature of C. albicans.

VOL. 54, 1990


http://mm

br.asm.org/

Dow

nloaded from



Two methods have been described by Kelly et al. tocircumvent this problem. One approach uses mitotic recom-bination induced with low-dose UV light to make an alteredgene homozygous (49). This procedure should be generallyapplicable to both single-copy and repeated sequences, butrequires advance knowledge of the phenotype of the desiredhomozygote. This approach was used to construct mutationsat the URA3 locus. In this case, the cloned URA3 gene wasdisrupted in vitro by insertion of the ADE2 gene into theURA3 coding sequences. A linear fragment from this con-struct with homology to the URA3 locus at both ends wasused to transform an ade2 auxotroph. As a consequence, thecreation of the ura3 mutant resulted in the loss of the othermarker in the transformation host. A second method usescotransformation of a linear altered gene with a selectableARS-containing plasmid (48). Cotransformation in C. albi-cans is quite extensive, and it is relatively easy to identifystrains carrying an appropriately marked gene from a collec-tion of such transformants. This method can be repeated tomutate the other homolog, but is less well suited to con-served gene families, as the second step might not be easilytargeted to the same locus on the other homolog. This typeof approach does not result in the loss of the originalselectable marker in the host and has been used in theisolation of a leu2 mutant.The availability of multiply marked C. albicans hosts and

the corresponding cloned genes permits the use of a differentselectable marker to disrupt each homolog with a differentgene. This procedure has been used by Kurtz and Marrinanto construct a hem3 mutant of C. albicans (57). This methodand the second method described above have the advantageof not requiring advance knowledge of the phenotype of themutant being constructed, except that it must be viable.

PARASEXUAL GENETICS

The first step in the development of the parasexual geneticsystem was the establishment of efficient procedures for theisolation of homozygous mutants. These mutants could ariseby one of at least three mechanisms. The simplest is uncov-ering a preexisting recessive heterozygous mutation bymitotic crossover or gene conversion. The second is muta-genesis of one homolog followed by uncovering of theinduced mutation. A third would be mutagenesis so exten-sive that both copies of the gene are mutated. In fact, nostudies have been done to determine the relative frequenciesof these three events, but analogy with other organismssuggests that they would occur with decreasing frequency aslisted. Because of the lack of a sexual cycle, the introductionof more than one marker usually involves at least two roundsof mutagenesis. A large variety of auxotrophs have beenisolated from C. albicans, but the enzymatic deficiencies ofonly a few are known.The earliest identification of the biochemical basis of a

mutation was the demonstration that red adenine-requiringmutants fell into two complementation groups, one of whichshowed intragenetic complementation while the other didnot (89). The adenine requirement of the first group wasrelieved by high CO2 concentrations. Based on these con-siderations, the mutations were called ade2 and adel, re-spectively, by analogy with S. cerevisiae. An arginine aux-otrophy was characterized by biochemical means as amutation in the ARG4 gene (25), and a chitobiase mutant hasbeen identified (43). Also, a fatty-acid auxotroph has beenreported to have a A9-desaturase deficiency (53). As de-scribed above, several defined mutations have been pro-

duced by molecular approaches. Recently, the defect in apyrimidine auxotroph was shown to be ura3 by complemen-tation with the cloned gene (29).The diploid nature of C. albicans allows one to establish

genetic linkage in a single strain without the need for geneticexchange (87). The simultaneous appearance of two or moremarkers in a single mutagenic treatment (such as exposure toUV light) is evidence for linkage, since they may be ex-pected to have been revealed by a single mitotic crossover.Multiply heterozygous auxotrophs (produced by serial re-version of homozygotes) can be subjected to UV irradiationto induce mitotic crossing over in order to give informationabout linkage. Two markers which often appear together insurvivors can be considered to be linked. However, thissimple analysis may be complicated by gene conversion,which may constitute as much as 10% of the recombinationevents.

Spheroplast Fusion

More extensive information about linkage can be obtainedfrom spheroplast fusion. In this technique, spheroplasts ofcells are fused with polyethylene glycol and regenerated onselective medium. Stable fusion products are relatively rare,and strong genetic selections are required to distinguish theresulting hybrids (commonly called fusants) from the paren-tal strains. The first report demonstrating the usefulness ofthis approach was by Sarachek et al. (101). A series ofreports from workers in other laboratories, using the tech-nique for complementation analysis and demonstration oflinkage, appeared soon thereafter (22, 46, 47, 88).

Spheroplast fusion is not a one-to-one process but in-volves the initial formation of large syncytia, which canapparently either undergo immediate nuclear fusion or be-come metastable heterokaryons (100, 102). These hetero-karyons can then undergo nuclear fusion to become tetra-ploids, they can reduce their nuclear number and segregateparental types, or they can undergo transfer of a limitedamount of genetic material from one nucleus to the other andsegregate recombinants which largely resemble one parent.Sarachek and Henderson (99) demonstrated that prototrophsconstructed by spheroplast fusion sometimes contained anuneven ratio of auxotrophic to prototrophic alleles.

Transfer of less than the full complement of chromosomeswas first postulated by Kakar et al. (47) and later studied bySarachek et al. (99, 103). These observations suggest thatspheroplast fusion is not a simple process; more evidence forthis view is the observation of elevated recombination insegregants from a fusion experiment (47). In these experi-ments, one parent was homozygous for tryptophan auxotro-phy and heterozygous for a linked lysine requirement, whilethe other parent was homozygous for prototrophy for thesetwo genes. Trp- Lys- progeny occurred in segregants fromthe heterokaryons produced. The only way these could havearisen would be via mitotic crossing over on a trp lyschromosome. These experiments cannot determine whetherthe event occurred during the fusion process or in the growthof the culture being used; however, Candida heterozygotesare normally extremely stable.The colonies which arise from protoplast fusion can be

considered to be of two types: heterokaryons, which segre-gate progeny lacking the selected markers (often includingentire parental phenotypes), and tetraploids or near-tetra-ploids, which are quite stable. Mitotic crossing over can beused to examine the stable fusants and extend the linkageinformation significantly. The fact that the fusants are tetra-

MICROBIOL. REV.


http://mm

br.asm.org/

Dow

nloaded from



ploid would be expected to make homozygous segregantsmuch less frequent than those from diploids; however, forunknown reasons, this does not seem to be a problem.Furthermore, the genome of the fusants can be reduced to adiploid (or near-diploid) state by heat shock. Treating thecells at 50 to 51°C for 1 to 2 min causes random loss ofchromosomes (36). An alternative method for reducing thechromosome number in fusants is to use one parent which isheterozygous for the partially dominant 5-FC resistanceallele (140). Seventy percent of resistant variants pickedfrom such hybrids have a diminished DNA content com-pared with the fusant, and about half of these are at or closeto the diploid level. Thus, one can generate a completeparasexual cycle in the organism by either of two methods.

Genetic MappingThe major use of spheroplast fusion to establish linkage

and gene order has come from the work of Poulter andco-workers (36, 86-89). Using mitotic recombination andheat shock, they have defined three linkage groups. Inaddition to these three chromosomes, several other linkagegroups have been defined by mutations which segregate inheat shock experiments independently of the defined chro-mosomes and of each other (E. Rikkerink and P. T. Magee,unpublished observations). The availability of several clonedgenes which correspond to mapped mutations has allowedthe assignment of bands on the electrophoretic karyotype tothe genetically defined chromosomes.The genetic map has been derived from experiments with

only a few strains, and although particular linkages seem tobe widely conserved (86), the variability of the electropho-retic karyotype suggests that considerable difference mayexist between strains. The extent of these differences can bedetermined only with more data. The most interesting com-parisons would be between fresh clinical isolates or strainsthat have especially interesting biological properties, such aselevated virulence or phenotypic instability, and the labora-tory strains being used to construct the genetic map. Ideally,one would like to perform parasexual crosses between thesetypes of strains.The requirement for auxotrophic markers in each parent

renders such experiments very difficult with nonlaboratorystrains. Recently, however, Goshorn and Scherer (29) havesucceeded in isolating a strain carrying a dominant mutationfor resistance to mycophenolic acid, one of the few antibi-otics effective against C. albicans, in a multiply auxotrophicbackground. Fusion of this strain, 1006, with a prototroph,can be selected on the basis of mycophenolic acid resistanceand prototrophy (29). These experiments have cast furtherlight on the nature of the primary fusants, mainly becausemycophenolic acid resistance appears dominant in a

monokaryon but recessive in a heterokaryon. The evidencefor this interesting property is the drug sensitivity of unstablefusion products that segregate parental types. Thus, anyfusion products selected with mycophenolic acid have notspent many generations before nuclear fusion, and pro-longed time in the heterokaryotic state is not necessary forthe production of fusion products. Nevertheless, so-calledexceptional progeny, lacking one or more chromosomesfrom either parent, arise. The data point to a high degree ofcoincidence for loss of different linkage groups. This occurs

only if they are lost from the same parent (A. Goshorn,personal communication). This agrees well with the notionthat a partial set of chromosomes would be transferred to an

intact nucleus in the exceptions. Transfer of chromosomes

between nuclei in heterokaryons has been observed in S.cerevisiae when karl mutants were used to block nuclearfusion (18). The exceptional fusion products may be usefulfor linkage analysis and strain construction in that recessivemarkers can be recovered without treatment with UV orheat shock.

GENETIC INSTABILITY ANDPHENOTYPIC VARIATION

Phenotypic variation (often called phenotypic instabilityor switching) is a phenomenon in C. albicans which involvesa heritable change in colony morphology, cell shape, or

alterations in other properties such as susceptibility toantifungal agents. Most significantly, the ability to undergophenotypic variation is itself a heritable trait.

Typically, strains which undergo a phenotypic transitionwill generate, at a relatively high frequency, variants whichform colonies that are wrinkled, contain raised rays, or haveother morphologies. Although phenotypic variation was

noted by many investigators, the significance of the phenom-enon was not generally recognized until the work of Slutskyet al. from the Soll laboratory (112, 113). These investiga-tions demonstrated the reversibility of the process, obtaineddata on the frequency of the changes, showed that the rangeof colony phenotypes was limited for any given switchingsystem, and provided evidence for the role of UV light as a

stimulator of the transition.The common characteristics of phenotypic transitions are

that they appear at a frequency about 2 orders of magnitudehigher than would be expected for ordinary mutation; theyare reversible, which rules out mitotic crossing over as a

mechanism for the switch; and they can affect a variety ofproperties, ranging from the frequency of mycelium forma-tion to cell shape and metabolism. Typically, they are

signaled by the appearance of colonies of abnormal morphol-ogy, at a frequency of 1% or less on a plate (Fig. 2). The cellsin these abnormal colonies often differ from their parents inthe frequency of mycelium formation (some grow almostexclusively as pseudomycelia) or other morphological char-acteristics. Phenotypic transitions have been observed instrains isolated from patients with active Candida infections;however, there is no strong evidence for the role of theseprocesses in pathogenesis (119, 120).One colony morphology phenotypic transition was studied

genetically by Pomes et al. (84). Thee investigators foundthat a parent with a smooth colony phenotype would yieldrough progeny after UV irradiation. This segregation was

originally attributed to the uncovering of a recessive gene bymitotic crossing over. However, the segregation pattern didnot fit this interpretation, since both rough and smoothcolonies gave rise to the opposite phenotype at high frequen-cies, and one or the other would have had to be homozygousfor the recessive gene and hence stable. The authors latershowed that treatment with the mitotic inhibitor benomyl,which causes chromosome loss via nondisjunction in manyfungi, would also lead to phenotypic instability in theirparent strain. Spheroplast fusion of two rough switchingstrains independently induced by benomyl but descendedfrom the same parent gave a smooth fusion product, indicat-ing that two complementation groups are represented (26).The authors demonstrated by segregant production that theparental auxotrophic markers were still present in the fu-sants. The fact that benomyl can induce phenotypic insta-bility is surprising, since most C. albicans strains appearresistant to the drug as measured by inhibition of growth.

VOL. 54, 1990


http://mm

br.asm.org/

Dow

nloaded from


MICROBIOL. REV.234 SCHERER AND MAGEE

.,..... __. . ........................._ M-z::

FIG. 2. Switching of colony morphology. Shown are a variety of colony types obtained after mild UV treatment of a transformant of strainSGY269. Both variant full colonies and sectors are seen. For additional colony types, see reference 112. Magnification, x4.5.

These results, however, are interesting in the light of thework of Suzuki et al. (127), who used pulsed-field gelelectrophoresis to show that a strain which gave rise toaberrant colonies had very frequent karyotypic rearrange-ments, although no specific pattern correlated exclusivelywith a particular phenotypic state. Similar results have beenobtained by Rustchenko-Bulgac et al. (96), using the samestrain studied by Slutsky et al. (112). A possible effect ofbenomyl would be nondisjunction leading to loss of a re-

pressing function which normally prevents switching. Theobservations that at least two examples of phenotypic insta-bility are recessive (see below) give this model some cre-dence.The transition which has been most carefully studied is the

white-opaque system (2, 3, 51, 91, 113), which was originallyfound in strain WO-1 but has since been observed in at leastthree other isolates. This transition involves a shift from theusual budding yeast shape (white cells) to an elongated,slipper-shaped cell which buds apically and at a character-istic angle in opaque cells (Fig. 3). Opaque cells are more

sensitive than white cells to sulfometuron-methyl, an anti-fungal agent, and to UV irradiation. They express both a

hypha-specific surface antigen and an antigen which appears

to be unique to opaque cells. Scanning electron micrographsshowed that the latter antigen is distributed in a punctatemanner on the cell surface (3). Opaque cells undergo thedimorphic transition only when anchored to a substratum(2). This is not observed with white cells or other C. albicansstrains. The hyphae arise from the center of the elongated

opaque cells, a position where they rarely bud. Variability inthe frequency of hyphal formation among different opaqueclones suggests that they may not be genetically identical,possibly as a consequence of switching. The extensivedifferences between the karyotype of WO-1 and those ofother strains further suggest a connection between chromo-some rearrangements and switching.Because white cells grow faster than opaque cells, the

frequency of the transition cannot be measured by countingthe number of opaque or white colonies in a given popula-tion. However, a Luria-Delbruck fluctuation experiment canbe used to give an accurate determination of the transitionfrequency (91). The opaque-to-white transition occurs atabout 5 x 10-4 per cell division at room temperature instrain WO-1; the white-to-opaque transition takes place at aslightly lower frequency. At higher temperatures, the white-to-opaque shift predominates. At 34°C all opaque cells giverise to white colonies. This shift is programmed by as little as6 to 8 h of exposure to the higher temperature (91).At least two phenotypic transitions have been shown by

spheroplast fusion experiments to be recessive to the stablestate. Goshorn and Scherer (29) have shown that fusions ofstrain 655, a wrinkled-colony-producing strain which givesrise to a variety of morphologies, with a stable smooth strain1006 (see above) leads to nonwrinkled fusants. Interestingly,many of the fusants had phenotypes distinguishable fromthat of either parent. Similar experiments have shown thatfusants of WO-1 (and auxotrophic derivatives of it) withstable strains are stable. Furthermore, heat-shock-induced


http://mm

br.asm.org/

Dow

nloaded from



FIG. 3. Opaque cells of C. albicans. Note the characteristicbudding angle, as well as the size and shape of the mother cell.Opaque cells are typically 15 to 20 ,um long. Magnification, x3,500.

chromosome loss, presumably leading to loss of the chro-mosome or chromosomes encoding a repressor, can lead toregaining of the white-opaque transition. Apparent linkage ofthe switching phenotype with the URA3 gene has suggestedthat one or more of the genes responsible for the switch inWO-1 lie on chromosome 3. These experiments also excludethe repressor function from chromosome 3, demonstratingthat at least two gene products are involved in the control ofswitching (W. Chu, personal communication).

It appears that it will be possible to analyze switchinggenetically, but the complicated nature of the system and ofCandida genetics will make this task challenging. For exam-ple, the fact that switching is recessive will complicateattempts to clone genes involved in the transition, sincerecipient strains will have to be specially constructed to lackthe repression function. The demonstration that at least twogenes are involved in the colony morphology switchingstudied by Nombela and co-workers (26, 84) implies that thesystem may be quite complicated. Finally, it is important toremember that there are potentially two genetic systemsinvolved in any instance of phenotypic instability: the genesinvolved in the transition, and the genes involved in produc-ing the alternate phenotype. The genes responsible for theswitching mechanism may be common to many or all of thephenotypic transitions, and it is only the target genes thatvary.

There are two reports about variation in C. albicansploidy. Suzuki et al. (126, 128) found that one clinical isolatecontained both diploid and tetraploid cells. Electron micros-copy of the culture showed that some diploid cells seemed to

be undergoing endomitosis, while some tetraploids showednuclear structures similar to those of meiosis II in S.cerevisiae. Therefore, the culture seemed to be in a dynamicstate with regard to ploidy. The fact that the tetraploidnature of the cells was reversible made it impossible to testthe hypothesis by genetic means. The relationship of theploidy shift reported by Suzuki et al. (127) to phenotypicinstability is unclear, but the authors draw some analogiesbetween the ploidy shift and the chromosomal rearrange-ments in switching cells.

VIRULENCE DETERMINANTS

Unlike bacteria, pathogenic fungi have not in general beenshown to have specific virulence factors.The strongest dataexist for the capsular polysaccharide of Cryptococcus neo-formans (93). Although C. albicans isolates vary signifi-cantly in virulence in animal models, there are relatively fewgenetic data on individual virulence determinants. Auxotro-phs have been shown to have reduced virulence (74, 108),and a fusion product of complementing auxotrophs wasvirulent as a tetraploid (40). However, these biosyntheticpathways cannot be considered virulence factors per se,because almost any mutant with reduced growth rates invivo might be expected to also have a reduction in virulence.It is also important to show that the presence of the virulencedeterminants correlates with pathogenic species or strains.Odds has suggested that the major candidates for viru-

lence factors are the ability to form hyphae (also called theyeast-to-hyphal-phase dimorphic transition), the ability toresist phagocytosis, the ability to adhere to epithelial cellsurfaces, and the ability to secrete an acid proteinase (79).More recently, an iC3b receptor has been proposed as a C.albicans virulence factor (28). To these one should add theability to grow well at 37°C.One very effective way to test the role of any of these

factors in pathogenicity is to generate two isogenic strains,one lacking the factor via mutation and the other normal, andto compare them in an animal model. Such experiments havebeen extremely difficult with Candida species in the pastbecause of the obligate diploid nature of the organism. Lossof a particular factor would presumably be due to a recessivemutation, and mitotic recombination is the most likely waythat recessive phenotypes appear. However, the product ofa mitotic crossover differs from its parent not only by thehomozygosity of the mutation, but also by the elimination ofany heterozygosity which is centromere distal to the cross-over point. This problem can be circumvented by generatingthe homozygote for the recessive gene and then picking arevertant, making the assumption that only one copy of thegene in question has changed and that no other informationhas been altered. The parent, the mutant, and the revertantare then tested for virulence. Only a few genetic experimentshave been carried out on potential virulence factors inCandida species. These have focused on three of the factorsmentioned above: the ability to form true hyphae, adher-ence, and the secretion of the proteinase.The experiments dealing with the dimorphic transition

have usually involved isolation of a variant which is blockedin the formation of either hyphae or yeast cells (13, 39, 108)and testing its virulence compared with that of the parentorganism. One set of such experiments indicated that anamycelial variant was more pathogenic than a strain(hOG301) unable to make yeast cells, although both wereable to kill mice (108). However, the two strains werederived from different parents, thus vitiating the conclusion.

VOL. 54, 1990


http://mm

br.asm.org/

Dow

nloaded from



ATCC 10261, the parent of hOG301, was more virulent thanthe mutant. In a later series of experiments from anotherlaboratory, hOG301 was found to be avirulent (39). Thedifference may be due to a variation in the size of theinoculum used. It is also important to remember that Can-dida infections are normally seen in hosts with some type ofcompromise, whereas the experimental infections are gener-ally done with large inocula (105 or more cells per animal) inhealthy animals. Interpretation of these experimental infec-tions with regard to human infections requires great caution.In any case, the general conclusion that this particularmutation led to reduced virulence seems clear. Buckley andco-workers isolated a variant which could no longer formgerm tubes (118); this strain was much less virulent than itsparent, and reversion for germ tube formation restored thevirulence (9, 118). A possible interpretation of all theseexperiments is that both the yeast and the hyphal forms arenecessary for efficient pathogenesis in mice. Additionalexperiments are needed to explore the role of the dimorphictransition in virulence.There have been two genetic tests of the role of adhesion

in virulence (10, 64). Calderone et al. (10) showed thatcirulenin-resistant mutants with reduced adherence in vitroare also specifically reduced in virulence in an endocarditismodel. Their mutants have altered mannoproteins (11, 104).Lehrer and his colleagues, using the same mutants in amouse vaginitis model, found that virulence was reduced inproportion to the reduction in adherence (64). However,cirulenin-resistant mutants can have pleiotropic properties(38), and not all are reduced in adherence. These studies hadthe advantage of working with spontaneous mutants unlikelyto have additional genetic defects; a difficulty is the lack of astrong selection to obtain revertants for further study.Most C. albicans strains secrete an aspartyl proteinase

with a very low pH optimum. Because levels of this enzymeseem to be elevated in virulent strains and because antibod-ies against it are found in many candidiasis patients, it wasproposed to be a virulence factor (123). Macdonald and Odds(68) were the first to find that a strain making reduced levelsof the enzyme was reduced in virulence. They showed thatthe variant was more easily phagocytized than the parent aswell. Because there were several differences between theproteinase-deficient variant and the parent, a second groupof workers repeated the experiments with a new set ofstrains (59). They showed that another proteinase-deficientmutant was reduced in virulence compared with the parent,whereas a reversion of the mutation led to regained virulencealong with the capacity to make the enzyme. This experi-ment ruled out the possibility that some of the other changesfound in the mutant, such as a reduced growth rate at 37°C,were responsible for the loss of virulence. Recently, onestrain that generated proteinase-negative mutants wasshown to be heterozygous for the prt mutation by thefrequency with which it generated Prt- derivatives bothspontaneously and after UV irradiation. For reasons that areunclear, the negative mutants reverted with a very highfrequency even under nonselective conditions (15). This mayindicate that the mutations were not in the proteinase geneper se, but in some gene which affected functions in additionto the proteinase. Proteinase mutants are often screened bytheir failure to grow with protein as the sole nitrogen source.As a result, some avirulent, "proteinase" mutants may haveother defects (20) such as the inability to transport peptides.Recently, Lott et al. have reported isolation of a C. albicansgene homologous to an S. cerevisiae aspartyl proteinaseused in that species for a different function (67). Sequence

analysis of the Candida protein and disruption of the clonedgene should easily determine whether the secreted protein-ase has been cloned; however, if the mutants that have beenstudied prove to lie elsewhere in the genome, the role of theproteinase in Candida virulence will become less certain.The proteinase therefore seems to be the best candidate

for a virulence factor so far identified. Pathogenesis is a verycomplex process, and a large number of functions are nodoubt involved. The genetic approach has proven to be themost efficient way to establish the role of individual viru-lence functions in other organisms. The same will no doubtbe true in C. albicans, but the complicated nature of thegenetic system requires that the experiments be carefullyconceived and very well controlled.

DRUG RESISTANCE

C. albicans is naturally resistant to a great variety ofantibiotics, severely limiting the set available for use againstinfections (93, 109). A number of additional antibiotics, notsuitable for use against infection, have proven useful in thelaboratory. As described above, C. albicans strains are quitesensitive to mycophenolic acid, and dominant resistantmutants can be obtained. Recessive nalidixic acid-resistantmutants have been described by Haught and Sarachek (32).Cerulenin-resistant mutants have been characterized bio-chemically (11, 104), but little is known about their genetics.In this section, the genetics of resistance to drugs in clinicaluse will be considered.

Fluoropyrimidines

The mechanism of action of the compounds in this groupdepends on their conversion to 5-fluoro-UMP. This moleculeis an inhibitor of thymidylate synthetase and has additionaltoxic effects after its incorporation into nucleic acids.Among the clinically useful compounds, most is knownabout the genetics of resistance to 5-FC. Recessive mutantshave been characterized extensively at the UMP-pyrophos-phorylase and cytosine deaminase loci by Whelan et al. (136,139). Candida species are also sensitive to 5-fluorouracil and5-fluorouridine, and these drugs have been used to elucidatethe biochemical defects in 5-FC-resistant mutants.Whelan et al. demonstrated extensive natural heterozy-

gosity for 5-FC resistance at a locus controlling UMP-pyrophosphorylase activity. Heterozygotes were distin-guishable from homozygotes for either resistance orsensitivity by their colony size on agar containing 50 ,ug of5-FC per ml (136). They found that of 137 clinical isolates, 78(57%) were homozygous for susceptibility, 51 (37%) wereheterozygous for resistance, and 8 (6%) were homozygousfor resistance (17). It is important to note that no data aregiven about the history of these strains, and some or mostmay have been exposed to the drug during treatment of thepatient from whom they were isolated. Thus, this heterozy-gosity may have been selected and the rate found may not bethe natural rate of occurrence.

S. cerevisiae mutants with dominant 5-FC-resistant muta-tions at several loci have been found (45), and it seemedlikely that such mutants would be obtained in a diploidorganism. Recently, such C. albicans mutants have beenrecovered (26, 29), but their biochemical defects have notyet been determined. In combination with appropriate reces-sive markers, these mutants can be used for selection offusion products with prototrophs, but the frequency of5-FC-resistant strains (and those that can mutate easily to

MICROBIOL. REV.


http://mm

br.asm.org/

Dow

nloaded from



resistance) may make this approach less generally applicablethan the use of the mycophenolic acid-resistant mutantsdescribed above.

C. albicans exhibits some sensitivity to 5-fluoro-oroticacid, and ura3 mutants are resistant to this drug. As a result,it should be possible to select for high-frequency loss ofURA3 function, a technique that has proven quite usefulwith S. cerevisiae (7).

Drugs That Target Sterol Pathways

Nystatin and amphotericin B both bind to ergosterol infungal cell membranes and make treated cells permeable tosmall molecules (34). Despite problems with solubility andtoxicity, these agents are widely used for treatment ofsurface and systemic infections, respectively (93). Resis-tance to these antibiotics in clinical strains is' extremelyinfrequent. Studies of resistance to polyene antibiotics by S.cerevisiae indicate that the major mechanism of resistance ismutation in certain steps in the pathway from lanosterol toergosterol, resulting in the incorporation of precursor sterolsin the cell membranes. In S. cerevisiae, the sterol composi-tion of the inner mitochondrial membrane appears to beimportant for proper function of that organelle (34). C.albicans mutants that accumulate 14-methyl sterols aredefective in the formation of hyphae (110). These factorsmay combine to select against mutants resistant to thepolyene antibiotics. The studies to date indicate that thesituation is quite similar in C. albicans (83). Mutants with asimilar pattern of accumulated sterols generally fail to com-plement, and only recessive mutants have been isolated.Among the most effective antifungal agents is a group of

compounds that target the cytochrome P450 system, whichis required for demethylation steps on the pathway fromlanosterol to ergosterol. These include the drugs micona-zole, ketaconazole, fluconazole, itraconazole, and the as yetunnamed SCH39304 (30, 77, 115, 131). Here the situation inC. albicans appears to be somewhat different from that in S.cerevisiae. There is evidence that Candida nonheme, cy-tochrome P450 mutants are different from their Saccharo-myces counterparts (5). Studies of mutants with mutations inthe heme pathway also reveal differences. Sequences capa-ble of complementing S. cerevisiae hem3 mutations havebeen cloned from C. albicans and used to disrupt the gene inC. albicans. Again, the phenotype was not the same as thatof the corresponding Saccharomyces mutant (57). Theseresults underscore the point that although general features ofCandida molecular biology may be discerned from studies ofmodel systems such as S. cerevisiae, certain critical featuresof the biology will have to be developed by genetic analysisof Candida species themselves.

CANDIDA SPECIES

Because of its importance as a common human pathogen,C. albicans has received more attention than several of theother Candida species. Several Candida species are grownon a large scale for the production of specific enzymes or asa source of feed protein. Among the most widely studied areCandida maltosa and Candida utilis. DNA probes have beenused to document a rare case of fungemia due to C. utilis (1).The alkane-utilizing C. maltosa has been the subject of moreextensive molecular analyses (107). An examination of the5S RNA sequences of various Candida species indicates thatthey may be no more similar to C. albicans than is S.cerevisiae. This is not surprising, as the genus Candida is

C. stellatoidea C. albicans

HJSi- ACTI ADEI SOR9ADE1 CDC21 HIS.:

ADE2CHR4LYS2CHIGC

- DFJ? 1

ADE2LYS2

SR9 -

CHR6 DFRI

FIG. 4. Karyotypes of C. albicans and C. stellatoidea. Thepositions of the bands are the relative order on contour-clampedhomogeneous electric field gels, and the distances shown are notproportional to mobility. Cloned genes in the central column labelbands of similar size in both species. Probes listed on the sideshighlight differences between the species. For example, the assign-ments of ADEI and HIS3 are reversed. C. stellatoidea has fiveadditional smaller chromosomes. Note that the probes for thesechromosomes also label sequences on the standard sized set. Thethree largest of the five extra chromosomes are very similar inmobility on gels. No probe has been found for one of the smallchromosomes. Its existence is inferred from a combination ofphysical and genetic experiments (92). For a discussion of thelargest band in the karyotype, see the text.

something of a taxonomic trash bin for asexual buddingyeasts. Several pathogenic fungi, previously thought to beimperfect, have been reclassified when their sexual cycleswere determined. In this section we would like to drawattention to other members of the Candida group, focusingon those of medical importance.

C. stellatoideaThe close relationship between C. albicans and C. stella-

toidea has been examined by Kwon-Chung et al. in greatdetail at the molecular level (60, 61). Strains typed as C.stellatoidea fall into two groups: type II appears to be simplySuc- variants of C. albicans, whereas type I has moreextensive differences with C. albicans. Although the ge-nomes of C. albicans and type I C. stellatoidea are con-served to the level of individual restriction sites, type Istrains have an extensively rearranged karyotype and sev-eral additional smaller chromosomes when compared withthe typical C. albicans set (Fig. 4). The assignment of C.albicans DNA probes onto the C. stellatoidea electropho-retic karyotype by Rikkerink et al. (92) has revealed thecomplexity of the rearrangements as well as extensive con-servation of linkage relationships. The smaller chromosomesmay be simple fragments of the larger ones or may beassembled from several fragments. The sucrose deficienciesin the two groups complement, whereas those tested withintype I fail to complement. This homogeneity of type I C.stellatoidea isolates extends to RFLPs seen when usingmitochondrial DNA and Ca3 probes. Type I C. stellatoideashares antigenic determinants with C. albicans serotype B,whereas the type II strains appear to be serotype A. Clearly,two distinct Suc- groups have separated from C. albicans inthe recent past.Type I C. stellatoidea strains differ from C. albicans in

several quantifiable phenotypes (60). They are more sensi-

VOL. 54, 1990


http://mm

br.asm.org/

Dow

nloaded from



tive to UV light and cycloheximide, have reduced rates ofgrowth in rich medium, and are less virulent in a mousemodel. In addition, they have the ability to produce thesecreted proteinase at lower pH than C. albicans and have ahighly wrinkled colony morphology reminiscent of thoseproduced by some C. albicans strains. C. stellatoidea israrely isolated from current patient populations, and existingstrains are almost all vaginal isolates. This species might beconsidered either a predecessor now being outcompeted bythe more virulent C. albicans or a highly specialized deriv-ative adapted to growth in the vagina.

Other Medically Important Candida SpeciesMany Candida species are known to cause human infec-

tions. Among those incapable of germ tube formation, C.parapsilosis and C. tropicalis are frequently encountered.Additional molecular and genetic analysis is required todetermine the precise taxonomic relationships of these im-perfect fungi to C. albicans. Sexual cycles have not beendetermined for these pathogenic Candida species.The DNA-based typing schemes described above for C.

albicans have been applied to both of these fungal patho-gens. Species-specific DNA probes have been developed forC. tropicalis (122), and mitochondrial DNA polymorphismshave been used to examine C. parapsilosis (12). DNAfingerprinting methods are readily applied to all of theCandida species.

C. parapsilosis and C. tropicalis have been the subjects ofboth molecular and genetic investigation. Genetic experi-ments suggest that both species are diploid. Natural het-erozygosity has been found in C. parapsilosis (137). C.tropicalis has been the subject of more extensive investiga-tions, in part due to interest in examination of peroxisomebiogenesis. Several genes have been cloned from C. tropi-calis, including those for peroxisomal enzymes (130) and thealkane-inducible cytochrome P450 (97). One approach toadvancing the genetics of these Candida species is the use ofinterspecific parasexual crosses (129). Corner and Poulter(14) have used interspecific complementation with C. albi-cans adenine auxotrophs to assign defects in mutants andhave developed enrichment procedures for auxotrophs byusing inositolless death. It has also been possible to use C.albicans ARS-containing cloning vectors for transformationof C. tropicalis adel and ade2 mutants (D. Sanglard, per-sonal communication).

CONCLUSIONS AND PROSPECTS

In the past several years, the molecular genetics of Can-dida species has undergone several major developments.The parasexual genetic system has developed to a point atwhich it can be effectively applied to biological questions.Physical mapping methods are permitting rapid creation of agenetic map for C. albicans. The development of a DNAtransformation system and allied procedures for chromo-somal mutagenesis has allowed the application of many of thetechniques previously available only in the major fungalgenetic systems to one of the most significant fungal patho-gens. Molecular methods have made great progress in clari-fying both the epidemiology and taxonomy of Candida species.One point not discussed above in connection with the C.

albicans genetic system is a wild-type or reference strain.Individual laboratories have chosen a variety of standardstrains or recent clinical isolates suitable for the problemsthey were addressing. Further development of C. albicans

genetics will require increased integration of these data.Several characteristics would be desirable in such a refer-ence strain. On the molecular level, high-frequency DNAtransformation and a karyotype which resolves the linkagegroups would be desirable. Aneuploidy or unusual levels ofheterozygosity would greatly complicate genetic analysis. Amultiply marked reference strain would facilitate genetics, asmapping procedures for C. albicans are far more efficientwith the markers in cis; however, such a strain would not belikely to exhibit virulence in animals. Clearly, this choice isnot a simple one and is complicated by the recent observa-tion that fundamental properties such as linkage relation-ships can show intraspecific variation.

Interest in the Candida system is driven in large part by itsimportance as a human pathogen. A by-product of this workhas been extensive examination of the diversity within thespecies and its significance. In attempting to learn howCandida species have adapted to life in a mammalian hostand evolved methods to evade or defeat the host defenses,much new knowledge will be gained about more generalquestions of evolution at the molecular level.

ACKNOWLEDGMENTS

We thank W. Chu, B. DiDomenico, M. Fling, J. Gorman, A.Goshom, J. Hicks, R. Kelley, J. Kwon-Chung, B. B. Magee, E.Rikkerink, E. Rustchenko-Bulgac, D. Sanglard, and C. Thrash-Bingham for sharing unpublished results. We are grateful to MartinDworkin and Irene Pech for critical readings of this manuscript.Our work has been supported by Public Health Service grants

A116567 (to P.T.M.) and AI23850 (to S.S.) from the NationalInstitutes of Health.

LITERATURE CITED1. Alsina, A., M. Mason, R. A. Uphoff, W. S. Riggsby, J. M.

Becker, and D. Murphy. 1988. Catheter-associated Candidautilis fungemia in a patient with acquired immunodeficiencysyndrome: species verification with a molecular probe. J. Clin.Microbiol. 26:621-624.

2. Anderson, J., L. Cundiff, B. Schnars, M. X. Gao, I. Mackenzie,and D. R. Soil. 1989. Hypha formation in the white-opaquetransition of Candida albicans. Infect. Immun. 57:458-467.

3. Anderson, J. M., and D. R. Soil. 1987. Unique phenotype ofopaque cells in the white-opaque transition of Candida albi-cans. J. Bacteriol. 169:5579-5588.

4. Baccanari, D. P., R. L. Tansik, S. S. Joyner, M. E. Fling, P. L.Smith, and J. H. Freisheim. 1989. Characterization of Candidaalbicans dihydrofolate reductase. J. Biol. Chem. 264:1100-1107.

5. Bard, M., N. D. Lees, R. J. Barbuch, and D. Sanglard. 1987.Characterization of a cytochrome P450 deficient mutant of Can-dida albicans. Biochem. Biophys. Res. Commun. 147:794-800.

6. Bish, J. T., and A. Sarachek. 1967. Influence of temperature andadenine concentration upon the cultural instability of a redadenine auxotroph of Candida albicans. Mycologia 59:671-686.

7. Boeke, J. D., F. LaCroute, and G. R. Fink. 1984. A positiveselection for mutants lacking orotidine-5'-phosphate decarbox-ylase activity in yeast: 5-fluoro-orotic acid resistance. Mol.Gen. Genet. 197:345-346.

8. Botstein, D., and R. W. Davis. 1982. Principles and practice ofrecombinant DNA research with yeast, p. 607-636. In J.Strathern, E. W. Jones, and J. R. Broach (ed.), The molecularbiology of the yeast Saccharomyces, vol. 2. Cold SpringHarbor Laboratory, Cold Spring Harbor, N.Y.

9. Buckley, H. R., L. Daneo-Moore, J. C. Ahrens, and J. D. Sobel.1986. Isolation of a germ-tube-forming revertant from Candidaalbicans B311V6. Infect. Immun. 53:13-15.

10. Calderone, R. A., R. L. Cihlar, D. D. Lee, K. Hoberg, andW. M. Scheld. 1985. Yeast adhesion in the pathogenesis ofendocarditis due to Candida albicans: studies with adherence-negative mutants. J. Infect. Dis. 152:710-715.

11. Calderone, R. A., and E. Wadsworth. 1988. Characterization of

MICROBIOL. REV.


http://mm

br.asm.org/

Dow

nloaded from



mannoproteins from a virulent Candida albicans strain and itsderived, avirulent strain. Rev. Infect. Dis. 10:S423-S427.

12. Camougrand, N., B. Mila, G. Velours, J. Lazowska, and M.Guerin. 1988. Discrimination between different groups of Can-dida parapsilosis by mitochondrial DNA restriction analyses.Curr. Genet. 90:125-132.

13. Cannon, R. D. 1986. Isolation of a mycelial mutant of Candidaalbicans. J. Gen. Microbiol. 132:2405-2407.

14. Corner, B. E., and R. T. Poulter. 1989. Interspecific comple-mentation analysis by protoplast fusion of Candida tropicalisand Candida albicans adenine auxotrophs. J. Bacteriol. 171:3586-3589.

15. Crandall, M., and J. E. Edwards. 1987. Segregation of pro-teinase-negative mutants from heterozygous Candida albi-cans. J. Gen. Microbiol. 133:2817-2824.

16. Cutler, J. E., P. M. Glee, and H. L. Horn. 1988. Candidaalbicans- and Candida stellatoidea-specific DNA fragment. J.Clin. Microbiol. 26:1720-1724.

17. Defever, K. S., W. L. Whelan, A. L. Rogers, E. S. Beneke,J. M. Veselenak, and D. M. Soil. 1982. Candida albicansresistance to 5-fluorocytosine: frequency of partially resistantstrains among clinical isolates. Antimicrob. Agents Chemo-ther. 22:810-815.

18. Dutcher, S. K. 1981. Internuclear transfer of genetic informa-tion in karl-llKARI-I heterokaryons in Saccharomyces cere-visiae. Mol. Cell. Biol. 1:245-253.

19. Dvorak, J. A., W. L. Whelan, J. P. McDaniel, C. C. Gibson,and K. J. Kwon-Chung. 1987. Flow cytometric analysis of theDNA synthetic cycle of Candida albicans. Infect. Immun.55:1490-1497.

20. Edison, A. M., and M. Manning-Zweerink. 1988. Comparisonof the extracellular proteinase activity produced by a low-virulence mutant of Candida albicans and its wild-type parent.Infect. Immun. 56:1388-1390.

21. Edwards, J. E. 1978. Severe candidal infections. Ann. Intern.Med. 89:91-106.

22. Evans, K. O., A. Adeniji, and D. 0. McClary. 1982. Selectionand fusion of auxotrophic protoplasts of Candida albicans.Antonie van Leeuwenhoek J. Microbiol. 48:169-182.

23. Fox, B. C., H. L. Mobley, and J. C. Wade. 1989. The use of aDNA probe for epidemiological studies of candidiasis in im-munocompromised hosts. J. Infect. Dis. 159:488-494.

24. Gaillardin, C., P. Fournier, F. Badar, B. Kudz, C. Gerbaud,and H. Heslot. 1983. Replication and recombination of 2-pLmDNA in Schizosaccharomyces pombe. Curr. Genet. 7:245-253.

25. Gibbons, G. F., and D. H. Howard. 1986. Arginine auxotrophsof Candida albicans deficient in arginosuccinate lyase. J. Gen.Microbiol. 132:263-268.

26. Gil, C., R. Pomes, and C. Nombela. 1988. A complementationanalysis by parasexual recombination of Candida albicansmorphological mutants. J. Gen. Microbiol. 134:1587-1595.

27. Gilium, A. M., E. T. H. Tsay, and D. R. Kirsch. 1984. Isolationof the Candida albicans gene for orotidine-5'-phosphate decar-boxylase by complementation of S. cerevisiae urea3 and E.coli pyrF mutations. Mol. Gen. Genet. 198:179-182.

28. Gilmore, B. J., E. M. Retsinas, J. S. Lorenz, and M. K.Hostetter. 1988. An iC3b receptor on Candida albicans: struc-ture, function, and correlates for pathogenicity. J. Infect. Dis.157:38-46.

29. Goshorn, A. K., and S. Scherer. 1989. Genetic analysis ofprototrophic natural variants of Candida albicans. Genetics123:667-673.

30. Graybill, J. R. 1989. New antifungal agents. Eur. J. Clin.Microbiol. Infect. Dis. 8:402-412.

31. Hahnenberger, K. M., M. P. Baum, C. M. Polizzi, J. Carbon,and L. Clarke. 1989. Construction of functional artificial mini-chromosomes in the fission yeast Schizosaccharomycespombe. Proc. Natl. Acad. Sci. USA 86:577-581.

32. Haught, M. A., and A. Sarachek. 1985. Transmission andexpression of mutations to nalidixic acid resistance amongproducts of protoplast fusion crosses of Candida albicans.Mutat. Res. 152:15-23.

33. Heinemann, J. A., and G. F. Spraque, Jr. 1989. Bacterial

conjugative plasmids mobilize transfer between bacteria andyeast. Nature (London) 340:205-209.

34. Henry, S. A. 1982. Membrane lipids of yeast: biochemical andgenetic studies, p. 101-158. In J. N. Strathern, E. W. Jones,and J. R. Broach (ed.), The molecular genetics of the yeastSaccharomyces. Cold Spring Harbor Laboratory, Cold SpringHarbor, N.Y.

35. Hieter, P., D. Pridmore, J. H. Hegemann, M. Thomas, R. W.Davis, and P. Philippsen. 1985. Functional selection and anal-ysis of centromeric DNA. Cell 42:913-921.

36. Hilton, C., D. Markie, B. Corner, E. Rikkerink, and R. T.Poulter. 1985. Heat shock induces chromosome loss in theyeast Candida albicans. Mol. Gen. Genet. 200:162-168.

37. Hinnen, A., J. B. Hicks, and G. R. Fink. 1978. Transformationof yeast. Proc. Natl. Acad. Sci. USA 75:1929-1933.

38. Hoberg, K. A., R. L. Cilhar, and R. A. Calderone. 1986.Characterization of cerulenin-resistant mutants of Candidaalbicans. Infect. Immun. 51:102-109.

39. Hubbard, M. J., D. Markie, and R. T. Poulter. 1986. Isolationand morphological characterization of a mycelial mutant ofCandida albicans. J. Bacteriol. 165:61-65.

40. Hubbard, M. J., R. T. Poulter, P. A. Sullivan, and M. G.Shepherd. 1985. Characterization of a tetraploid derivative ofCandida albicans ATCC 10261. J. Bacteriol. 161:781-783.

41. Ito, H., Y. Fakuda, K. Murota, and A. Kimura. 1983. Trans-formation of intact yeast cells treated with alkali cations. J.Bacteriol. 153:163-168.

42. Jenkinson, H. F., G. P. Schep, and M. G. Shepherd. 1988.Cloning and expression of the 3-isopropylmalate dehydroge-nase gene from Candida albicans. FEMS Microbiol. Lett.49:285-288.

43. Jenkinson, H. F., and M. G. Shepherd. 1987. A mutant ofCandida albicans deficient in P-acetylglucosaminidase (chito-biase). J. Gen. Microbiol. 133:2097-2106.

44. Jones, E., and G. R. Fink. 1982. Amino acid and nucleotidemetabolism, p. 181-299. In J. N. Strathern, E. W. Jones, andJ. R. Broach (ed.), The molecular genetics of the yeastSaccharomyces. Cold Spring Harbor Laboratory, Cold SpringHarbor, N.Y.

45. Jund, R., and F. Lacroute. 1970. Genetic and physiologicaspects of resistance to 5-fluoropyrimidines in Saccharomycescerevisiae. J. Bacteriol. 102:607-615.

46. Kakar, S. N., and P. T. Magee. 1982. Genetic analysis ofCandida albicans: identification of different isoleucine-valine,methionine, and arginine alleles by complementation. J. Bac-teriol. 151:1247-1252.

47. Kakar, S. N., M. Partridge, and P. T. Magee. 1983. A geneticanalysis of Candida albicans: isolation of a wide variety ofauxotrophs and demonstration of linkage and complementa-tion. Genetics 104:241-255.

48. Kelly, R., S. M. Miller, and M. B. Kurtz. 1988. One-step genedisruption by cotransformation to isolate double auxotrophs inCandida albicans. Mol. Gen. Genet. 214:24-31.

49. Kelly, R., S. M. Miller, M. B. Kurtz, and D. R. Kirsch. 1987.Directed mutagenesis in Candida albicans: one-step genedisruption to isolate ura3 mutants. Mol. Cell. Biol. 7:199-208.

50. Kelly, R., S. M. Miller, M. H. Lai, and D. R. Kirsch. 1990.Cloning and characterization of the 2,3-oxidosqualene cyclasecoding gene of Candida albicans. Gene 187:177-183.

51. Kennedy, M. J., A. L. Rogers, L. R. Hanselmen, D. R. Soll, andR. J. Yancey. 1988. Variation in adhesion and cell surfacehydrophobicity in Candida albicans white and opaque pheno-types. Mycopathologia 102:149-156.

52. Kirsch, D. R., M. H. Lai, and J. O'Sullivan. 1988. Isolation ofthe gene for cytochrome P450 LlAl (lanosterol 14a-demethy-lase) from Candida albicans. Gene 68:229-237.

53. Koh, T. Y., M. S. Marriott, J. Taylor, and E. F. Gale. 1977.Growth characteristics and polyene sensitivity of a fatty acidauxotroph of Candida albicans. J. Gen. Microbiol. 102:105-110.

54. Kurtz, M. B., M. W. Cortelyou, and D. R. Kirsch. 1986.Integrative transformation of Candida albicans using a clonedCandida ADE2 gene. Mol. Cell. Biol. 6:142-149.

55. Kurtz, M. B., M. W. Cortelyou, S. M. Miller, M. Lai, and D. R.

VOL. 54, 1990


http://mm

br.asm.org/

Dow

nloaded from



Kirsch. 1987. Development of autonomously replicating plas-mids for Candida albicans. Mol. Cell. Biol. 7:209-217.

56. Kurtz, M. B., D. R. Kirsch, and R. Kelly. 1988. The moleculargenetics of Candida albicans. Microbiol. Sci. 5:58-63.

57. Kurtz, M. B., and J. Marrinan. 1989. Isolation of hem3mutants from Candida albicans by sequential gene disruption.Mol. Gen. Genet. 217:47-52.

58. Kwon-Chung, K. J., and W. B. Hill. 1970. Studies on the pinkadenine-deficient strains of Candida albicans. I. Cultural andmorphological characteristics. Saboraudia 8:48-55.

59. Kwon-Chung, K. J., D. Lehman, C. Good, and P. T. Magee.1985. Genetic evidence for role of extracellular proteinase invirulence of Candida albicans. Infect. Immun. 49:571-578.

60. Kwon-Chung, K.-J., W. S. Riggsby, R. A. Uphoff, J. B. Hicks,W. L. Whelan, E. Reiss, B. B. Magee, and B. L. Wickes. 1989.Genetic differences between type I and type II Candidastellatoidea. Infect. Immun. 57:527-532.

61. Kwon-Chung, K.-J., B. L. Wickes, and W. G. Metz. 1988.Association of electrophoretic karyotype with virulence formice. Infect. Immun. 56:1814-1819.

62. Lai, M. H., and D. R. Kirsch. 1989. Nucleotide sequence ofcytochrome P450 LlAl (lanosterol 14 a-demethylase) fromCandida albicans. Nucleic Acids Res. 17:804.

63. Lasker, B. A., G. F. Carle, G. S. Kobayashi, and G. Medoff.1989. Comparison of the separation of Candida albicans chro-mosome-sized DNA by pulsed-field gel electrophoresis tech-niques. Nucleic Acids Res. 17:3783-3793.

64. Lehrer, N., E. Segal, R. L. Cihlar, and R. A. Calderone. 1986.Pathogenesis of vaginal candidiasis: studies with a mutantwhich has reduced ability to adhere in vitro. J. Med. Vet.Mycol. 24:127-131.

65. Lerner, C. W., and M. L. Tapper. 1984. Opportunistic infec-tions complicating acquired immune deficiency syndrome.Medicine (Baltimore) 63:155-164.

66. Loft, T. J., P. Boiron, and E. Reiss. 1987. An electrophoretickaryotype for Candida albicans reveals large chromosomes inmultiples. Mol. Gen. Genet. 209:170-174.

67. Loft, T. J., L. S. Page, P. Boiron, J. Benson, and E. Reiss. 1989.Nucleotide sequence of the Candida albicans aspartyl protein-ase gene. Nucleic Acids Res. 17:1779.

68. Macdonald, F., and F. C. Odds. 1983. Virulence for mice of a

proteinase-secreting strain of Candida albicans and a protein-ase-deficient mutant. J. Gen. Microbiol. 129:431-438.

69. Magee, B. B., T. D. D'Souza, and P. T. Magee. 1987. Ribo-somal DNA restriction fragment length polymorphism can beused to identify species and biotypes of medically importantyeasts. J. Bacteriol. 169:1639-1643.

70. Magee, B. B., Y. Koltin, J. Gorman, and P. T. Magee. 1988.Assignment of cloned Candida albicans genes to bands on theelectrophoretic karyotype. Mol. Cell. Biol. 8:4721-4726.

71. Magee, B. B., and P. T. Magee. 1987. Electrophoretic karyo-types and chromosome numbers in Candida species. J. Gen.Microbiol. 133:425-430.

72. Magee, P. T., E. H. A. Rikkerink, and B. B. Magee. 1988.Methods for the genetics and molecular biology of Candidaalbicans. Anal. Biochem. 175:361-372.

73. Mann, C., and R. W. Davis. 1983. Instability of dicentricplasmids in yeast. Proc. Natl. Acad. Sci. USA 80:228-232.

74. Manning, M., C. B. Snoddy, and R. A. Fromtling. 1984.Comparative pathogenicity of auxotrophic mutants of Candidaalbicans. Can. J. Microbiol. 30:31-35.

75. Mason, M. M., B. A. Lasker, and W. S. Riggsby. 1987.Molecular probe for identification of medically important Can-dida species and Torulopsis glabrata. J. Clin. Microbiol.25:563-566.

76. Matthews, R., and J. Burnie. 1989. Assessment of DNAfingerprinting for rapid identification of outbreaks of systemiccandidiasis. Br. Med. J. 298:354-357.

77. McIntyre, K. A., and J. M. Galgiani. 1989. In vitro suscepti-bilities of yeasts to a new antifungal triazole, SCH39304:effects of test conditions and relation to in vivo efficacy.Antimicrob. Agents Chemother. 33:1095-1100.

78. Merz, W. G., C. Connelly, and P. Hieter. 1988. Variation of

electrophoretic karyotype among clinical isolates of Candidaalbicans. J. Clin. Microbiol. 26:842-845.

79. Odds, F. C. 1987. Candida infections: an overview. Crit. Rev.Microbiol. 15:1-5.

80. Olaiya, A. F., and S. J. Sogin. 1979. Ploidy determination ofCandida albicans. J. Bacteriol. 140:1043-1049.

81. Olivo, P. D., E. J. McManus, W. S. Riggsby, and J. M. Jones.1987. Mitochondrial DNA polymorphism in Candida albicans.J. Infect. Dis. 156:214-215.

82. Olson, M. V. 1989. Separation of large DNA molecules bypulsed-field gel electrophoresis. A review of the basic phenom-enology. J. Chromatogr. 470:377-383.

83. Pesti, M., and L. Ferenczy. 1982. Protoplast fusion of hybridsof Candida albicans. Several mutants differing in nystatinresistance. J. Gen. Microbiol. 128:123-128.

84. Pomes, R., C. Gil, and C. Nombela. 1985. Genetic analysis ofCandida albicans morphological mutants. J. Gen. Microbiol.131:2107-2113.

85. Poulter, R. T. 1987. Natural auxotrophic heterozygosity inCandida albicans. Crit. Rev. Microbiol. 15:97-101.

86. Poulter, R. T., and V. Hanrahan. 1983. Conservation of geneticlinkage in nonisogenic isolates of Candida albicans. J. Bacte-riol. 156:157-162.

87. Poulter, R. T., V. Hanrahan, K. Jeffrey, D. Markie, M. G.Shepherd, and P. A. Sullivan. 1982. Recombination analysis ofnaturally occurring diploid Candida albicans. J. Bacteriol.152:969-975.

88. Poulter, R. T., K. Jeffery, M. J. Hubbard, M. G. Shepherd, andP. A. Sullivan. 1981. Parasexual genetic analysis of Candidaalbicans by spheroplast fusion. J. Bacteriol. 146:833-840.

89. Poulter, R. T., and E. H. Rikkerink. 1983. Genetic analysis ofred, adenine-requiring mutants of Candida albicans. J. Bacte-riol. 156:1066-1077.

90. Riggsby, W. S., L. J. Torres-Bauza, J. W. Wills, and T. M.Townes. 1982. DNA content, kinetic complexity, and the ploidyquestion in Candida albicans. Mol. Cell. Biol. 2:853-862.

91. Rikkerink, E. H., B. B. Magee, and P. T. Magee. 1988. Opaque-white phenotype transition: a programmed morphological transi-tion in Candida albicans. J. Bacteriol. 170:895-899.

92. Rikkerink, E. H., B. B. Magee, and P. T. Magee. 1990. Extrachromosomes and gene duplication in Candida stellatoidea.Infect. Immun. 58:949-954.

93. Rippon, J. W. 1982. Medical mycology, 2nd ed. The W. B.Saunders Co., Philadelphia.

94. Rosenbluh, A., M. Mevarech, Y. Koltin, and J. A. Gorman.1985. Isolation of genes from Candida albicans by complemen-tation in Saccharomyces cerevisiae. Mol. Gen. Genet. 200:500-502.

95. Russell, P. J., J. A. Welch, E. M. Rachlin, and J. A. McCloskey.1987. Different levels of DNA methylation in the yeast andmycelial forms of Candida albicans. J. Bacteriol. 169:4393-4395.

96. Rustechenko-Bulgac, E. P., F. Sherman, and J. B. Hicks. 1990.Chromosomal rearrangements associated with morphologicalmutants provide a means for genetic variation of Candidaalbicans. J. Bacteriol. 172:1276-1283.

97. Sanglard, D., C. Chen, and L. C. Loper. 1987. Isolation of thealkane inducible cytochrome P450 (P450alk) gene from theyeast Candida tropicalis. Biochem. Biophys. Res. Commun.144:251-257.

98. Sarachek, A., and J. A. Brammer. 1976. Differentiation ofpathogenic Candida species by their recovery characteristicsfollowing ultraviolet irradiation. Antonie van Leeuwenhoek J.Microbiol. 42:165-180.

99. Sarachek, A., and L. A. Henderson. 1988. Variations forsusceptibilities to ultraviolet induced cellular inactivation andgene segregation among protoplast fusion hybrids of Candidaalbicans. Cytobios 55:171-184.

100. Sarachek, A., and D. D. Rhoads. 1983. Effects of growthtemperatures on plating efficiencies and stabilities of hetero-karyons of Candida albicans. Mycopathologia 83:87-95.

101. Sarachek, A., D. D. Rhoads, and R. H. Schwartzhoff. 1981.Hybridization of Candida albicans through fusion of proto-plasts. Arch. Microbiol. 129:1-8.

MICROBIOL. REV.


http://mm

br.asm.org/

Dow

nloaded from



102. Sarachek, A., and D. A. Weber. 1984. Temperature-dependentinternuclear transfer of genetic material in heterokaryosis ofCandida albicans. Curr. Genet. 8:181-187.

103. Sarachek, A., and D. A. Weber. 1986. Segregant-defectiveheterokaryons of Candida albicans. Curr. Genet. 10:685-693.

104. Saxena, A., C. F. Hammer, and R. L. Cihlar. 1989. Analysis ofmannans of two relatively avirulent mutant strains of Candidaalbicans. Infect. Immun. 57:413-419.

105. Scherer, S., and D. A. Stevens. 1987. Application of DNAtyping methods to epidemiology and taxonomy of Candidaspecies. J. Clin. Microbiol. 25:675-679.

106. Scherer, S., and D. A. Stevens. 1988. A Candida albicansdispersed, repeated gene family and its epidemiologic applica-tions. Proc. Natl. Acad. Sci. USA 85:1452-1456.

107. Schunck, W. H., E. Kargel, B. Gross, B. Wiedmann, S.Mauersberger, K. Kopke, U. Kiessling, M. Strauss, M. Gaestel,and H. G. Muller. 1989. Molecular cloning and characteriza-tion of the primary structure of the alkane hydroxylatingcytochrome P-450 from the yeast Candida maltosa. Biochem.Biophys. Res. Commun. 161:843-850.

108. Shepherd, M. G. 1985. Pathogenicity of morphological andauxotrophic mutants of Candida albicans in experimentalinfections. Infect. Immun. 50:541-544.

109. Shepherd, M. G., R. T. Poulter, and P. A. Sullivan. 1985.Candida albicans: biology, genetics, and pathogenicity. Annu.Rev. Microbiol. 39:579-614.

110. Shimokawa, O., Y. Kato, and H. Nakayama. 1986. Accumula-tion of 14-methyl sterols and defective hyphal growth inCandida albicans. J. Med. Vet. Mycol. 24:327-336.

111. Singer, S. C., C. A. Richards, R. Ferone, D. Benedict, and P.Ray. 1989. Cloning, purification, and properties of Candidaalbicans thymidylate synthase. J. Bacteriol. 171:1372-1378.

112. Slutsky, B., J. Buffo, and D. R. Soll. 1985. High-frequencyswitching of colony morphology in Candida albicans. Science230:666-669.

113. Slutsky, B., M. Staebell, J. Anderson, L. Risen, M. Pfaller, andD. R. Soll. 1987. "White-opaque transition": a second high-frequency switching system in Candida albicans. J. Bacteriol.169:189-197.

114. Smith, H. A., H. S. Allaudeen, M. H. Whitman, Y. Koltin, andJ. A. Gorman. 1988. Isolation and characterization of a -

tubulin gene from Candida albicans. Gene 63:53-63.115. Smith, K. J., D. W. Warnock, C. T. Kennedy, E. M. Johnson,

V. Hopwood, J. Van Cutsem, and H. Vanden Bossche. 1986.Azole resistance in Candida albicans. J. Med. Vet. Mycol.24:133-144.

116. Snell, R. G., I. F. Hermans, R. J. Wilkins, and B. E. Corner.1987. Chromosomal variations in Candida albicans. NucleicAcids Res. 15:3625.

117. Sneil, R. G., and R. J. Wilkins. 1986. Separation of chromo-somal DNA molecules from C. albicans by pulsed field gelelectrophoresis. Nucleic Acids Res. 14:4401-4406.

118. Sobel, J. D., G. Muller, and H. R. Buckley. 1984. Critical roleof germ tube formation in the pathogenesis of candidal vagini-tis. Infect. Immun. 44:576-580.

119. Soll, D. R. 1988. High-frequency switching in Candida albicansand its relations to vaginal candidiasis. Am. J. Obstet. Gyne-col. 154:997-1001.

120. Soll, D. R., R. Galask, S. Isley, T. V. Rao, D. Stone, J. Hicks,J. Schmid, K. Mac, and C. Hanna. 1989. Switching of Candidaalbicans during successive episodes of recurrent vaginitis. J.Clin. Microbiol. 27:681-690.

121. Soll, D. R., C. J.Langn, J. McDowell, J. Hicks, and R.Galask. 1987. High-frequency switching in Candida strains iso-lated from vaginitis patients. J. Clin. Microbiol. 25:1611-1622.

122. Soll, D. R., M. Staebell, C. Langtim, M. Pfaller, J. Hicks, andT. V. Gopala Rao. 1988. Multiple Candida strains in the courseof a single systemic infection. J. Clin. Microbiol. 26:1448-1459.

123. Staib, F. 1969. Proteolysis and pathogenicity of Candida albi-cans strains. Mycopathol. Mycol. Appl. 37:345-348.

124. Stevens, D. A., F. C. Odds, and S. Scherer. 1990. Application ofDNA typing methods to Candida albicans epidemiology andcorrelations with phenotype. Rev. Infect. Dis. 12:258-266.

125. Sundstrom, P., D. Smith, and P. S. Sypherd. 1990. Sequenceanalysis and expression of the two genes for elongation factorla from the dimorphic yeast Candida albicans. J. Bacteriol.172:2036-2045.

126. Suzuki, T., T. Kanbe, T. Kuroiwa, and K. Tanaka. 1986.Occurrence of a ploidy shift in a strain of the imperfect yeastCandida albicans. J. Gen. Microbiol. 132:443-453.

127. Suzuki, T., I. Kobayashi, T. Kanbe, and K. Tanaka. 1989. Highfrequency variation of colony morphology and chromosomereorganization in the pathogenic yeast Candida albicans. J.Gen. Microbiol. 135:425-434.

128. Suzuki, T., S. Nishibayashi, T. Kuroiwa, T. Kanbe, and K.Tanaka. 1982. Variance of ploidy in Candida albicans. J.Bacteriol. 152:893-896.

129. Suzuki, T., A. L. Rogers, and P. T. Magee. 1986. Inter- andintra-species crosses between Candida albicans and Candidaguillermondii. Yeast 2:53-58.

130. Szabo, L. J., G. M. Small, and P. B. Lazarow. 1989. Thenucleotide sequence of POX18, a gene encoding a smalloleate-inducible peroxisomal protein from Candida tropicalis.Gene 75:119-126.

131. Van't Wout, J. W., H. Mattie, and R. van Furth. 1989.Comparison of the efficacies of amphotericin B, fluconazole,and itraconazole against a systemic Candida albicans infectionin normal and neutropenic mice. Antimicrob. Agents Chemo-ther. 33:147-151.

132. Vollrath, D., and R. W. Davis. 1987. Resolution of DNAmolecules greater than 5 megabases by contour-clamped ho-mogeneous electric fields. Nucleic Acids Res. 15:7865-7876.

133. Warnock, D. W. 1984. Typing of Candida albicans. J. Hosp.Infect. 5:244-252.

134. Weaver, J. C., G. I. Harrison, J. G. Bliss, J. R. Mourant, andK. T. Powell. 1988. Electroporation: high frequency of occur-rence of a high permeability state in erythrocytes and intactyeast. FEBS Lett. 229:30-34.

135. Whelan, W. L., E. S. Beneke, A. L. Rogers, and D. R. Soll.1981. Segregation of 5-fluorocytosine resistant variants byCandida albicans. Antimicrob. Agents Chemother. 29:726-729.

136. Whelan, W. L., and D. Kerridge. 1984. Decreased Activity ofUMP pyrophosphorylase associated with resistance to 5-fluo-rocytosine in Candida albicans. Antimicrob. Agents Chemo-ther. 26:570-574.

137. Whelan, W. L., and K. J. Kwon-Chung. 1988. Auxotrophicheterozygosities and the ploidy of Candida parapsilosis andCandida krusei. J. Med. Vet. Mycol. 26:163-171.

138. Whelan, W. L., and P. T. Magee. 1981. Natural heterozygosityin Candida albicans. J. Bacteriol. 145:896-903.

139. Whelan, W. L., D. Markie, and K. J. Kwon-Chung. 1986.Complementation analysis of resistance to 5-fluorocytosine inCandida albicans. Antimicrob. Agents Chemother. 29:726-729.

140. Whelan, W. L., D. Markie, K. G. Simpkin, and R. M. Poulter.1985. Instability of Candida albicans hybrids. J. Bacteriol.161:1131-1136.

141. Whelan, W. L., R. M. Partridge, and P. T. Magee. 1980.Heterozygosity and segregation in Candida albicans. Mol.Gen. Genet. 180:107-113.

142. Whelan, W. L., and D. R. Soil. 1982. Mitotic recombination inCandida albicans: recessive lethal alleles linked to a generequired for methionine biosynthesis. Mol. Gen. Genet. 187:477-485.

143. Wills, J. W., B. A. Lasker, K. Sirotkin, and W. S. Riggsby.1984. Repetitive DNA of Candida albicans: nuclear and mito-chondrial component. J. Bacteriol. 157:918-924.

144. Wills, J. W., W. B. Troutman, and W. S. Riggsby. 1985.Circular mitochondrial genome of Candida albicans contains alarge inverted duplication. J. Bacteriol. 164:7-13.

VOL. 54, 1990


http://mm

br.asm.org/

Dow

nloaded from


Genetics of Candida albicans - mmbr.asm.org · GENETICS OF C. ALBICANS 227 NATURALHETEROZYGOSITYAND...

Documents

Transcript of Genetics of Candida albicans - mmbr.asm.org · GENETICS OF C. ALBICANS 227 NATURALHETEROZYGOSITYAND...