Multilocus Sequence Typing Reveals Intrafamilial Transmission and Microevolutions of Candida albicans Isolates from the Human Digestive Tract

Candida albicans is a human commensal that is also responsiblefor superficial and systemic infections. Little is known aboutthe carriage of C. albicans in the digestive tract and the genomedynamics that occur during commensalisms of this diploid species.The aim of this study was to evaluate the prevalence, diversity,and genetic relationships among C. albicans isolates recoveredduring natural colonization of the digestive tract of humans,with emphasis on Crohn's disease patients who produce anti-yeastantibodies and may have altered Candida sp. carriage. Candidasp. isolates were recovered from 234 subjects within 25 familieswith multiple cases of Crohn's disease and 10 control families,sampled at the oral and fecal sites. Prevalences of Candidasp. and C. albicans carriage were 53.4% and 46.5%, respectively,indicating frequent commensal carriage. No differences in prevalenceof carriage could be observed between Crohn's disease patientsand healthy subjects. Multilocus sequence typing (MLST) of C.albicans isolates revealed frequent colonization of a subjector several members of the same family by genetically indistinguishableor genetically close isolates. These latter isolates differedby loss-of-heterozygosity events at one or several of the MLSTloci. These loss-of-heterozygosity events could be due to eitherchromosome loss followed by duplication or large mitotic recombinationevents between complementary chromosomes. This study was thefirst to jointly assess commensal carriage of C. albicans, intrafamilialtransmission, and microevolution. The high frequency of eachof these events suggests that the digestive tract provides animportant and natural niche for microevolutions of diploid C.albicans through the loss of heterozygosity.

INTRODUCTION

Top

Introduction

Candida albicans is currently the most important opportunisticfungal pathogen of humans, responsible for both superficialand systemic infections (7). Clinical manifestations of C. albicansinfections include superficial candidiasis infections (cutaneouscandidiasis, oropharyngeal candidiasis, and vulvovaginitis)that are frequent but usually benign in immunocompetent hosts.They also include severe infections in hospitalized patients,in particular, candidemia and disseminated candidiasis, thatare associated with high mortality rates. These systemic infectionsoccur in numerous patients with severe underlying diseases orcritical illnesses that need aggressive diagnosis or treatmentprocedures. As a consequence, C. albicans is the leading causeof nosocomial fungal infections (7).

Despite being a fungal pathogen, C. albicans is carried withoutsymptoms by a large fraction of the population. Indeed, C. albicanscolonizes mucosal surfaces of healthy subjects and is consideredto be a component of the normal digestive and genital floras.In this regard, intestinal colonization is recognized as a keycomponent of further development of both superficial and systemicC. albicans infections (8, 9, 24). However, frequency and chronologyof carriage are only partially known, and much of the biologyof C. albicans in the commensal stage remains to be understood.Early molecular epidemiological studies have shown that healthysubjects can be colonized simultaneously or sequentially bydifferent strains of C. albicans, indicating that carriage isa dynamic process (46). During the life of an individual, C.albicans isolates can persist, evolve through minor geneticvariations (referred to as microevolutions), or be replacedby other isolates (16-18).

The genetic mechanisms that underlie these microevolutions andtheir role in the diversification of C. albicans populationsand in the adaptive response to different host environmentshave not been investigated. C. albicans is a diploid organismthat has no known full sexual cycle (2). Several studies indicatethat complementary chromosomes show a high level of allelism,and mitotic recombinations between complementary chromosomesare a probable source of genetic microevolutions (10, 11, 15,47). In addition, the C. albicans genome contains genes thatare homologous to those necessary for mating and meiosis inthe yeast Saccharomyces cerevisiae, and it has been shown thatmating-type-compatible C. albicans diploid strains are ableto mate and form tetraploids (13, 19, 43). However, meioticdivisions have not been observed. Instead, tetraploids can contributea parasexual cycle yielding diploid progenies (1). This parasexualcycle may represent a source of microevolutions through chromosomereassortments due to random distribution and/or mitotic recombination.

Molecular typing methods provide insights into the genetic diversityof a species as well as the mechanisms underlying the acquisitionof such diversity. Molecular epidemiology studies, mainly carriedout through fingerprinting of genomic DNA with the mildly repetitivesequence Ca3, have revealed that commensal and infecting isolatesof C. albicans exhibit a high level of genetic diversity andare distributed within five major genetic groups, or clades(namely, clades I, II, III, E, and SA), with different geographicdistributions (3, 28, 30). However, fingerprinting with theCa3 probe does not provide the resolution necessary to inferthe mechanisms that are at the origin of genetic microevolutions.In contrast, multilocus sequence typing (MLST) can be used forthe overall analysis of the C. albicans population and for thedetection of minute genetic changes that are indicative of evolutionaryprocesses. MLST of C. albicans is a highly discriminatory methodbased on the analysis of nucleotide polymorphisms within thesequences of six or seven PCR-generated 400- to 500-bp internalfragments of housekeeping genes (loci) (5, 6). For each locus,the different sequences are assigned as distinct genotypes (representedby integers), and for each isolate, the combination of the genotypesat each of the sequenced loci defines a profile referred toas a diploid sequence type (DST). Because MLST relies only onnucleotide sequencing, it generates highly standardized datathat can be exchanged through a web-based database (http://calbicans.mlst.net)(4). Importantly, it has been shown by us and others that MLSTgroups C. albicans strains into clades that are superimposablewith those revealed by Ca3 fingerprinting, although analysisof large strain collections suggests the occurrence of new clades(39). Furthermore, MLST provides nucleotide sequence data thatcan be used to understand the genetic processes that underliethe evolutionary dynamics of strains within the overall populationand within clades (40) (M. E. Bougnoux, C. Pujol, S. Morand,C. Bouchier, D. Soll and C. d'Enfert, Abstr. 7th ASM Conf. CandidaCandidiasis, abstr. 72, 2004).

The human digestive tract is probably among the most relevantsources to investigate the natural history of these evolutionaryprocesses. In this regard, in the present study, we have takenadvantage of a collection of C. albicans strains obtained aspart of a large epidemiological study of patients with Crohn'sdisease (CD) and their families (44). CD is a severe inflammatorybowel disease whose etiology is poorly understood. Predispositionto CD is thought to combine genetic and environmental/endoluminalfactors (26). Among these factors, it has been established that50 to 70% of the CD patients produce anti-yeast antibodies;these are antibodies reacting with S. cerevisiae mannan andare named anti-Saccharomyces cerevisiae antibodies (ASCA) (31,35). ASCA-positive subjects represent only 5% of the controlpopulation (27), while they are found in 20% of relatives ofCD patients (37). In patients, ASCA are associated with severeileal forms and are used for differential diagnosis/stratification(22, 45). The endogenous immunogen for ASCA is unknown, butrecent clinical and experimental evidence has shown the productionof such an immunogen by C. albicans under certain growth conditionsincluding pathogenic development (A. Standaert-Vitse, T. Jouault,P. Vandewalle, C. Mille, M. Seddik, B. Sendid, J. M. Mallet,J. F. Colombel, and D. Poulain, unpublished data). Mycologicalsampling was therefore included in the above-mentioned study(44) in order to test the hypothesis that C. albicans couldbe one component of the human endogenous flora triggering ASCAand to evaluate a possible link between C. albicans colonizationand CD.

Here, we have used this collection to define the prevalenceof digestive commensal carriage of Candida spp. and to analyze,using MLST, the potential relationships that exist between C.albicans isolates that are carried by a single individual orby different individuals within the same family. Our resultsshow that the rate of cross-transmission within families washigh. Furthermore, we show that isolates within an individualor a family can differ through microevolution of one or severalof the sequenced MLST loci. Importantly, these microevolutionsare always associated with a loss of heterozygosity (LOH), suggestinga predominant contribution of mitotic recombination events duringthe evolution of commensal isolates.

MATERIALS AND METHODS

Top

Materials and Methods

Population and strains. Candida sp. strains were collected in the framework of a case-controlstudy aimed at analyzing the correlation between Candida carriagein the digestive tract and the presence of ASCA and other geneticand environmental factors associated with CD (12, 44). Thisstudy included 177 subjects from 25 families (i.e., parentsand first and second generations of children) from northernBelgium. Each family included one to seven individuals withCD. There were 79 subjects with CD and 98 unaffected familymembers. Fifty-seven subjects from 10 families with no casesof CD but living in the same geographical area were used ascontrols. A specific mycological sampling procedure was includedin the study. A mouth swab and a stool sample were obtainedfrom each subject and cultured extemporaneously on CHROMagarCandida medium (Bio-Mérieux, Marcy-l'Etoile, France).After growth, for each plate, all colonies exhibiting differentcolors were further identified either by Bichro-Latex-Albicans(Fumouze, Levallois, France) for C. albicans or by the API 32Csystem (Bio-Mérieux, Marcy-l'Etoile, France) for otherspecies. For each sample, one Candida sp. clone was stored at–80°C in Cryo-bille tubes (AES, Combourg, France).C. albicans isolates were then subcultured for DNA extraction.

MLST for Candida albicans. C. albicans isolates were typed using the original MLST schemedescribed previously by Bougnoux et al. (5). Methods for DNApreparation, amplification, and determination of nucleotidesequences have been presented in detail previously (5). Sequenceanalysis was performed using ABI PRISM SeqScape software, version2.0 (Applied Biosystems, CA). Each strain is characterized bya DST resulting from the combination of the genotypes obtainedat the six following loci: ACC1, VPS13, GLN4, ADP1, RPN2, andSYA1. Genetic relatedness between the investigated strains wasevaluated by constructing an unweighted-pair group method witharithmetic averages dendrogram using START software (http://outbreak.ceid.ox.ac.uk),which compares genotypic profiles (14). In addition, each strainwas introduced into the original MLST online database (http://calbicans.mlst.net),which allowed us to compare the strains to those previouslytyped by this method, and deposited in the database (4).

Sequencing of haplotypes. PCR-generated DNA fragments corresponding to loci where microevolutionshad been detected through MLST were cloned into the pCR2.1-TOPOvector (Invitrogen Life Technologies, France). Inserts in oneto eight pCR2.1-TOPO derivatives obtained for each individualcloning procedure were amplified by PCR, and the resulting productswere directly sequenced for the identification of haplotypes.Sequence analysis was performed as described above.

Identification of C. albicans genetic groups. The genetic group of each isolate was defined through a comparisonof the MLST profile of the isolate with those previously determinedfor 24 C. albicans reference strains from group I, 14 strainsfrom group II, 15 strains from group III, 10 strains from groupSA, 9 strains from group E, and 5 outliers, as determined usingDNA fingerprinting with the Ca3 probe (30). Each strain wasassigned to one of the genetic groups that constitute the C.albicans population.

Determination of C. albicans serotype. Serotypes A and B were determined by using the Iatron agglutinationkit (Mitsubishi Kagaku Iatron Inc., Tokyo, Japan) accordingto the manufacturer's instructions. For strains with equivocalagglutination patterns, serotypes were confirmed by phospholipomannanpolyacrylamide gel serotype-specific migration patterns in Westernblots using anti-ß-mannoside antibodies (42).

RESULTS

Top

Results

Prevalence of carriage of Candida spp. and C. albicans in the digestive tract of patients with Crohn's disease and healthy subjects. This study included 177 subjects from 25 families (i.e., parentsand first and second generations of children) from northernBelgium, all including one to seven individuals with CD, and57 subjects from 10 families with no cases of CD but livingin the same geographical area. For each subject, Candida strainswere recovered from the mouth and feces. One hundred twenty-five(53.4%) of the 234 subjects were colonized by one or severalspecies of Candida (111 subjects [88%] were colonized by onlyone species). Fifty-eight subjects (25%) were colonized at onlyone of the two sites studied (mouth [30 subjects] and feces[28 subjects]), whereas 67 subjects (29%) were colonized atboth sites. C. albicans was identified in 88% of the Candidacarriers (109/125), C. glabrata was identified in 9% of thecarriers (12/125), C. parapsilosis was identified in 4% of thecarriers (5/125), and another Candida species (including C.norvegensis, C. sake, C. utilis, C. tropicalis, and C. lusitaniae)was identified in 7% of the carriers (9/125). C. albicans carrierswere identified in all of the 35 families except 2 (one controlfamily and one family with subjects with CD). Overall, the rateof digestive carriage of C. albicans was 46.5%.

Prevalence rates of digestive carriage of Candida spp. and C.albicans in the two different types of families are presentedin Table 1. There was a trend towards an increased prevalenceof carriage of Candida spp. and of C. albicans in the CD families(prevalence ratios of 1.17 and 1.21, respectively). However,this trend was not found to be statistically significant. Prevalencewas not significantly different in individuals with CD (49.3%;39/79) and in healthy members of their families (48%; 47/98),and it was not significant in the 57 members of the 10 controlfamilies (40.3%; 23/57) (Table 1). Similarly, prevalence ofcarriage of Candida spp. and of C. albicans was not found tobe significantly different between ASCA-positive and ASCA-negativeCD subjects (data not shown).

View this table:
[in this window]
[in a new window]
TABLE 1. Prevalence of digestive carriage of Candida spp. and C. albicans among the 234 subjects from the 35 families studied

MLST analysis of C. albicans isolates. Taken together, 166 C. albicans isolates were obtained: 30 wereobtained from subjects colonized in the oral cavity only, 24were obtained from subjects colonized in the intestinal tractonly, and 112 (56 pairs) were obtained from subjects colonizedconcomitantly in the oral cavity and in the intestinal tract.MLST analysis focused on 82 C. albicans isolates obtained from56 different carriers distributed within 19 families (26 carrierswere colonized at both sites). As shown in Table 2, 46 differentDSTs were identified: 26 (57%) corresponded to a single isolate,whereas the remaining 20 were shared by multiple isolates (1by nine isolates, 2 by four isolates, 5 by three isolates, and12 by two isolates). Two DSTs included isolates obtained fromsubjects from different families: DST 61 was shared by six subjectsdistributed between families F1 (one subject), F6 (one subject),F12 (two subjects), and F16 (two subjects), and DST 95 was sharedby three subjects distributed between families F15 (one subject)and F17 (two subjects) (Table 2). The 18 remaining DSTs sharedby several isolates were specific to either one subject or onefamily (Table 2 and Fig. 1).

View this table:
[in this window]
[in a new window]
TABLE 2. Characteristics of C. albicans isolates from the 56 carriers studied

View larger version (13K):
[in this window]
[in a new window]
FIG. 1. Dendrogram of the genetic relationships among 82 isolates of C. albicans based on MLST data. The dendrogram was constructed by using the unweighted-pair group method with arithmetic averages using the START software, which compares genotypic profiles (14). Strains were assigned to clades (I, II, III, III', E, and SA) on the basis of their coclustering with a set of reference C. albicans strains that had been previously typed using the Ca3 probe and MLST. Strains are designated according to the following nomenclature: family_subject_origin_number_zip code. For origin, M indicates mouth and F indicates feces. The linkage distance is indicated at the bottom.

A mixed dendrogram (not shown) was generated using the MLSTdata from reference strains and the 82 isolates from the 19families mentioned above. Since the clades of the referencestrains were known, the 82 isolates could be assigned, on thebasis of their coclustering characteristics, to one of six C.albicans clades that have been previously defined: I, II, III,III', SA, and E (3, 29, 30). Clades III and III' have been definedon the basis of MLST data and include strains that are foundwithin clade III when Ca3 fingerprinting was used (Bougnouxet al., Abstr. 7th ASM Conf. Candida Candidiasis 2004). Twelveisolates (15%) did not group into any of the six clades, 19were grouped into clade I (23%), 4 were grouped into clade II(5%), 11 were grouped into clade III (13%), 5 were grouped intoclade III' (6%), 2 were grouped into clade SA (2.5%), and 29were grouped into clade E (35%). Figure 2 illustrates the clusteringof the 82 isolates in the different clades. In 11 (79%) of the15 families that included at least two carriers of C. albicans,isolates belonged to different genetic groups (two to four differentgroups) (Table 2 and Fig. 2). In the 26 carriers positive bymouth and feces, only 2 (8%) had isolates from two differentgenetic groups (subject 1 in family F1 [isolates C1 from groupIII' and C2 from group E] and subject 30 in family F12 [isolatesC68 from group III and C69 from group E]) (Table 2 and Fig.1).

View larger version (72K):
[in this window]
[in a new window]
FIG. 2. Graphical representation of C. albicans familial carriage and genetic relationships among C. albicans strains. DSTs and clades of C. albicans strains carried at the oral (top) or fecal (bottom) site, zip code area of residence, and family are represented for each carrier whose isolates have been typed by MLST. Zip codes are color coded in order to indicate proximity (e.g., all zip codes between 8310 and 8970 are color coded in green). Subjects with isolates with identical DSTs at both sites are shaded in gray. Strains with identical DSTs and that are shared by two or more subjects are linked by a thick line, while closely related strains of different DSTs (see the text) are linked by a thin line.

The 82 isolates were also subjected to serotyping. Seventy-fourisolates were serotype A, and 8 isolates were serotype B. SerotypeA isolates were present in all clades (Table 2). Serotype Bisolates were found only in clades I and II. However, this mayresult from the small size of the sample.

Analysis of familial C. albicans isolates. In 15 out of the 19 families, C. albicans isolates originatedfrom different members. In these families, 22 (43%) of the 52C. albicans carriers had one isolate with a DST identical tothat of an isolate from another member of the same family, suggestingintrafamilial transmission. Indeed, in 17 instances, the subjectsthat shared C. albicans strains with the same DST lived in thesame zip code area, while in five other instances, they livedin a contiguous zip code area (subjects 5 and 6 from familyF3 lived in zip code area 3800 and 3806, and subjects 15, 17,and 19 from family F6 lived in zip code areas 8650, 8657, and8810) (Table 2 and Fig. 1), thus suggesting likely frequentencounters.

Family F12 provides a good illustration of the situations thatwere observed in this study (Table 2 and Fig. 2). Here, sixC. albicans carriers were studied (subjects 28 to 33), fiveof which were carriers at both sites. Of these carriers, threecarried isolates with identical DSTs at both sites (subjects28, 29, and 31). Two carriers had isolates with identical DSTs(subjects 31 and 32). Three carriers had isolates with DSTsthat differed by one or two loci only (subjects 28, 30, and33). Subject 33 was of particular interest, since the subjectcarries two strains that differ at two loci only. These datasuggested that in this family, colonization by a single isolateat both sites was frequent, that colonization could be associatedwith microevolution as indicated by the occurrence of strainswith closely related DSTs, and that intrafamilial transmissionwas also frequent.

Taken together, the results presented in Table 2 and Fig. 2show 20 carriers with isolates with identical DSTs obtainedfrom both sites (77% of the carriers with isolates obtainedat both sites), 11 examples where two or more individuals inone family carry an isolate with an identical DST (73% of thefamilies with more than one C. albicans carrier), 4 exampleswhere one individual carries isolates with DSTs that differby one or two loci only (67% of the cases where a carrier harborsisolates with different DSTs), and 4 examples where two or moreindividuals within the same family carry isolates with DSTsthat differ by one or two loci only (27% of the families withmore than one C. albicans carrier). For each family, carrierswith identical or microevolved DSTs always resided within closezip code areas. These data suggest that the trend observed infamily F12, i.e., a high frequency of colonization by isolatesof identical or closely related DSTs at both sites and of intrafamilialtransmission, is common within the samples we have studied.

Characterization of the microevolutions at the MLST loci. Of particular interest to us was the identification of isolatesthat differed at one or two of the MLST loci within an individualor a family, since these isolates may provide a hint as to themolecular mechanisms that are involved in the dynamics of thegenome during C. albicans commensalism. We therefore investigatedthe nature of the microevolved genotypes within four pairs ofstrains obtained from a single individual (C5 [DST 372] andC6 [DST 57] from subject 3 in family F2, differing at the ACC1locus; C17 [DST 98] and C18 [DST 99] from subject 9 in familyF4, differing at the ADP1 locus; C78 [DST 373] and C79 [DST313] from subject 33 in family F12, differing at the ACC1 andVPS13 loci; and C161 [DST 84] and C162 [DST 150] from subject44 in family F17, differing at the ACC1 locus).

Sequence analysis of variant genotypes within each of thesefour pairs of isolates showed that variations between isolatesresulted from a LOH at one or several of the polymorphic sitesof one or two loci (Fig. 3A). In two instances, the differencesbetween the two genotypes within a locus were limited to a singleLOH (C5 versus C6, AAC1, nucleotide 29, A/T versus T/T; C18versus C17, ADP1, nucleotide 125, A/G versus G/G) that may resultfrom either a point mutation or a mitotic recombination event.In the three other instances, the divergence resulted from twoor three LOHs that were never separated by a heterozygous polymorphicsite (Fig. 3A).

View larger version (55K):
[in this window]
[in a new window]
FIG. 3. Characteristics of the variant genotypes between pairs of closely related C. albicans isolates. For each pair of isolates, the subjects, family, isolate numbers, and clade are indicated along with the sequence at the polymorphic sites within the microevolved locus. Two statuses can be observed at the investigated loci: a heterozygous state (Het) where at least one of the polymorphic sites is heterozygous (boldface type) and a homozygous state where all polymorphic sites are homozygous. Sites where a loss of heterozygosity has occurred between the two related isolates are shown in underlined boldface type. Panel A presents the data obtained for six pairs of isolates, each obtained from a single subject, while panel B presents the data obtained for four pairs of isolates obtained, each from two subjects within the same family.

For the four microevolved loci (ACC1 locus for subjects 3, 33,and 44 and VPS13 for subject 33), the PCR products that hadbeen sequenced by MLST were subcloned, and inserts in eightplasmids for each product were sequenced in order to definethe haplotypes that when combined would yield the genotype observedby MLST. Analysis of the haplotypes in each of the three pairsof isolates showed that a first haplotype was shared by bothisolates, while the other haplotypes differed between the twostrains at the positions corresponding to the LOH observed byMLST (data not shown). Because LOH always occurred at the polymorphicsites identified within a large collection of strains by MLSTand the distance between the sites of LOH was large (20 to 360bp), it is likely that these LOH events reflect mitotic recombinationevents between complementary chromosomes.

We also investigated four pairs of strains obtained from twoindividuals within the same family (C32 [DST 126] and C34 [DST133] from subjects 14 and 15 in family F6, differing at theACC1 locus; C64 [DST 41] and C68 [DST 314] from subjects 28and 30 in family F12, differing at the SYA1 locus; C64 [DST41] and C79 [DST 313] from subjects 28 and 33 in family F12,differing at the ACC1 locus; and C135 [DST 92] and C142 [DST312] from subjects 39 and 40 in family F15, differing at theADP1 locus). The divergence always resulted from an LOH at onelocus with the replacement of two, three, or six heterozygousnucleotide sites by homozygous sites (Fig. 3B). Analysis ofthe haplotypes in two of the four pairs of isolates (C32/C34and C64/C68) showed a pattern similar to that observed in thepairs described above, i.e., one haplotype shared by the twostrains and two haplotypes each specific for a strain and differingby mutations corresponding to the LOH (data not shown). Again,LOH events were never separated by a heterozygous position andwere located at polymorphic sites, indicating that they werelikely to result from a mitotic recombination event.

DISCUSSION

Top

Discussion

An aim of this study was to evaluate the prevalence, diversity,and genetic relationships among C. albicans isolates recoveredduring natural colonization of the digestive tract of humans.For this purpose, we have screened oral and fecal samples from234 subjects within 35 different families for Candida spp. Theobserved prevalence of commensal carriage (oral and/or fecal)of Candida spp. was 53.4%, indicating that digestive carriageof Candida spp. was very common in our samples. This prevalenceis high in comparison with the data from the literature thatindicate oral and anorectal carriage with a median prevalenceof 24% and 27%, respectively (23). The concomitant analysisof two digestive sites (oral and fecal) used here is likelyto allow a better evaluation of the digestive carriage of Candidaspp. from healthy individuals and may explain our observation.Discrepancies between our results and those from other studiesmay also reflect differences in sampling procedures and thresholdsused to distinguish colonized and uncolonized subjects. Ourresults also showed that few Candida species were identifiedin the digestive tract. A total of 91% of the carriers werecolonized by only one Candida species, with 88% being colonizedby C. albicans. C. glabrata was the second most frequently isolatedspecies, colonizing 9% of the carriers. Because of the largenumber of subjects studied, our data are particularly valuableand emphasize the predominant role of C. albicans in the normaldigestive flora of humans (87.8% of all isolates identified).This contrasts with the observations that have been previouslymade in hospitalized patients. There, the epidemiology of digestivecarriage of Candida spp. is strikingly different, with non-C.albicans isolates representing ca. 40% to 60% of all Candidaisolates (9, 34, 41).

One of the main characteristics of our study was that 25 outof the 35 families included members affected by CD: 79 (34%)of the 234 subjects suffered from CD. Our results do not indicateany significant difference in the prevalence of commensal Candidacarriage between individuals with CD and healthy members oftheir families or members of the 10 control families. Similarly,the distribution of species isolated as well as the high prevalenceof C. albicans were not significantly different in the two groupsof families studied (control families and families with CD members).Therefore, cross-transmission that may be higher in familieswhere members suffer from severe intestinal disorders seemsto have a low impact on the prevalence of commensal Candidacarriage.

Finally, carriage of Candida spp. and C. albicans was not significantlydifferent between ASCA-positive and ASCA-negative CD subjects.However, we have independently found a difference between CDpatients and their healthy relatives and healthy controls intheir ability to mount ASCA and anti-C. albicans mannan antibodyresponses in relation to C. albicans colonization (A. Standaert-Vitse,T. Jouault, P. Vandewalle, C. Mille, M. Seddik, B. Sendid, J.M. Mallet, J. F. Colombel, and D. Poulain, Abstr. 13th UnitedEur. Gastroenterol. Wk., abstr. 159, 2005). This latter observationwould confirm a possible link between tolerance to C. albicansand CD. However, the nature of the alteration in CD patientsunderlying this tolerance and its relation to the immune responseremain to be established.

MLST was used to evaluate the diversity and genetic relationshipsbetween 82 C. albicans isolates recovered during natural colonizationof the digestive tract of 56 subjects within 19 families. Theseisolates could be assigned to 46 unique DSTs, indicating significantredundancy within our sample (see below). Despite this redundancy,we observed DSTs that belonged to almost all of the major cladesthat have been defined through molecular typing using the Ca3probe and/or MLST (3, 29, 30) (Bougnoux et al., Abstr. 7th ASMConf. Candida Candidiasis 2004).

This suggests that the geographic specificity that has beenproposed for the different clades (clades I, II, and III; cladeE; and clade SA, enriched for isolates from North America, Europe,and South Africa, respectively) should be reconsidered. A similarconclusion has been reached recently by Tavanti et al. (39)following the typing of 416 C. albicans isolates by MLST. Becauseour panel of strains, that used by Tavanti et al. (39), andthose used for Ca3 typing have their own geographic bias, thegeographical specificity of clades will have to be reevaluatedwhen the typing of a large and geographically unbiased collectionof isolates is completed. This will be facilitated by the natureof MLST that permits archiving of data generated by differentlaboratories worldwide in a central web resource. Interestingly,however, isolates of clade E and clade I formed the majorityof those characterized in this study. Overall, isolates of cladesE and I represented 58.5% of all isolates, and DSTs belongingto clade E and I represented 46.8% of all DSTs. If intrafamilialtransmission was taken into account by grouping isolates ofidentical or closely linked DSTs into a single class (see below),the predominance of clades E and I remained (53.7%). A similardistribution between clades was observed for CD patients andhealthy subjects. Similarly, intrafamilial transmission wasnot associated with a specific clade, suggesting that isolatesof the different clades did not differ in their ability to betransferred between family members or to colonize individualswith CD.

As mentioned above, a significant number of strains with identicalDSTs were observed in our study (55 isolates corresponding to20 DSTs). Colonization of different sites in an individual bythe same isolate has been reported previously (17, 46). Therefore,it is not surprising that in 19 subjects, two isolates withidentical DSTs were obtained, indicating colonization of thetwo sampled digestive sites by the same or very closely relatedC. albicans isolates. Additionally, we observed 10 instanceswhere isolates with identical DSTs were shared by several individuals(two or three individuals) within a family. Interestingly, theseindividuals always lived in the same household or same zip codearea or in contiguous zip code areas, strongly suggesting intrafamilialtransmission through frequent encounters. Adding to this, weobserved the occurrence of isolates that differed at one ortwo loci only and that were likely to have evolved from a commonancestor in several families (see below). These isolates couldbe harbored by a single individual (five instances) or by severalindividuals (two to four individuals) within the family (fourinstances). Here again, these latter individuals lived neareach other. It is therefore likely that these observations reflectother events of intrafamilial transmission where microevolutionhas occurred in one individual and only one microevolved isolatehas been transmitted or where the ancestor has been transmittedand microevolution has occurred in the recipient individual.It is worth stressing here that sampling was simultaneous forall members of a family, indicating that strains evolved inparallel but differently, depending on the host harboring them.These findings demonstrated that intrafamilial transmissionis frequent and is a potential source of contamination of humans,which is in agreement with previously published data (16, 21).

The observation of strains that differed at only one or twoof the sequenced loci is indicative of microevolutions withinthe digestive tract. The fact that these strains differed throughmicroevolutions is supported by our analysis of the haplotypesat each of the differing loci in pairs of related strains. Inall instances, microevolutions were due to an LOH at one orseveral of the polymorphic sites within the sequenced locus.These LOH events can be explained by point mutations at thepolymorphic sites or by mitotic recombination events betweenthe two original haplotypes, resulting in a full or partialloss of heterozygosity. Although point mutations might accountfor cases where only one polymorphic site has evolved, we stronglyfavor mitotic recombination events in the other cases, becausethe LOH events encompass several polymorphic sites (if not all)within the locus. This hypothesis is supported by previous studiesthat have identified LOH events as a source of microevolutionsin C. albicans (10, 11). However, the extent of these LOH eventsremains to be investigated. In this regard, these closely relatedisolates provide an invaluable resource to analyze the dynamicsof the C. albicans genome that occur naturally during commensalismin the digestive tract. As mentioned above, the occurrence ofisolates with identical DSTs or differing through LOH eventswas frequent in our study, and this may be explained by thefamilial nature of the sampling. While most typing studies ofC. albicans populations have reported a striking diversity withinthe species (6, 25, 38), some previous studies that addressedconfined environments with a high probability of transmissionhave shown a lower diversity but have not evaluated the geneticorigin of this phenomenon (5, 20, 32, 36). On the basis of ourdata, it is likely that these samples also contained a significantproportion of microevolved C. albicans isolates that contributedto reduce the overall diversity of the sample.

In summary, data presented in this study indicate that commensalcarriage of Candida spp. and more specifically C. albicans issimilar in CD patients and healthy subjects. This establishesthat the presence of ASCA is not solely influenced by the Candidaload but probably results from an alteration of the immune responseoccurring in CD patients (Standaert-Vitse et al., Abstr. 13thUnited Eur. Gastroenterol. Wk. 2005). Independently, our datashow that the digestive tract is an important site for commensalismand associated intrafamilial transmission as well as microevolutionsof C. albicans. While it was suggested in a recent study thatLOH at the mating-type-like locus could be the result of antifungalpressure (33), our data suggest that at least at other loci,and in the population studied, LOH is a phenomenon that occursnaturally in the natural history of C. albicans carriage. Futurestudies should aim at detailing the nature and molecular basisof these microevolutions and evaluating their impact on thefitness of C. albicans isolates in the context of the digestivetract.

ACKNOWLEDGMENTS

The technical help of Arnaud Manier and Diana Sharafi for sequencedetermination was greatly appreciated.

This work was supported by the Ministère de la Rechercheet de la Technologie (Programme de Recherche Fondamentale enMicrobiologie, Maladies Infectieuses et Parasitaires-RéseauInfections Fongiques) and Pasteur Genopole-Ile-de-France.

FOOTNOTES

* Corresponding author. Mailing address: Unité Postulante Biologie et Pathogénicité Fongiques, Institut Pasteur, 25 rue du Docteur Roux, 75724 Paris Cedex 15, France. Phone: 33 1 40 61 32 57. Fax: 33 1 45 68 89 38. E-mail: bougnoux{at}pasteur.fr.

REFERENCES

Top