JOURNAL OF CLINICAL MICROBIOLOGY, Oct. 1994, p. 2364-2371 Vol. 32, No. 100095-1137/94/$04.00+0Copyright © 1994, American Society for Microbiology

Molecular Typing of Cryptococcus neoformans Serotype DClinical Isolates

FRAN(COISE DROMER,1* ASHOK VARMA,2 OLIVIER RONIN,'SIMONE MATHOULIN,' AND BERTRAND DUPONT'

Unite de Mycologie, Institut Pasteur, 75724 Paris Cedex 15, France,1 and Clinical Mycology Section,Laboratory of Clinical Investigation, National Institute ofAllergy and Infectious

Diseases, Bethesda, Maryland 208922

Received 15 April 1994/Returned for modification 14 May 1994/Accepted 30 June 1994

Cryptococcus neoformans serotype A is responsible for the majority of cryptococcal infections in AIDSpatients. In France, approximately 17% of the patients are infected with serotype D, regardless of their humanimmunodeficiency virus status. In a retrospective study of 273 patients, we found that serotype D was unevenlydistributed in France. We wondered if this was related to the yeast's genetic background. We used karyotypingand DNA fingerprints generated by UT-4p to analyze 40 serotype D clinical isolates. We found an extensivepolymorphism, with only two conserved karyotypes from drug-addicted patients living in the same area.Although highly variable, the DNA fingerprints were classified into 10 groups. Four pairs of isolates wereidentical; three of these pairs were from patients living in the same area, but there was no other correlationwith the geographical area. The two isolates with identical karyotypes belonged to the same fingerprint group.Five of the six isolates that made up fingerprint group I were recovered from drug-addicted patients (P < 0.002;chi-square), and all five isolates found in fingerprint group III were from male homosexual patients (P < 0.02).Finally, five of the seven isolates from patients with cryptococcal pneumonia were classified as fingerprintgroup V (P < 0.04). These results suggest that there are possible relationships between characteristics of theisolates and body localization or even risk factors. Results of the present study warrant other studies onisolates of all serotypes and on isolates from clinical and environmental sources.

Cryptococcus neoformans, an encapsulated yeast, is respon-sible for disseminated meningoencephalitis mostly in patientswith cellular immune defects (6, 32). Cryptococcosis occurs in6 to 10% of the patients with AIDS in the United States (8, 21).It is the most common life-threatening fungal infection in thesepatients (8). The yeast exists in two varieties: C. neofornansvar. neoformans, which corresponds to serotypes A and D, andC. neoformans var. gattii, which corresponds to serotypes B andC (1). Using standard immunological methods of serotyping,Kwon-Chung, Bennett, and colleagues (1, 12) have describedan uneven distribution of the serotypes around the world. Theyshowed that C. neofornans var. neoformans is found worldwideand that C. neoformans var. gattii is limited to tropical andsubtropical areas. However, C. neofornans var. neoformans iscultured from almost all AIDS patients with cryptococcalinfections, whatever the country of diagnosis (15), changing thepreviously known epidemiology of the disease. In all reports ofC. neofornans serotypes in patients with AIDS, serotype A isthe predominant serotype, being found in about 99% ofpatients with AIDS (3, 17, 22) even in areas where C. neofor-mans var. gattii is still able to cause infection in non-AIDSpatients. The disappearance of infections caused by serotypesB and C when the immune defect is related to the humanimmunodeficiency virus (HIV) remains unexplained.

If immunological studies were essential for the first descrip-tion of the epidemiology, they cannot help in the analysis ofstrain variability. More discriminatory techniques are needed.Molecular typing techniques, including karyotyping, restrictionfragment length polymorphisms, and DNA fingerprinting, have

* Corresponding author. Mailing address: Unite de Mycologie, 25rue du Dr. Roux, 75724 Paris Cedex 15, France. Phone: (33) 1 40 61 3255. Fax: (33) 1 45 68 84 20.

therefore been recently applied as epidemiological tools (5, 14,15, 18-20, 25, 31). These techniques have been describedmostly with reference strains (or sometimes strains fromculture collections) of the four serotypes or with clinical andenvironmental isolates collected in the United States andtherefore mostly with strains of the A or B serotype. All studiesdealing with the epidemiology of cryptococcal disease wereissued from countries where strains of serotype D are presentin low numbers or are even absent. Before the AIDS epidemicserotype D was described to be responsible for fewer than 5%of the infections in the United States, whereas it was a morefrequent cause of infection in Europe, especially in Denmarkand Italy (12). Therefore, very little is known about thevirulence of serotype D in AIDS patients and the geneticpolymorphism of this serotype compared with that of serotypeA.Through the national survey on cryptococcosis initiated in

France in 1985, we were able to analyze 273 cases of crypto-coccosis diagnosed between January 1990 and December 1992and for which we had the corresponding infecting isolates. Ofthese, 17% were serotype D. We thus showed that serotype Dis able to infect patients with AIDS since more than 80% of thepatients with cryptococcosis were HIV infected (10). Thegeographical area of diagnosis was of importance, with, forexample, 67% of serotype D strains infecting patients diag-nosed with cryptococcosis in Brittany and 32% in southwesternFrance, in comparison with 11% in Paris and suburbs (Ile-de-France) and 10% in southeast France.Because of the peculiarities in the distribution of serotype D

compared with that of serotype A, we wondered if strainsimilarities among the serotype D isolates could be demon-strated to be associated with a given geographical area or a siteof yeast isolation, for example. Chromosome size variability inC. neoformans has been used to determine the relatedness of

2364

on Novem

ber 10, 2020 by guesthttp://jcm

.asm.org/

Dow

nloaded from

MOLECULAR TYPING OF C. NEOFORMANS SEROTYPE D 2365

TABLE 1. Characteristics of patients and corresponding isolates analyzed by molecular typing methods

Isolate Patient

No. Site from which it Year of Sexb Risk Risk factor Ethnic Regionwas cultured' isolation factor for HIVC origin'

CSFCSFCSF, bloodSkinCSF, bloodCSFCSF, lung, skinBloodLungCSF, blood, urinesSkinCSF, unnesLymph nodeCSF, lungCSFCSFCSFCSFCSF, urinesCSFCSFCSFLungCSFCSFCSFCSFLungCSF, LungBloodCSFBlood, urinesCSFCSF, Blood, lungCSFCSF, blood, liverUrines, lungCSFCSF, lungCSF, blood, lung, skinCSFCSFbloodCSF, bloodCSF

199119921992199119921992199219921992199019921992199119921992199219901990199119911991199119911991199219931991199019901991199119911991199219921992199119911991199119921992199219921993

M unknownM HIVM HIVM noneM HIVM HIVM HIVM CorticM HIVM HIVF noneM HIVM HIVM HIVM HIVM HIVM HIVF HIVM HIVM HIVM HIVM HIVM HIVM HIVM HIVM HIVM HIVM HIVF HIVM HIVM HIVM HIVM HIVM HIVM HIVM HIVM HIVM HIVM HIVM HIVM HIVM HIVM HIVM HIVM HIV

a The body site from which the isolate used for molecular typing methods was obtained is in italics, unless there was a lack of precision about the body site from whichit was obtained when the isolate was sent to the National Reference Center for Mycosis.

b M, male; F, female.c NA, not available; Drug, drug addiction; Homo, homosexual; Hetero, heterosexual; Bi, bisexual; Neg, HIV negative.d Cauc: caucasian; NA, not available.e IDF, Ile-de-France.

strains (19, 26). Varma and Kwon-Chung (30) recently de-scribed a probe that is able to hybridize to repetitive genomicDNA. They showed that this probe, UT-4p, hybridizes toseveral restriction fragments of AccI-digested genomic DNAand generates DNA fingerprints which allow determination ofthe genetic diversity and relatedness of C. neoformans isolates.Karyotyping and DNA fingerprinting were then performedwith the UT-4p probe to characterize clinical isolates of the Dserotype cultured from various body sites in AIDS and non-

AIDS patients diagnosed with cryptococcosis in different partsof France.

(This work was presented in part at the 2nd International

Conference on Cryptococcus and Cryptococcosis, Milan, Italy,19 to 23 September 1993.)

MATERIALS AND METHODS

Isolates. Table 1 describes the characteristics of the serotypeD isolates that were selected for the present study. Clinicalinformation concerning the patients was available through the

national survey on cryptococcosis. The data and the isolates

were collected at the National Reference Center for Mycosis,Pasteur Institute, Paris, France. For convenience, the cities

where the different cases were diagnosed were grouped into

DID24D40D7D31D36D39D47D48D6D26D33D20D21D32D34D2D3D4D5D13D15D16D17D42D49D8D9D10DllD14D18D19D22D23D25D27D28D29D30D35D37D38D41D45

NANADrugNegHomoHomoHeteroNegNANANegDrugHomoNADrugHomoDrugTransfusionTransfusionHomoDrugHomoHomoDrugDrugTransfusionDrugHomoHeteroHomoNAHeteroHomoDrugHomoNADrugHomoBiHomoHomoNADrugTransfusionHomo

NACaucCaucCaucCaucCaucCaucArabianArabianCaucCaucCaucCaucNACaucCaucCaucCaucCaucCaucCaucCaucCaucCaucCaucCaucCaucCaucCaucCaucCaucNACaucCaucCaucCaucCaucCaucCaucCaucCaucCaucBlackCaucCauc

CenterCenterCenterEastEastEastEastEastEastSoutheastSoutheastSoutheastBrittanyBrittanyBrittanyBrittanySouthwestSouthwestSouthwestSouthwestSouthwestSouthwestSouthwestSouthwestSouthwestSouthwestIDFIDFIDFIDFIDFIDFIDFIDFIDFIDFIDFIDFIDFIDFIDFIDFIDFIDFAlpes

VOL. 32, 1994

on Novem

ber 10, 2020 by guesthttp://jcm

.asm.org/

Dow

nloaded from

2366 DROMER ET AL.

FIG. 1. Map of France showing the main regions and the cities where the patients with cryptococcosis were diagnosed.

the regions represented in Fig. 1 (Brittany, Paris, the suburbsof Paris, called Ile-de-France, eastern France, the Alps, South-eastern France, Southwestern France, and the center).

Serotyping of isolates. Identification of the clinical isolatesas C. neoformans was always done by the laboratory where theyeast was first isolated and was usually done on Sabourauddextrose agar. C. neofornans was identified by urease produc-tion in substrate urea-indole at 37°C, positive India ink prep-

aration, and sugar assimilations with commercially availablestrips (API 20AUX or ID32C; bioMerieux, Marcy-l'Etoile,France). When required, the identification was checked at theNational Reference Center for Mycosis. Serotyping was per-formed by a previously described method (9), combiningdiagnostic media and immunofluorescence assay with themonoclonal antibody El (11).

CHEF electrophoresis. Agarose blocks containing C. neofor-mans chromosomes from selected isolates were prepared by aprocedure described before (13). Saccharomyces cerevisiaechromosomes (Bio-Rad, Richmond, Calif.) were used as mo-lecular size standards. Electrophoresis in 0.7% agarose wascarried out on the contour-clamped homogeneous electric field(CHEF) DRII gel electrophoresis system (Bio-Rad) at 100 Vwith a dually ramped switch time of 100 to 200 s for 16 h and200 to 300 s for 32 h. The buffer temperature was kept constantat 12°C throughout the run.

Southern blot analysis. DNAs were extracted from selectedC. neoformans isolates as described before (30). Completedigestion of the DNA with the restriction endonuclease AccI,electrophoresis in 0.8% agarose, transfer, and hybridizationwith the radiolabelled plasmid extracted from UT-4p were

J. CLIN. MICROBIOL.

on Novem

ber 10, 2020 by guesthttp://jcm

.asm.org/

Dow

nloaded from

MOLECULAR TYPING OF C. NEOFORMANS SEROTYPE D 2367

SOUTH WEST IDF EASTSOUTH

CENTER BRITTANY EAST ALPES

m u to Nern4 W- M 8 In _- _- _- Vacoca a000c

lU_

0oW- g- e 000000 00

0 rs 0 o-0Cjaj cca a a a ajaj0 a a cj a

qw 0O'-- Crw

0a 00

Um

FIG. 2. Karyotypes of 40 serotype D clinical isolates recovered from patients diagnosed in different parts of France. The regions were as follows:Southwest (Bordeaux), Southeast (Nice and Nimes); Ile-de-France (IDF) (Paris and suburbs), East (Nancy and Besanqon), Brittany (Rennes andBrest), Center (Tulle and Clermont-Ferrand), Alps (Lyon). The size of the chromosomes from S. cerevisiae are shown on the ordinate. Heavy linesrepresent at least two chromosomes at the same location under the electrophoresis conditions. Details about the isolates are available in Table 1.

carried out as reported before (30). A 1-kb ladder (Gibco-BRL, Gaithersburg, Md.) was used as a molecular size stan-dard.

Statistical analysis. The distributions among the DNAfingerprint groups according to the risk factor for cryptococ-cosis or the site of culture were analyzed by using EPI INFOsoftware (Centers for Disease Control and Prevention, At-lanta, Ga., and World Health Organization, Geneva, Switzer-land) by the chi-square test.

RESULTS

Karyotyping of the D isolates. We compared serotype Disolates using their karyotyping patterns. Figure 2 shows thereconstructed CHEF patterns obtained for 40 isolates col-lected from various body sites of patients coming from differ-ent geographical areas. The number of separable bands variedfrom 6 to 12, with their sizes ranging from .2.2 kb toapproximately 630 bp. Some isolates had one or two heavilystained bands that probably corresponded to two chromo-somes that could not be separated under our electrophoresisconditions. Two isolates (D23 and D25) had a small chromo-some of about 245 bp. The identification of these isolates as C.neoformans was confirmed by using commercially available ID32C strips (bioMerieux).

Extensive polymorphisms were found when analyzing thekaryotyping patterns of the serotype D isolates. There was no

apparent identifiable pattern which related a strain to a givengeographical area, a site of yeast isolation, the patient's HIVstatus, or the year of diagnosis. The patterns were unique toeach isolate except for two isolates that exhibited the same

pattern. These isolates (D2 and D13) were cultured frompatients diagnosed with cryptococcosis almost a year apart inBordeaux (southwestern France). Both patients had drugaddiction as the predisposing factor for HIV infection. It was

impossible to obtain more information about potential com-

mon environmental sources of infection for the two patients,like friendship, sharing of syringes, or visiting the same places,because both patients died shortly after they were diagnosedwith cryptococcosis. For another patient, the isolate that was

recovered after 5 months of antifungal therapy with flucon-azole was found to be identical to the patient's initial isolate(Fig. 3).

Fingerprinting of serotype D isolates with UT-4p. Compar-ison of the patterns was made easier by the fact that electro-phoresis of the 31 digested DNAs was performed on the same

day by using the same power supply. Likewise, transfer andhybridization were performed during one unique experiment.The migrations of the bands in the standard ladder were

therefore similar in all gels. By using the probe, the reproduc-ibility of the patterns had previously been checked with variousisolates, including those from a culture collection, clinicalisolates cultured from different body sites of the same patient,and isolates obtained after experimental infection of mice (30).The reproducibility was checked during this experiment withreference strains of the four serotypes. The patterns of theserotype D clinical isolates were then checked only once.

Hybridization with the radiolabelled plasmid UT-4p gener-ated highly variable restriction fragment patterns (Fig. 4). Allisolates had the double band (2.1 and 2.3 kb) characteristic ofserotype D (30). All except three isolates were indistinguish-able by their lower fragments; the other three isolates (D14,D26, and D28) exhibited only one band instead of two at about1.8 kb. Pattern variations among the isolates were thereforeseen essentially for restriction fragments ranging from about

2.5 kb to about 10 kb. In that range, three to six DNA

fragments were hybridized to UT-4p.We arbitrarily chose to separate the isolates into groups

according to the number and the sizes of the bands larger than

1600-

1125-

700-

245-

VOL. 32, 1994

__

on Novem

ber 10, 2020 by guesthttp://jcm

.asm.org/

Dow

nloaded from

2368 DROMER ET AL.

EASI SOL1H EAST BRITTANY;br- I~ --I

12-

8 - O"

3-

6-_5- a*;

-4-

a*1';-

cr m r. co co m

on a aon n Q n Q a Q n

... oms

_w Am

SOUTIH WES-T

s- S.,"

WfOM"

~~~~~.I6Ma m 40

1 6 6332T;M3- 4 T1 T1 (i I

L.E. DE RANC[

FIG. 3. Gel with ethidium bromide-stained chromosome bandsfrom clinical isolates of C. neoformans serotype D. Molecular sizestandards are provided by chromosomes from S. cerevisiae. IsolatesD35a and D35b were cultured from the same patient at the time ofdiagnosis and 5 months after antifungal therapy with fluconazole,respectively.

2.5 kb (Table 2). This classification was made blindly andended with 10 groups (groups I to X). Five groups (II, IV, VII,VIII, and X) contained only one isolate each. The other fivegroups included between 2 and 10 different isolates. Variabilitywithin a group was limited to small variations in the size of thehigher-molecular-mass bands, especially in group I, which hadtwo strains with identical patterns (isolates D2 and D17) andlittle variation among the remaining members of this group,including isolate D13. We also found isolates with identicalpatterns in group V (isolates D21 and D27 and isolates D10and D22), and group IX was composed of two identical isolates(isolates D3 and D5).

Except for isolates D21 and D27, which were cultured frompatients who lived in southwestern France and Ile-de-France,respectively (and for whom no information about possibletravel to the other patient's region was available), the pairs ofisolates with identical patterns were found for patients living inthe same geographical area. However, there was no correlationbetween the pattern group and the geographical area ofdiagnosis. The number of patients with a risk factor forcryptococcosis independent of HIV infection was too small toanalyze any possible connection, as was the number of non-Caucasian patients.We then looked at the possible relationship between the

type of cryptococcal isolate infecting the patient and the riskfactor for HIV infection or the body site where the isolate wasfound. Group I comprised six isolates, with five isolatesrecovered from drug-addicted patients (P < 0.002; chi-square),whereas group III contained only five isolates, all from malehomosexual patients (P < 0.02). There was no apparent linkbetween a pattern group and localization of the isolate incerebrospinal fluid, urine, or skin. On the other hand, a smallnumber of DNA fingerprint groups contained isolates recov-

ered from blood or lungs. Five of the seven isolates culturedfrom patients with cryptococcal pneumonia, whether or notthey were associated with meningitis or with other body

8- _ _ __ "

4-

as- am ft-____400, _1.7 _-u _ " bw0

- A-

8 1' 1 18I i>. 2 8 ?'4 3

FIG. 4. Hybridization patterns of serotype D clinical isolates.Genomic DNA was extracted, digested with the endonuclease AccI,electrophoresed on 0.8% agarose, and hybridized with labelled plas-mid DNA from UT-4p. Molecular size standards were obtained withthe 1-kb ladder.

locations, were in group V (P < 0.04). Interestingly, amongthese five isolates, isolates D10 and D22 and isolates D21 andD27 had identical DNA fingerprints.

DISCUSSION

Strain variations among C. neoformans serotype A isolateshave been demonstrated by using biochemical (4, 28), immu-nological (9, 23, 24, 27), and molecular typing (18, 19, 31)techniques. The small number of infections caused by C.neoformans serotype D in most of the world except Europeexplains that characterization of serotype D isolates has beenlimited to a few isolates from culture collections. Since sero-type D isolates infect AIDS and non-AIDS patients in thesame proportion in France (personal data), it is unlikely that aparticularly virulent strain is selected by the immune defectinduced by HIV. This has already been demonstrated withserotype A isolates (7, 15). On the other hand, our finding thatinfections related to serotype D were more often diagnosed inpatients living in certain parts of France could suggest differ-ences between serotypes D and A. It also raises the possibilityof a genetic association with an environmental source or a

given tropism, for example. Serotyping with our monoclonalantibody does not provide enough variation in patterns (9) toexplore this hypothesis. We thus chose to use two differentmolecular typing methods since their sensitivities for serotypeD isolates have never been compared.For the karyotyping of the isolates, we used migration

0.i

0)

cnl

J. CLIN. MICROBIOL.

on Novem

ber 10, 2020 by guesthttp://jcm

.asm.org/

Dow

nloaded from

MOLECULAR TYPING OF C. NEOFORMANS SEROTYPE D 2369

TABLE 2. Distribution of 31 isolates of C. neoformans serotype D by DNA fingerprint patterns generated by UT-4p

Fingerprint No. of No. and site of Risk factor for HIVgroupa isolates Isolate no. isolationb infectionc Geographical area

I 6 D2, D13, D17, D18, D32, D33 5 C, 2 U, 1 B 5 Drug, 1 Hetero 3 Southwest, 1 IDFd, 1Brittany, 1 Southeast

II 1 D20 LN Homo BrittanyIII 5 D15, D19, D23, D29, D35 5 C, 1 L 4 Homo, 1 Bi 1 Southwest, 4 IDFIV 1 D30 C, B, L, S Homo IDFV 10 D7, D8, D10, Dll, D16, D21, 6 C, 5 L, 3 B, 4 Homo, 3 Drug, 1 5 IDF, 2 Brittany, 1

D22, D27, D31, D34 1 U, 1 S Hetero, 1 NA, 1 Neg Southwest, 2 eastVI 3 D6, D14, D25 3 C, 2 B, 1 U 3 NA 2 IDF, 1 SoutheastVII 1 D28 C Homo IDFVIII 1 D4 C Transfusion SouthwestIX 2 D3, D5 2 C 1 Homo, 1 Transfusion 2 SouthwestX 1 D26 S 1 Neg Southwest

a DNA fingerprint groups were arbitrarily designed on the basis of similarities in the number of bands above 2.5 kb.b Number of isolates that were cultured from the following sites: cerebrospinal fluid (C), blood (B), urine (U), lung (L), skin (S), and lymph node (LN).See footnote c to Table 1.

d IDF, Ile-de-France.

conditions slightly different from those used recently by Perfectet al. (19) but found the same number of chromosome bands(between 6 and 12). With the serotype D isolates, we alsoconfirmed the known chromosome size variability of C. neo-formans var. neoformans, as opposed to the chromosome sizesimilarity previously found for clinical isolates of C. neofor-mans var. gattii (16). Among the 40 isolates analyzed, two hadone band at the place of the minichromosome previouslydescribed by Varma and Kwon-Chung (29). As expected, wedid not find a specific pattern associated with the ethnic originof the patients, the geographical area of diagnosis, or theclinical site of infection, a finding already noted for serotype A(19).The karyotyping technique allows for the rapid screening of

selected isolates. It is reproducible and easy to perform.However, the results were disappointing since the patternswere so dissimilar that it prevented any kind of grouping whenlooking at a small number of isolates from different patients.We found that only 2 of the 40 serotype D clinical isolates hada similar karyotype, and Perfect et al. (19) described 5 isolateswith two conserved karyotypes of the 46 serotype A isolatesthat they studied. While these isolates were from individualswith no common background, our isolates were cultured fromtwo drug-addicted patients who lived in the same geographicalarea. This may suggest that karyotyping could be useful forlarge epidemiological studies involving hundreds of isolatesand might then detect similarities more often. Nevertheless, itis a convenient technique for studying isolates recovered fromthe same patient for a prolonged period of time or forcomparing the isolates recovered during a relapse(s) and theinitial isolate (26).When isolates have a conserved karyotype or when one

works with a limited number of them, other degrees ofresolution are necessary to look for relatedness. The DNAfingerprints generated with UT-4p were, in our hands, helpfulin allowing us to validate our serotyping method using themonoclonal antibody El and to distinguish between isolates ofthe D serotype. In the first report on UT-4p, serotype Dexhibited the most variable DNA fingerprint of the fourserotypes examined. In the present study, only four pairs ofisolates had identical patterns. We arbitrarily and blindlydesigned groups on the basis of similarities in the number ofbands larger than 2.5 kb and ended up with 10 groups eachcontaining 1 to 10 isolates. The purpose of this classificationwas to look for a correlation between a group and character-

istics of the patient or the isolate. We found several isolatesthat could fit in the same group, with small variations in thesizes of the bands. We did not collect enough isolates fromnon-AIDS or non-Caucasian patients to make conclusions onthese points. Other studies have reported that there are nodifferences in the C. neoformans isolates that infect AIDS andnon-AIDS patients (7, 15, 31). The geographical area wherethe patient was diagnosed bore no relationship to any groupexcept for isolates with identical patterns, which were obtainedfrom patients coming from the same area in three of the fourcases. The two isolates with a conserved karyotype were foundin the same fingerprint group, group I. One of these isolates(isolate D2) had a DNA fingerprint identical to that of anotherisolate (isolate D17). All three isolates were recovered fromdrug-addicted patients living in the same area. However, we donot have enough information to confirm that this cluster ofisolates with apparently the same DNA sequences came froma common exposure. The heterogeneity of isolates comingfrom the same area was also recently reported for serotype Aisolates by Varma and collaborators (31). The DNA fingerprintpatterns generated by UT-4p for 79 isolates recovered fromclinical and environmental sources in California, Berlin, Zaire,and the National Institutes of Health Clinical Center weredivided into nine distinct groups. The most heterogeneouspatterns were found for the 19 isolates collected in California,which belonged to six distinct groups, whereas the 21 isolatesfrom Berlin belonged to five groups, but most of the isolates(15 of 21 isolates) belonged to one group. There was littleoverlapping, but considering the distance it was surprising tohave a group containing isolates from all four locations. Travelhabits do not account for this phenomenon, since environmen-tal isolates were also found in the same group. Experiments onserotype D isolates cultured from the environment in Franceare under way to see if the distribution of serotypes in theenvironment will match that found for clinical isolates and alsoif the heterogeneity described for serotype A with isolates fromenvironmental sources will be found on a smaller scale. Wealso want to see if patients can be infected with strains of C.neoformans genetically related to strains found in the environ-ment, as recently described by Currie et al. (7) with isolatesfrom New York City.

If geographical areas did not seem to influence the type ofisolates infecting the patients, there might be a link with therisk factor for HIV infection, especially for isolates fromdrug-addicted patients, which were mostly found to be in group

VOL. 32, 1994

on Novem

ber 10, 2020 by guesthttp://jcm

.asm.org/

Dow

nloaded from

2370 DROMER ET AL.

I. However, all except three patients diagnosed with crypto-coccosis between 1990 and 1992 lived in different geographicalareas; the three exceptions lived in southwestern France.Further studies are thus needed on a larger number of clinicalisolates to confirm these results. We also found that isolatesfrom patients with cryptococcal pneumonia were more likely tobe in fingerprint group V. We did not check that the finger-prints generated with UT-4p were identical for serotype Disolates cultured from different body sites of the same patient,but we can speculate that this is the case from the resultsobtained with serotype A. Knowing that extrameningeal local-izations are of poor prognostic value during cryptococcalmeningitis (6), it may be important to find out if isolates differin terms of their tropisms for a given organ or, more likely, ifdissemination is dependent only on the ability of the host tocontrol the infection. The fact that UT-4p might provide agenetic marker of tropism or virulence has yet to be demon-strated.

It is difficult to draw conclusions about the sensitivity ofmolecular typing methods. If the purpose of one technique isto be discriminatory, then karyotyping is the most sensitivetechnique. If, on the other hand, one wants to see therelatedness of isolates, then a highly sensitive technique is notdesirable since each isolate tested will have a different pattern.Fingerprinting with a well-defined probe that hybridizes with arepetitive element present in multiple copies on all chromo-somes is more appropriate and seems more powerful. This iswhy we would do karyotyping to compare isolates culturedfrom the same patient over time and favor fingerprinting withprobes like UT-4p or CNRE-1 (7, 25) to study a limitednumber of isolates, as in the present study. Other techniquessuch as random amplified polymorphic DNA analysis warrantevaluation with C. neofornans isolates (2).Our results, although limited by the number of isolates

tested, raise the question of the peculiarities of the strainsassociated with particular body locations during cryptococcalinfections or risk factors for HIV infection. Studies on a largernumber of clinical isolates and on environmental isolates fromprecise geographical areas are necessary to confirm our results.

ACKNOWLEDGMENTS

The participation of members of the Societe Francaise de MycologieMedicale and the Societe de Pathologie Infectieuse de Langue Fran-,aise is greatly acknowledged. The following people are especiallythanked: in the southwest, B. Couprie (Bordeaux); in the southeast, M.Gari-Toussaint (Nice) and A. Delage (Nimes); in Ile-de-France, I.Hilmarsdottir, J.-L. Poirot, F. Traore, C. Hennequin, and D. Basset(Paris), N. Godineau (Saint Denis), C. Nauciel (Garches), A. Chap-man (Suresnes), 0. Patey (Villeneuve Saint Georges); in the East, L.Kures (Nancy) and T. Barale (Besancon); in Brittany, C. Guigen(Rennes) and 0. Masure (Brest); in the center, J. Verger (Tulle) andM. Cambon (Clermont-Ferrand); and in the Alps: M.-A. Piens (Lyon).

REFERENCES1. Bennett, J. E., K. J. Kwon-Chung, and D. H. Howard. 1977.

Epidemiologic differences among serotypes of Cryptococcus neo-formans. Am. J. Epidemiol. 105:582-586.

2. Bostock, A., M. N. Khattak, R. Matthews, and J. Burnie. 1993.Comparison of PCR fingerprinting, by random amplification ofpolymorphic DNA, with other molecular typing methods forCandida albicans. J. Gen. Microbiol. 139:2179-2184.

3. Bottone, E. J., I. F. Salkin, N. J. Hurd, and G. P. Wormser. 1987.Serogroup distribution of Cryptococcus neoformans in patientswith AIDS. J. Infect. Dis. 156:242.

4. Brandt, M. E., S. L. Bragg, and R. W. Pinner. 1993. Multilocusenzyme typing of Cryptococcus neoformans. J. Clin. Microbiol.31:2819-2823.

5. Casadevall, A., L. F. Freundlich, L. Marsh, and M. D. Scharff.

1992. Extensive allelic variation in Cryptococcus neoformans. J.Clin. Microbiol. 30:1080-1084.

6. Chuck, S. L., and M. A. Sande. 1989. Infections with Cryptococcusneoformans in the acquired immunodeficiency syndrome. N. Engl.J. Med. 321:794-799.

7. Currie, B. P., L. F. Freundlich, and A. Casadevall. 1994. Restric-tion fragment length polymorphism analysis of Cryptococcus neo-formans isolates from environmental (pigeon excreta) and clinicalsources in New York City. J. Clin. Microbiol. 32:1188-1192.

8. Dismukes, W. E. 1988. Cryptococcal meningitis in patients withAIDS. J. Infect. Dis. 157:624-628.

9. Dromer, F., E. Gueho, 0. Ronin, and B. Dupont. 1993. Serotypingof Cryptococcus neofornans by using a monoclonal antibodyspecific for capsular polysaccharide. J. Clin. Microbiol. 31:359-363.

10. Dromer, F., 0. Ronin, A. Varma, and B. Dupont. 1993. Mono-clonals for serotyping Cryptococcus neofonnans, abstr. L16, p. 48.2nd International Conference on Cryptococcus and Cryptococco-sis, Milan.

11. Dromer, F., J. Salamero, A. Contrepois, C. Carbon, and P. Yeni.1987. Production, characterization, and antibody specificity of amouse monoclonal antibody reactive with Cryptococcus neofor-mans capsular polysaccharide. Infect. Immun. 55:742-748.

12. Kwon-Chung, K. J., and J. E. Bennett. 1984. Epidemiologicdifferences between the two varieties of Cryptococcus neoformans.Am. J. Epidemiol. 120:123-130.

13. Kwon-Chung, K. J., J. C. Edman, and B. L. Wickes. 1992. Geneticassociation of mating types and virulence in Cryptococcus neofor-mans. Infect. Immun. 60:602-605.

14. Kwon-Chung, K. J., A. K. Varma, and D. H. Howard. 1988.Ecology and epidemiology of Cryptococcus neoformans: a recentstudy of isolates in the United States, p. 107-112. In J. M.Torres-Rodriguez (ed.), Proceedings of the 10th Congress of theInternational Society for Human and Animal Mycology. J. R.Prous Science Publisher, Barcelona.

15. Kwon-Chung, K. J., A. Varma, and D. H. Howard. 1990. Ecologyof Cryptococcus neoformans and prevalence of its two varieties inAIDS and non-AIDS associated cryptococcosis, p. 103-113. In H.Vanden Bossche et al. (ed.), Mycoses in AIDS patients. PlenumPress, New York.

16. Kwon-Chung, K. J., B. L. Wickes, L. Stockman, G. D. Roberts, D.Ellis, and D. H. Howard. 1992. Virulence, serotype, and molecularcharacteristics of environmental strains of Cryptococcus neofor-mans var. gattii. Infect. Immun. 60:1869-1874.

17. Levitz, S. M. 1991. The ecology of Cryptococcus neoformans andthe epidemiology of cryptococcosis. Rev. Infect. Dis. 13:1163-1169.

18. Meyer, W., T. G. Mitchell, E. Z. Freedman, and R. Vilgalys. 1993.Hybridization probes for conventional DNA fingerprinting used assingle primers in the polymerase chain reaction to distinguishstrains of Cryptococcus neoformans. J. Clin. Microbiol. 31:2274-2280.

19. Perfect, J. R., N. Kebatchi, G. M. Cox, C. W. Ingram, and C. L.Beiser. 1993. Karyotyping of Cryptococcus neofornans as anepidemiological tool. J. Clin. Microbiol. 31:3305-3309.

20. Perfect, J. R., B. B. Magee, and P. T. Magee. 1989. Separation ofchromosomes of Cryptococcus neofornans by pulse field gel elec-trophoresis. Infect. Immun. 57:2624-2627.

21. Powderly, W. G. 1993. Cryptococcal meningitis and AIDS. Clin.Infect. Dis. 17:837-842.

22. Rinaldi, M. G., D. J. Drutz, A. Howell, M. A. Sande, C. B. Wofsy,and W. K. Hadley. 1986. Serotypes of Cryptococcus neoformans inpatients with AIDS. J. Infect. Dis. 153:642.

23. Small, J. M., T. G. Mitchell, and R. W. Wheat. 1986. Strainvariation in composition and molecular size of the capsularpolysaccharide of Cryptococcus neoformans serotype A. Infect.Immun. 54:735-741.

24. Spiropulu, C., R. A. Eppard, E. Otteson, and T. R. Kozel. 1989.Antigenic variation within serotypes of Cryptococcus neoformansdetected by monoclonal antibodies specific for the capsular poly-saccharide. Infect. Immun. 57:3240-3242.

25. Spitzer, E. D., and S. G. Spitzer. 1992. Use of a dispersedrepetitive DNA element to distinguish clinical isolates of Crypto-

J. CLIN. MICROBIOL.

on Novem

ber 10, 2020 by guesthttp://jcm

.asm.org/

Dow

nloaded from

MOLECULAR TYPING OF C. NEOFORMANS SEROTYPE D 2371

coccus neoformans. J. Clin. Microbiol. 30:1094-1097.26. Spitzer, E. D., S. G. Spitzer, L. F. Freundlich, and A. Casadevall.

1993. Persistence of initial infection in recurrent Cryptococcusneoformans meningitis. Lancet 341:595-596.

27. Todaro-Luck, F., E; Reiss, R. Cherniak, and L Kaufman. 1989.Characterization of Cryptococcus neoformans capsular glucuron-oxylomannan polysaccharide with monoclonal antibodies. Infect.Immun. 57:3882-3887.

28. Turner, S. H., and R. Cherniak. 1991. Multiplicity in the structureof the glucuronoxylomannan of Cryptococcus neofornans. In J. P.Latge and D. Boucias (ed.), Fungal cell wall and immune re-

sponse. Springer Verlag, Heidelberg.29. Varma, A., and K. J. Kwon-Chung. 1991. Characterization of

Cryptococcus neoformans minichromosome, abstr. PS1.43, p. 73.

XI Congress of the International Society for Human and AnimalMycoses, Montreal.

30. Varma, A., and K. J. Kwon-Chung. 1992. DNA probe for straintyping of Cryptococcus neoformans. J. Clin. Microbiol. 30:2960-2967.

31. Varma, A., D. Swinne, J. E. Bennett, and K. J. Kwon-Chung. 1993.Epidemiology of Cryptococcus neoformans, abstr. L15, p. 46. 2ndInternational Conference on Cryptococcus and Cryptococcosis,Milan.

32. Zuger, A., E. Louie, R. S. Holzman, M. S. Simberkoff, and J. J.Rahal. 1986. Cryptococcal disease in patients with the acquiredimmunodeficiency syndrome. Diagnostic features and outcome oftreatment. Ann. Intern. Med. 104:234-240.

VOL. 32, 1994

on Novem

ber 10, 2020 by guesthttp://jcm

.asm.org/

Dow

nloaded from