Molecular Identification of Cryptococcus neoformans Serotypes

Abstract

Cryptococcus neoformans is a fungal pathogen that causes life-threatening infections primarily in immunocompromised hosts. Based on the genetic characteristics and serologic properties of capsular polysaccharides, three varieties and five serotypes have been defined: C. neoformans var. neoformans (serotype D), C. neoformans var. grubii (serotype A), hybrid serotype AD, and C. neoformans var. gattii (serotypes B and C). Epidemiologic features, such as geographic distribution and ecologic niche, and clinical characteristics have been shown to be associated with serotypes. At the present time, serotyping is based on agglutination tests with either commercial or “homemade” antisera or on immunofluorescence assays using a monoclonal antibody directed against the capsule polysaccharide. In this paper, we describe two molecular methods (PCR-restriction enzyme analysis and length polymorphism analysis) for C. neoformans serotype identification. Both are based on the sequence characteristics of a fragment of the CAP59 gene required for capsule biosynthesis. Testing of 72 C. neoformans strains including representatives of the five serotypes demonstrated the reliability of these methods.

Cryptococcus neoformans is a basidiomycetous encapsulated yeast that causes life-threatening infections, mainly in immunocompromised hosts and particularly in human immunodeficiency virus-infected patients (1, 13). Based on the genetic characteristics and serologic properties of capsular polysaccharides (CPS) (9), three varieties and four nonhybrid serotypes have been defined, namely, C. neoformans var. neoformans (serotype D), C. neoformans var. grubii (serotype A), and C. neoformans var. gattii (serotypes B and C) (4). Serotypes correspond to different sexual teleomorphs, namely, Filobasidiella neoformans and Filobasidiella bacillispora for serotypes A and D and serotypes B and C, respectively. Moreover, hybrid strains that most likely correspond to either diploid or aneuploid organisms have been characterized as serotype AD strains (11, 14, 17, 20). Some epidemiologic properties, such as geographic distribution, have been associated with serotypes, as serotypes A, D, and AD are found worldwide while serotypes B and C are restricted mainly to tropical and subtropical regions (1). While serotype A and D strains are usually isolated from pigeon droppings and cause disease mainly in immunocompromised hosts, serotype B and C strains are more commonly isolated from eucalyptus trees and more often infect hosts with normal immune status, causing cryptococcoma (1). The need for prolonged treatment and surgery and sequelae including impaired visual acuity and a higher mortality rate have also been found to be more frequently associated with serotype B strains than with serotype A strains (2, 18).

Serotyping is presently performed by using agglutination with commercial (Crypto Check kit; Iatron Labs, Tokyo, Japan) or “homemade” antisera components (9) or a combination of canavanine-glycine-bromothymol (CGB) blue agar diagnostic medium and a direct immunofluorescence assay using E1 monoclonal antibody directed against CPS (3). These methods are widely used but may be either inapplicable or unreliable in the case of noncapsulated or nontypeable isolates, respectively (8, 16). Moreover, cases of misidentification have already been reported (3, 19). In the present paper, we report two molecular methods, based on the sequence characteristics of a fragment of the CAP59 gene required for capsule biosynthesis (5). Both methods allow for the rapid and reliable identification of all C. neoformans serotypes.

MATERIALS AND METHODS

Strains.

Seventy-two C. neoformans strains, including 33 reference strains, representing the five serotypes and one Cryptococcus uniguttulatus strain were tested (Table 1 ). The identification of all strains was confirmed by using the ID32C auxanogram panel (BioMérieux, Marcy l'Etoile, France). Strains were serotyped by agglutination with homemade antisera based on the procedure described by Ikeda et al. (9) or with a commercial kit (Crypto Check kit; Iatron Labs, Tokyo, Japan). Some strains have been serotyped by the Centre National de Référence Mycologie et Antifongiques at the Pasteur Institute of Paris by combining CGB diagnostic medium with a direct immunofluorescence assay using E1 monoclonal antibody specific for CPS, as described by Dromer et al. (3). Isolates were stored at −80°C and were grown on yeast-peptone-glucose agar plates for 2 days at 30°C before testing.

TABLE 1.

C. neoformans strains tested in this study and results of immunoserotyping and molecular identification of the serotype

Strain no.	Strain name	Isolationa		Immunologic serotyping		Molecular identification
		Origin	Site	Methodb	Result	REA patternc	Fragment size(s) (bp)d
1	H99	United States	CSF	Commercial kit	A	A	394
2	IHEM 11752	Congo	CSF	Commercial kit	A	A	400
3	IHEM 11753	Congo	CSF	Commercial kit	A	A	394
4	IHEM 11756	Congo	CSF	Commercial kit	A	A	394
5	IHEM 11757	Congo	CSF	Commercial kit	A	A	394
6	IHEM 11876	Congo	CSF	Commercial kit	A	A	394
7	IHEM 11877	Congo	CSF	Commercial kit	A	A	394
8	IHEM 11880	Congo	CSF	Commercial kit	A	A	394
9	IHEM 11882	Congo	CSF	Commercial kit	A	A	394
10	Tou	Mali	CSF	Homemade Ab	A	A	394
11	Fat	Congo	CSF	Homemade Ab	A	A	400
12	coc 119	Cochin, India	CSF	Homemade Ab	A	A	394
13	333	Gabon	CSF	Homemade Ab	A	A	394
14	2598 P	Ivory Coast	CSF	Homemade Ab	A	A	394
15	g 57497	Gabon	CSF	Homemade Ab	A	A	394
16	Oug 1	Uganda	CSF	Homemade Ab	A	A	394
17	PP38	Cambodia	CSF	Homemade Ab	A	A	394
18	RV66254	Rwanda	CSF	Homemade Ab	A	A	394
19	3/768	Congo	CSF	Homemade Ab	A	A	394
20	Lim	France	CSF	Homemade Ab	D	A	394
21	C3	Cambodia	CSF	Homemade Ab	A	A	394
22	C6	Cambodia	CSF	Homemade Ab	A	A	394
23	C7	Cambodia	CSF	Homemade Ab	A	A	394
24	Oug 7	Uganda	CSF	Homemade Ab	A	A	394
25	PP29	Cambodia	CSF	Homemade Ab	A	A	394
26	RV68535	Rwanda	CSF	Homemade Ab	A	A	394
27	RV68528	Rwanda	CSF	Homemade Ab	A	A	394
28	Tnn Ci	France	CSF	Homemade Ab	A	A	394
29	g 59546	Gabon	CSF	Homemade Ab	A	A	394
30	RV66254	Rwanda	CSF	Homemade Ab	A	A	394
31	C57	Cambodia	CSF	Homemade Ab	A	A	394
32	C68	Cambodia	CSF	Homemade Ab	A	A	394
33	Oug 3	Uganda	CSF	Homemade Ab	AD	A	394
34	Oug 9	Uganda	CSF	Homemade Ab	A	A	394
35	PP43	Cambodia	CSF	Homemade Ab	A	A	394
36	241	Gabon	CSF	Homemade Ab	A	A	394
37	Tnn Mb	France	Blood	mAb IF	A	A	394
38	Tnn Da	France	Blood	mAb IF	A	A	394
39	Tnn Mo	France	CSF	mAb IF	A	A	394
40	Bel	Belgium	CSF	Commercial kit	A	A	394
41	IHEM 4157	Belgium	NS	Commercial kit	AD	AD	382, 394
42	IHEM 13877	Belgium	CSF	Commercial kit	AD	AD	382, 394
43	IHEM 14440	Belgium	CSF	Commercial kit	AD	AD	382, 394
44	IHEM 16079	Belgium	Voice prosthesis	Commercial kit	AD	AD	382, 394
45	IHEM 16513	Peru	CSF	Commercial kit	AD	AD	382, 394
46	IHEM 17861	Peru	CSF	Commercial kit	AD	AD	382, 394
47	RV 66423	Rwanda	CSF	Homemade Ab	A	D	382
48	JEC21	NS	NS	Commercial kit	D	D	382
49	B 3501 A	NS	NS	Commercial kit	D	D	382
50	NIH D52	United States	CSF	Commercial kit	D	D	382
51	IHEM 15038	Peru	NS	Commercial kit	D	D	382
52	IHEM 18620	Belgium	Skin	Commercial kit	D	D	382
53	IHEM 16711	Belgium	Skin	Commercial kit	D	D	382
54	IHEM 4158	Italy	NS	Commercial kit	D	D	382
55	AKD 30	Ivory Coast	CSF	Homemade Ab	A	D	380
56	T 6073	France	CSF	Homemade Ab	D	D	382
57	2815	Cambodia	CSF	Homemade Ab	A	D	382
58	Tnn Re	France	BAL fluid	mAb IF	D	D	382
59	JLP34	France	CSF	mAb IF	D	D	382
60	JLP23	France	Urine	mAb IF	D	D	382
61	IHEM 11758	Congo	CSF	Commercial kit	D	D	380
62	C9	Cambodia	CSF	Homemade Ab	A	D	382
63	IHEM 14934	Brazil	NS	Commercial kit	B	B	373
64	IHEM 4161e	United States	Sputum	Commercial kit	B	B	373
65	IHEM 11796e	Zaire	CSF	Commercial kit	B	B	373
66	IHEM 4170	Malaysia	NS	Commercial kit	B	B	373
67	IHEM 4164	China	CSF	Commercial kit	B	B	371
68	WM 276	Australia	Environment	Commercial kit	B	B	371
69	VanR265	Canada	BWF	Commercial kit	B	B	380
70	RV68522	Rwanda	CSF	Homemade Ab	D	B	371
71	IHEM 10079	NS	CSF	Commercial kit	C	C	366
72	IHEM 4159e	United States	CSF	Commercial kit	C	C	366

^aNS, not specified; CSF, cerebrospinal fluid; BAL, bronchoalveolar lavage; BWF, bronchial wash fluid; Congo, Democratic Republic of the Congo.

^bThe commercial kit used was the Crypto Check kit (Iatron Labs, Tokyo, Japan). Homemade Ab, agglutination with homemade antisera (9); mAb IF, immunofluorescence with monoclonal antibody specific for CPS (3).

^cREA pattern deduced from in silico analysis (Table 2).

^dPCR fragments are supposed to be 8 bases longer than the sequences due to a 7-base tail added to the reverse primer used to promote the addition of an extra A to the amplified fragment.

^eType strain.

CAP59 gene amplification.

Sequences of CAP59 genes from the four nonhybrid serotypes of C. neoformans were retrieved from the GenBank database and the Cryptococcus neoformans Serotype B Sequencing Project of the Broad Institute of Harvard and the Massachusetts Institute of Technology (http://www.broad.mit.edu). Sequences were then aligned using ClustalX (version 1.83), and two primers (CH-Cap 59F, 5′-CCTTGCCGAAGTTCGAAACG, and CH-Cap 59R, 5′-AATCGGTGGTTGGATTCAGTGT) were designed from the conserved regions of the CAP59 genes. DNA was extracted using a rapid method based on thermal shock and the chelation of components other than nucleic acids by using a resin suspension, as previously described (7). The PCR was carried out with a 50-μl reaction volume containing 1× PCR buffer, 1.6 μM (each) primers (Eurogentec, Liege, Belgium), 0.2 mM (each) deoxynucleoside triphosphates (equimolar concentrations of dATP, dCTP, dGTP, and dTTP), and 0.5 U of Taq polymerase (New England BioLabs Inc.). A touchdown amplification program was performed as follows: 7 min at 94°C; three cycles of 30 s at 94°C, 30 s at 60°C, and 30 s at 72°C; three cycles of 30 s at 94°C, 30 s at 58°C, and 30 s at 72°C; three cycles of 30 s at 94°C, 30 s at 55°C, and 30 s at 72°C; 28 cycles of 30 s at 52°C and 30 s at 72°C; and a final 15-min extension at 72°C.

Restriction enzyme analysis (REA).

Restriction maps of unique sites obtained using DNA Strider version 1.4x-4e (15) were used to select restriction enzymes allowing for the identification of the various serotypes (Table 2). In a first-line identification, PCR products were subjected to separate restriction procedures with BsmFI and HpaII according to the instructions of the manufacturer (New England BioLabs Inc.). Digested fragments were separated by electrophoresis in agarose gel (3% in Tris-borate-EDTA buffer) stained with ethidium bromide (0.5 μg/ml) at 90 V for 3 h. PCR products with patterns compatible with the B or C serotype were further digested with the AgeI enzyme (New England BioLabs Inc.).

TABLE 2.

Expected sizes (bp) of CAP59 gene fragments from different serotypes before and after restriction with BsmFI, HpaII, and AgeI enzymes

Assay or fragment	Expected size(s) (bp) for serotypea:
	A	B	C	D
Native amplicon	387	373	359	375
REA with BsmFI	206, 191	NRS	NRS	NRS
REA with HpaII	NRS	270, 101	256, 101	211, 162
REA with AgeI	NRS	269, 100	NRS	NRS

^aNRS, no restriction site; serotype A, strain H99 (GenBank accession number AF337639); serotype B, strain R265 (http://www.broad.mit.edu/annotation/genome/cryptococcus_neoformans_b/Assembly.html); serotype C, strain IHEM 4159 (sequence obtained in the present study); serotype D, JEC21 strain (GenBank accession number AE017341).

Fragment length polymorphism analysis.

For fragment length analysis, a similar amplification protocol was performed but with a fluorolabeled (6-carboxyfluorescein dye; Applied Biosystems, Courtaboeuf, France) forward primer. A 7-base tail used to promote the addition of an extra A to the amplified fragment was added to the reverse primer in order to facilitate sizing. Each PCR product was run with an internal molecular marker (GenScan HD400Rox, ABI) for size determination on a DNA sequencer (ABI 3100; Applied Biosystems).

In the case of conflicting results, a second serotyping determination was performed with the Crypto Check commercial kit (Iatron Labs, Tokyo, Japan).

Based on fragment length results, we selected certain PCR products to be sequenced (Table 3). Fragments of a serotype AD strain (IHEM 13877) not digested by either BsmFI or HpaII and corresponding to the D or A allele, respectively, were gel purified (Gel Band purification kit; Amersham Biosciences, Freiburg, Germany) before sequencing. Direct sequencing was performed using a Big Dye terminator protocol as recommended by the manufacturer (Applied Biosystems). Double-strand sequencing was performed using the same primers as those used for PCR amplification. After manual correction, sequences were compared using the BLASTN software on the GenBank database limited to the organism C. neoformans.

TABLE 3.

Comparison of the partial sequence of the CAP59 gene with sequences in the GenBank database by using BLASTN software

Strain	Serotype	Fragment length (bases)	Results (best hit)	Accession no.
IHEM 11752	A	392	Best score obtained when aligned with 2 fragments from serotype A strain M9253 (accession no. AB066120)	EF392812
			Partial 5′ end: 99% identity over a 203-nucleotide alignment
			Partial 3′ end: 97% identity over a 131-nucleotide alignment
			Better score for partial 3′ end obtained with sequence from serotype A strain H99: 98% identity over a 173-nucleotide alignment
IHEM 13877	AD	387 (A allelea)	99% identity over a 382-nucleotide alignment with sequence from serotype A strain H99 (accession no. AF337639)	EF392816
		375 (D alleleb)	98% identity over a 375-nucleotide alignment with sequence from serotype D strain JEC21 (accession no. AF337639)	EF392813
IHEM 11796	B	373	99% identity over a 324-nucleotide alignment with sequence from serotype B strain M9244 (accession no. AB066134)	EF392816
IHEM 4159	C	359	99% identity over a 312-nucleotide alignment with sequence from serotype C strain M9247 (accession no. AB066126)c	EF392815

^aNo restricted fragment after digestion with HpaII.

^bNo restricted fragment after digestion with BsmFI.

^cComplete sequence of the fragment not available in the GenBank database.

Nucleotide sequence accession number.

The sequence of IHEM 4159 obtained in the present study has been deposited in GenBank with accession number EF392815.

RESULTS

All tested C. neoformans strains but not the C. uniguttulatus strain gave positive amplification.

PCR-REA.

As expected from in silico analysis, enzymatic restriction with BsmFI and HpaII gave three different patterns for serotypes A, D, and B and C (Fig. 1). Serotype AD strains exhibited a mixed A and D restriction profile. All strains were tested, and no profile variation among strains of a given serotype was observed. Serotype D and serotypes B and C, while having similar profiles (unique cutting site for HpaII and no cutting site for BsmFI), can be easily differentiated based on the size difference of the digested fragments. Serotypes B and C can be further differentiated by AgeI digestion and gel electrophoresis. Two fragments corresponding to serotype B strains (one cutting site) were detected, while serotype C amplicons remained unchanged (no cutting site) (Fig. 1).

FIG. 1.

Agarose gel (3%) electrophoresis of native amplicons of the CAP59 gene fragments and restriction profiles of C. neoformans serotypes.

Length polymorphism analysis.

Because in silico analysis predicted length polymorphism corresponding to serotypes in the amplified fragments (Table 2), we designed a fluorescent fragment length analysis method using a DNA sequencer. Seven different alleles accounting for eight different genotypes were detected (Table 1). Fragments amplified from the strains of serotypes A, B, C, and D used as a reference for in silico analysis were sized with the predicted length plus or minus 1 base, which is the instrument resolution provided by the manufacturer.

A small degree of intraserotype polymorphism was seen: serotype A strains exhibited an allele of either 394 bp (n = 38) or 400 bp (n = 2); serotype D strains had an allele of either 380 bp (n = 2) or 382 bp (n = 14); and serotype B strains had an allele of 371 bp (n = 3), 373 bp (n = 4), or 380 bp (n = 1), while the two serotype C strains tested had an allele of 366 bp (Table 1). The six serotype AD strains exhibited two alleles, of 382 and 394 bp. Fragments with sizes different from the predominant size for a given serotype were sequenced. Table 3 summarizes the results of the BLASTN comparison of these sequences with sequences in the GenBank database. In all cases, the results of the sequence comparison agreed with the molecular identifications. Sequence alignment showed polymorphism among and within serotypes, mainly in the intronic part the fragments, located between nucleotide positions 200 and 280 (data not shown).

Analysis of discrepancies between conventional and molecular serotype identifications.

In a first analysis, we found seven cases of discrepancies in identifications (strains 20, 33, 47, 55, 57, 62, and 64). These discrepancies were seen only for strains for which the serotype had been determined by a homemade agglutination method. For these strains, a new serotype determination was performed using the Crypto Check kit, which agreed with the PCR-REA and length polymorphism identifications in every case.

DISCUSSION

Serotyping has been useful for the description of the epidemiologic and clinical characteristics associated with C. neoformans. More recently, it has played a major role in the investigation of the endemic infection on Vancouver Island (Canada) (10). Serotyping has no diagnostic value, but differences in treatments and outcomes corresponding to serotypes have been reported (2, 18). Thus, considering that the Crypto Check kit (Iatron, Japan) is no longer available and in view of the relative complexity of obtaining monoclonal antibodies, new techniques need to be developed for the typing of C. neoformans strains.

Both methods developed in this work allow the reliable identification of the five serotypes of C. neoformans by testing a unique genomic region. Our assays based on molecular analysis of a fragment of the CAP59 gene appeared to be sensitive and specific, providing interpretable results for all C. neoformans strains tested but no amplification for a C. uniguttulatus strain.

PCR-REA is an easy-to-use approach and requires only common molecular materials. Its value for distinguishing among fungal species such as Malassezia spp. has already been emphasized (6). As restriction sites were chosen within conserved regions of the fragments, all tested strains exhibited the restriction pattern predicted by in silico analysis. A flow chart for serotype identification by combining separate digestions with BsmFI and HpaII, followed by subsequent digestion with AgeI when the pattern suggests a B or C serotype, can therefore be proposed (Table 2). Serotype AD strains exhibited a mixed A and D restriction profile. Our results were in accordance with those of previous studies suggesting that AD serotype strains are diploid or aneuploid and possess, at least for a number of genes including the CAP59 gene, both serotype A-like and serotype D-like alleles (14, 17, 19, 20). This indication was further confirmed by a sequence analysis performed with the purified fragments (Table 3).

Length polymorphism analysis showed different lengths for different serotypes, with the exception of two serotype D strains (no. 55 and 61) and a serotype B strain (no. 69), which all exhibited a 380-bp allele. Note that these particular cases concerned only strains with alleles of the minority size within the serotype considered. These difficult cases can be solved by identifying the serotype by using either REA with AgeI (only serotype B is restricted) or the results of culture on CGB blue agar, on which serotype B but not serotype D strains produce a color change in the medium (12). Fragment length polymorphism may be easier to perform but requires a DNA sequencer, which is now more frequently available in teaching hospitals in industrialized countries. Following PCR amplification, results can be obtained within 1 h and are entirely objective.

Overall, results from both molecular methods fit well with serotype identifications obtained with immunologic procedures, but the homemade technique is probably less reliable. Both methods presented in this study appear to be reliable alternatives to immunologic methods.