Isolation and Characterization of a Species-Specific DNA Fragment for Identification of Candida (Torulopsis) glabrata by PCR

Abstract

A PCR specific for Candida glabrata that amplifies a mitochondrial rRNA gene fragment was developed by analysis of C. glabrata-specific agarose gel bands, which were generated by arbitrarily primed PCR. The expected PCR product was successfully amplified with genomic DNA from 95 C. glabrata isolates but not from a number of other fungal isolates.

In recent years, distinct shifts in the distribution of Candida species resulting in an increase in hospital-acquired infections due to Candida (Torulopsis) glabrata and Candida species other than C. albicans have been reported (17, 21, 23). In addition, C. glabrata and other emerging yeasts are often innately resistant to antifungal agents, specifically the azoles (9, 24). Since conventional phenotyping systems for yeasts may be unreliable in the identification of C. glabrata and the conventional cultivation and diagnosis of yeasts by morphologic and biochemical techniques require a minimum of 2 to 3 days, there is an obvious need to introduce modern molecular methods for fast and specific identification of this yeast species (4, 16). Although arbitrarily primed PCR (AP-PCR) was developed primarily for genotyping purposes, random-primer-based applications for the identification of microorganisms have been described (5, 7, 22). The applicability of AP-PCR, however, is hampered fundamentally by nonstringent conditions. To establish a conventional PCR based on stringent reaction conditions, knowing the target gene sequences is necessary. In an attempt to design specific oligonucleotide primers for the detection of C. glabrata by conventional PCR, AP-PCR-derived agarose gel electrophoresis patterns of different C. glabrata genotypes were analyzed and compared with the patterns of other yeast species. Bands covering all C. glabrata genotypes, but absent in patterns of other Candida species, were selected for further analyses. Subsequently, the DNA from these bands was isolated and sequenced directly. From these sequences, putative C. glabrata-specific PCR primer pairs were designed and tested for sensitivity and specificity on a large panel of C. glabrata and other fungal isolates.

Fungal strains and DNA extraction.

The fungal isolates were maintained on Kimmig's agar (Merck, Darmstadt, Germany) supplemented with 0.1 g of chloramphenicol per liter and 0.1 g of tetracycline per liter. All Candida isolates were tested in their anamorphic form. The test isolates were identified by use of the ID 32 C identification system for yeasts (bioMérieux, Marcy-l'Etoile, France) and were confirmed by standard taxonomic procedures (1, 18). Nucleic acid (NA) extracts were prepared from 10 ml of an 18-h culture in yeast extract-peptone-glucose broth with shaking at 37°C. Yeast cells were pelleted by centrifugation at 5,000 × g for 10 min, resuspended in 600 μl of sorbitol buffer with 200 U of lyticase, and incubated at 30°C for 0.5 h. Spheroplasts were centrifuged at 5,000 × g for 5 min and resuspended in 180 μl of ATL buffer from the QIAamp tissue kit (Qiagen, Hilden, Germany). Subsequently, NAs were extracted with the QIAamp tissue kit, following the manufacturer's recommendations. NA samples were eluted with distilled water and adjusted to a final concentration of 1 μg/ml according to A₂₆₀ values.

Analysis of AP-PCR bands and generation of specific primers.

In preliminary studies, genotyping of C. glabrata isolates was performed by an optimized and standardized AP-PCR protocol, resulting in three major genotypes (3). AP-PCR with random primer AP50-1 (5′-GAT TCA GAC C-3′) was done as described previously, although with drastically prolonged ramp times (7 min) (3, 8). Putative C. glabrata species-specific bands (approximately 0.35, 1.8, and 2.8 kbp) (Fig. 1) were excised and NAs were extracted from agarose gel by using Ultrafree-DA (Millipore, Bedford, Mass.). Subsequently, DNA was sequenced directly using the Taq DyeDeoxy terminator cycle sequencing kit (Applied Biosystems, Foster City, Calif.) and the ABIPRISM 310 genetic analyzer automated sequencing system (Applied Biosystems) (data not shown). From the derived sequence information, a number of putative C. glabrata-specific primer pairs were designed by using Primer Premier version 4.04 software (Premier Biosoft International, Palo Alto, Calif.) and were tested for their applicability in PCR for C. glabrata detection (data not shown). The primer pair CG-R31-1 (5′-AAG AAG GCT GCC TGT TGT AAT G-3′)–CG-R31-2 (5′-CAC TTA TCT AAA CAA CGG TGG C-3′), which was derived from a fragment designated CGR31, was further studied. Using this primer pair, a 978-bp product was amplified (Fig. 2).

FIG. 1.

Results of agarose gel electrophoresis of AP-PCR with random primer AP50-1 demonstrating patterns of different C. glabrata genotypes compared with other Candida species. Lane M, DNA molecular size marker (1-kb/100-bp DNA ladder); lane 1, C. glabrata genotype A; lane 2, C. glabrata genotype B; lane 3, C. glabrata genotype C; lane 4, C. albicans; lane 5, C. guilliermondii; lane 6, C. kefyr; lane 7, C. krusei; lane 8, C. norvegensis; lane 9, C. parapsilosis; lane 10, C. tropicalis. Sizes are marked in kilobase pairs on the left.

FIG. 2.

PCR-generated 978-bp product obtained with C. glabrata-specific primer pair. Lane M, DNA molecular size marker (1-kb/100-bp DNA ladder); lane 1, C. glabrata genotype A; lane 2, C. glabrata genotype B; lane 3, C. glabrata genotype C; lane 4, C. tropicalis; lane 5, C. krusei; lane 6, C. parapsilosis; lane 7, C. guilliermondii; lane 8, C. albicans; lane 9, C. kefyr; lane 10, C. inconspicua; lane 11, C. lusitaniae; lane 12, C. lambica; lane 13, S. cerevisiae; lane 14, no DNA template. Sizes are marked in base pairs on the left.

Analysis of 978-bp CGR31 fragment.

The 978-bp PCR product of the CGR31 fragment was sequenced as described above. The NA sequence alignment of the 978-bp fragment by FASTA program searches of the EMBL Data Library showed 58.1% homology to a part of a 3.8-kb DNA fragment which contained, in addition to two unassigned open reading frames (ORFs), the gene encoding a putative mitochondrial ribosomal protein of Saccharomyces cerevisiae designated MRP-L6p (12). The nucleotide sequence of the 978-bp CGR31 fragment enclosed an ORF, presumably beginning before the start of the fragment and extending to base pair 226, which codes for a polypeptide of 75 amino acids whose deduced amino acid sequence showed 84.0% identity to that of the S. cerevisiae ORF product P32899. This ORF was found to be neighbored by MRP-L6 (12). Disruption of MRP-L6 led to the phenotype of a mitochondrial translation-defective yeast mutant, suggesting that the MRP-L6 gene is coding for an essential component of yeast mitochondrial ribosomes (12).

C. glabrata-specific PCR of CGR31 fragment.

Amplification reactions of C. glabrata-specific PCR using primers CG-R31-1 and CG-R31-2 were carried out in a volume of 100 μl of a PCR mixture containing 10 μM (each) dATP, dCTP, dGTP, and dTTP and 1 μg of DNA template. The master mixture contained 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl₂, 50 pmol of each primer, and 2.5 U of DNA polymerase. The amplification was performed in an automated thermocycler with a hot bonnet (Hybaid, Teddington, United Kingdom). The optimized thermal cycling conditions were 30 cycles of denaturation at 95°C for 1 min (5 min for the first cycle), annealing at 58°C for 1 min, and polymerization at 72°C for 2 min. Amplified products (10 μl) were resolved by 2% agarose gel electrophoresis at 150 V for 1.5 h. The gel was stained with ethidium bromide and exposed to UV light (254-nm wavelength) to visualize the amplified products. To avoid possible contamination, all reactions were done as described previously (19).

Sensitivity and specificity of C. glabrata-specific PCR.

A total of 173 yeast isolates, including 95 reference and clinical isolates of C. glabrata as well as 78 isolates of other yeast species, were studied (Tables 1 and 2). The clinical isolates were obtained from human (n = 84) and animal (n = 9) specimens collected in different German and European medical centers (3). Each of the 53 human C. glabrata isolates was from a different subject. The remaining 40 human isolates were from 13 patients and were obtained from diverse specimens or at different times of isolation. In tests of the clinical and reference strains of C. glabrata, the expected amplicon was successfully generated from all human and animal C. glabrata isolates tested (Table 1). With regard to the three major genotypes of C. glabrata, no differences in PCR products were observed (Fig. 2). The primer pair amplified the predicted PCR products in reaction mixtures with about 10 pg of total DNA. To investigate the specificity of PCR for putative C. glabrata isolates, 66 isolates of 16 other Candida species were tested (Table 2). Since C. glabrata is specifically related to S. cerevisiae (2), 12 isolates of this species were included. Additionally, 14 mold strains were encompassed. No PCR products were amplified when DNA isolates from fungi other than C. glabrata were used as templates (Fig. 2; Table 2).

TABLE 1.

Number and source of C. glabrata isolates tested and results of PCR amplification

Source	No. of isolates positive/no. tested
Reference strains	2/2
Humans
Blood	16/16
Normally sterile tissues	2/2
Respiratory tract specimens	14/14
Oral and gastrointestinal specimens	24/24
Urogenital specimens	17/17
Cutaneous and other specimens	11/11
Total human isolates	84/84
Animals	9/9
Total from all sources	95/95

TABLE 2.

Specificity controls used in the study

Species	No. negative/no. tested
	Reference strainsa	Clinical strainsb	Total
Yeasts
Candida albicans	2/2	18/18	20/20
Candida dubliniensis	0	3/3	3/3
Candida famata	0	3/3	3/3
Candida guilliermondii	1/1	4/4	5/5
Candida inconspicua	0	3/3	3/3
Candida kefyr (Kluyveromyces marxianus)	1/1	1/1	2/2
Candida krusei (Issatchenkia orientalis)	2/2	4/4	6/6
Candida lambica	0	3/3	3/3
Candida lusitaniae	0	4/4	4/4
Candida norvegensis	0	2/2	2/2
Candida parapsilosis	2/2	2/2	4/4
Candida pelliculosa	0	1/1	1/1
Candida rugosa	0	1/1	1/1
Candida tropicalis	2/2	3/3	5/5
Candida utilis	0	2/2	2/2
Candida valida	0	2/2	2/2
Saccharomyces cerevisiae	1/1	11/11	12/12
Total yeast strains tested	11/11	67/67	78/78
Molds
Acremonium strictum	0	1/1	1/1
Aspergillus flavus	1/1	1/1	2/2
Aspergillus fumigatus	1/1	1/1	2/2
Aspergillus niger	1/1	2/2	3/3
Aspergillus terreus	1/1	1/1	2/2
Emericella nidulans	1/1	0	1/1
Fusarium proliferatum	1/1	0	1/1
Paecilomyces lilacinus	1/1	0	1/1
Scedosporium apiospermum	0	1/1	1/1
Total mold strains tested	7/7	7/7	14/14
Total yeast and mold strains tested	18/18	74/74	92/92

^aYeast reference strains used in the study included C. albicans ATCC 36801 and ATCC 44374, C. guilliermondii ATCC 90877, C. kefyr DSM 11954, C. krusei ATCC 6258 and ATCC 90878, C. parapsilosis ATCC 22019 and DSM 11955, C. tropicalis ATCC 28707 and ATCC 90874, and S. cerevisiae DSM 70449. Mold reference strains used in the study included A. flavus DSM 1959, A. fumigatus DSM 819, A. niger DSM 2143, A. terreus DSM 1958, E. nidulans DSM 820, F. proliferatum DSM 840, and P. lilacinus DSM 846. 

^bAll clinical isolates were collected in the diagnostic laboratory of the Institute of Medical Microbiology, University of Münster. 

Compared with species-specific structural genes, rRNA genes are attractive targets for amplification-based detection assays, since these genes are present at a high copy number, thus increasing the sensitivity of the PCR (15). Furthermore, rRNA genes are composed of regions of higher and lower evolutionary conservation, thereby enabling amplification at different taxonomic levels. Consequently, nuclear ribosomal genes have been widely used as a target for detection of fungal microorganisms by molecular methods (13, 15, 20). Although mitochondrial ribosomal sequences seem to offer the same diagnostic advantages, their use as diagnostic targets is as yet more uncommon. Regarding fungal pathogens, several PCR applications based on mitochondrial ribosomal gene sequences have been developed for Pneumocystis carinii (6, 10, 14). Recently, the use of conserved and variable regions of different domains of mitochondrial small-subunit rRNA of Agrocybe spp. was proposed as a specific molecular marker for differentiation of Basidiomycota (11). To our knowledge, the present study reports the first PCR application of a mitochondrial ribosome sequence as a target to identify one of the medically important Candida species.

As shown in our study, the selection and analysis of taxon-specific bands within patterns of AP-PCR may offer an alternative way to obtain NA sequences which could be used to deduce specific diagnostic primers for conventional high-stringency PCR. In summary, the newly described PCR will facilitate rapid, sensitive, and specific identification of C. glabrata isolates. Since rapid detection of putative azole-resistant yeasts is of clinical importance, further research is desirable to apply this molecular approach to other emerging yeasts and to adjust it to direct PCR examination of clinical specimens.

Nucleotide sequence accession number.

For the sequence of the 978-bp PCR product of the CGR31 fragment, the GenBank accession number AJ289782 has been assigned.