Rapid molecular differentiation and genotypic heterogeneity among Candida parapsilosis and Candida orthopsilosis strains isolated from clinical specimens in Kuwait

Although Candida albicans is the most commonly isolated Candida species from bloodstream infections, Candida parapsilosis has recently emerged as an important nosocomial pathogen causing invasive candidiasis in critically ill hospitalized patients and low-birth-weight infants (Pappas et al., 2003; Almirante et al., 2006; Bassetti et al., 2006; Pfaller & Diekema, 2007; Yong et al., 2008). In Kuwait, C. parapsilosis is the second most frequently isolated Candida species from bloodstream infections, in both adult and paediatric patients (Mokaddas et al., 2007; Asadzadeh et al., 2008; Al-Sweih et al., 2009). The species is frequently found on the skin and thus can be transmitted exogenously on the hands of healthcare professionals (Bonassoli et al., 2005). Infections usually occur in association with catheter/prosthetic devices, where C. parapsilosis is capable of forming biofilms (Weems, 1992).

It has been known for some time that C. parapsilosis represents a complex of genetically different strains/species and consists of at least three genetically distinct groups (I, II and III) (Lin et al., 1995; Roy & Meyer, 1998). Detailed molecular analyses including multilocus sequence typing studies have recently recognized C. parapsilosis groups II and III as separate species, Candida orthopsilosis and Candida metapsilosis, within the C. parapsilosis complex, whilst group I isolates have been proposed to retain the name C. parapsilosis (Tavanti et al., 2005). Although the exact prevalence of C. orthopsilosis and C. metapsilosis infections remains unknown, several recent studies indicate a significant contribution of these two newly described species towards human diseases, including bloodstream infections (Lasker et al., 2006; Kocsube et al., 2007; Tavanti et al., 2007; Yong et al., 2008; Gomez-Lopez et al., 2008; Lockhart et al., 2008a). A major impediment of the lack of this information is the application of tedious and complex techniques that are required for the identification of C. orthopsilosis and C. metapsilosis strains among clinical isolates (Tavanti et al., 2005; Kocsube et al., 2007; Yong et al., 2008). In this study, we have developed a simple PCR-based method for rapid identification of clinical isolates of C. parapsilosis, C. orthopsilosis and C. metapsilosis by using species-specific primers derived from unique sequences within the internally transcribed spacer (ITS) 1 and ITS2 of the ITS1–5.8S rRNA–ITS2 region. Using this protocol, we analysed 114 phenotypically identified isolates of C. parapsilosis recovered from various clinical specimens and the results of these studies are presented here.

Reference strains, clinical isolates and DNA isolation. The reference or well-characterized clinical strains of Candida and other fungal species were obtained from the American Type Culture Collection (ATCC) or the Centraalbureau voor Schimmelcultures (CBS). These included C. parapsilosis (ATCC 22019), C. orthopsilosis (ATCC 96139), C. metapsilosis (ATCC 96143), Lodderomyces elongisporus (clinical isolate), C. albicans (ATCC 76615), Candida dubliniensis (CD36), Candida tropicalis (ATCC 750), Candida glabrata (ATCC 15545), Candida krusei (ATCC 6258), Candida haemulonii (CBS 5149), Aspergillus fumigatus (CBS 113.26/ATCC 1028), Aspergillus flavus (CBS 113.2), Cryptococcus neoformans (ATCC 90112), Trichosporon asahii (CBS 2530) and Trichosporon jirovecii (CBS 6864). A total of 114 randomly selected C. parapsilosis strains isolated from the bloodstream (n=66) and other clinical specimens (n=48) from 1996 to 2007 in Kuwait and identified by CHROMagar Candida and the VITEK 2 yeast identification system (bioMérieux) were analysed. The cultures were grown on Sabouraud dextrose agar (Difco; Becton-Dickinson) and the genomic DNA from reference and clinical Candida species and other fungal pathogens was isolated as described previously (Ahmad et al., 2002; Khan et al., 2008a).

Primer construction and PCR amplification. Nine primer pairs were synthesized for pan-fungal, C. parapsilosis-complex or Candida species-specific amplification of various target sequences (Table 1). The primer pair CPDET and CTS1R (Ahmad et al., 2002) was used to confirm that all clinical isolates used in this study belonged to the C. parapsilosis complex. The species specificity of the primer pairs CPAF+CPAR, CORF+CORR and CMEF+CMER for C. parapsilosis, C. orthopsilosis and C. metapsilosis, respectively, was indicated by BLAST searches (). PCR amplification of the secondary alcohol dehydrogenase (SADH) gene followed by BanI digestion of amplicons to generate three unique RFLP patterns for the three species of the C. parapsilosis-complex strains was also carried out by using the SADHF1+SADHR1 (Tavanti et al., 2005) or SADHF2+SADHR2 primers (Table 1). The pan-fungal primers ITS1+ITS4 (Ahmad et al., 2005) and NL-1+NL-4 (Kurtzman & Robnett, 1997) were used for PCR-based sequencing of the ITS (ITS1–5.8S rRNA–ITS2) region and the D1/D2 regions of the 28S rRNA gene, respectively. The primers IGSF and IGSR were designed for PCR-RFLP analyses of the intergenic spacer (IGS1) between the 28S and 5S rRNA genes from C. parapsilosis-complex isolates.

Table 1. Specific features and sequences of the PCR primers used in this study

PCR amplification of various gene targets was carried out as follows. The reactions in a total volume of 50 µl contained 1x AmpliTaq PCR buffer I, 1 U AmpliTaq DNA polymerase (Applied Biosystems), 10 pmol each of forward and reverse primer (Table 1), 2 µl genomic DNA and 0.1 mM dNTPs. The cycling conditions included an initial denaturation at 95 °C for 5 min, followed by 30 cycles of 95 °C for 1 min, 63 °C (for the CPDET+CTS1R, CPAF+CPAR, CORF+CORR and CMEF+CMER primer pairs), 60 °C (for the ITS1+ITS4 and NL-1+NL-4 primer pairs) or 50 °C (for the SADHF1+SADHR1 or SADHF2+SADHR2 primer pairs) for 30 s and 72 °C for 1 min, with a final extension at 72 °C for 10 min. The reactions for amplification of the IGS1 region in a final volume of 50 µl contained 15 µl 3.3x rTth DNA polymerase buffer, 5 µl 10x magnesium acetate, 2 U rTth DNA polymerase (Applied Biosystems), 10 pmol each of IGS1F and IGS1R primers, 2 µl genomic DNA and 0.1 mM dNTPs. The cycling conditions comprised one cycle of 95 °C for 3 min and 96 °C for 1 min, followed by 35 cycles of 95 °C for 1 min, 60 °C for 1 min and 72 °C for 3 min, with a final extension period of 10 min at 72 °C. Inclusion of the spike to 96 °C for 1 min during denaturation of the genomic DNA resulted in improved amplification of the IGS region, as has also been observed previously for higher G+C DNA regions (Ahmad et al., 2007). For species-specific detection of C. parapsilosis, C. orthopsilosis and C. metapsilosis strains, PCR products (20 µl) were run on 2 % (w/v) agarose gels, as described previously (Ahmad et al., 2004). An amplicon was obtained only with the corresponding species-specific primer pair. For other PCRs, a sample (5 µl) was run on a 2 % (w/v) agarose gel to confirm amplification of a DNA fragment of the expected size. The remaining PCR-amplified fragments were purified by using QIAquick PCR purification columns (Qiagen) according to the kit instructions. The purified amplicons were either digested with restriction enzymes to generate RFLP patterns (for the SADH gene and the IGS1 region) or subjected to DNA sequencing.

RFLP analyses of the SADH gene fragment. SADH amplicons obtained with the SADHF1+SADHR1 or SADHF2+SADHR2 primers from reference and clinical isolates were digested with BanI. The reactions in a final volume of 25 µl contained 10 µl purified amplicon, 2.5 µl 10x NE Buffer 4, 2 µl (10 U) BanI (New England Bio-Laboratories) and 10.5 µl sterile distilled water and were incubated at 37 °C for 3 h. The products were separated on 2 % agarose gels to generate RFLP patterns. The 717 bp amplicons obtained with the SADHF1+SADHR1 primers yielded two fragments of 521 and 196 bp, a single undigested fragment of 717 bp, and four fragments of 375, 189, 93 and 60 bp from C. parapsilosis, C. orthopsilosis and C. metapsilosis, respectively (Tavanti et al., 2005). The 310 bp amplicons generated with SADHF2+SADHR2 designed in this study should yield two fragments of 166 and 144 bp, a single undigested fragment of 310 bp, and two fragments of 268 and 42 bp from C. parapsilosis, C. orthopsilosis and C. metapsilosis, respectively.

RFLP analysis of the IGS1 region DNA fragment. Genotypic heterogeneity among clinical C. parapsilosis or C. orthopsilosis strains was determined by restriction digestion of IGS1 region amplicons with a battery of frequently cutting restriction enzymes to generate RFLPs. The restriction digestions containing 10 µl of the purified amplicon, 2.5 µl appropriate 10x NE Buffer, 2 µl (10 U) of each restriction enzyme (DdeI, RsaI, HinfI, MspI or AluI; New England Bio-Laboratories) and 10.5 µl sterile distilled water were carried out at 37 °C for 4 h and the products were separated on 1 % agarose gels.

DNA sequencing of the ITS region and the D1/D2 region of the 28S rRNA gene. The results of species-specific identification of clinical isolates were confirmed by DNA sequencing of the ITS region and the D1/D2 region of the 28S rRNA gene and both strands were sequenced. The amplicons obtained with the ITS1+ITS4 primers were sequenced by using ITS1FS, ITS2, ITS3 and ITS4RS as sequencing primers as described previously (Ahmad et al., 2005; Al-Sweih et al., 2005; Al-Obaid et al., 2006). The amplicons obtained with the NL-1+NL-4 primers were sequenced by using NL-1, NL-2A, NL-3A and NLR3R as sequencing primers as described previously (Khan et al., 2008b). Reverse complements were generated and aligned with forward sequences using CLUSTAL W () during sequence assembly (Khan et al., 2008b). BLAST searches were performed for species identification. Pairwise comparisons and multiple sequence alignments were also performed with CLUSTAL W.

Species-specific amplification of C. parapsilosis, C. orthopsilosis and C. metapsilosis
The species specificity of the CTS1R+CPDET primers for detection of DNA from C. parapsilosis-complex strains has been described previously (Ahmad et al., 2002). PCR amplification with the CPAF+CPAR primers (Table 1) yielded an amplicon of ∼379 bp with DNA from the reference strain of C. parapsilosis (Fig. 1, lane 1), but not from C. orthopsilosis (Fig. 1, lane 4) or C. metapsilosis (Fig. 1, lane 7), as expected. Similarly, PCR amplification with CORF+CORR and CMEF+CMER yielded amplicons of the expected size (367 and 374 bp, respectively) only with DNA from C. orthopsilosis (Fig. 1, lane 5) and C. metapsilosis (Fig. 1, lane 9), respectively, but not from the other two C. parapsilosis-complex species. No amplicon was obtained with any of the three primer pairs using DNA isolated from reference strains of other Candida or other common fungal species (data not shown). These results established the specificity of the three primer pairs for species-specific detection of C. parapsilosis, C. orthopsilosis and C. metapsilosis. PCR with C. parapsilosis-complex-specific (CPDET+CTS1R) primers (Ahmad et al., 2002) yielded an amplicon of ∼90 bp from all clinical C. parapsilosis isolates (n=114) identified by phenotypic methods (CHROMagar Candida and VITEK 2 yeast identification system) (data not shown). In uniplex PCR with primers CPAF+CPAR, CORF+CORR or CMEF+CMER (Table 1), five and 109 isolates yielded an amplicon with C. orthopsilosis-specific (Fig. 1b, lane 2) and C. parapsilosis-specific (Fig. 1b, lanes 4, 7 and 10) primer pairs, respectively, whilst no isolate yielded an amplicon with C. metapsilosis-specific primers.

(16K): Fig. 1. (a) Agarose gel of PCR-amplified products obtained from C. parapsilosis (ATCC 22019) (lanes 1–3), C. orthopsilosis (ATCC 96139) (lanes 4–6) and C. metapsilosis (ATCC 96143) (lanes 7–9) reference strains amplified with primers CPAF+CPAR (lanes 1, 4 and 7), CORF+CORR (lanes 2, 5 and 8) and CMEF+CMER (lanes 3, 6 and 9). (b) Agarose gel of amplicons from four representative clinical isolates amplified individually with primers CPAF+CPAR (lanes 1, 4, 7 and 10), CORF+CORR (lanes 2, 5, 8 and 11) and CMEF+CMER (lanes 3, 6, 9 and 12). As shown in lanes 1–3, 4–6, 7–9 and 10–12, each isolate yielded an amplicon with only one of the three species-specific primer pairs. Thus the isolate in lanes 1–3 was identified as C. orthopsilosis, whilst the isolates in lanes 4–6, 7–9 and 10–12 were identified as C. parapsilosis. Lane M is a 100 bp DNA ladder and the positions of migration of the 100, 300 and 600 bp fragments are indicated.

Identification based on PCR-RFLP analyses of the SADH gene
To validate the results of our protocol, PCR-RFLP analyses of the SADH gene were also carried out as described by Tavanti et al. (2005). SADH amplicons obtained with primers SADHF1+SADHR1 from the reference C. parapsilosis, C. orthopsilosis and C. metapsilosis strains yielded RFLP patterns as described previously (Tavanti et al., 2005). However, a PCR product was obtained only from 75 isolates, whilst five strains did not yield an amplicon during initial analyses of 80/114 C. parapsilosis-complex strains isolated in Kuwait. The latter five isolates were easily amplified with C. parapsilosis-complex-specific or species-specific primers (Table 1) and were identified as C. parapsilosis strains, indicating that lack of amplification was most likely due to differences in primer-binding sequences in the SADH gene. These isolates also could not have belonged to group IV advocated by Iida et al. (2005) or L. elongisporus (Lockhart et al., 2008b) due to differences in the target sequences of the CPAF and CPAR primers. Lockhart et al. (2008a) also reported recently that 47/1976 C. parapsilosis-complex strains isolated from diverse geographical locations around the world could not be speciated, probably due to lack of amplification of the SADH gene with the SADHF1+SADHR1 primers. Thus a new primer pair (SADHF2+SADHR2) was designed for SADH gene amplification. The target region of the SADH gene was carefully selected within the region originally amplified by Tavanti et al. (2005) to contain variable patterns of BanI restriction sites in the resulting amplicons from C. parapsilosis, C. orthopsilosis and C. metapsilosis, as described in Methods. The shorter SADH gene fragment (∼310 bp) obtained with the SADHF2+SADHR2 primers from the reference C. parapsilosis, C. orthopsilosis and C. metapsilosis strains yielded species-specific RFLP patterns, as described in Methods. The SADH gene was amplified with the SADHF2+SADHR2 primers from all of the 114 phenotypically identified C. parapsilosis-complex isolates (data from six isolates are shown in Fig. 2a, lanes 1–6). The BanI-generated RFLP patterns identified 109 isolates (data from four isolates are shown in Fig. 2b, lanes 1–4) with two fragments of 166 and 144 bp as C. parapsilosis and five isolates (data from two isolates are shown in Fig. 2b, lanes 5 and 6) with the original undigested fragment of ∼310 bp as C. orthopsilosis.

(13K): Fig. 2. Agarose gel of amplicons of the SADH gene using primers SADHF2+SADHR2 (a) and RFLP patterns of the amplicons from six selected clinical C. parapsilosis-complex strains (b). The digested fragments in lanes 1–4 are characteristic for C. parapsilosis, whilst those in lanes 5 and 6 are characteristic for C. orthopsilosis. Lane M is a 100 bp DNA ladder and the positions of migration of the 100, 300 and 600 bp fragments are indicated.

DNA sequencing studies
Direct DNA sequencing of the ITS region was carried out to confirm the identification of all five C. orthopsilosis strains (isolates 301/97, 304/97, 709/04, 1782/06 and 1056/04). Two distinct genotypes were identified. The ITS region sequences of three C. orthopsilosis isolates (301/97, 304/97 and 1056/04) were identical and differed at three nucleotide positions with the sequences from the other two isolates (insertion of T in isolates 709/04 and 1782/06). A pairwise comparison of ITS region sequences showed fewer differences with the corresponding sequence from the reference C. orthopsilosis strain than with sequences from reference C. parapsilosis or C. metapsilosis strains (Table 2). Four distinct genotypes among the clinical C. dubliniensis strains were also identified recently based on ITS region sequences (Gee et al., 2002; Al-Sweih et al., 2005). The D1/D2 region sequences of two C. orthopsilosis isolates (301/97 and 709/040) representing the two ITS region-based genotypes exhibited a single nucleotide difference with each other and were nearly identical (zero or one nucleotide difference) with the sequence from the reference C. orthopsilosis strain. Two randomly selected C. parapsilosis strains were also sequenced in the ITS region (isolates 113/98 and 119/98) and D1/D2 region (isolates 113/98 and 426/04). The ITS and D1/D2 region sequences of both C. parapsilosis isolates were identical and showed 100 % identity with the corresponding sequences from the reference C. parapsilosis strain.

Table 2. Nucleotide differences within the ITS region (ITS1–5.8S rRNA–ITS2) among clinical C. parapsilosis and C. orthopsilosis isolates in Kuwait in comparison with reference strains

Thus species-specific amplification of the ITS region and PCR-RFLP analyses of the SADH gene were concordant for all clinical isolates in Kuwait and identified 109 isolates as C. parapsilosis and five isolates as C. orthopsilosis. None of the C. parapsilosis-complex strains in Kuwait was identified as C. metapsilosis. Our results are in agreement with recent reports showing that some clinical isolates identified as C. parapsilosis by conventional carbon-assimilation methods, such as the VITEK 2 Yeast ID system, are in fact isolates of C. orthopsilosis (Tavanti et al., 2007; Gomez-Lopez et al., 2008; Lockhart et al., 2008a). The absence of C. metapsilosis among 114 C. parapsilosis-complex strains in Kuwait also supports the infrequent occurrence of this species, particularly among bloodstream isolates (Kocsube et al., 2007; Tavanti et al., 2007). The exact prevalence of C. orthopsilosis and C. metapsilosis strains among clinical Candida species isolates or candidaemia cases is not known, as only a few studies have been carried out in the last couple of years. A comparison of the prevalence of C. parapsilosis, C. orthopsilosis and C. metapsilosis among phenotypically characterized C. parapsilosis-complex strains in various studies is presented in Table 3. Only two studies detected the presence of both of the newly described species, i.e. C. orthopsilosis and C. metapsilosis, among C. parapsilosis-complex strains (Gomez-Lopez et al., 2008; Lockhart et al., 2008a). Similar to our results, Tavanti et al. (2005, 2007) also did not detect any C. metapsilosis strains, whilst Kocsube et al. (2007) identified 7.6 % (2/26) of C. parapsilosis-complex isolates as C. metapsilosis but did not find any C. orthopsilosis strains (Table 3). It is not known at present whether C. orthopsilosis and C. metapsilosis occupy any particular niche in the human body.

Table 3. Occurrence of C. parapsilosis, C. orthopsilosis and C. metapsilosis among phenotypically identified C. parapsilosis-complex strains in various studies

Genotypic heterogeneity among clinical C. orthopsilosis isolates
The genotypic heterogeneity among 81 randomly selected C. parapsilosis and all five C. orthopsilosis strains was studied further by PCR-RFLP analyses of the IGS1 region. Only RsaI and HinfI yielded discriminatory RFLP patterns of the IGS1 region among C. parapsilosis and C. orthopsilosis strains, respectively. All 81 C. parapsilosis strains yielded an IGS1 amplicon of ∼2200 bp (data not shown). Forty-nine isolates yielded one pattern (designated pattern A) (Fig. 3a, lanes 1–4), whilst 32 isolates exhibited a second pattern (designated pattern B) (Fig. 3a, lanes 5–7) during RFLP analyses of IGS1 with RsaI. Almost equal numbers of C. parapsilosis strains isolated from the blood yielded pattern A (23/47, 49 %) and pattern B (24/47, 51 %). However, the vast majority (26/34, 76 %) of strains isolated from other body sites yielded pattern A (Table 4). All five C. orthopsilosis strains yielded a shorter (∼1700 bp) IGS1 region amplicon (data not shown). Thus the length of the IGS1 region was also species-specific. Other yeasts such as Trichosporon species also exhibit IGS1 length variations (Sugita et al., 2002; Rodriguez-Tudela et al., 2007). Three C. orthopsilosis isolates (Fig. 3b, lanes 1, 2 and 4) yielded one pattern (designated pattern C), whilst two isolates (Fig. 3b, lanes 3 and 5) exhibited a second pattern (designated pattern D) during HinfI-generated RFLP analyses (Table 4). Application of tedious and technically demanding amplified fragment length polymorphism analyses has also recently shown genotypic heterogeneity and identified four major clusters among 13 C. orthopsilosis strains from Italy (Tavanti et al., 2007). All three C. orthopsilosis strains yielding the pattern C RFLP also exhibited the same genotype based on ITS region sequences and were isolated from the bloodstream (Table 4). Whether these findings have any significance with regard to the virulence of this genotype remains to be determined. Although only two C. parapsilosis strains were sequenced, the PCR-RFLP results of the IGS1 region were consistent with earlier observations showing greater sequence variations among C. orthopsilosis than C. parapsilosis clinical isolates (Tavanti et al., 2005, 2007).

(32K): Fig. 3. Representative agarose gel of restriction digestion of IGS1 region amplicons with RsaI from seven selected C. parapsilosis strains (a) and with HinfI from the five C. orthopsilosis strains (b). In (a), the RFLP patterns in lanes 1–4 were classified as pattern A, whilst those in lanes 5–7 were classified as pattern B. In (b), the RFLP patterns in lanes 1, 2 and 4 were classified as pattern C, whilst those in lanes 3 and 5 were classified as pattern D. Lane M, BstEII-digested λ DNA ladder, with fragment sizes indicated (bp).

Table 4. Genotypic heterogeneity among 81 randomly selected C. parapsilosis and all five C. orthopsilosis strains isolated in Kuwait by PCR-RFLP patterns of the 28S–5S rRNA IGS region

In conclusion, we have described a simple amplification protocol for rapid discrimination among C. parapsilosis, C. orthopsilosis and C. metapsilosis clinical isolates. We have also reported the occurrence of C. orthopsilosis among clinical isolates of the C. parapsilosis complex from the Arabian Gulf region for the first time. Three out of five C. orthopsilosis strains were isolated from the bloodstream and belonged to a single genotype. The absence of C. metapsilosis indicates that this species probably has a limited clinical significance. Technical support received from Daad Farhat and Rachel Chandy is gratefully acknowledged. This study was supported by Research Administration grant YM04/06 and the College of Graduate Studies, Kuwait University.