MTL genotypes, phenotypic switching, and susceptibility profiles of Candida parapsilosis species group compared to Lodderomyces elongisporus

Reference isolates of Candida parapsilosis (n = 8), Candida metapsilosis (n = 6), Candida orthopsilosis (n = 7), and Lodderomyces elongisporus (n = 11) were analyzed to gain insight into their pathobiology and virulence mechanisms. Initial evaluation using BBL Chromagar Candida medium misidentified L. elongisporus isolates as C. albicans. Polymerase chain reaction analysis of isolate MTL idiomorphs revealed that all C. parapsilosis isolates were MTLa homozygous and no MTL α1, α2, a1, or a2 gene was detected in L. elongisporus isolates. For C. orthopsilosis, two isolates were MTLa homozygous and five were MTL-heterozygous. Similarly, one C. metapsilosis isolate was MTLα homozygous whereas five were MTL-heterozygous. Isolate phenotypic switching analysis revealed potential phenotypic switching in the MTLα homozygous C. metapsilosis isolate, resulting in concomitant elongated cell formation. Minimum inhibitory concentrations of fluconazole (FLC) and FK506, alone or in combination, were determined by checkerboard assay, with data analyzed using the fractional inhibitory concentration index model. Synergistic or additive effects of these compounds were commonly observed in C. parapsilosis and L. elongisporus isolates. No killer activity was observed in the studied isolates, as determined phenotypically. No significant difference in virulence was seen for the four species in a Galleria mellonella model (P > 0.05). In conclusion, our results demonstrated phenotypic switching of C. metapsilosis CBS 2315 and that FLC and FK506 represent a promising drug combination against C. parapsilosis and L. elongisporus. The findings of the present study contribute to our understanding of the biology, diagnosis, and new possible treatments of the C. parapsilosis species group and L. elongisporus.

Introduction metapsilosis strains were shown to exhibit killer activity, in contrast to C. parapsilosis and C. orthopsilosis [39].
Published studies strongly suggest that Galleria mellonella provides a good alternative model for studying virulence in several fungi, including the major human fungal pathogens Aspergillus spp. [40], Candida spp. [41], and Cryptococcus spp. [42]. As the emerging resistance in the entire group of C. parapsilosis has led to difficult-to-treat infections [8,9], it is therefore important to understand the virulence potential of C. parapsilosis and its relatives. Furthermore, the development of drug combinations to address emerging antifungal resistance is critical for the management of patients, particularly in the case of invasive diseases [43][44][45][46][47][48]. Accordingly, Sun et al. [43] suggested the antifungal potential of a combination therapy with calcineurin pathway inhibitors (i.e., FK506, also known as tacrolimus) as a replacement for the activity of several azoles to combat azole resistance in C. albicans.
In the present study, we analyzed C. parapsilosis and its relatives to further characterize (i) the accuracy of their identification on BBL Chromagar Candida medium, (ii) their MTL genotypes and the occurrence of phenotypic switching, (iii) their killer activity and virulence in the G. mellonella model, and (iv) the ability of FK506 to enhance the isolate susceptibility to fluconazole (FLC).

Isolates
The strains used in this study, i.e., C. metapsilosis (n = 6), C. orthopsilosis (n = 7), C. parapsilosis (n = 8), and L. elongisporus (n = 11), as well as their origin, identification number, and place of origin, are listed in S1 Table. C. parapsilosis species group isolates were provided by the culture collection of Centraalbureau voor Schimmelcultures (housed at Westerdijk Fungal Biodiversity Institute, Utrecht, the Netherlands) and L. elongisporus isolates were from the collection of the Molecular Genetics and Microbiology Department (Duke University, Durham, NC). All strains were subcultured on yeast extract-peptone dextrose agar (YEPD; Difco, Detroit, MI), at 37˚C for 3 d, prior to analysis. The identities of all strains were verified prior to this study by sequencing the internal transcribed spacer region.

Growth on chromogenic medium
All strains were inoculated in parallel onto BBL Chromagar Candida medium (Becton Dickinson and Company, Sparks, MD) and Sabouraud dextrose agar (SDA; Merck, Darmstadt, Germany), and incubated at 37˚C for 72 h [15]. Following the incubation period, the isolates were evaluated on the basis of colony color. C. albicans SC5314 and C. tropicalis YJM57 were used as control strains.
Genomic DNA isolation, polymerase chain reaction (PCR) amplification, and DNA sequencing For genomic DNA isolation, all strains were collected directly from YEPD plates after 2 d of growth. Genomic DNA isolation was performed using the MasterPure Yeast DNA Purification Kit (Epicentre Biotechnologies, Madison, WI) according to the manufacturer's instructions.
All PCR assays were conducted in a PTC-200 automated thermal cycler (BioRad, Hercules, CA); 300 ng DNA was used as a template for amplification in a 25 0μL reaction mixture containing 10 pM of each primer, 2 mM of each nucleotide (dATP, dCTP, dGTP, and dTTP), 2.5 μL 10 × Ex Taq buffer, 0.125 mL ExTaq polymerase (TaKaRa, Shiga, Japan), and an appropriate volume of distilled water. The primers and their sequences are specified in S2 Table. The following conditions were used for standard PCR amplification: an initial 5 min denaturation at 94˚C; followed by 36 cycles of denaturation for 1 min at 94˚C, annealing for 1 min at 57˚C, and an extension for 1 min at 72˚C. The amplification was completed with a final extension period of 10 min at 72˚C. For amplification using degenerate primers, a touchdown protocol was applied (5 min at 94˚C; 24 cycles of 45 s at 94˚C, 45 s at 66˚C-54˚C step-down at 0.5˚C every cycle, 1 min at 72˚C; 16 cycles of 45 s at 94˚C, 45 s at 54˚C, 1 min at 72˚C, and a final extension step of 10 min at 72˚C). Sterile water instead of DNA served as a negative control in each assay. PCR products were analyzed on 1% agarose gels.
Amplicons to be sequenced were purified using the QIAquick PCR Purification Kit (Qiagen, Germantown, MD) as recommended by the manufacturer. Both strands of PCR products were sequenced using BigDye Terminator version 3.1 cycle sequencing ready reaction mix (Applied Biosystems, Foster City, CA). Sequencing products were resolved using an ABI 3130 automated sequencer (Applied Biosystems) and the sequences were assembled using Sequencher 4.8. software (Gene Code Corporation, Ann Arbor, MI).

Switching test
To examine the occurrence of phenotypic switching, the isolates were grown on plates of supplemented Lee's agar medium for 5 d and then plated on synthetic complete (SC) medium containing 5 μg/mL phloxine B; they were then further grown for 7 d at 26˚C and 30˚C [30]. C. albicans WO-1 white and C. albicans WO-1 white opaque strains obtained from the Duke University Molecular Genetics and Microbiology Department collection were used as control strains.
Killer activity/sensitivity assay C. parapsilosis species group and L. elongisporus isolates were assayed to determine their killer/ sensitivity phenotypes. The killer sensitivity of C. parapsilosis species group and L. elongisporus strains was assayed by mixing each isolate with YEPD-MB agar [0.3% yeast extract, 0.3% malt extract, 0.5% peptone, 2% glucose, 2% agar, and 0.003% methylene blue (MB); adjusted to pH 4.5 with 0.1 M citrate-phosphate buffer] to a final concentration of 10 6 cells/mL and by streaking the known killer and sensitive isolates on the surface of plates. The plates were incubated at 26˚C for up to 72 h. The sensitivity test was considered positive if killer strains showed a clear inhibition zone surrounded by a blue halo [39]. RNA extraction was performed using TRIzol reagent (Thermo Fisher Scientific, Waltham, MA) according to a protocol provided by the manufacturer. S. cerevisiae strains were used as controls. Products were analyzed on 1% agarose gels.

Evaluation of virulence using the G. mellonella model
Isolates were pre-grown on YPD agar for 24 h at 37˚C and then harvested by gentle scraping of the colony surfaces with plastic loops and washed three times in sterile phosphate-bufferedsaline (PBS). Cell suspensions were counted using a hemocytometer, and cell density adjusted to 10 6 cells/μL with sterile PBS. The virulence of each isolate was tested in 15 G. mellonella larvae. Cell suspensions in sterile PBS (4 μL) were injected via the last left rear proleg, using a 100 μL Hamilton syringe with dispenser. The syringe was rinsed several times with 70% ethanol, followed by a PBS rinse, prior to injecting each larva. The control group of larvae was inoculated with sterile PBS. Inoculated larvae were incubated at 37˚C and the number of dead animals was monitored daily [4].

Determination of antimicrobial drug resistance
FLC (Sigma, St. Louis, MO) and FK506 (Sigma) were diluted in sterile water according to the Clinical and Laboratory Standards Institute (CLSI) protocol [49]. Serial two-fold dilutions of each drug were prepared in RPMI 1640. Synergy testing of FLC and FK506 against C. parapsilosis species group and L. elongisporus strains was assessed by the checkerboard method [50]. The test was performed as a microdilution assay, in duplicate for each fungal strain. The minimum inhibitory concentrations (MIC) were interpreted according to the CLSI guidelines [49].
To evaluate the effect of the combinations of FLC and FK506, the fractional inhibitory concentration (FIC) was calculated for each antifungal agent in every combination. The following formulas were used to calculate the FIC index: FIC of drug A, MIC (drug A in combination)/ MIC (drug A alone); FIC of drug B, MIC (drug B in combination)/MIC (drug B alone); FIC index, sum of FIC of drug A and FIC of drug B. Antifungal combinations were evaluated based on FIC index ranges, as follows: synergistic, if 0.5; additive, if > 0.5 but < 1; no effect, if ! 1 but < 4; and antagonistic if ! 4 [51,52].

Statistical analysis
The Kaplan-Meier test was performed to assess the statistical significance of differences in survival among groups. Survival curves were analyzed using Minitab v. 16.1 software with the logrank (Mantel-Cox) test; P < 0.05 was considered statistically significant.

Results and discussions
In this study, the identifying characteristics, MTL genotypes, phenotypic switching, and susceptibility profiles of C. parapsilosis species group and L. elongisporus reference isolates were analyzed to gain insight into their pathobiology and virulence mechanisms.

Chromogenic medium
The adequacy of BBL Chromagar Candida medium for the initial identification of reference strains was evaluated. All C. parapsilosis and C. metapsilosis isolates formed light pink colonies; C. orthopsilosis formed ivory colonies; and L. elongisporus formed blue/green colonies similar to C. tropicalis and C. albicans on this medium. Representative colonies on the chromogenic medium are shown in Fig 1. The utility of chromogenic media for the screening and initial identification of yeasts in polyfungal clinical materials is well established [53][54][55][56]. C. parapsilosis characteristically produces pink or lavender colonies on BBL CHROMagar Candida medium, whereas L. elongisporus isolates were reported to form colonies with a distinct turquoise color [15]. However, C. parapsilosis and C. metapsilosis isolates formed light pink colonies, rendering their presumptive identification problematic.

Determination of MTL genotypes
Sexual reproduction is important in the evolution of fungal pathogens. In particular, pathogenic fungi are hypothesized to restrict sexual reproduction so as not to disrupt well-adapted pathogenic genotypes; however, they maintain the potential for sexual reproduction to cope with stressful conditions such as antimicrobial therapy [57]. The MTL locus is responsible for the determination of cell identity and regulation of mating in C. albicans and related species [25]. In C. albicans, there are two versions of MTL: MTLa which harbors the transcription factor genes a1 and a2; and MTLα, which encodes the α1 and α2 regulatory elements [24][25][26]. Both idiomorphs also possess a or α versions of additional genes that have no known function in mating, such as PAB, OBP, and PIK [20]. In C. albicans, α1 and a2 activate the α-and a-specific genes, respectively; the a1/α2 heterodimer plays a role both in the regulation of mating and in white/opaque switching, by repressing mating and limiting switching to the opaque phase in a/α cells by inhibiting White-opaque regulator 1 (WOR1), which is regulated by other factors as well [29,58,59]. In comparison, C. orthopsilosis has an MTL locus very similar to that of C. albicans and harbors both MTLa and MTLα idiomorphs, whereas only the MTLa idiomorph has been identified to date in C. parapsilosis [27].
Mating type genes of the isolates were determined by PCR. As no MTL α1, α2, a1, or a2 genes have been detected in L. elongisporus to date and no C. parapsilosis MTLα idiomorph is known, degenerate primers to amplify the MTL transcription factor genes a1, a2, α1, and α2 were designed to search for the presence of these genes. In addition, specific primers were designed to amplify the a2 gene and a1 pseudogene in C. parapsilosis. PCR screening of eight C. parapsilosis isolates revealed only the presence of a1 and a2 genes (Fig 2A), indicating that these isolates all bear the MTLa idiomorph. An MTLα idiomorph was not detected using the degenerate primers.
L. elongisporus is the only close relative of C. parapsilosis reported to exhibit a homothallic sexual state [13]. However, in the sequenced strain and in seven additional isolates of L. elongisporus, a genomic region that is syntenic to the MTL locus of other Candida species only has the MTLa versions of PAB, OBP, and PIK genes, and does not contain any transcription factor genes [25]. Considering the genome analysis and molecular data, it was suggested that L. elongisporus might not have a sexual cycle or the sexual cycle might function in a manner independent of MTL [25].
In this study, for L. elongisporus, primers were designed to bind within PIKa and orf19.3202 (encoding a hypothetical protein ortholog of C. albicans CAALFM_C501730WA, located just outside the MTL locus in C. albicans and other closely related species), respectively [25], to check for the presence of mating genes between these sites. In the absence of any such genes, a 933 bp PCR product was predicted, as in the sequenced isolate [25]. All L. elongisporus isolates yielded 933 bp amplicons, indicating the absence of an MTL idiomorph (Fig 2B1 and 2B2). Furthermore, neither an a nor α gene was detected in the isolates using degenerate primers.
In C. orthopsilosis, α1, α2, a1, and a2 genes were screened using specific primers. Only a1and a2-specific PCR product bands were obtained from isolates CBS 107.41 and CBS 107.42; all the expected gene PCR product bands were obtained from the remaining five isolates. This indicated a/a homozygosity of the former two isolates and a/α heterozygosity of the remainder (Fig 2C1 and 2C2). In comparison, in a study by Sai et al. [27], only two of 16 C. orthopsilosis isolates were found to be MTL heterozygous, whereas nine were MTLa homozygous, and five were MTLα homozygous.
According to Pryszcz et al. [60], the MTLα locus of C. metapsilosis is very similar to that of C. albicans in structure, encoding the genes MTLα1, MTLα2, OBPα, PIKα, and PAPα. The MTLa locus, however, harbors MTLα2, OBPα, and PIKα in addition to the a-specific genes MTLa1 and MTLa2. For C. metapsilosis isolates, primers designed by Pryszcz et al. [60] were used to amplify the α1, α2, a1, and a2 genes. Whereas PCR products for each gene were obtained from five C. metapsilosis isolates, only the α1 and α2 products were obtained from CBS 2315. Thus, CBS 2315 carried the α/α genotype and the other five isolates were a/α ( Fig  2D). Similarly, 10 out of 11 C. metapsilosis isolates analyzed by Pryszcz et al. [60] were MTL heterozygous.

Phenotypic switching within the isolate set
Previous studies have identified different colony and cell morphologies of Candida spp., especially C. albicans and C. parapsilosis isolates. Although pseudohyphae formation has been observed in C. parapsilosis and C. albicans as well as in C. orthopsilosis, it was not observed in C. metapsilosis [4,61,62]. In contrast, in the present study we detected phenotypic switching in C. metapsilosis CBS 2315, the isolate carrying the MTLα/MTLα genotype, at 37˚C. The switching led to the formation of pink colonies in the presence of phyloxine B, which contained elongated cells revealed by scanning electron microscopy (Fig 3). Killer activity of the isolates A total of 32 isolates were analyzed by the killer/sensitivity phenotype test. Briefly, 22 out of 32 isolates produced weak blue halos around their colonies; nine isolates did not produce such halos. C. parapsilosis CBS 2915 produced a weak blue halo around both the killer and sensitive strains. However, no RNA band was observed on a 1% agarose gel, indicating the absence of killer activity due to a dsRNA virus.
C. parapsilosis has been reported to be a killer yeast [63], although killer strains were reported to represent less than 3% of clinical isolates of the species [64]. Killer activity was found to be expressed at 25˚C, whereas isolates of C. parapsilosis and C. orthopsilosis did not show this activity at 25˚C [62]. Because the killer toxin is thermolabile [34], wild-type killers exhibit very little killing activity at 30˚C and are normally tested at 20˚C. Therefore, in the present study, the isolates were analyzed at 26˚C, 30˚C, and 37˚C; however, no killer activity was detected.
Virulence of the isolates in the G. mellonella model Virulence of the 32 study isolates was compared using the G. mellonella model (S3 Table). We observed no significant differences among the C. parapsilosis species group and their closely related species L. elongisporus (P > 0.05); however, significant differences were detected with the PBS control group (P < 0.05; Fig 4).
C. orthopsilosis was reported by Gago et al. [4] to represent the most virulent species of the C. parapsilosis species group in the G. mellonella model, followed by C. parapsilosis and C. metapsilosis, with a median survival time of 2.3, 2.6, and 4.5 d, respectively. It has been previously suggested that these scores could not be related to the growth rate of Candida spp. [65]. Notably, however, C. metapsilosis strains were more effectively phagocytosed by G. mellonella hemocytes than in C. parapsilosis and C. orthopsilosis (P < 0.05) [4]. Furthermore, hyphae or pseudohyphae formation was less frequent in C. metapsilosis isolates than in C. parapsilosis and C. orthopsilosis (P < 0.05) [4]. In another study, one oral and one systemic isolate of C. parapsilosis, both susceptible to FLC and amphotericin B, killed the G. mellonella larvae within 18 h and 21 h, respectively, which suggested that the clinical origin of the strain is not important for virulence (P = 0.6) [4]. However, in contrast to these previous reports [4,65], which did not differ with respect to inoculum size in comparison with the present study (1 × 10 6 cells/larvae), we did not observe a difference in virulence using this model among the studied fungi.
Antimicrobial resistance of the isolates MIC values of FLC and FK506 were determined in synergy tests using the checkerboard assay ( The intrinsic antifungal resistance of Candida spp. constitutes a major issue related to the therapeutic management of infections and has required the utilization of combinatorial therapy, such as with FK506. For example, Chen et al. [44] reported that posaconazole exhibits an in vitro and in vivo synergistic antifungal activity with caspofungin or FK506 against C. albicans isolates. Cruz et al. [45] also observed that FK506 was synergistic with FLC against azole-resistant C. albicans mutants, against other Candida species, or when combined with different azoles. Notably, Li et al. [46] observed that a combination of FLC and FK506 might represent a promising approach toward overcoming the intrinsic resistance of Candida krusei to FLC. Denardi et al. [47] also investigated the in vitro interaction of FK506 and four azole compounds against 30 clinical FLC-susceptible or FLC-resistant Candida glabrata isolates using the microdilution checkerboard method. In particular, they detected a promising synergistic effect against FLCresistant C. glabrata isolates of FK506 combined with ketoconazole (77%), itraconazole (73%), voriconazole (63%), and FLC (60%). In contrast, FK506 showed no activity against 30 clinical FLC-susceptible and FLC-resistant Trichosporon asahii isolates, with MICs ! 64 μg/ml. However, a pronounced synergistic interaction of FK506 in combination with amphotericin B (96.7%) and caspofungin (73.3%) was observed, although low rates of synergism were observed with FLU (40%) and itraconazole (10%) [48].
In the present study, we observed synergistic and additive, or indifferent, effects of FLC and FK506 against C. parapsilosis and related species including L. elongisporus; however, antagonistic activity was not observed. Previously, nine human L. elongisporus isolates were tested against FLC, amphotericin B, caspofungin, anidulafungin, and micafungin, all of which exhibited low MICs, as determined by the CLSI microdilution method [15]. Consistent with these findings [15], we observed low FLC MICs for L. elongisporus, for which no established breakpoint values are currently available. The present study also demonstrated that synergistic and additive effects of FLC and FK506 were more apparent against C. parapsilosis (87.5%) and L. elongisporus (81.8%) than other isolates, and that the combination was likely to have no effect against C. orthopsilosis (71.4%) and C. metapsilosis (66.7%). However, a limitation of the present study is that only a small number of isolates was tested. Hence, no universal conclusion may be reached regarding the data obtained herein.
In conclusion, we determined the MTL genotypes of a set of reference isolates of the C. parapsilosis species group and detected an MTLα homozygous C. metapsilosis isolate that underwent phenotypic switching and produced elongated cells. Furthermore, we observed no significant difference in virulence among the four species, using a G. mellonella model. We suggest that the FLC/FK506 combination may be promising as a therapeutic strategy against L. elongisporus and C. parapsilosis isolates, but not against C. orthopsilosis and C. metapsilosis. In vitro assessment in an experimental model is required to verify the efficacy of this drug combination. Overall, these new data may be used to guide strategies for combating these pathogens in the clinic.
Supporting information S1