Deciphering Colonies of Phenotypic Switching-Derived Morphotypes of the Pathogenic Yeast Candida tropicalis

Phenotypic switching generates fungal colonies with altered morphology and allows pathogens to adapt to changing environments. This study investigated the structure and genetic factors of switched morphotypes colonies in Candida tropicalis. Morphotypes of C. tropicalis comprised the clinical strain 49.07 that exhibited smooth colony phenotype and switched (crepe and rough) morphotypes that showed colonies with marked structural variations, including wrinkled surface, depressions areas, and irregular edges (structured morphology). The morphotypes were analyzed for the presence and distribution of the extracellular matrix (ECM) at the ultrastructural level-SEM. The composition of the ECM and the percentage of hyphae in colonies were evaluated. The expression of EFG1 (Enhanced filamentous growth protein 1), WOR1 (White-opaque regulator 1), and BCR1 (Biofilm and cell wall regulator 1) in the morphotypes was measured by RT-qPCR. Colonies of the switched variants exhibited distinct arrangements of ECM compared to the smooth phenotype (clinical strain). In addition, rough variant colonies showed higher amounts of total carbohydrates and proteins in ECM (p < 0.05). Switched (crepe and rough) colonies exhibited a higher percentage of hyphae throughout their development (p < 0.05). The mRNA expression levels of EFG1, WOR1, and BCR1 in the rough morphotype were significantly higher than they were in the smooth morphotype. In addition, there was a positive correlation between the expression of these genes and filamentation (hyphae formation) of the rough morphotype (r2 > 0.9472, p < 0.05). Structural variations in switched morphotypes colonies of C. tropicalis seem to be associated with increased hyphae growth and the amount and distribution of ECM. Switched colonies have distinct expressions of the EFG1, WOR1, and BCR1 master regulators genes.


Introduction
Candida tropicalis is an opportunistic yeast pathogen with an ability to cause both superficial and systemic infections in humans [1]. This species has been the second most prevalent Candida species in bloodstream infections, particularly in tropical regions, promoting high mortality [2][3][4][5]. The success of C. tropicalis as a human pathogen can be attributed to its vast repertoire of virulence determinants, such as adhesion, morphogenesis, biofilm-forming ability, and secretion of lytic enzymes [1]. C. tropicalis is a polymorphic fungus, existing in a unicellular yeast cells form, pseudohyphae, and hyphae [1], where hyphal growth plays a vital role in the pathogenicity of this species [6].
A notable feature of C. tropicalis is its ability to undergo phenotypic switching. This phenomenon is associated with a reversible change in colony morphology at rates higher than somatic mutation rates [7]. Phenotypic switching confers plasticity within isogenic populations and enables pathogens to adapt to a constantly changing microenvironment [7][8][9]. Switching systems that comprise white-opaque and white-gray-opaque transitions, as well as multiple forms of reversible switch phenotypes, were described for C. tropicalis [10][11][12][13][14][15].
Soll et al. [10] first described that C. tropicalis possesses a varied repertoire of switch phenotypes. In previous studies we described that phenotypic switching promotes changes in the colony architecture of C. tropicalis clinical isolates, leading to the development of colonies with structured morphological patterns exhibiting depressions and elevations areas, wrinkled surface, and irregular edges [11,14,16]. Ultrastructural analysis of colonies of structured morphological patterns revealed the presence of high amounts of extracellular matrix (ECM), suggesting a possible role for ECM in C. tropicalis switching events [11]. For the yeast Saccharomyces cerevisiae, the ability to form an abundant ECM is one of the features typical for colonies with structured architecture [17]. This morphological pattern was characterized as ''biofilm-like'' due to the presence of relevant amounts of ECM [11,17]. In addition, colonies of switched phenotypes showing structured morphological patterns were associated with higher percentages of filamentous growth compared to colonies of unstructured (smooth) phenotype [14,16].
Phenotypic switching can translate into changes in the virulence of the pathogen. Switched variant phenotypes of C. tropicalis have altered virulence, showing variations in hemolytic activity, adhesion to biotic and abiotic surfaces, biofilm formation, and recognition by Galleria mellonella hemocytes [11,14,16,18,19].
Although the presence of hyphae and ECM in switch colonies of C. tropicalis may be associated with the architecture of these colonies [11,16], the composition of the matrix and the genetic factors that mediate multiple forms of switch phenotypes, including structured colonies remain unknown. In contrast, molecular mechanisms that regulate the white-opaque and tristable switching system in C. tropicalis are reported [12,15]. Therefore, we evaluated hyphae production throughout colony development, as well as the presence, distribution, and composition (total carbohydrates and proteins) of the ECM in C. tropicalis colonies of smooth phenotype and switched phenotypes of structured morphological patterns (crepe and rough). The expression of transcriptional regulatory genes (WOR1, EFG1 and BCR1) by colonies cells of these morphological patterns was also investigated.

Candida tropicalis Morphotypes and Culture Conditions
Morphotypes employed in this study comprise a C. tropicalis clinical strain (49.07), obtained from a patient admitted at a tertiary-care hospital at Londrina-Parana State [20], and two switch variants (crepe and rough) that arose from the 49.07 strain [14]. The strain 49.07 exhibited a smooth dome colony with a flatly convex profile (hemispherical shape), and was characterized by colony of ''smooth morphology''. The variants (crepe and rough) exhibited marked differences in colony morphologies with wrinkled surface, depressions areas, and irregular edges, and were characterized by colonies of ''structured morphology'' [16] (Fig. 1, I and II). SEM analysis were made as previously described by our group, using the FEI Quanta 200 Scanning Electron Microscope at 30 kV [16]. These morphotypes were obtained as a stock culture from the Fungal Genetics Laboratory, The State University of Londrina-Brazil. The morphotypes were stored as frozen stocks with 15% (w/v) glycerol at -80°C and cultured on yeast extractpeptone-dextrose (YPD) (DIFCO) agar plates at 28°C for 96 h.

Matrix Composition of the Colonies
The extracellular matrix (ECM) of the colonies was extracted as described by Azeredo et al. [21], with modifications. Prior to the extraction procedure, portions of the colonies were pretreated with glutaraldehyde (GTA). Portions (0.3 g wet weight) of each of the phenotypes were incubated at 4°C for 3 h and 30 min in 30 ml of glutaraldehyde (1.8% GTA). After this period, the samples were centrifuged at 9000 g for 10 min, resuspended in phosphate-buffered saline (PBS) and centrifuged again. Then the final cell pellet was once again resuspended in PBS. The suspensions were then sonicated for 30 00 , 1 0 and 2 0 . The tubes were kept on ice during sonication.
Subsequently, the supernatant was filtered (nitrocellulose filter, 0.2 lm) and stored at -20°C. The cell pellets were freeze-dried to determine the dry weight of cells after extraction of matrix. Portions of 0.3 g (wet weight) of each phenotype (parental and variants) were also freeze-dried to determine the control dry weight (dry weight of cells not subjected to matrix extraction). At the end of the extraction procedure, 30 ml of matrix suspension were obtained. The experiments were performed in triplicate, in three independent experiments.
The total carbohydrate content of the colony ECM was estimated according to Dubois et al. [22], using glucose as a standard. The protein content of the matrix was measured by employing the BCA kit (Bicinchoninic Acid, Sigma-Aldrich, St Louis, USA), using bovine serum albumin (BSA) as a standard. The content of both components was relativized with the dry weight of the colonies.

Determination of the Percentage of Hyphae in Colonies with Different Phenotypes
Throughout colony development, the percentage of hyphae was determined at 48, 72, and 96 h of culture. At each time, three colonies from each morphological group were suspended in PBS (1x) and the percentage of hyphae was determined by direct counting of 1000 cells per experiment in a Neubauer counting chamber under a light microscope (OlemanÒ). Three independent experiments were conducted. Each hyphae was considered 1 multicellular unit.

Gene Expression
EFG1 (Enhanced filamentous growth protein 1), WOR1 (White-opaque regulator 1) and BCR1 (Biofilm and cell wall regulator 1) expression was determined by quantitative PCR (qPCR). Cells of colonies with typical morphology were selected, transferred to microtubes, frozen in liquid nitrogen, and their RNA was extracted using the Trizol-chloroform method (Thermo Fisher Scientific, USA). RNA was quantified and the quality was assessed using a NanoDrop spectrophotometer (ThermoScientific, Loughborough, UK). Complementary DNA (cDNA) was synthesized from 200 ng of extracted RNA using an RT-PCR kit (Invitrogen, Carlsbad, CA, USA) in a GeneAmpÒ PCR (Eppendorf, Gradient Mastercycler) following the manufacturer's instructions. The cDNA obtained was stored in a freezer at -20°C. Primers used for qPCR are described in Table 1.
Each sample was analyzed in duplicate by real-time PCR performed on a StepOnePlus TM Real-Time PCR System (Applied BiosystemsÒ). The reaction was performed using PlatinumÒ SYBRÒ Green qPCR Supermix-UDG (Invitrogen, Carlsbad, CA, USA) with a final volume of 20 ll. Gene expression data were normalized by the expression of the constitutive b-Actin gene (ACT1). Once the data were normalized, the relative gene expression levels were analyzed by the equation 2 -(D-Dct) .

Statistical Analysis
The paired t-test was used to compare the means, considering statistical significance *p \ 0.05, **p \ 0.01 and ***p \ 0.001, by GraphPad Prism 5 software. Pearson's correlation, determined by the R statistical software, was used to correlate the variables (p \ 0.05).

Evaluation of Extracellular Matrix and Morphogenesis in Smooth and Switched Colonies
As shown in Fig. 1, colonies of smooth morphology (clinical strain) exhibited homogeneous distribution of extracellular matrix, similar to a thin film, covering the cells (Fig. 1-Smooth III). In contrast, colonies of crepe and rough (structured morphology) had a greater amount of ECM. In crepe variant colonies, the matrix was arranged in ''lamellae'', filling many of the depressions along the colony surface ( Fig. 1-Crepe III). In rough variant colonies, the matrix appeared in a polarized form, arranged in dense networks of  Fig. 1-Rough III). The ECM preparations obtained were evaluated for total carbohydrate and protein content by colorimetric methods. Comparative analyses between phenotypes showed that the total carbohydrate and protein content in rough variant colonies (structured phenotype) (33.74 ± 12.47 mg carbohydrate and 3.47 ± 1.42 mg protein/g colony dry weight) is higher than that observed in the smooth phenotype (clinical strain) (13.75 ± 11.05 mg carbohydrate and 1.11 ± 0.26 mg protein/g colony dry weight) (p \ 0.05). The crepe variant exhibited no differences from the parental strain (p [ 0.05) (Fig. 2a).
To evaluate morphogenesis throughout the growth of switched colonies, the number of hyphae was counted at 48, 72, and 96 h. The proportion of hyphae in the colonies did not increase over time (p [ 0.05) (Fig. 2b). Colonies of smooth morphology, showed small amounts of hyphae (\ 1%) during all analyzed times (Fig. 2b) (p [ 0.05). In crepe and rough colonies, on the other hand, the percentage of hyphae increased (p \ 0.05), as observed in Fig. 2b, c. The crepe phenotype showed intermediate hyphal formation compared to the other morphotypes, with percentages ranging from 2.6 ± 0.4% to 3.1 ± 0.3% (p \ 0.05). The rough variant showed the highest percentage of hyphae ([ 4.5 ± 0.5%), when compared to the smooth phenotype (p \ 0.05, Fig. 2).
Expression of WOR1, EFG1, and BCR1 Varied Among Switched Morphotypes of C. tropicalis The expression levels of transcriptional regulatory genes (WOR1, EFG1 and BCR1) were quantified from colony cells of smooth, crepe and rough morphotypes. Switched colonies showed distinct gene expression profiles. The rough morphotype showed a significant increase in the expression of EFG1 (11.3 ± 1.1) and WOR1 (2.7 ± 0.3), important transcriptional regulators of filamentous growth when compared to the smooth phenotype. BCR1 expression (4.6 ± 0.5), a regulator of biofilm formation, was also higher in the rough morphotype than in the smooth strain (p \ 0.05) (Fig. 3). Differently, the crepe morphotype showed a decrease in the expression levels of EFG1 (0.02 ± 0) and BCR1 (0.01 ± 0) compared to the expression of the smooth phenotype (P \ 0.001), while no significant difference was observed on the level of WOR1 expression between crepe and smooth morphotypes (Fig. 3). In addition, the expression of all three transcription factors was positively correlated (r 2 [ 0.9472, p \ 0.05) to the hyphal formation profile of the rough morhotype.

Discussion
In this study, we evaluate the structural and genetic factors associated with switching-derived morphotypes of C. tropicalis. During the ultrastructural analysis, we observed that, although present in colonies of all morphotypes (smooth, crepe and rough), extracellular matrix exhibits differences in abundance and distribution (Fig. 1, III). These variations suggest an association with the complexity of the phenotypes since both crepe and rough variants, characterized by colonies of structured morphology [16], have dense matrix networks covering and connecting the cells, in contrast to that observed for the smooth phenotype. This evidence reinforces the possible role and biological importance of the matrix in maintaining the architecture of structured colonies derived from phenotypic switching in C. tropicalis, classified as biofilm-like colonies [11]. For the yeast Saccharomyces cerevisiae, the formation and architecture of structured biofilm colony phenotype are related to the abundant presence of ECM [17].
Currently, studies related to ECM in colonies of Candida species are largely unexplored. In the present study, we showed for the first time that the amounts of total carbohydrates and proteins in ECM extracted from colonies were variable between switched phenotypes of C. tropicalis (Fig. 2a) indicating that macromolecular components may vary in matrix composition in a morphotype-dependent fashion. ECM extracted from biofilms of C. tropicalis has been evaluated [23][24][25]. According to these authors, the amounts of carbohydrate and protein in the C. tropicalis biofilm matrix are variable compared to other Candida species. In addition, it was demonstrated that the composition of the matrix contributes to drug resistance in C. tropicalis biofilms [23].
In addition, our data also revealed that structured phenotypes vary in cellular differentiation throughout colony development. The switch phenotypes (crepe and rough) exhibited the highest percentages of hyphae compared to the smooth phenotype. The high proportion of hyphae was maintained throughout 48, 72 and 96 h of culture (Fig. 3a, c). These results suggest that phenotypic switching may act on the yeast-to-hyphae transition at the early stages of colony development reflected in structured colonies (crepe and rough), corroborating our previous findings [16]. For C. albicans white-opaque switching system, opaque cells do not undergo filamentation under conditions that induce hyphae formation in white cells, suggesting that switching may have an effect on infection as hyphal growth is a virulence determinant that facilitates tissue invasion [26].
Using RT-PCR, we analyzed the expression of EFG1, WOR1 and BCR1 transcription factors by colonies cells of C. tropicalis morphotypes. Our data showed higher levels of gene expression in the rough than in the smooth colonies, being more than 11-fold for EFG1 and near three-fold for WOR1 (Fig. 3). Both genes are important regulators of the white-opaque transition [27][28][29], and play key roles in regulating cell differentiation of tristable switching [30] in C. tropicalis. The expression level of BCR1, a regulator of biofilm formation, was 4.6-fold higher in the rough than in the smooth colonies. In C. tropicalis, BCR1 also acts as a regulator of the white-opaque transition [29], and as an activator of filamentation [28]. Here, we show for the first time that the expression levels of EFG1, WOR1 and BCR1 were increased in a switch morphotype (rough variant) of C. tropicalis that is characterized by colonies of structured morphology, as illustrated in Fig. 1. In addition, gene expression of all three transcription regulators correlated to the high capacity of the rough morphotype to form hyphae throughout colony development.
In contrast, although no significant difference was observed in the level of WOR1 expression between the crepe and smooth colonies, the former had a higher percentage of hyphae in relation to the smooth morphotype. According to Zhang et al. [28] and Porman et al. [31], WOR1 overexpression led to an increase in filamentous growth in C. tropicalis, however, deletion of WOR1 had no prominent effect on filamentation, suggesting that WOR1 is not an essential gene for filamentation in C. tropicalis [28]. In the crepe variant, the phenotypic switching transiently induced the expression of EFG1, since its expression was high in colonies at 24 h of culture (data not shown) and absent in colonies at 48 h of culture (Fig. 3). Therefore, switched morphotypes of C. tropicalis characterized by colonies of structured morphologies (crepe and rough) seem to have distinct gene expression profiles throughout their development.
Yeast-filamentous growth transition is a strategy by which Candida species may increase their virulence. In previous studies, we showed that structured switch phenotypes of C. tropicalis with higher amounts of filamentous forms exhibit greater success in colonizing biotic and abiotic surfaces [19], higher biofilm formation [14] as well as elevated virulence against G. mellonella larvae [18] compared to smooth counterpart morphotypes.
In conclusion, our data suggest that structured morphological patterns of switched colonies in C. tropicalis may be due, at least in part, to changes in the matrix arrangements and increased filamentous growth. In addition, switched colonies have distinct gene expression profiles. The transcriptional factors EFG1, WOR1 and BCR1 are likely to influence filamentation in colonies of the rough variant of C. tropicalis. This is the first study to evaluate genetic factors that may mediate the switching event in C. tropicalis that originate colonies of structured morphology.