Investigation of Candida parapsilosis virulence regulatory factors during host-pathogen interaction

Invasive candidiasis is among the most life-threatening infections in patients in intensive care units. Although Candida albicans is the leading cause of candidaemia, the incidence of Candida parapsilosis infections is also rising, particularly among the neonates. Due to differences in their biology, these species employ different antifungal resistance and virulence mechanisms and also induce dissimilar immune responses. Previously, it has been suggested that core virulence effecting transcription regulators could be attractive ligands for future antifungal drugs. Although the virulence regulatory mechanisms of C. albicans are well studied, less is known about similar mechanisms in C. parapsilosis. In order to search for potential targets for future antifungal drugs against this species, we analyzed the fungal transcriptome during host-pathogen interaction using an in vitro infection model. Selected genes with high expression levels were further examined through their respective null mutant strains, under conditions that mimic the host environment or influence pathogenicity. As a result, we identified several mutants with relevant pathogenicity affecting phenotypes. During the study we highlight three potentially tractable signaling regulators that influence C. parapsilosis pathogenicity in distinct mechanisms. During infection, CPAR2_100540 is responsible for nutrient acquisition, CPAR2_200390 for cell wall assembly and morphology switching and CPAR2_303700 for fungal viability.


Results
Identification of C. parapsilosis virulence regulatory genes. In order to identify virulence regulatory factors, THP-1 monocytes were infected with C. parapsilosis cells using a multiplicity of infection (MOI) of 5. Following co-incubation, host cells were removed after 1 and 6 hours post-infection and fungal RNA was isolated for whole transcriptome analysis using Illumina-based sequencing (see Materials and Methods). In addition, we also considered yeast cells incubated in the same medium but in the absence of THP-1 cells as a control. We used a state-of-the-art pipeline (see Materials and Methods) to analyze the RNA sequencing reads. Our results show clear changes in expression upon incubation with THP-1 cells and during the monitored time course, with an increase in expression of fungal genes during host interaction, suggesting their involvement in virulence (see Materials and Methods, Supplementary Fig. S2 and Supplementary Table S1). A total fold change greater than 4 in gene expression (log2fold change greater than 2) was used to select genes for further analyses (Supplementary Table S1). The set of up-regulated genes includes several uncharacterized ORFs with hypothetical regulatory functions ranging from transcriptional factors to protein kinases, according to orthology-based functional assignment. Furthermore, the set of up-regulated genes included an additional ORF, CPAR2_501400, for which orthology-based functional assignment suggests a role in cell wall beta 1,6-glucan assembly. Based on their expression profiles and their putative biochemical activities, we aimed to generate a set of deletion mutant strains ( Table 1).

Preparation of deletion mutant strains.
In order to study the role of the up-regulated hypothetical regulators in C. parapsilosis virulence, we established a deletion mutant collection of the identified ORFs. ORF deletion was performed using the fusion PCR method 20 , as adapted to C. parapsilosis by Holland et al. 15 . Due to its high specificity and speed, this method allowed us to generate a set of deletion mutant strains of the selected genes, each represented by two parallel homozygous deletion mutant strains. Following confirmation of the null mutants, strains were prepared for phenotypic assays.

Systematic screen of the generated mutant strains. Viability testing and response to stressors.
During invasion, fungal cells need to adapt to restrictive environmental conditions in order to survive and disseminate throughout a host. Such conditions include the presence of alternative nitrogen and carbon sources, elevated temperature, and a shift from neutral pH to slightly basic or acidic [21][22][23] . Pathogenic species also developed strategies to degrade antimicrobial components present in the host serum 19,24 . Therefore, the deletion mutant strains were first analyzed in terms of fitness and viability under various conditions. Among the tested mutant strains, three showed a general growth defect on YPD complex media, indicating an adverse effect on fungal viability ( Fig. 1/A). Furthermore, five showed reduced growth on minimal media (YNB + glucose), three on complex media set to pH 8, three in the presence of bovine serum albumin (BSA), and five on fetal bovine serum (FBS) supplemented plates, suggesting a potential defect in adaptation to alkaline conditions, in alternative energy source utilization and serum protein degradation ( Fig. 1/A). The response of each deletion mutant strain was also examined in the setting of various restrictive environmental conditions 25 . Survival of each mutant strain was evaluated in the presence of different stressors in liquid media. During fungal infection, phagocytic cells first distinguish cell wall surface components of the invaders via pattern recognition receptors, thus cell wall assembly significantly influences virulence 26 . To assess alterations in cell wall homeostasis calcofluor white, congo red and caffeine were used 27,28 . To identify mutants with potential glycosylation defects, Hygromycin B was also applied 29 . According to our results, eight strains showed altered response to the aforementioned stressors ( Fig. 1/B). To evaluate the responses of the mutants to oxidative damage, we used H 2 O 2 as stressor, to which two strains were susceptible ( Fig. 1/B). In order to examine cell membrane integrity the membrane stressor SDS was used. During the analyses three mutant strains showed altered susceptibility to the membrane disturbing agent ( Fig. 1/B). Altogether, these data suggest that 13 of the 19 examined genes may be involved in viability and stress tolerance regulation.
Morphology change. Although C. parapsilosis is unable to form true hyphae, the morphogenic shift to pseudohyphal forms has been associated with virulence 13 . When examining the ability of the mutants to form pseudohyphae, we found that CPAR2_200390Δ/Δ and CPAR2_501400Δ/Δ strains showed a remarkably different phenotype compared to the wild type. Yeast cells of CPAR2_200390Δ/Δ rapidly transitioned into extremely long and aggregating pseudohyphae, while CPAR2_501400Δ/Δ cells remained phenotypically locked in a yeast form (Fig. 2). Thus, these data suggest that CPAR2_200390 and CPAR2_501400 may be involved in the regulation and maintenance of morphology.
Biofilm formation and adhesive properties. C. parapsilosis effectively forms biofilms on intravenous catheters, prostheses, and other indwelling medical devices 13 . In order to examine the biofilm forming abilities of each deletion mutant strain we applied the FDA metabolic assay (Supplementary Fig. S3 and Supplementary Table S2). Our results demonstrated that three strains showed a different biofilm profile when compared to the CLIB 214 wild-type strain ( Fig. 3/A). Strains CPAR2_200390Δ/Δ, CPAR2_209520Δ/Δ and CPAR2_501400Δ/Δ displayed a lower capacity for biofilm formation, suggesting that the corresponding genes influence biofilm formation in C. parapsilosis. Adhesion to various biotic and abiotic surfaces is a critical step for biofilm formation. While several key regulators of adhesion have been identified in C. albicans, only a few have been described in C. parapsilosis 17 . In order to search for such regulators, we tested adhesion to polystyrene plastic, the substrate used for the aforementioned biofilm formation assay (Supplementary Fig. S4 and Supplementary Table S2.). According to our results, six of the examined strains (CPAR2_100540Δ/Δ, CPAR2_200390Δ/Δ, CPAR2_300080Δ/Δ, Determining gene function -selected mutant strains. Following the systematic screening of our mutant strains, three interesting ORFs with ≥6 altered phenotypes were selected for further in-depth analysis as potential regulators of distinct virulence-affecting mechanisms during host-pathogen interactions. These were nutrient acquisition (CPAR2_100540), morphology switch, cell wall reassembly (CPAR2_200390), and viability (CPAR2_303700).
Nutrient acquisition by CPAR2_100540. During the general characterization of the deletion mutant strains, the CPAR2_100540Δ/Δ strain produced smooth colony morphology ( Fig. 4/A), showed a mild growth defect on minimal media ( Fig. 1/A) and appeared to be defective in terms of adhesion ( Fig. 3/B) when compared to the wild type. CPAR2_100540Δ/Δ cells were also more susceptible to oxidative stress ( Fig. 4/B) and alkaline environmental conditions ( Fig. 4/C) than the wild type. While the inability to grow on pH 8 suggests a failure in trace element (e.g. ferric or ferrous iron) acquisition, oxidative stress susceptibility may indicate defects in respiration. The orthologous gene of CPAR2_100540 in C. albicans (HAP5, 68.0% amino acid sequence identity) is known to play a role in both processes.
Iron acquisition: C. albicans Hap5, an indispensable subunit of the CBF (CCAAT-binding factor) complex, is required for the expression of the essential iron reductase FRP1 under iron-limited conditions 30,31 . Loss of CaHAP5 results in low FRP1 expression levels when cells are grown in inducing media with either pH 8 or an iron chelator 30 . Furthermore, iron chelators inhibit the growth of Cahap5Δ/hap5Δ cells 30 . Similarly, C. parapsilosis CPAR2_100540Δ/Δ has decreased growth at pH 8 ( Fig. 4/C) and on iron chelator (BPS)-supplemented media that also contained hemin (low-iron source) ( Fig. 4/D). Gene expression analysis suggests that iron acquisition by C. parapsilosis is also dependent on a ferric reductase (CPAR2_402880) similar to C. albicans FRP1. CPAR2_402880 expression in C. parapsilosis is also induced under both iron-limited conditions in the wild type strain ( Fig. 4/E). In contrast, CPAR2_402880 expression levels remain relatively low in the CPAR2_100540Δ/Δ mutant strain, thus further supporting a role of CPAR2_100540 in C. parapsilosis iron acquisition ( Fig. 4

/E).
Alternative carbon source utilization: Carbon utilization via respiration is tightly related to available iron 32 , therefore the phenotypic traits observed in this section might be also linked to the altered iron homeostasis. Previous reports suggested that HAP5 is involved in alternative carbon source utilization in S. cerevisiae and C. albicans 33,34 . Alternative energy source metabolism is usually associated with altered regulation of respiratory chain element encoding genes such as cytochrome c (CYC) subunits and/or cytochrome c oxidases (COX) 33,34 . Thus, growth deficiencies on alternative carbon sources are often related to defects in the respiratory chain 35 . Susceptibility of CPAR2_100540Δ/Δ cells to oxidative stressors further suggests a respiratory chain defect. Furthermore, gene expression analysis results suggest that the observed phenotype may be due to the misregulation of ORFs equivalent to CYC1 (CPAR2_407500) and COX4 (CPAR2_207710) of C. albicans ( Fig. 4/H). Therefore, the obtained data led us to the conclusion that CPAR2_100540 is also involved in alternative carbon source utilization via regulating elements of the respiratory chain.
Contribution of CPAR2_100540 to virulence: Competition for the available iron sources and the ability to utilize alternative energy sources in the host are considered important virulence traits of pathogenic fungi 36 . To investigate the effects of the CPAR2_100540 ORF on virulence, we used both in vitro and in vivo infection models. Killing assays performed with J774.1 macrophage-like cells indicated that more of CPAR2_100540Δ/Δ yeast cells were killed when compared to the wild-type strain both at 18 h and 24 h of the interaction ( Fig. 4/I). Due to structural and functional similarities between G. mellonella and the mammalian innate immune system, this non-vertebrate model is frequently used as an alternative for virulence studies of Candida species 37 . Following inoculation with C. parapsilosis strains, survival of the individual larvae was monitored for 7 days. The survival results showed that the loss of CPAR2_100540 ORF resulted in significantly decreased virulence compared to the wild type strain (  an irregular colony morphology ( Fig. 5/B) and hyper-filamentation ( Fig. 5/C), a phenotype set similar to that observed in C. albicans following the removal of its orthologous gene, SPT3 (81.8% amino acid sequence identity). C. albicans SPT3 is a negative regulator of filamentous growth 38 . The obtained phenotypic traits suggested functional homology between SPT3 and the CPAR2_200390 ORF, supporting the inclusion of CPAR2_200390 as a regulator of morphogenesis in C. parapsilosis. Interestingly, during the characterization of the CPAR2_200390Δ/Δ mutant strain we also observed low adhesive and biofilm forming capability (   [39][40][41][42] . Real-time PCR analyses revealed altered chitinase and chitin synthase expression profiles for the mutant strain in both YPD complex and 10% FBS supplemented DMEM-medium when compared to the wild type ( Fig. 5/G). These data suggest that the CPAR2_200390 ORF is also involved in chitin homeostasis regulation.
Role of CPAR2_200390 in virulence: During fungal infection, cell wall assembly has a major impact on recognition by immune cells, while a morphology switch is associated with host cell and tissue disruption 36 . In killing assays performed with J774.1 macrophage-like cells, the CPAR2_200390Δ/Δ strain had increased survival compared to wild-type cells both at 18 h and 24 h of interaction ( Fig. 5/H). Interestingly however, G. mellonella infection studies suggested that CPAR2_200390Δ/Δ cells are avirulent in vivo, as survival rates were similar to that of PBS infected larvae (Fig. 5/I). Furthermore, following infection of BALB/c mice, the mutant cells had reduced CFUs present in the spleen 7 days after infection, compared to wild type-infected animals (Fig. 5/J). Hence, CPAR2_200390 plays a role in morphology transition and cell wall homeostasis maintenance, and also contributes to virulence. Fitness regulation by CPAR2_303700. Viability regulation: Notable features of the CPAR2_303700Δ/Δ mutant strain included a growth defective phenotype (Fig. 6/A), low adhesive properties (Fig. 3/B), and susceptibility to the oxidative stressor H 2 O 2 ( Fig. 6/B) and to decreased temperature ( Fig. 6/C). The closest characterized orthologue of CPAR2_303700 is S. cerevisiae CGI121, encoding a subunit of the KEOPS/EKC complex [43][44][45] . According to preliminary searches, this factor is conserved across the entire Candida clade and Cpar2_303700 also contains conserved domains of the CGI121 superfamily (NCBI conserved domain search). According to the performed in silico data analyses (Supplementary Materials and Methods), the two proteins share similar secondary and tertiary structures ( Fig. 6/D, Supplementary Table S3) and Cpar2_303700 is likely to form a stable conformation with the closest interacting partner of Cgi121 in the S. cerevisiae KEOPS/EKC complex, Bud32 (Fig. 6/E, Supplementary  Table S4.), suggesting that Cpar2_303700 and Cgi121 may also share similar properties. Structural similarity and the predicted interaction between Cpar2_303700 and Bud32 was supported by measurable values such as low average RMSD values used for distance comparison (0.61 ± 0.04 Å in case of structure comparison and 0.4 ± 0.2 Å in the predicted complex) and negative overall energy levels (Supplementary Table S4). In silico 7 H-bonds were observed between the residues of Cpar2_303700 and Bud32 that were stabilized by several hydrophobic interactions, further supporting a stable interaction (Fig. 6/E). These data also suggest that Cpar2_303700 might be a potential member of the C. parapsilosis KEOPS/EKC complex.
CPAR2_303700 and virulence: In A. nidulans, members of the KEOPS/EKC complex are associated with metabolic processes that are known to influence the virulence properties of C. albicans. These mechanisms include the regulation of TUP1's function as a key yeast to filamentous growth switch regulating transcriptional factor as well as the complex's involvement in amino acid and carbon source acquisition [46][47][48][49] . Sensitivity to oxidative stress and to temperature changes is also known to impact the pathogenicity of Candida species 36 . Our results indicate that the CPAR2_303700Δ/Δ mutant strain was more susceptible to killing by J774.1 macrophage-like cells, in particular at 18 h of interaction when compared to the wild type strain (Fig. 6/F). When examining survival rates of G. mellonella larvae we found that the mutant cells were less virulent compared to the wild type ( Fig. 6/G). The intermediate phenotype of the 30700 revertant (RI) strain could be due to the gene dosage effect or to local regulatory effects present at the target locus used for reintegration. BALB/c infection results further suggested reduced virulence of the mutant strain as a reduction was also observable in the number of CFUs recovered from the spleen (p = 0.0571) and the kidney (p = 0.0571) 7 days after the infection (Fig. 6/H). These data suggest that the CPAR2_303700 ORF could also contribute to C. parapsilosis virulence.

Discussion
To date, several kinase inhibitors are in use against certain diseases, mainly acting as ATP competitive inhibitors, and lead transcriptional factor inhibitory drugs have already passed preclinical studies, either interfering with the dimerization, co-factor or DNA binding properties of their targets 50,51 . Thus, transcriptional regulators are attractive 'ligands' for novel drug design in various diseases. Due to the complexity of pathogenic mechanisms, multifunctional regulatory factors such as transcriptional factors and kinases are as important as virulence factors themselves in determining the outcomes of host-pathogen interactions. Therefore, targeting the virulence regulatory machinery of pathogenic fungal species has been proposed as an innovative method for effectively combating fungi while concomitantly avoiding the toxicity and resistance mechanisms of the currently available antifungal compounds 2 .
In this study, we used a pre-selective approach for target selection in order to identify regulators potentially contributing to the virulence of an emerging human fungal pathogen, C. parapsilosis in various routes. To identify such regulatory genes, whole RNA sequencing was performed shortly following host-pathogen interaction. Among the ORFs upregulated after infection, transcriptional factor-and kinase-encoding genes were selected for further examination and a targeted deletion mutant collection was constructed. These deletion mutant strains were tested under various conditions designed to recapitulate environments the fungus would encounter under certain host conditions or that would likely influence C. parapsilosis pathogenicity in an indirect manner. As a result we found that 84% of the characterized mutant strains had a phenotype different from wild-type ( Supplementary Fig. S5, Supplementary Table S5). Furthermore, 32% of the tested mutant strains showed multiple phenotypic defects suggesting that the examined regulators have pleiotropic effects (Supplementary Fig. S5). Out of the characterized ORFs, three with potentially distinct regulatory roles in virulence were further examined: CPAR2_100540, CPAR2_200390 and CPAR2_303700.
During infection, pathogens continually compete with host cells for available iron sources. Although iron compounds are usually carried by high affinity proteins in the host, limiting access to these trace elements, pathogenic fungi have developed various strategies in order to acquire these vital compounds 52 . In addition, carbon utilization via respiration is tightly linked to available iron -due to the presence of iron containing enzymes -, that further highlights the role of this trace element in viability 32 . Depending on the site of invasion, invading fungi face a limited access to glucose and often only alternative carbon sources are available for utilization. Under such circumstances, alternative carbon and nitrogen source metabolic pathways are activated, which usually require additional energy and, thus, alter the function of the mitochondrial respiratory chain 49,53 . In C. albicans both iron acquisition and alternative carbon source metabolism are influenced by the transcriptional factor, Hap5, a subunit of the conserved CBF complex. Hap5 controls iron homeostasis by regulating the ferric reductase FRP1, and also influences carbon source utilization via regulating respiratory chain elements cytochrome c and cytochrome c oxidases 30,34 . According to our results, the identified CPAR2_100540 ORF, an orthologous gene of HAP5, plays a similar role in both processes in C. parapsilosis. Moreover, results of the performed virulence studies indicate that the identified regulator further contributes to the virulence of this species in both in vitro and in vivo infection models. It is noteworthy to mention that S. cerevisiae HAP5, is also involved in alternative carbon source utilization, although, while CYC and COX gene expression was downregulated in the S. cerevisiae hap5Δ/hap5Δ strain, upregulation was observed in the respective mutant of C. albicans, which underscores the variability in gene regulation between different yeast species and supports further investigation of these processes in other fungi 33,34 . In the present work, our functional studies show that CPAR2_100540 is comparable to that of C. albicans HAP5; thus, this identified regulator is likely required for C. parapsilosis virulence via the regulation of nutrient acquisition and alternative carbon source utilization.
Yeast to pseudohyphal growth transition promotes C. parapsilosis invasion through several mechanisms, including host cell and tissue disruption, tissue penetration and biofilm formation 13 . In this study, loss of the CPAR2_200390 ORF resulted in a phenotypic feature set previously observed after the removal of SPT3 in C. albicans. Spt3, a subunit of the evolutionally conserved SAGA complex, is known to play a key role in the morphological switching of both S. cerevisiae and C. albicans 38 . Although while SPT3 deletion resulted in a yeast-locked phenotype in S. cerevisiae, hyper-filamentous growth was observed in C. albicans, suggesting opposite regulatory functions in the two species 38 . The obtained phenotype set of the CPAR2_200390Δ/Δ strain indicated that CPAR2_200390 function is equivalent to that of C. albicans SPT3. Interestingly, our further analyses revealed that CPAR2_200390 also influences the chitin homeostasis of the cell wall, adhesive properties and biofilm formation, suggesting a pleiotropic effect on C. parapsilosis virulence. These features have not been previously associated with SPT3. Furthermore, our virulence studies underscored the influence of this regulator on C. parapsilosis pathogenicity, although the outcomes were inconsistent, as the CPAR2_200390Δ/Δ strain was significantly attenuated in vivo, however less efficiently killed by murine macrophages in vitro, when compared to the wild type. These results suggest that at the cellular level, clearance of CPAR2_200390 Δ/Δ is less dependent on macrophages due to weak or no preference (decreased phagocytic activity, data not shown), although could be the result of an alternative host response. In total, these data confirm that CPAR2_200390 regulates virulence in C. parapsilosis by mechanisms that affect mainly morphogenesis and cell wall assembly.
Several conditions, such as temperature and oxidative stress, are known to influence fungal viability, which in turn indirectly affects virulence. Removal of CPAR2_303700 resulted in various defects, mainly including a general growth defective phenotype coupled with sensitivity to low temperatures, suggesting the regulator's involvement in fungal fitness. The closest ortholog to the identified C. parapsilosis ORF is S. cerevisiae's CGI121, and there is no ortholog yet characterized among other Candida species. As part of the highly conserved KEOPS/EKC complex, Cgi121 is involved in functions such as transcription co-activation, tRNA modification, and telomere maintenance, although not required for survival in S. cerevisiae [43][44][45] . Although the performed in silico data analyses do not confirm functional similarity between Cgi121 and Cpar2_303700, they suggest that Cpar2_303700 is a protein structurally similar to Cgi121. Thus Cpar2_303700 may also be a subunit of the evolutionarily conserved KEOPS/EKC complex in C. parapsilosis.
Interestingly, disruption of 3 of the 5 KEOPS/EKC subunits in S. cerevisiae resulted in a serious growth defective phenotype and temperature sensitivity, features that are especially common among telomere defective strains 43 . Mutants defective in telomere maintenance are also hypothesized to be more susceptible to oxidative stress, although this phenotype was not reported in with CGI121. In addition, removal of CGI121 rescued growth defective phenotypes of telomere defective mutant strains, which led to the conclusion that Cgi121 promotes telomere uncapping 43 . The evolutionally conserved hypothetical function and phenotypical features of CPAR2_303700Δ/Δ and Cgi121Δ/Δ strains suggest that there might be opposite regulatory mechanisms between C. parapsilosis and S. cerevisiae, similarly to what was observed in case of HAP5 and SPT3 34,38 , although this hypothesis needs to be supported by further experiments and analyses. Although the exact function of CPAR2_303700 is not yet determined, according to our results, the identified ORF nevertheless contributes to virulence regulation, possibly via an indirect manner.
Taken together, in this study we identified several interesting C. parapsilosis mutant strains with relevant virulence-determining features and a relatively high yield of mutants with multiple defects that indicate the respective regulatory gene's function in virulence. Out of the characterized genes, we describe three transcription regulators that influence C. parapsilosis pathogenicity in distinct ways and thus contribute to our better understanding of virulence regulation in this species. These include Cpar2_100540, a transcriptional factor regulating nutrient acquisition, Cpar2_200390 involved in morphology switch and cell wall assembly, and Cpar2_303700, a protein kinase regulating fungal viability. Although more in-depth studies of the identified regulatory processes in C. parapsilosis are required, these data further support the idea that the origin of virulence in pathogenic species can be due to alterations in signaling and regulatory networks. RNA sequencing and data analysis. RNA sequencing libraries were prepared according to protocols provided by Illumina. Sequencing was carried out in an Illumina HiSeq2000 sequencer with a depth of >20 million reads per sample at the CRG genomics facility. We sequenced paired-end 50bp-long reads RNAseq data for the above-mentioned time-points. The reference genomes for the reference strain CDC317 and the human hg19 were obtained from the Candida Genome Database (CGD) 54 and UCSC (UCSC Genome Browser) 55 , respectively. Gene annotations were obtained from the gencode project, version 18 and from CGD. Read mapping and alignment were carried out using tophat2 v2.0.9 56 with default sensitivity and specificity conditions, based on the bowtie2 2.1.0 short read mapper 57 . Transcript abundance estimation was obtained via FluxCapacitor v1.5.2 with automatic annotation mapping. Gene normalization for thresholding was carried out via RPKM (Reads Per Kilobase of transcript per Million mapped reads) obtained from Flux Capacitor output. A minimum threshold of RPKM ≥10 was used to eliminate low expression transcripts and limit noise. Differential expression analysis was carried out via the DESeq2 v1.10.1 Differential Expression analysis package in R. Supplementary   Table S6. Two of the 19 deletion mutants were also part of the Holland deletion collection and analysis (CPAR2_202040 and CPAR2_100540). Homozygous mutant strains were generated according to methods described by Holland et al. using the C. parapsilosis CLIB 214 double auxotrophic strain (CPL2H1) 15  Response to oxidative, cell wall and other stressors. Cells from overnight cultures were collected, washed 3× with PBS and diluted with fresh YPD to the desired concentration, and 95 μl of 5 × 10 4 cells/ml suspension was plated into each well of 96-well plates. In liquid media: 5 µl of the stress inducing reagents was added to each well from a serial of two-fold dilutions. Detailed information about the applied concentrations of stress inducing reagents is freely accessible and listed in Supplementary Materials and Methods file. Survival rates were normalized to the values of cells grown in YPD media only and compared to the wild-type strain. Survival curves were determined after end-point measurements performed at OD 600 following 18 h of incubation at 30 °C. At least two individual experiments were performed using at least three parallels.

Preparation of C. parapsilosis mutant strains. Strains used in this study are listed in
For validation of CPAR2_100540Δ/Δ and CPAR2_200390Δ/Δ altered phenotypes, additional stressor containing YPD solid media were also used (Supplementary Materials and Methods). At least two individual experiments were performed to confirm obtained phenotypes.

Pseudohypha formation.
For the experiments, 100 µl of 10 7 cells/ml suspension of each strain was added into 24-well cell culture plates, containing 1 ml YPD or DMEM (Gibco) supplemented with 10% heat-inactivated FBS and 1% Penicillin-Streptomycin. Strains were grown at 37 °C and the morphologies were documented at 24 and 48 h of incubation using Leica light microscope (DMI4000B, Leica Microsystems) and HITACHI S-4700 cold field emission scanning electron microscope (see Supplementary Materials and Methods for sample preparation). Three individual experiments were performed to confirm abnormal phenotypes.
Cell wall composition assay (microscopy). To examine cell wall composition of C. parapsilosis mutant cells, ConA-FITC (Sigma-Aldrich), calcofluor white (Sigma-Aldrich) and WGA-TRITC (Sigma-Aldrich) were used. A detailed protocol of staining procedure is available in Supplementary Materials and Methods. Stained strains were imaged with a Zeiss Observer Z1 fluorescence microscope. At least three individual experiments were performed to confirm alterations in cell wall assembly.
Adhesion and biofilm assay. Cultures were grown overnight (12-16 h) at 30 °C with shaking in YNB medium. Cells were washed with PBS and adjusted to 10 5 cells/ml. In a microtiter plate, 100 µL of adjusted cultures were aliquoted per well and statically incubated at 37 °C for 90 min for adhesion to occur. Each well was washed once with 100 µL of PBS, and metabolic activity assay (FDA) was performed to assess adhesion capacity. For adhesion assays, N = 36 for wild-type and N ≥ 12 for mutant strains was used from at least 2 independent experiments. In case of biofilm formation, following adhesion of seeded cells, each well was washed once with 100 µL of PBS, and 100 µL of fresh YNB media was added per well. Plates were incubated at 37 °C with a wet paper towel for 48 h for biofilms to form. After biofilm formation, each well was washed with 100 µL of PBS and biofilm levels were assessed using a metabolic activity assay (FDA). For biofilm assays, at least 2 independent experiments were performed for each mutant strain.
Fluorescein diacetate (FDA) Assay. To each tested well, 100 µL of FDA solution (40 µg/mL, Sigma-Aldrich) was added. Plates were incubated at 37 °C for 1 h, protected from light, and supernatants were transferred to a new fluorescence-compatible plate and read at 485/535 nm (ex/em) in a Victor plate reader.

Gene expression analyses.
For total RNA isolation, the Ambion, Ribopure TM -Yeast RNA isolation kit (Invitrogen) was applied. Prior to examining expression levels in BPS, pH8, amino acid supplemented YNB and FBS-containing DMEM medium, 50 µl of 2 × 10 7 cells from overnight cultures were transferred into the respective media and incubated overnight at 37 °C. For gene expression analysis during biofilm formation, cells were grown in 1 ml of 0.5% glucose-supplemented YNB solution for 48 h.
For cDNA synthesis, the RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific) was applied according to the manufacturer's guide. For expression analyses 100ng cDNA were used. PCR conditions were as follows: 95 °C for 3 minutes, followed by 50 cycles of 95 °C for 10 s, and 60 °C for 30 s. Primers used for gene expression studies are listed in Supplementary Dataset 1. TUB4 housekeeping gene was used as an internal control. Data were normalized to wild-type gene expression levels. Three individual experiments were performed to confirm alterations in gene expression.

Interaction between C. parapsilosis and J774.1 macrophage-like cells.
Killing assays were performed according to a standardized protocol with minor modifications 59 . Briefly, 5 × 10 4 J774.1 macrophages per well (American Type Culture Collection) were infected with C. parapsilosis cells in a 5:1 ratio of yeast:macrophage-like cells. Control wells contained yeast cells only. Interactions were left for 0, 2, 6, 18, and 24 h at 37 °C, with 10% CO 2 . Contents were collected, washed, and passed through a 26 G needle. Suspensions were plated on Sabouraud dextrose agar plates and incubated at 30 °C for 24-48 h. Killing efficiency was calculated as described previously 59 . Graph values were normalized to the number of viable fungal cells in untreated wells at 24 h (100%). At 2 h N ≥ 5, 6 h N = 8, and at 18 h and 24 h N = 4 from 2 independent experiments were used.
Galleria mellonella infection assay. Infection was achieved with 10 µl of 5 × 10 8 C. parapsilosis cells/ml inoculated in the last pro-leg and 20 caterpillars (N = 20) were used per strain. Groups of PBS sham-infected and uninfected (witness group) larvae were also used. The Galleria were maintained at 30 °C and survival was monitored daily.
In vivo murine infection model. Female Balb/c mice (Jackson Laboratories, 6-8 weeks of age, NCI) were injected intraperitoneally with 10 7 cells (N ≥ 6 per yeast strain). After 3 days of infection, 3 mice per condition were sacrificed and the remaining mice were sacrificed at day 7 post-infection. Liver, kidney, and spleen were SCIENTIFIC REPORTS | (2018) 8 Statistical analyses. Unpaired, two-tailed t-tests were applied for gene expression analysis, Kruskal-Wallis test with Dunn's multiple comparisons test was performed for FDA assays, 2-way ANOVA and Dunnett's test for multiple comparisons was applied for J774.1, and non-parametric Mann-Whitney tests were performed on BALB/c mice data assessment, and Mantel-Cox (Log-rank) tests were used for survival data evaluation. After unpaired, two-tailed t-tests, the mean ± SEM values, whereas in case of non-parametric tests, medians and the interquartile range are represented on the graphs. Statistical significance was determined by GraphPad Prism v5.0 or v6.0software. Significant differences were considered at P-values of ≤ 0.05(*p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; ****p ≤ 0.0001).