The C. tropicalis PMR1 and OCH1 sequences were identified following a standard BLAST analysis at the NCBI website¹, using the protein sequences of C. albicans Pmr1 (GenBank accession code XP_720380) and Och1 (GenBank accession code AOW28617) as a query. The putative ortholog of C. albicans Pmr1 was the product encoded by the locus EER31186 (GenBank accession code EER31186). The open reading frame (ORF) spans 2760 bp with no putative introns identified and is predicted to encode a polypeptide of 919 amino acids, with 88 and 95% identity and similarity to C. albicans Pmr1, respectively. The putative protein is predicted to fold in nine transmembrane domains, to contain the canonical motif ³⁵²DKTGTLT that includes the aspartic acid involved in the phosphorylation of P-type ATPases (Lutsenko and Kaplan, 1995; Bates et al., 2005), the type A and C cation transport domains, and a putative dehalogenase domain; all these traits already identified in the C. albicans Pmr1 (Bates et al., 2005). For the case of OCH1, the putative ortholog of C. albicans Och1 was the gene product of the locus EER33436 (GenBank accession code EER33436), whose ORF spans 1131 bp with no intron predicted. The encoded protein spans 376 amino acids, with 72 and 85% identity and similarity to C. albicans Och1, respectively, is predicted to be a type-II transmembrane protein belonging to the glycosyltransferase family 32, and contains the sequence ¹⁸⁰DXD, a common signature sequence of proteins that use divalent cations, such as Mn²⁺ as a cofactor during the transference of monosaccharides from a nucleotide-activated sugar to oligo and polysaccharides (Bates et al., 2006). To test these hypotheses generated by bioinformatic means, we next complemented C. albicans pmr1Δ and och1Δ null mutants (Bates et al., 2005, 2006) with the corresponding C. tropicalis ORFs under the control of the strong C. albicans ACT1 promoter (Barelle et al., 2004). As previously reported (Bates et al., 2005, 2006), the C. albicans pmr1Δ and och1Δ null mutants showed defects in the cell wall composition, with low mannan and phosphomannan content and increased glucan levels (Table 1). Chitin content was only modified in the och1Δ null mutant but not in pmr1Δ cells (Table 1). Moreover, in line with the reduction in mannan content, both mutant strains showed increased cell wall porosity (Table 1). Expression of C. tropicalis PMR1 or OCH1 in the corresponding C. albicans null mutant restored this phenotype to levels comparable to those found in the wild-type (WT) control strain (Table 1). Collectively, these data suggest that C. tropicalis PMR1 and OCH1 are the functional orthologs of C. albicans PMR1 and OCH1, respectively.

Next, to generate a C. tropicalis strain lacking either PMR1 or OCH1 the SATI flipper methodology was used to disrupt both alleles of these two genes. The disruption cassettes were constructed in the pSFS2 plasmid (Reuß et al., 2004) and contained 1500 bp regions upstream and downstream the targeted ORF for homologous recombination. For both genes, disruption of the first allele was confirmed by PCR, cells underwent marker recycling, used in the second round of transformation with the same disruption cassette, and again transformation marker recycling, generating the pmr1Δ and the och1Δ null mutants (Table 2). To generate reintegrant control strains, either PMR1 or OCH1 ORF plus regulatory regions were cloned into pSFS2 and this construction used to transform the pmr1Δ or och1Δ null mutant, respectively. The integration of constructions into the targeted loci was confirmed by PCR (data not shown).

Both the pmr1Δ and och1Δ null mutants showed defects in the cell morphology, forming cell aggregates, which were more prominent in the former (Figure 1). The reintegration of one allele of the disrupted gene reverted the phenotype of the null mutant strains, with a cell morphology similar to that displayed by the WT control strain (Figure 1). In addition, both mutant strains showed defects in the colony morphology when grown on a solid culture medium, forming colonies with irregular borders and a wrinkled surface (Figure 1). Again, the reintegrant control strains showed a phenotype similar to the WT strain (Figure 1). These changes were accompanied by changes in the doubling time at 28°C, with the null mutant strains tending to grow slower than the WT and reintegrant control strains (doubling time of the WT control strain 1.49 ± 0.19 h, pmr1Δ strain 2.25 ± 0.24 h, pmr1Δ + PMR1 strain 1.47 ± 0.25 h, och1Δ strain 2.22 ± 0.27 h, and och1Δ + OCH1 strain 1.52 ± 0.12 h, p < 0.05 when compared the null mutant strain with the WT strain or the corresponding reintegrant strain), and both the pmr1Δ and och1Δ null mutants, but not the WT or reintegrant control strains, were unable to grow at 42°C (data not shown).

To assess the role of either PMR1 or OCH1 in the C. tropicalis wall composition, cell walls were isolated, cleaned, acid-hydrolyzed, and analyzed by High-Performance Anion-Exchange Chromatography with Pulsed Amperometric Detection (HPAEC-PAD). The cell wall composition of the WT control strain was similar to that recently reported for C. tropicalis, with a large amount of glucan, followed by mannan, and a small proportion of chitin (Hernandez-Chavez et al., 2018; Table 3). The pmr1Δ null mutant showed about 10-fold more chitin content than the WT control strain, a significant 5-fold reduction in the mannan levels, and no changes in the glucan content (Table 3). The och1Δ null mutant showed a 3-fold reduction in the mannan content, no changes in the chitin levels and a significant 1.3-fold increment in the glucan content (Table 3). Interestingly, the changes in the content of these three cell wall components were different when both null mutants were compared (Table 3). In both cases, the reintegrant control strains restored the content of these wall components to levels similar to those found in the WT control strain (Table 3). The phosphomannan content and the porosity to DEAE-dextran are wall parameters that have been correlated with defects in the protein glycosylation pathways (Mora-Montes et al., 2007, 2010; Cheng et al., 2011; Navarro-Arias et al., 2016, 2017; Perez-Garcia et al., 2016; Lozoya-Perez et al., 2019). The phosphomannan content was reduced in both null mutants, and this was accompanied by an increment in the cell wall porosity (Table 3). Again, both reintegrant strains showed a phenotype similar to the WT strain (Table 3). To further explore defects in the mannosylation pathways in either the pmr1Δ or the och1Δ null mutant, cells were subjected to treatment with endo H or β-elimination to remove N-linked or O-linked mannans, respectively (Navarro-Arias et al., 2016, 2019; Lozoya-Perez et al., 2019). The N-linked and O-linked mannan content in the WT control strain was similar to that recently reported (Navarro-Arias et al., 2019), with most of the cell wall mannan bound to proteins by N-linkages (Figure 2). The pmr1Δ null mutant showed a low level of both mannans, whereas the och1Δ null mutant had a reduction of N-linked mannans and an increment in the O-linked mannan content (Figure 2). The mannan content was different when both null mutant strains were compared (Figure 2). The reintegrant control strains displayed a phenotype similar to the WT strain, confirming that the defects found in the mannosylation pathways are associated with the disruption of the genes under analysis (Figure 2). To further confirm the decreased content of cell wall mannan in both mutants, we labeled these oligosaccharides with fluorescein isothiocyanate-concanavalin A conjugate (Con A-FITC) and inspected the cells under fluorescent microscopy. The lectin strongly bound to the mannan of the WT cells, but a weak interaction of this with the cell wall of either the pmr1Δ or the och1Δ null mutant was observed, confirming of findings of decreased mannan content in these mutants (Figure 3 and Supplementary Figure S1). The reintegrant control strains showed a similar ability to bind Con A-FITC than the WT cells (Figure 3 and Supplementary Figure S1). Since heat inactivation of cells artificially exposes β1,3-glucan and chitin at the cell wall surface (Gow et al., 2007; Mora-Montes et al., 2011; Estrada-Mata et al., 2015; Navarro-Arias et al., 2016, 2017, 2019; Perez-Garcia et al., 2016; Hernandez-Chavez et al., 2018); we next repeated the experiment but using heat-killed (HK) cells for fluorescent labeling. This cell treatment did not have an impact on the ability of Con A-FITC to bind the cell wall (Figure 3 and Supplementary Figure S1). When these experiments were performed with β-eliminated cells we observed a small reduction in the ability of the lectin to bind cell walls, but this was not significant (P > 0.05; Figure 3). When endo H-treated cells were used for interaction with Con A-FITC there was a significant reduction in the binding of the lectin by both the live and HK WT and reintegrant control cells (Figure 3). In the case of both mutants, the treatment with endo H decreased the binding of Con A-FITC, but this was not significant (Figure 3).

Next, we analyzed the organization of the structural polysaccharides chitin and β1,3-glucan within the cell wall, using fluorescence labeling with fluorescein isothiocyanate-wheat germ agglutinin conjugate (WGA-FITC) and an IgG Fc-Dectin-1 chimera, respectively (Graham et al., 2006; Mora-Montes et al., 2011; Marakalala et al., 2013; Estrada-Mata et al., 2015; Navarro-Arias et al., 2016, 2017, 2019; Hernandez-Chavez et al., 2018). Both lectins barely bound to chitin and β1,3-glucan localized in the cell wall of the WT and the re-integrant control strains (Figure 3 and Supplementary Figures S2, S3). Cells from either the pmr1Δ or the och1Δ null mutant were significantly more labeled than the WT or reintegrant control cells with both lectins, suggesting exposure of chitin and β1,3-glucan at the cell wall surface (Figure 3 and Supplementary Figures S2, S3). The HK WT and reintegrant control cells showed an increased ability to bind both lectins (Figure 3 and Supplementary Figures S2, S3), suggesting that the modest interaction between the lectins and live cells was due to inaccessibility of lectins to polysaccharides, i.e., these are in the inner part of the cell wall. The chitin labeling in both the pmr1Δ and och1Δ null mutants was similar to that observed in live cells, and β1,3-glucan labeling was increased upon heat inactivation (Figure 3). As in the case of mannan labeling, the β-elimination treatment did not affect the chitin or β1,3-glucan labeling (Figure 3). Upon incubation with endo H, the lectin binding to both polysaccharides from live and HK null mutant strains was similar to the untreated cells, but an increased binding to both chitin and β1,3-glucan was observed in the walls of live WT and reintegrant control cells (Figure 3). Collectively these data confirm that both polysaccharides are exposed at the cell surface of the pmr1Δ and och1Δ null mutants. In addition, they are in line with the increment in glucan content quantified by HPAEC-PAD.

It has been previously reported that Candida mutants in either the PMR1 or the OCH1 genes have increased susceptibilities to cell wall perturbing agents (Bates et al., 2005, 2006; Navarro-Arias et al., 2016; Perez-Garcia et al., 2016). Therefore, we analyzed the effect of this kind of compounds on the growth of the C. tropicalis pmr1Δ and och1Δ null mutants. Congo red, a compound that disturbs the glucan within the cell wall (Kopecka and Gabriel, 1992), had a modest effect on the growth of the WT and reintegrant control strains but both the pmr1Δ and och1Δ null mutants showed increased susceptibility to this stressor (Figure 4). The WT and reintegrant control strains showed a dose-dependent increased susceptibility to both Calcofluor white and hygromycin B, and this effect was more prominent in the presence of the latter (Figure 4). Calcofluor white binds to chitin; while increased susceptibility to hygromycin B has been reported in yeast cells with defects in the protein glycosylation pathways (Dean, 1995; Mora-Montes et al., 2011). Again, the pmr1Δ and och1Δ null mutants had increased susceptibility to these perturbing agents (Figure 4). Interestingly, when compared to the ability of both mutants to grow in presence of these three cell wall perturbing agents, the och1Δ null mutant was more affected than the strain were PMR1 was disrupted (P < 0.05 in all cases; Figure 4). Altogether, these data indicate that loss of either PMR1 or OCH1 affects the C. tropicalis protein glycosylation and the cell wall composition, organization, and fitness.

Next, we assessed the importance of cell wall mannan on the C. tropicalis-host interaction, analyzing the ability of this fungal species to stimulate cytokine production by human PBMCs. We first compared the cytokine profile induced by the WT strain treated with endo H or β-elimination, to remove N-linked or O-linked mannans, respectively (Navarro-Arias et al., 2016; Perez-Garcia et al., 2016). Moreover, since in other Candida species β1,3-glucan is one of the major fungal ligands to stimulate cytokine production and is normally buried behind the mannan layer, and therefore inaccessible to engage with dectin-1 (Gow et al., 2007; Estrada-Mata et al., 2015; Navarro-Arias et al., 2016, 2017, 2019; Perez-Garcia et al., 2016), we also compared the ability of live and HK cells to stimulate cytokine production. Live cells stimulated low levels of TNFα, IL-1β, IL-6, and IL-10, but PMBCs interacting with HK cells produced a significantly higher level of these cytokines (Figure 5). The experimental setting where PBMCs were pre-incubated with laminarin and challenged with HK cells stimulated cytokine levels similar to those with the live cells (Figure 5). Thus, the high cytokine levels induced by HK cells were associated with the engagement of dectin-1 with its ligand, as reported for other fungal cells (Estrada-Mata et al., 2015; Navarro-Arias et al., 2016, 2017, 2019; Perez-Garcia et al., 2016). Removal of O-linked mannans via β-elimination did not affect the TNFα, IL-1β, or IL-6 profiles, but live cells with no O-linked mannans on the surface tended to simulate higher IL-10 levels than untreated cells (Figure 5). The trimming of N-linked mannans with endo H only affected the ability of live cells to stimulate IL-10, which was higher, when compared with the levels stimulated with untreated cells (Figure 5). The endo H-treated HK cells were not able to stimulate a strong production of TNFα, IL-1β, or IL-6; while the levels of IL-10 were not affected when compared with those stimulated with the untreated HK cells (Figure 5). It is possible to hypothesize that treatment with endo-H could expose cell wall proteins to the extracellular compartment, promoting detachment of these macromolecules from the wall, accounting for the changes in the ability to stimulate cytokine production. However, these is unlikely, as the wall protein content of cells was not affected upon treatment with endo-H (132.4 ± 21.4, 144.1 ± 17.2, 122.9 ± 13.9, and 136.8 ± 15.6 μg of protein mg of cell wall^–1 for live, live + endo H-, HK, HK + endo H-treated cells, respectively).

In similar experiments, the ability of the pmr1Δ and och1Δ null mutant to stimulate cytokine production by human PBMCs was also assessed. Live cells of either the pmr1Δ or och1Δ null mutant stimulated higher TNFα, IL-6, IL-1β, and IL-10 when compared with the WT live cells (Figure 6). For the three pro-inflammatory cytokines quantified, the stimulated levels using HK mutant cells were similar to those observed in the system with live cells and not sufficient enough to be comparable to those stimulated with the WT HK cells (Figure 6). For IL-10, the cytokine level stimulated with HK cells was comparable with that observed with live mutant cells, but slightly higher, and therefore statistically similar to the levels stimulated with the WT HK cells (P = 0.627 when WT and och1Δ null mutant cells are compared; P = 0.295 when WT and pmr1Δ null mutant cells are compared). Both live and HK cells of the reintegrant control strains for OCH1 and PMR1 showed an ability to stimulate cytokine production similar to WT control cells (Figure 6). It is worthy of note that it was expected that the cytokine profile stimulated with endo-H-treated WT cells was similar to that observed with cells lacking OCH1 but results from Figures 5, 6 clearly showed this was not the case. Since Och1 is a protein localized within the Golgi complex, it is expected that the N-linked mannan core is properly assembled and transferred to glycoproteins, as demonstrated in C. albicans (Bates et al., 2006). Therefore, the cell surface of an och1Δ null mutant strain will display the N-linked mannan core on the wall surface, which contrasts with the surface of endo-H-treated cells where all the N-linked mannan, including the oligosaccharide core, is absent. Thus, we next compared the ability to stimulate cytokine production of the null mutants subjected to endo H treatment. For both live and HK cells treated with endo H, the levels of the proinflammatory cytokines TNFα, IL-6, and IL-1β were similar to those produced in the system where live untreated cells were included (Figure 6). For IL-10 stimulation, the och1Δ null mutant did not show any difference in the ability to stimulate this cytokine when compared to the system where endo H was not included (Figure 6). In the case of the pmr1Δ null mutant, live cells treated with endo H stimulated significantly higher levels of this cytokine, when compared to the system that not included the glycosidase (Figure 6). The reintegrant control strains for both genes showed a similar ability to stimulate cytokine production to that observed with the WT control cells (Figure 6). To further explore the differences between endo-H-treated WT cells and the och1Δ null mutant we performed interaction experiments in the presence of the N-linked mannan core Man₉GlcNAc₂ (M9) (Mora-Montes et al., 2004) and measured the cytokine production. Human PBMCs pre-incubated with M9 and then challenged with live WT cells stimulated similar cytokine levels than cells without the preincubation step; however, in similar experiments using HK cells the amount of TNFα, IL-6 and IL-1β was significantly diminished, when compared with the system without M9 (Figure 7 and data not shown). For the case of IL-10, pre-incubation with M9 did not affect the cytokine levels stimulated by either live or HK cells (Figure 7). In the case of both live and HK och1Δ null mutant cells, the levels of TNFα, IL-6, and IL-1β were significantly reduced upon pre-incubation with M9 (Figure 7 and data not shown). As the WT cells, the levels of IL-10 stimulated with either live and HK och1Δ null mutant cells were not affected by pre-incubation with M9 (Figure 7). The reintegrant control strain stimulated similar cytokine levels to those observed with the WT strain (Figure 7). The pmr1Δ null mutant showed a similar ability to stimulate cytokine production than the och1Δ null strain (data not shown). This antagonistic ability of M9 during cytokine production suggested the involvement of mannose receptor (MR) in the PBMC-yeast interaction, as this pattern recognition receptor senses branched oligosaccharides with alpha glycosidic bonds, as M9 (Netea et al., 2006). Therefore, we explored the contribution of this receptor in the C. tropicalis recognition. Human PBMCs preincubated with anti-MR antibodies were challenged with either WT or mutant cells. The anti-MR antibodies did not affect the ability to produce IL-10 but negatively affected the TNFα, IL-6, and IL-1β levels stimulated with HK WT cells (Figure 7 and data not shown). For the case of the och1Δ null mutant, the presence of anti-MR antibodies within the system had little impact on the production of IL-10 but reduced the production of TNFα, IL-6, and IL-1β stimulated with live cells (Figure 7 and data not shown). Although the pro-inflammatory cytokine levels were reduced in the presence of anti-MR and HK cells of the och1Δ null mutant, this was not statistically significant (Figure 7 and data not shown). The control system with an isotype-matched antibody or with the reintegrant control strain shown cytokine levels similar to those stimulated with the WT cells.

Finally, we assessed the contribution of dectin-1 during the interaction with och1Δ null cells. As in the case of the WT strain, the TNFα, IL-6, IL-1β, and IL-10 levels stimulated by either live or HK mutant cells was significantly reduced (Figure 7 and data not shown). Similarly, the cytokine production stimulated by live or HK pmr1Δ cells was negatively affected by pre-incubation with laminarin (data not shown), stressing the importance of dectin-1 engagement with β1,3-glucan during the C. tropicalis-human PBMC interaction.

To assess the relevance of protein mannosylation in the C. tropicalis virulence, we first infected larvae of G. mellonella and evaluated the survival rate of these animals. The animals infected with WT and the reintegrant control strains showed similar survival curves, with the total of the animal population dead after 8 days of infection when injected with the WT strain (Figure 8). For the case of the reintegrant strains, the total of the animal population succumbed after 10 days post-infection, and this difference was not statistically significant (Figure 8; P = 0.19 and 0.27, when the curve associated to WT cells is compared with either pmr1Δ + PMR1 or och1Δ + OCH1, respectively). The survival curves of animals infected with either the pmr1Δ or the och1Δ null mutant were similar and showed virulence attenuation, with 50% of the animal population surviving at the end of the observation period (Figure 8). The fungal burdens in the inoculated animals indicated a defect in the null mutant strains to colonize the host tissues, as this was lower than that associated to the WT or the reintegrant control strains (1.3 × 10⁷ ± 0.6 × 10⁷ cells mL^–1, 0.7 × 10⁷ ± 0.2 × 10⁷, 1.1 × 10⁷ ± 0.4 × 10⁷cells mL^–1, 0.6 × 10⁷ ± 0.4 × 10⁷, 1.0 × 10⁷ ± 0.6 × 10⁷cells mL^–1 for WT, pmr1Δ, pmr1Δ + PMR1, och1Δ, och1Δ + OCH1, respectively. P < 0.05 when compared to the null mutant strains with the WT control strain).

We next used a murine model of non-lethal disseminated candidiasis as previously reported (Ifrim et al., 2014; Navarro-Arias et al., 2016; Perez-Garcia et al., 2016), to evaluate the virulence in this in vivo setting. After 3 days of infection, BALB/c mice infected with the pmr1Δ null mutant had lower fungal burdens in the spleen, kidneys, brain, and liver, when compared to either the WT control cells or the reintegrant control strain (Figure 9). Similarly, the fungal burden of animals infected with the och1Δ null mutant strain was lower to that observed in animals infected with the WT strain, but interestingly, the och1Δ + OCH1 reintegrant control strain behaved like the null mutant strain, and fewer colony-forming units were recovered from the spleen, kidneys, brain, and liver of infected animals (Figure 10). Collectively, these results indicate that PMR1 and OCH1 have a significant role in the interaction of C. tropicalis with either G. mellonella or the mouse.

Organisms used in this study are listed in Table 2. Unless otherwise indicated, cells were maintained and propagated at 28°C in YPD medium [2% (w/v) bacteriological peptone, 1% (w/v) yeast extract, 2% (w/v) glucose]. The medium was supplemented with 2% (w/v) agar when the solid plates were required. During gene disruption, the SAT1 marker was recycled by growing cells at 28°C in liquid YEP [2% (w/v) bacteriological peptone and 1% (w/v) yeast extract]. To prepare cells for interaction with the host and cell-wall analysis, the strains were grown at 30°C in 500 mL flasks containing 100 mL of fresh medium with shaking at 200 rpm until reaching the logarithmic growth phase. Cells were inactivated by incubating at 56°C for 60 min. The confirmation of the presence of non-viable cells was performed on YPD plates incubated for 4 days at 28°C. For chemical remotion of O-linked mannans, cells were β-eliminated in 10 mL of NaOH 0.1 N and incubated at room temperature during 18 h, as previously described (Diaz-Jimenez et al., 2012). Cell suspensions were neutralized with HCl 0.1 N, and pelleted at 2000 × g, for 5 min. The supernatant was saved for O-linked mannans analysis (see below). The enzymatic N-linked mannan removal was performed with 25 U endo H (New England Biolabs) as reported (Mora-Montes et al., 2012). In both cases, cells were washed twice with sterile phosphate-buffered saline (PBS) and viability was confirmed by quantifying the colony-forming units before and after the treatment. The loss of viability after β-elimination and endo H treatment was 1.8 ± 0.9% and 3.3 ± 4.1%, respectively). To disperse cell aggregates, 2 units mL^–1 chitinase (Sigma) was added to medium when required (Bates et al., 2006; Perez-Garcia et al., 2016).

The PMR1 ORF was amplified by PCR using the primer pair 5′-AAGCTTATGAGTGATAACCCTTATGAACTA-3′ and 5′-GCTAGCTTATACTCCATATGTATAATTATTAGTATAAATAG T-3′ (underlined sequences correspond to restriction sites for HindIII and NheI, respectively); while OCH1 ORF was amplified with the primer pair 5′-AAGCTTATG CGTCTGAAGGATATCA-3′ and 5′- GCTAGCTTAATCTTCC ATTTCTGGCAT (underlined sequences correspond to restriction sites for HindIII y NheI, respectively). The amplicons were cloned into pCR^®2.1-TOPO^® (Invitrogen), and subcloned into the HindIII and NheI sites of pACT1 (Barelle et al., 2004), generating pACT1-CtPMR1 and pACT1-CtOCH1. A C. albicans pmr1Δ null mutant (Bates et al., 2005) was transformed with StuI-digested pACT1-CtPMR1, generating strain HMY185; whereas a C. albicans och1Δ null mutant (Bates et al., 2006) was transformed with StuI-digested pACT1-CtOCH1, generating the strain HMY148. Since the OCH1 ORF contains an internal recognition site for HindIII, after cloning into pCR^®2.1-TOPO^® (Invitrogen), the codon ⁴⁵⁰AAA⁴⁵³ was modified to ⁴⁵⁰AAG⁴⁵³ by site-directed mutagenesis using the Phusion Site-Directed Mutagenesis Kit (Thermo) and the primer pair 5′- CCTGATGTTTCAAGGCTTATAAAATTATGCCAAAATC-3′ and 5′- GGCATAATTTTATAAGCTTTGAAACATCAGGAA TTTC-3′. The mutation was confirmed by DNA sequencing, before cloning into pACT1.

PMR1 and OCH1 were deleted using the SAT1 flipper method, as previously described (Hernandez-Chavez et al., 2018). For both genes, the 1500 bp upstream and the 1500 bp downstream regions of the ORF were amplified by PCR and cloned into the ApaI-XhoI and NotI-SacI sites of pSFS2 (Reuß et al., 2004), respectively, generating the corresponding disruption cassettes. The primer pairs used for the generation of the disruption cassettes were: for the PMR1 upstream region 5′-GGGCCCAGTTGATTGAAAAATTTTGGCCAGG-3′ and 5′- CTCGAGAGGGAGTCTGTTGAAGGAGTG; for the PMR1 downstream region 5′-GCGGCCGCACTTCCAGAAG TGTGGTATGTGT-3′ and 5′- GAGCTCACTACGCATACCT TCCAACCA-3′; for the OCH1 upstream region 5′-GGGCCCACACAACCGTCTCTATCAGG-3′ and 5′-CTCGAG TGTTAACTTCTGAATTGGTCTATGA-3′; and for the OCH1 downstream region 5′- GCGGCCGCTCTAATGTTCACAA TTAAAACTTGTATT-3′ and GAGCTCGATTTGTTAATTAA AGAAGATGGTGT-3′. The disruption cassettes were excised from the vector by digesting with ApaI and SacI and used for cell transformation of strain ATCC^® MYA-3404 (referred to as WT control strain). The selection of transformed cells was performed in YPD plates added with 200 μg mL^–1 nourseothricin (Goldbio). The integration of the disruption cassettes in the targeted locus was analyzed by PCR with primer pairs aligning inside and outside of the recombinations regions. Cells were grown for 2 days at 28°C in YEP broth added with 2% (w/v) maltose to recycle the SAT1 marker. The loss of SAT1 was confirmed by plating cells in YPD supplemented with 10 μg mL^–1 nourseothricin and PCR. The second allele of PMR1 and OCH1 was disrupted in the second round of transformation with the same cassettes and recycling of the SAT1 marker.

To reintegrate PMR1 or OCH1 into the C. tropicalis null mutant generated, the ORF plus 1000 bp upstream and 500 bp downstream was amplified by PCR using the following primer pairs: 5′- GGGCCCCCAATGTTATTACGTCCTGGTGG-3′ and 5′- CTCGAGGAGGGAGAAAAGGGGAGGATAC-3′ for PMR1, and 5′- GGGCCCGGTAATGCTTCACCATCATCAACT-3′ and 5′- CTCGAGATCCACGGAAAGAACCGCAA-3′ for OCH1. Either the 2789 bp fragment corresponding to OCH1 or PMR1, a 4517 bp amplicon, were cloned into the ApaI-XhoI sites of pSFS2, and then a 700 bp fragment downstream of the disrupted locus was cloned into the NotI-SacI sites of the same vector. These downstream fragments were generated by PCR using the primer pairs 5′-GCGGCCGCATTGCTTCATTTGGAGGTGTT G-3′ and 5′-GAGCTCCTACGCATACCTTCCAACCA-3′ for PMR1; and 5′-GCGGCCGCCCTGGAGGCAAGTGCTTTTCA-3′ and 5′-GAGCTCACGGATTTGTATTTGTCAAGAGCCA-3′ for OCH1. These reintegration constructions were linearized with ApaI and SacI and used to transform the null mutant strains. Upon recycling of the SAT1 marker, confirmation of the reintegration of either PMR1 or OCH1 in one of the native loci was confirmed by PCR.

For the analysis of the composition, grown-cells in YPD were disrupted in a Precellys homogenizer (Bertin), the homogenate was centrifuged, the pellet saved, thoroughly washed with deionized water, and the cell walls were cleansed and acid-hydrolyzed as described previously (Mora-Montes et al., 2007). The acid-hydrolyzed samples were analyzed by HPAEC-PAD in a carbohydrate analyzer system (Thermo), using the columns and separation conditions previously reported (Plaine et al., 2008). The cell wall protein content was measured upon alkali hydrolysis of the cell wall and the Bradford method, as previously described (Mora-Montes et al., 2007).

The cell wall porosity was determined by the relative porosity of polycations, as reported earlier (De Nobel et al., 1990). Cells were grown in YPD broth at 28°C until reaching the mid-log phase, washed twice with PBS, and aliquots containing 1 × 10⁸ cells were suspended in either 10 mM Tris–HCl, pH 7.4 (buffer A), buffer A plus 30 μg/mL poly-L-lysine (Mw 30–70 kDa, Sigma) or buffer A plus 30 μg/mL DEAE-dextran (MW 500 kDa, Sigma), and incubated for 30 min at 30°C with constant shaking at 200 rpm. Cell suspensions were pelleted by centrifuging, the supernatants saved, further centrifuged, and used to measure the absorbance at 260 nm. The cell wall porosity was calculated as described previously, using the porosity to poly-L-lysine for data normalization (De Nobel et al., 1990). To quantify the cell wall phosphomannan, the cellular ability to bind Alcian blue was analyzed as described (Hobson et al., 2004).

Localization of mannan, chitin, and β1,3-glucan within the cell wall was analyzed by fluorescent microscopy using 1 mg mL^–1 Con A-FITC (Sigma), 1 μg mL^–1 WGA-FITC (Sigma) and 5 μg/mL IgG Fc-Dectin-1 chimera conjugated with anti-Fc IgG-FITC for mannan, chitin, and β1,3-glucan staining, respectively (Graham et al., 2006; Mora-Montes et al., 2011; Marakalala et al., 2013). Cells were examined by fluorescence microscopy using a Zeiss Axioscope-40 microscope and an Axiocam MRc camera. The fluorescence was quantified from the pictures acquired with Adobe Photoshop^TM CS6 using the formula: [(total of green pixels-background green pixels) × 100]/total pixels. The experiment was performed three independent times, with a total of 300 cells analyzed per strain.

Cells grown at the mid-log phase in YPD were incubated overnight at 37°C with 25 U endo H (New England Biolabs) to trim cell wall N-linked mannans (Mora-Montes et al., 2012). Cells were also β-eliminated with 0.1 N NaOH, as described (Diaz-Jimenez et al., 2012). Then, cells were pelleted, and the supernatant used to determine the mannose content with the phenol-sulfuric-acid method (Dubois et al., 1956). In both cases, mannan release was confirmed by HPAEC-PAD as reported (Perez-Garcia et al., 2016).

Cell growth in the presence of cell wall perturbing agents was performed as previously described (Navarro-Arias et al., 2016). Briefly, overnight grown cells in YPD were washed with deionized water, separated using a syringe with a 32-gauge needle and suspended at an OD_600nm = 1.0. Then, fresh YPD broth was inoculated with the fungal cells at an OD_600nm = 0.01, and aliquots containing 95 μL were placed into 96-well plates. The cell wall perturbing agents, in a final volume of 5 μL, were added to each well. Mock wells contained 5 μL of the vehicle only and were used to normalize the results. The final OD_600nm for each well was quantified after 24 h incubation at 30°C. Calcofluor white and Congo red were from Sigma; whereas hygromycin B from GoldBio. Growth data were normalized as a percentage of those generated with the mock interactions.

The use of human PMBCs was approved by the Ethics Committee from Universidad de Guanajuato (permission number 17082011), and peripheral blood samples were withdrawn from healthy adult volunteers after information of the study was disclosed and a written consent form signed. The Ethics Committee at the University of Szeged approved the experimentation with mice (permission number XVI./3646/2016).

Upon peripheral blood was collected in tubes containing EDTA as anticlotting, PBMCs were isolated by differential centrifugation using Histopaque-1077 (Sigma), as reported (Endres et al., 1988). To stimulate cytokine production, 100 μL containing 5 × 10⁵ PBMCs in RPMI 1640 Dutch modification (added with 2 mM glutamine, 0.1 mM pyruvate and 0.05 mg mL^–1 gentamycin; all reagents from Sigma) were plated onto round-bottom 96-well microplates, and then 100 μL containing 1 × 10⁵ fungal cells freshly harvested or treated were added to each well. The plates were incubated for 24 h at 37°C with 5% (v/v) CO₂. In some experiments, PBMCs were pre-incubated for 1 h at 37°C with 200 μg mL^–1 laminarin (Sigma), 10 μg mL^–1 M9 (Man₉GlcNAc₂; Sigma), 10 μg mL^–1 anti-MR (Invitrogen, Cat. No. Mab-Hmr), or 10 μg mL^–1 IgG1 antibodies (Santa Cruz Biotechnology) before interaction with fungal cells. All these reagents were LPS free, as tested with the Limulus amebocyte lysate (Sigma, data not shown). Nonetheless, interactions were conducted in the presence of 5 μg mL^–1 polymyxin B (Sigma) (Navarro-Arias et al., 2016). The plates were centrifuged for 10 min at 3000 × g at 4°C, and the supernatants were collected and stored at −20°C until used.

The concentrations of TNFα, IL-6 and IL-10 were quantified with the ABTS ELISA Development kits from Preprotech; while IL-1β levels were measured with a DuoSet ELISA Development kit (R&D Systems).

The G. mellonella infection was performed as previously reported (Perez-Garcia et al., 2016). Briefly, 2 × 10⁷ yeast cells contained in 10 μL PBS were passaged through a syringe with a 32-gauge needle, the last left pro-leg of the larva sanitized and fungal cells directly injected into the hemolymph, using a Hamilton syringe and a 26-gauge needle. Upon infection, the animals were kept at 37°C and survival was monitored daily. Animal death was defined by extense melanization and irritability absence. Each experimental group contained 30 larvae. Mock infections with animals injected with PBS were included as a control. The content of colony-forming units was determined by incubating serial dilutions of the hemolymph on YPD plates for 28°C for 72 h.

A non-lethal experimental model of disseminated candidiasis was performed as reported (Ifrim et al., 2014; Perez-Garcia et al., 2016). Animal groups containing 8 female Balb/c WT mice (older than 12 weeks old and 20 to 25 g of weight) were injected via the lateral tail vein with 100 μL of sterile PBS containing 1 × 10⁶ C. tropicalis yeast cells, previously passaged through a syringe with a 29-gauge needle. As a control group, eight mice were injected with 100 μL of sterile PBS. Animals having access to sterile water and a normal diet ad libitum were monitored daily and showed no signs of infection such as weight loss, lethargy, ruffled fur or rapid shallow breathing during 3 days following the intravenous injection. At this point, animals were euthanatized, and the liver, brain, kidneys, and spleen were removed, weighed, and separately homogenized with a tissue grinder. The fungal burden was determined by incubating serial dilutions on YPD agar plates for 28°C for 48 h.

Statistical analysis was performed using GraphPad Prism 6 software. Growth data in the presence of cell wall perturbing agents were analyzed by two-way ANOVA. Cytokine stimulation using human PMBCs was performed in duplicate with seven healthy donors, whereas the rest of the in vitro experiments were performed at least thrice in duplicate. Data represent the cumulative results of all experiments performed. The Mann–Whitney U test or unpaired t-test was used to establish statistical significance (see figure legends for details), with a significance level set at P < 0.05. Experiments with G. mellonella were performed three times, with a total of 30 larvae per group (10 larvae for each experiment). Results were analyzed using the Log–rank test and arranged in survival curves using Kaplan-Meier charts. The statistical significance was set at P < 0.05.