Strains used in this work are listed in Table 1. Cells were growth in YPD medium (1% [w/v] yeast extract, 2% [w/v] gelatin peptone, 2% [w/v] dextrose) at 28°C and 200 rpm. When solid medium was required, 2% (w/v) agar was added. Yeast transformants were selected in SD medium (0.76% [w/v] yeast nitrogen base with ammonium sulfate without amino acids, 2% [w/v] dextrose and 0.077% [w/v] complete supplement mixture minus uracil] with 50 μg/mL uridine when required. For selection of transformants using the CRISPR-Cas9 strategy, cells were grown on YPD agar supplemented with 200 μg/mL nourseothricin (ClonNAT, WERNER BioAgents, Jena, Germany).

The mini-ura-blaster technique was used for gene disruption as follows. Specific primer pairs for each gene, containing complementary sequences to the 5′- and 3′- regions of the target ORF (see Supplementary Material, Table S1) were used to amplify by PCR the disruption cassette from the pDDB57 plasmid (Wilson et al., 2000). The strain CAI4, a Ura⁻ mutant derived from the clinical isolated SC5314 (Fonzi and Irwin, 1993), was transformed sequentially with the disruption cassettes and the URA3 marker recycled by growing transformants on SC medium supplemented with 1 mg/mL 5-fluoroorotic acid and uridine. The plasmid CIp10 was used to restore URA3 at the RPS1 locus, as previously described (Murad et al., 2000). For gene disruption using the CRISPR-Cas9 strategy, C. albicans strain BWP17 (Wilson et al., 1999) was used as parental strain. Plasmid for CaCas9 Solo system (pV1200 vector) was adopted in this work (Vyas et al., 2015). Constructions of knockout vectors, marker recycling and verification of CRISPR-mutagenized loci were performed according to the previous method (Vyas et al., 2015). The oligonucleotide sequences used in this study are listed in Table S2.

The ORF of each MNN4-like gene-family, plus ~1,000 bp upstream and ~600 bp downstream were amplified by PCR using primers listed in Table S3. The PCR product was cloned into NotI sites of the CIp10 plasmid (Murad et al., 2000). The null mutants generated in this study, a mnn4Δ null strain (Hobson et al., 2004) and a double mnt3Δ, mnt5Δ null strain were transformed with the constructions generated (see Table 1). Before transformation, constructions were digested with StuI, and confirmation of plasmid insertion into the RPS1 locus was performed by PCR.

Total RNA was isolated from yeast cells using TRIzol (Invitrogen), and further purified with the kit RNeasy (Qiagen), following the manufacture's instruction. The cDNA was synthesized using the SuperScript system (Invitrogen). Following synthesis, cDNA purification was conducted after RNA degradation using adsorption chromatography, as described (Trujillo-Esquivel et al., 2016). Absence of genomic DNA in the cDNA preparations was confirmed by amplification of the ACT1 gene which contains an intron of 658 bp (data not shown).

All primer pairs used in qPCR reactions are listed in Table S4. The reaction mixtures were prepared using the SYBR® Green PCR Master Mix (Thermo Fisher Scientific) and analyzed in a StepOnePlus Real-Time PCR System (Applied Biosystems). All reactions produced a single amplicon, with a uniform melting curve, as determined by the dissociation profile of the products. Relative quantification was determined by calculating 2^−ΔΔCT (Livak and Schmittgen, 2001). Expression data were normalized using RPP2B, a housekeeping gene previously used as a control in expression assays (Nailis et al., 2006).

The phosphomannan content was determined by measuring the binding of the cationic dye Alcian Blue to the cell surface, as previously described (Hobson et al., 2004). Briefly, cells in exponential growth phase were collected, washed twice with deionized water and adjusted to an OD₆₀₀ of 0.2. Aliquots of 1 mL were pelleted, and cells suspended in 1 mL of 30 μg/mL Alcian Blue (in 0.02 M HCl) and incubated at room temperature for 15 min. The cell suspension was then centrifuged and the supernatant saved and used to quantify the content of Alcian Blue absorbance at 620 nm. The concentration of the free (non-bound) dye was then determined and used to calculate the amount of Alcian Blue bound to cells as described (Hobson et al., 2004).

Cells from overnight cultures grown in YPD at 28°C and 200 rpm were used to inoculate fresh YPD broth and cell growth was monitored every 30 min by absorbance at 600 nm. Yeast-hypha dimorphism was evaluated by incubating 5 × 10⁶ cells/mL in RPMI 1640 (Sigma) supplemented with 10% (v/v) fetal bovine serum (Sigma) for 4 h at 37°C and 200 rpm. Cell preparations were inspected by bright-field microscopy to evaluate the percentage of yeast cells, pseudohyphae and hyphae. Two hundred cells were counted per strain. Alternatively, hypha formation was stimulated by growing cells on solid Spider medium (Liu et al., 1994). The robustness of the cell wall was tested by assessing the sensitivity of yeast cells to a range of cell wall perturbing agents. Cells in the exponential growth phase were collected, washed twice with deionized water, adjusted to an OD₆₀₀ of 0.05 and seeded into a 96-well plate containing YPD plus doubling dilutions of: SDS, Calcofluor White, Congo Red, and hygromycin B. Plates were incubated at 28°C for 24 h and then read at OD₆₀₀. The highest concentrations tested for each agent were: 0.25% (v/v) SDS, 100 μg/mL Calcofluor White, 100 μg/mL Congo Red, and 500 μg/mL hygromycin B.

The RAW 264.7 (ATCC® TIB-71™) murine cell line was cultured in DMEM media (Sigma), supplemented with 10% (v/v) fetal bovine, at 37°C and 5% (v/v) CO_2. After reaching 90% confluence, cells were detached using trypsin (Sigma) in DMEM medium and sub-seeded into 6-well plates for further growth. Cells were again harvested by trypsin detachment and the concentration adjusted to 2 × 10⁵ cells/mL in DMEM medium. Yeast cells were grown in YPD at 28°C with reciprocal shaking at 200 rpm until they reached exponential growth phase and were then washed twice with PBS, and labeled simultaneously with Acridine Orange (1 mg/mL, Sigma) as reported (Abrams et al., 1983). The fungal cells were washed twice with PBS and resuspended at a cell density of 1 × 10⁷ cell/mL. Phagocytosis assays were carried out in a total volume of 800 μL of fresh DMEM, in 6-well plates. The macrophage:yeast cell ratio was set at 1:3, and interactions were incubated for 2.5 h at 37°C under a CO₂ atmosphere. Macrophages were washed once with PBS and detached from plates using trypsin. Cells were washed twice with PBS by centrifuging at 200 × g for 10 min at 4°C, and resuspended in PBS containing 1.25 mg/mL Trypan Blue as an external fluorescence quencher, as described previously (Santos et al., 2015). Macrophages were kept on ice until they were analyzed by flow cytometry.

Flow cytometry was performed in a MoFlo XDP system (Beckman Coulter) collecting 50,000 events that were singlet events, and then gated for macrophage cells. Fluorescence was recovered from the compensated FL1 (green) and FL3 (red) channels using macrophage cells without any labeling. Phagocytosis of yeast cells was assessed by counts in the green (recently phagocytosed cells) and red (cells within acidified phagolysosomes) fluorescence channels.

Statistical analyses were performed using GraphPad Prism 7 software. All the experiments were performed with three biological replicates in duplicate. Data represent cumulative results of all experiments performed and are shown as means ± S.D. The Mann-Whitney U-test was used to establish statistical significance, which was set at P < 0.05.

The C. albicans MNN4 gene has been previously characterized, and encodes a protein of 997 amino acids, with a putative signal peptide, a transmembrane domain, and an N-terminal region of lysine/glutamic acid repeats that is essential for enzyme activity (Hobson et al., 2004). Near the C-terminal end is a domain that is common to members of the LicD family of proteins, which are involved in phosphocholine metabolism (Zhang et al., 1999). This domain is present in proteins that belong to the nucleotidyltransferase superfamily, which includes enzymes that catalyze the transfer of nucleoside monophosphate to the hydroxyl group of an acceptor, using a nucleoside triphosphate as the donor (Kuchta et al., 2009). Bioinformatics analyses of putative encoded products of MNN4-like genes indicated that all of them have the LicD domain, a putative signal peptide, and a transmembrane domain close to the N-terminus (Figure 1A). These analyses did not identify lysine/glutamic acid repeats in the other MNN4 genes, but some revealed the presence of glutamine repeats, of 2–4 amino acids: For Mnn45 these repeats were closer to the N-terminus, whereas in Mnn41, Mnn42, and Mnn47 they were closer to the C- terminus. No glutamine repeats were identified in Mnn43, Mnn44, and Mnn46 (data not shown). All the seven family members had similarity scores of ~50% to Mnn4, varying between 43% for Mnn41 and Mnn45 to 58% for Mnn42 and Mnn47 (Table 2). The putative amino acid sequences of the MNN4-like genes were used to generate a phylogenetic tree (Figure 1B), using the “Neighbor-joining tree without distance correction” algorithm (Sievers et al., 2011). This analysis showed three main groups: one composed of Mnn43, Mnn44, Mnn45, and Mnn46; the second of Mnn4, Mnn41, and Mnn42; while Mnn47 was isolated as a single outlier (Figure 1B). This dendrogram was used to design subsequent sequential gene disruption strategies.

The C. albicans strain CAI4, a Ura⁻ derived strain from the clinical isolate SC5314 (Fonzi and Irwin, 1993), was used as genetic background for the generation of single null mutants, using the mini-Ura-blaster strategy (Wilson et al., 2000). These null mutants were used as background to generate double, triple, or quadruple null mutants. To avoid problems due to ectopic expression of URA3 (Brand et al., 2004), the resulting null mutants were transformed with the StuI-linearized CIp10 plasmid (Murad et al., 2000), introducing URA3 at the neutral RPS1 locus. The strain CAI4 transformed with the StuI-linearized CIp10 plasmid (NGY152) (Brand et al., 2004) was used as a wild-type (WT) control strain. The re-integrant control strains were constructed introducing the gene of interest at the RPS1 locus, as described in Materials and Methods. Furthermore, to generate a mutant strain lacking all eight MNN4-like genes, the CRISPR-Cas9 system (Vyas et al., 2015) was used in the BWP17 genetic background (Wilson et al., 1999). None of the strains generated in this study (see Table 1) displayed changes in the growth rate, ability to undergo dimorphism, cell and colony morphologies, and the sensitivity to the cell wall perturbing agents Congo Red, Calcofluor White, SDS, and hygromycin B (data not shown). Mutants derived independently in parallel lineages in which the order of disruption of individual genes differed, all generated the same final phenotype.

Next, we tested the ability of the single null mutants to bind Alcian Blue, a dye that specifically binds the negative charge provided by the phosphate group of phosphomannan (Hobson et al., 2004). The WT control cells efficiently bound Alcian Blue whilst the mnn4Δ was incapable of binding this cationic dye (Figure 2). Compared to the WT control strain, no significant changes were observed in the ability of the mnn41Δ, mnn43Δ, mnn44Δ, mnn45Δ, and mnn46Δ null mutants to bind the dye. However, the mnn42Δ null mutant displayed a significant reduction in Alcian Blue binding, which was restored in the re-integrant control strain (Figure 2). Although it was not significant at P > 0.05, the mnn47Δ null mutant showed a reduction in the ability to bind the dye (P = 0.27). Since Figure 1 indicates Mnn41, Mnn42, and Mnn47 are closest paralogs to Mnn4, and because Mnn42 is required for full cell wall phosphomannosylation (Figure 2), we then generated a double null mutant lacking both MNN41 and MNN47. This double mutant showed binding levels comparable to those found in the WT control strain, indicating MNN42 was enough to sustain the normal cell wall phosphomannosylation (Figure 2). A triple mnn41Δ, mnn42Δ, and mnn47Δ null mutant was significantly attenuated in the ability to bind Alcian Blue compared to the WT control cells or the single mnn42Δ null mutant (Figure 2), indicating these family members may operate in tandem in cell wall phosphomannosylation. Complementation of this strain with MNN41 did not increase Alcian Blue binding, but complementation with either MNN42 or MNN47, resulted in a significant elevation in Alcian Blue binding in the triple null mutant (Figure 2), suggesting that MNN42 and MNN47 can play a significant role in the C. albicans cell wall phosphomannosylation. Mnn43, Mnn44, Mnn45, and Mnn46 belong to a different clade of the Mnn4 family (Figure 1). Therefore, we generated a quadruple null mutant lacking all four of these genes. However, this quadruple mutant did not show significant changes in the ability to bind Alcian Blue. The CRISPR-Cas9 single mutants reproduced the same Alcian Blue patterns generated by mini-Ura-Blaster disruption (data not shown). The CRISPR Cas9 generated octuple mutant, lacking all the MNN4-like genes and MNN4 was unable to bind Alcian Blue, showing a phenotype similar to that observed in the mnn4Δ null mutant (Figure 2). These data indicate that some, but not all the members of the MNN4-like gene family participate in the C. albicans cell wall phosphomannosylation, and confirmed the dominant role of MNN4 during phosphomannan synthesis in cells grown under standard laboratory conditions.

We next performed experiments in which transformed additional alleles of each member of the MNN4-like gene family into mutant strains lacking a functional MNN4, to assess whether this impacted on the ability to bind Alcian Blue. A Ura⁻ strain lacking MNN4 (strain CDH7, see Table 1) was transformed with the StuI-linearized CIp10 plasmid (Murad et al., 2000) harboring alleles of the MNN4-like gene family, generating a series of plasmids with an extra copy of each of these genes (see Table 1). All of these genes were expressed under the control of their own regulatory upstream sequences. We first analyzed whether the resulting increase in gene dosage impacted the expression levels of other MNN4 genes. It was shown (Table 3) that in the mnn4Δ null mutant background the expression of other family members was not significantly affected with two exceptions - MNN42 expression almost doubled and MNN47 expression was increased by about 70%. All strains showed a significantly increased expression of the supplemented gene, when extra copies were transformed into the mnn4Δ null mutant background (Table 3). Therefore, incremental increases in the gene dose had a positive impact on the expression level of each member of the MNN4-like gene family. Next, we measured the ability of these strains to bind Alcian Blue and found that none of the complemented transformants were capable of restoring wild-type levels of Alcian Blue binding (Figure 3). However, the incremental increase in the gene dosage of MNN41, MNN42, MNN46, or MNN47 resulted in strains that had significantly higher levels of dye bound, compared to the mnn4Δ null mutant (Figure 3). Therefore, increased expression of some, but all MNN4-like genes, was capable of partially complementing the mnn4Δ null mutant.

Using the same strategy, complementation experiments were performed using the double mnt3Δ, mnt5Δ null mutant background, which lacks the phosphomannosyltransferases required for 50% of the normal level of total cell wall phosphomannan (Mora-Montes et al., 2010). The Ura⁻ strain carrying this dual disruption was used as the parent for complementing transformations with all individual MNN4-like genes (Table 1). Expression analyses indicated that loss of both MNT3 and MNT5 did not affect the expression of other MNN4-like gene family members (Table 3). Upon insertion of an additional MNN4 allele, the expression of the relevant family member was significantly increased (Table 3). When extra copies of the genes encoding MNN41, MNN42, MNN43, MNN44, MNN45, or MNN47 were introduced in the double mnt3Δ, mnt5Δ null background, the resulting increased expression of each gene increased the ability of the double null mutant to bind Alcian Blue (Figure 3). However, increased MNN46 expression in this double mutant did not affect Alcian Blue binding (Figure 3).

Because disruption of the cell wall phosphomannosylation is related to defective phagocytosis of C. albicans by macrophages (McKenzie et al., 2010), we then assessed whether gene disruption or gene supplementation of the MNN4-like gene family members impacted C. albicans-macrophages interactions. As reported previously (McKenzie et al., 2010), the mnn4Δ null mutant strain exhibited a 50% reduction in phagocytosis compared to WT control cells (Figures 4, 5). The mnn42Δ null mutant also showed a significant reduction in uptake by macrophages, that was restored to WT levels in the re-integrant control strain (Figure 4). The triple mnn41Δ, mnn42Δ, mnn47Δ null mutant also displayed a reduction in the phagocytosis when interacting with human macrophages, but reintroduction of a single disrupted gene failed to complement this phenotype (Figure 4). Although the parental strain BPW17 displayed increased phagocytosis compared to the NGY52 (WT) control strain (P = 0.0932, when compared WT and BPW17), the octuple null mutant showed a 50% reduction in the uptake by macrophages (Figure 4). The other null mutant strains generated in this work were phagocytosed to a similar extent to WT control cells (data dot shown).

When the interaction of macrophages and cells harboring an extra copy of the members of the MNN4-like gene family was analyzed, we observed that the cell uptake followed a similar trend to that described in the ability of binding Alcian Blue. Extra copies of either MNN41, MNN42, or MNN46 in the mnn4Δ null background significantly increased yeast cell phagocytosis (Figure 5). Extra copies of all the family members, with the exception of MNN46, positively influenced the phagocytosis of cells in a genetic background lacking both MNT3 and MNT5 (Figure 5). Collectively, these data indicate the members of the MNN4-like gene family influence the C. albicans-human macrophage interaction and that phosphomannan content correlated positively with the degree of phagocytosis of the yeast cells by macrophages.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.