Candida albicans resistance to hypochlorous acid

Time-course assays were carried out to define the doses and kinetics with which HOCl kills C. albicans. Cells were exposed to different concentrations of HOCl and then plated on agar medium to determine the viable colony-forming units (CFUs). (Note that dilutions of NaOCl were prepared, but since the pKa is ~7.4, there will be a mix of HOCl and ClO⁻. For simplicity, we will describe this mix as HOCl in the text and figures.) As shown in Fig. 1A, significant killing occurred at >10 µM HOCl. This is interesting since it is a much lower concentration than is needed for H₂O₂ to kill C. albicans, which is in the mM range (24). It was also interesting that HOCl acted very quickly. Treatment of cells for 15 min with 10 µM HOCl resulted in ~42% killing, and treatment with 20 µM HOCl resulted in ~99% killing. Additional loss of viability continued to occur at later time points through 60 min. In fact, there was even a significant drop in viability for cells treated for 60 min with 5 µM HOCl. These doses are expected to be in the range of HOCl concentrations that cells experience in the phagosome (19), although it has been challenging to estimate the concentration of HOCl in the phagosome because HOCl is short-lived due to its ability to rapidly react with cellular components.

We predicted that the PM would be a critical target for the highly reactive HOCl, as the PM would be an initial point of contact after HOCl is created in the phagosome. We therefore examined the integrity of the PM after HOCl treatment by staining cells with SYTOX Green, a membrane-impermeable fluorescent stain that binds nucleic acids (25). Interestingly, exposure of cells to 20 µM HOCl for as short as 2.5 min resulted in about 66% stained cells, indicating a high level of PM permeabilization (Fig. 1B). The percent of stained cells did not increase much with longer times of incubation. Some of our control studies indicated that this might be due to HOCl interfering with the SYTOX Green assay. Nonetheless, these studies show a very rapid permeabilization of the C. albicans PM. Damage to the PM is expected to exacerbate the effects of HOCl by facilitating the entry of HOCl and other oxidized products into the cytoplasm, where essential functions can be perturbed.

The ability of a combination of copper and HOCl to kill cells was examined since it has been shown that copper is pumped into the phagosome (26) and that copper is known to be redox-active and can attack the plasma membrane (27). Interestingly, a mix of CuSO₄ and HOCl showed more substantial effects than either one alone, indicating a synergistic effect on cell killing (Fig. 1C).

The PM forms a critical barrier around cells that also participates in a wide range of dynamic functions essential for virulence, including secretion, endocytosis, morphogenesis, and cell wall synthesis (28). To better define how HOCl impacts PM function, we analyzed a set of mutants that were defective in endocytosis (rvs161Δ; rvs167Δ), MCC/eisosomes (sur7Δ; pil1Δ lsp1Δ), phospholipid flippase (drs2Δ), actin cytoskeleton (arp2Δ arp3Δ), and mannosyl transferase (och1Δ) that affects the PM and cell wall (27, 29–31) (Fig. 2). Disk diffusion (halo) assays, rather than assays in which cells are spotted onto an agar plate with a fixed concentration of a chemical, were used to test the mutants for sensitivity to the oxidants HOCl and H₂O₂ so that differences between strains could be quantified (Fig. 2A and B). The HOCl solution was spotted directly on the plate since we found that paper disks that are often used in disk-diffusion-type assays had unpredictable effects, with some disks quenching the effects of HOCl while others from the same batch had no effect. The results showed a trend in which all of the PM mutants showed larger zones of growth inhibition caused by HOCl, but only the arp2Δ arp3Δ strain reached statistical significance by analysis of variance (ANOVA; P < 0.0001). In contrast, the effects of H₂O₂ were often distinct from those of HOCl (Fig. 2B). For example, the sur7Δ, pil1Δ lsp1Δ, rvs161Δ, and rvs167Δ mutants all showed significantly increased susceptibility to H₂O₂.

A set of mutants carrying mutations in known antioxidant enzymes was tested next for sensitivity to HOCl to determine whether any of these pathways play a critical role in resistance to HOCl (Fig. 3A). These mutants were defective in processes that included catalase, superoxide dismutases, flavodoxin-like proteins, the HOG MAP kinase pathway, and the Cap1 transcription factor that regulates the expression of many antioxidant genes (32–34). Although many of the mutants showed a trend toward increased susceptibility to HOCl, only the HOG pathway mutants ssk2Δ, pbs2Δ, and hog1Δ showed a statistically significant difference by ANOVA (P < 0.01). Interestingly, although these results identified a key role for the HOG map kinase pathway, only a limited effect at most was caused by the cap1Δ mutation.

Analysis of the susceptibility of these antioxidant mutants to H₂O₂ gave a very different pattern of results. As expected, the cap1Δ mutant showed very strong killing by H₂O₂ compared to the other strains, consistent with its role in regulating the expression of a broad range of antioxidant genes (Fig. 3B). In addition, the cat1Δ catalase mutant and sod1Δ superoxide dismutase mutants showed significantly increased susceptibility to H₂O₂ by ANOVA. In contrast, the HOG pathway mutants showed only a slight trend toward increased susceptibility to H₂O₂ that was only significant for the pbs2Δ mutant at one of the doses of H₂O₂ that were used (Fig. 3B). These results highlight the different effects HOCl and H₂O₂ have on cells and that they appear to be countered by different antioxidant mechanisms.

To better understand how C. albicans responds to HOCl, we carried out RNA-seq analysis to identify the genes that are regulated by this oxidant. For comparison, we also analyzed the effects on transcription when cells were treated with H₂O₂ or with benzoquinone, which are chemically very different from HOCl. H₂O₂ is a peroxide that can oxidize a wide range of compounds, and benzoquinone can be converted to a semiquinone that generates reactive oxygen species (34). Cells were treated for 15 min with a sublethal dose of the oxidants that was determined to be the highest concentration that caused less than 1% reduction in CFUs. Interestingly, the patterns of gene regulation in response to these oxidants were very different (Fig. 4; Table S1). HOCl only induced 173 genes above the twofold cutoff, whereas H₂O₂ induced 826 genes and benzoquinone induced 1,146 genes. Only 106 genes were induced in common, although that represents more than half of the genes induced by HOCl. Principal component analysis revealed that the patterns of gene expression were also qualitatively very different. Transcriptomes from cells treated with sublethal doses of HOCl (1 or 5 µM) clustered very close to the untreated control cells, consistent with a small number of changes in gene expression (Table S1). In contrast, principal component analysis indicated that the patterns of gene expression affected by H₂O₂ and benzoquinone were very different from HOCl and from each other.

Gene ontology (GO) term analysis showed significant similarities in gene expression profiles regulated by all three oxidants for the categories of cell redox homeostasis, regulation of biological quality, and drug transport. These are consistent with cells counteracting oxidative stress, degrading damaged proteins, and pumping out toxic molecules. The gene expression profiles from cells treated with H₂O₂ and benzoquinone also showed similar GO term profiles for five other categories relating to response to oxidative stress and proteasomal processing. A comparison of a set of well-studied antioxidant genes, including those coding for catalase, superoxide dismutase, glutathione, and thioredoxin, showed a trend that they were induced by all three types of oxidative stress but were induced more weakly by HOCl (Fig. 5). Perhaps this contributes to the observed phenotype that a cap1Δ mutant was highly susceptible to H₂O₂ but not HOCl (Fig. 3).

The experiments described above examined C. albicans cells treated with 5 µM HOCl for 15 min. In view of the lower number of genes induced by HOCl compared to H₂O₂ and benzoquinone, we compared RNA-seq profiles from cells treated with 5 µM HOCl with those treated with a higher dose of HOCl (10 µM) after a 15 min incubation and a longer 30 min incubation. The results showed that both changes significantly increased the number of up-regulated genes (Fig. 6A). The strongest change was in cells treated with 10 µM HOCl for 30 min, which resulted in 287 upregulated genes.

There were 131 genes induced in common for cells incubated with 10 µM HOCl for 15 or 30 min (Fig. 6B). This represents 65% of the genes upregulated at 15 min and 46% of genes upregulated at 30 min. Analysis of the genes induced in common by GO term analysis showed that the major categories included response to oxidative stress and regulation of biological quality, consistent with cells attempting to deal with damage from HOCl. Genes specifically upregulated at 15 min were mapped to similar GO term categories, although the specific genes were different. In contrast, genes that were specifically induced at 30 min mapped to a distinct group of eight GO terms, all of which related to amino acid synthesis. This may be a reaction to help synthesize new proteins. It is also interesting to speculate that an increase in amino acids may have the beneficial effect of providing amine groups to react with HOCl, thereby protecting cellular proteins. Although the chloramines that would be created are still toxic, they are less so than HOCl.

Inspection of the genes induced by HOCl revealed the presence of four understudied genes that are implicated in reversing damage caused by HOCl (Fig. 7A). Two genes, MXR1 (C2_00,960C; orf19.2028) and SRX1 (C2_05,060C; orf19.3537), function to reverse oxidative damage to sulfur-containing amino acids that are a major target of HOCl. MXR1 encodes methionine-S-sulfoxide reductase, and SRX1 encodes a sulfiredoxin that reduces cysteine-sulfinic acid groups. The other two genes include AYS1 (C3_02,360C; orf19.1608), which encodes an enzyme with similarity to arylsulfitases that cleave sulfate esters, and TRX1 (CR_10,350C; orf19.7611), which encodes thioredoxin, a key antioxidant in eukaryotic cells. All of these genes are broadly conserved, although AYS1 is absent from species related to Saccharomyces cerevisiae and Candida glabrata. Similar results were reported previously for the induction of C. albicans genes by H₂O₂ (35) and the MXR1, SRX1, and AYS1 genes were reported to be induced after exposure to neutrophils (36).

Testing the killing by HOCl revealed that the ays1Δ, srx1Δ, mxr1Δ, and trx1Δ mutants all displayed a trend toward increased susceptibility to HOCl, although this only reached statistical significance by ANOVA for the mxr1Δ and trx1Δ mutants (Fig. 7B). Once again, the pattern for H₂O₂ was distinct from HOCl. Only trx1Δ showed significantly increased susceptibility to both H₂O₂ and HOCl. Interestingly, the srx1Δ mutant was significantly more susceptible to H₂O₂ but not HOCl, whereas the opposite was true for the mxr1Δ mutant. The ays1Δ did not show a significant increase in susceptibility to H₂O₂, although it did show a trend toward increased killing by HOCl. Thus, although MXR1, SRX1, and TRX1 all promote resistance to oxidation, they have differential effects in resisting HOCl and H₂O₂.

The genotypes of the C. albicans strains used are described in Table 1. Cells were grown in rich YPD medium (2% yeast extract, 1% peptone, 2% dextrose, 80 mg/L uridine) or a synthetic medium containing yeast nitrogen base, 2% dextrose, amino acids, and uridine if necessary (45).

Homozygous deletion mutants lacking the AYS1, MXR1, SRX1, and TRX1 genes were constructed using transient expression of CRISPR-Cas9 in C. albicans strain SN152 (47), essentially as described previously (52, 53). Cassettes for CaCAS9 expression, single guide RNA (sgRNA) expression, and a repair template with the selectable marker were co-transformed into cells. The CaCAS9 gene was codon optimized for expression in C. albicans (54). The CaCAS9 expression cassette, which was codon optimized for expression in C. albicans, was PCR amplified from the plasmid pV1093 (kindly provided by Dr. Valmik Vyas) (54). The cassettes for the sgRNA expression were constructed by PCR using the plasmid template pV1093 and 20-bp target sequences for each gene that were defined previously by Vyas et al. (55). The sgRNA was used to target Cas9 to make a DNA double-strand break at specific target sites (Table S2) (52). Repair templates were constructed by PCR using primers with ~80 bases of homology to the sequences upstream or downstream from the target region to amplify CmLEU2 on plasmid pSN40 (47). The oligonucleotide primers are listed in Table S2. PCR was conducted with Ex Taq polymerase (TaKaRa Bio, Inc.). PCR products were purified by extraction with a phenol/chloroform/isoamyl alcohol mixture (25:24:1). DNA was introduced into cells by the lithium acetate method (56). Homozygous deletion mutants were identified by PCR amplification of genomic DNA using primers that flanked the 5′ and 3′ ends of the genes as well as internal primers. Four independent isolates for each mutant were examined to verify that they displayed the same phenotype. To generate complemented strains, the corresponding wild-type gene containing 500 bp upstream and 350 bp downstream was PCR amplified and then inserted into Sma I-cleaved pDIS3 by gap repair in S. cerevisiae strain W3031A, as described previously (57). The wild-type gene was amplified using primers containing 80 bp of homology to the ends of Sma I-digested pDIS3 to facilitate the gap repair. The resulting plasmid was digested with Sfi I to release the wild-type gene and the NAT1 selectable marker, flanked by sequences corresponding to the NEUT5L locus, and then transformed into the C. albicans deletion strain to integrate the wild-type gene at NEUT5L. The oligonucleotides used to construct the C. albicans strains are described in Table S2.

For the assessment of cell viability by CFU assay following incubation in HOCl, an overnight dilution series was set up in minimal medium and incubated at 30°C with rotation. The next day, log-phase cells were washed twice and resuspended in sterile deionized H₂O. Cells were diluted to a final concentration of 1 × 10⁶ cells/ml in a reaction volume of 1 ml containing 1 mM sodium phosphate pH 7.4 buffer and 0, 2.5, 10, or 20 μM HOCl. Following incubation for 15, 30, or 60 min at 30°C with rotation, 30 μL of reaction was added to 10 ml of sterile deionized H₂O, and then 100 μL was spread onto YPD agar plates. After incubation for 48 h at 30°C, colony-forming units were counted.

SYTOX Green (Invitrogen, Molecular Probes, Eugene, OR, USA) is a membrane-impermeable nucleic acid stain that can be used to assay plasma membrane integrity (25). For the analysis, cells were grown in synthetic medium overnight at 30°C to log phase, washed in sterile H₂O, and diluted to 1 × 10⁶ cells/ml. Following incubation in 20 µM HOCl at 30°C for the indicated time, 500 μM methionine was added to a final concentration of 45 μM to quench the reaction. The cells were washed in sterile H₂O, SYTOX Green was added to a final concentration of 2.5 nM, and the cells were incubated at room temperature for 5 min. Cells were washed again in sterile H₂O and then analyzed by fluorescence microscopy. Images were obtained using an Olympus BH2 microscope equipped with a Zeiss AxioCam digital camera. The percent of stained cells was determined by counting 50–200 cells in three independent experiments.

Halo assays used to quantify the sensitivity of C. albicans cells to HOCl and other oxidants were carried out with strains that were grown overnight in YPD medium at 30°C with rotation. The cells were harvested by centrifugation, resuspended in sterile H₂O at a density of 1.0 × 10⁶ cells/ml, and then 250 μL was spread onto the surface of a synthetic medium agar plate. After allowing the cell mixture to dry on the plate, 5 µL of the indicated concentration of HOCl or H₂O₂ was spotted directly onto the agar surface. The plates were incubated at 30°C for 48 h, and then the diameters of the zones of growth inhibition (halos) were measured and the plates were photographed. Paper discs, often employed in this type of disc diffusion halo assay for the application of a chemical onto the surface of an agar plate, were not used since we found that paper discs had differential effects on HOCl that altered the uniformity of the zones of growth inhibition. The assays were carried out in duplicate on at least three independent days. The average change in the zone of growth inhibition was then assessed for statistical significance by ANOVA using GraphPad Prism. The comparison between the wild-type and mutant strains was assessed using a Dunnett test.

C. albicans cells were freshly grown on YPD medium and then were grown in minimal BYNB medium with dextrose. A liquid culture was grown at 30°C overnight to saturation, diluted, and then kept in log phase growth overnight at 30°C. The cultures were then adjusted to 0.1 × 10⁷ cells/ml and were grown until they reached 1 × 10⁷ cells/ml. Ten-milliliter aliquots of cells were then incubated in the presence or absence of the indicated concentration of H₂O₂, HOCl, or benzoquinone at 30°C for the indicated time. Cells were then quick chilled and washed with ice-cold water, and pellets were quick-frozen with liquid nitrogen and reserved for subsequent analysis.

The extraction of RNA, preparation of cDNA, and sequencing reactions were conducted at GENEWIZ, LLC. (South Plainfield, NJ, USA). The RNA was extracted from a frozen cell pellet of 10⁸ cells using an RNeasy Plus Universal mini kit following the manufacturer’s instructions (Qiagen, Germantown, MD, USA). The RNA samples were quantified using a Qubit 2.0 fluorometer (Life Technologies), and RNA integrity was assessed using an Agilent TapeStation 4200 (Agilent Technologies, Santa Clara, CA, USA). RNA samples were prepared for sequencing using the NEBNext Ultra II RNA Library Prep Kit for Illumina following the instructions of the manufacturer (NEB, Ipswich, MA, USA). Briefly, samples were first enriched for mRNA using Oligo(dT) beads. The mRNA samples were then fragmented for 15 min at 94°C and then used as a template for cDNA synthesis. The ends of the cDNA fragments were repaired and then adenylated at 3′ ends. Universal adapters were ligated to the cDNA, followed by index addition and library enrichment by limited-cycle PCR. The sequencing libraries were then validated using an Agilent TapeStation (Agilent Technologies). They were then quantified with a Qubit 2.0 fluorometer (Invitrogen) and by quantitative PCR (KAPA Biosystems). The sequencing libraries were pooled, clustered on one lane of a flowcell, and then loaded on an Illumina HiSeq instrument (4,000 or equivalent) according to the manufacturer’s instructions and sequenced using a 2 × 150 bp paired end configuration. Image analysis and base calling of the data were conducted using HiSeq Control Software. The raw sequence data .bcl files generated from the Illumina HiSeq were converted into fastq files and de-multiplexed using Illumina’s bcl2fastq 2.17 software. One mismatch was allowed for index sequence identification.

Sequence data were then subjected to quality profiling, adapter trimming, read filtering, and base correction for raw data using fastp, an all-in-one FASTQ preprocessor (58). The high-quality paired-end reads were mapped to the C. albicans SC5314 genome (Candida Genome Database; Assembly 22) using HISAT2 (59). The read alignments obtained in the previous step were assembled with StringTie (60) and used to estimate transcript abundances. The absolute mRNA abundance of the samples was expressed as fragments per kilobase of transcript per million mapped reads. Analysis of differential gene expression was conducted using the DESeq2 (60) package (61) from Bioconductor (62) on R. GO term analysis was carried out at the ShinyGO web site (http://bioinformatics.sdstate.edu/go) (63).