Intrinsically High Resistance of Candida glabrata to Hydrogen Peroxide and Its Reversal in a Fluconazole-Resistant Mutant

LETTER

Candida glabrata is the second most common cause, after Candida albicans, of mucosal and invasive fungal infection in the United States, Europe, and elsewhere (1–4). Other than its high capacity for acquired resistance to widely used azole antifungals (e.g., fluconazole [FLC]), the factors that contribute to C. glabrata infection remain incompletely understood (5–7). An important arm of innate immunity involves reactive oxygen species hydrogen peroxide (H₂O₂), produced by both phagocytes and the diverse bacteria that colonize the oral and vaginal mucosa (8–12). We report here that C. glabrata exhibits intrinsically high H₂O₂ resistance relative to C. albicans, correlated with higher whole-cell catalase activity. Intriguingly, in the previously characterized FLC-resistant mutant F15 with the transcription factor Pdr1 gain-of-function mutation P927L (13), catalase gene expression decreased (14) and correspondingly, H₂O₂ susceptibility increased. We describe how this phenotype was exploited to select for H₂O₂-resistant revertants of F15, which correspondingly reverted to FLC susceptibility associated with the Pdr1 loss-of-function mutation. This genetic approach should prove useful for further studies of Pdr1 structure/function and its relationship to H₂O₂ resistance and virulence in C. glabrata.

Log phase cultures of C. glabrata (18 strains) and C. albicans (8 strains) were tested for H₂O₂ susceptibility by broth microdilution assays in yeast extract-peptone-dextrose (YPD) medium at 35°C for 24 h as described (13). A median H₂O₂ MIC of 32 mM was observed for C. glabrata, compared to 4 mM for C. albicans (Table 1 and data not shown). To assess survival following brief exposure, cultures (8 strains for each species) were incubated for 1 h with 64 mM H₂O₂, diluted, and plated onto YPD agar; the colonies were counted after 72 h. Relative to the H₂O₂-free control, C. glabrata exhibited median 74 to 77% survival, compared to ≤1% for C. albicans (Table 1).

TABLE 1.

C. glabrata demonstrates high H₂O₂ resistance relative to C. albicans, correlated with higher whole-cell catalase activity

Species	Strain	H2O2 MIC (mM)	Survival (%)a	Catalase activityb
C. glabrata	C20	32	80	81
	38326	32	71	81
	417	32	100	81
	444	32	83	81
	448	32	74	81
	038	16	17	27
	2001	16	74	27
	376	8	77	27
C. albicans	90028	4	≤1	27
	2-76	4	≤1	9
	LL	4	≤1	9
	I-59	4	≤1	9
	66027	4	≤1	9
	12-99	2	≤1	9
	11006	1	≤1	9
	SC5314	4	≤1	3

^aCultures were incubated for 1 h with 64 mM H₂O₂, diluted, and plated onto YPD agar; the colonies were counted after 72 h.

^bHighest fold-dilution of the cell suspension continuing to yield visible O₂ bubbling 5 min after the addition of 30% H₂O₂.

To assess the whole-cell catalase activity, cell suspensions (3 × 10⁹/mL) were diluted 3-, 9-, 27-, and 81-fold in H₂O, combined with an equal volume of 30% H₂O₂, and visually monitored for active O₂ bubbling at 5 min. The median dilution yielding detectable bubbling was 81-fold for C. glabrata versus 9-fold for C. albicans (Table 1). Analogous results were obtained for subgroups of C. glabrata and C. albicans strains using a bioassay method in which the supernatants of cultures incubated for 1 h with 8 mM H₂O₂ were assayed for inhibitory activity versus an H₂O₂-hypersensitive Saccharomyces cerevisiae yap1Δ strain (data not shown).

C. glabrata mutant F15 was derived from parent strain 66032 (FLC MICs, ≥128 and 16 μg/mL, respectively) by a one-step selection for FLC resistance in our lab (13). The gain-of-function pdr1-P927L mutation in F15 results in upregulated expression of multidrug transporter genes, including CDR1. Intriguingly, microarray analysis of this mutant revealed a 2.5-fold decrease in expression of the catalase gene CTA1 (14). Consistent with this transcriptional data, F15 demonstrated a 2-fold reduction in H₂O₂ MIC (from 32 to 16 mM) in broth microdilution assays.

To confirm the role of the pdr1-P927L mutation in the diminished H₂O₂ resistance of F15, the mutated allele was amplified and used to transform strain 20089 Δpdr1 using previously described methods (14), with selection on FLC-containing plates. Following characterization, this 200989pdr1-P927L construct was tested for H₂O₂ susceptibility, which again demonstrated a 2-fold reduction (from 32 to 16 mM) in MIC relative to its parent strain. Thus, the reduction in H₂O₂ resistance associated with the pdr1-P927L mutation is not strain specific. However, it appears to be mutation specific. First, the fluconazole-resistant mutant F26 isolated in parallel with F15 (13) exhibits the gain-of-function mutation pdr1-R889G but unchanged H₂O₂ resistance (MIC, 32 mM). Second, we tested 19 FLC-resistant C. glabrata clinical isolates with diverse Pdr1 mutations characterized by Ferrari et al. (15), and 17 exhibited unaltered H₂O₂ susceptibility relative to their FLC-susceptible parents. The exceptions were the DSY753/DSY754 and DSY2253/DSY2254 susceptible/resistant strain pairs, in which H₂O₂ resistance increased (MIC, 4/21 mM) and decreased (MIC, 32/21 mM), respectively; however, this result was not reproduced in the laboratory constructs SFY110/SFY111 and SFY102/SFY103, carrying the same Pdr1 alleles (for all four: MIC, 21 mM).

To further characterize the link between H₂O₂ and FLC susceptibilities associated with the pdr1-P927L mutation, F15 (H₂O₂ MIC, 16 mM; FLC MIC, >128 μg/mL) was subjected to a one-step selection for revertants to H₂O₂ resistance (3 × 10⁶ cells plated on 28 mM H₂O₂-containing YPD and incubated for 72 h at 35°C). Of 12 H₂O₂-resistant revertants (MIC, 32 mM), 8 similarly reverted to FLC susceptibility (MIC, 2 to 16 μg/mL). RNA analysis of a representative revertant demonstrated decreased CDR1 expression (data not shown). Consistent with this, in 2 of the 8 revertants, amplification and sequencing (14) of the Pdr1 gene revealed, in addition to the P927L mutation of F15, the putative loss-of function mutations R875stop (AGA=R to TGA=stop, truncating the C terminus by 232 residues; FLC MIC, 2 μg/mL) and stop1107L (TGA=stop to TTA=L, extending the C terminus by 3 residues; FLC MIC, 4 μg/mL).

In conclusion, while the high capacity of C. glabrata for acquired resistance to widely used azole antifungals, mediated by the Pdr1 gain-of-function mutation, may be its primary “virulence factor,” the in vitro data presented here indicate that H₂O₂ resistance also warrants consideration as a contributor to C. glabrata virulence. The data are consistent with but significantly extend a previous report that was limited to 2 clinical isolates each of C. glabrata and C. albicans and employed nonstandard H₂O₂ susceptibility assays (16). Our additional data demonstrating linkage in the FLC-resistant strain F15 between the Pdr1 mutation P927L and reduced H₂O₂ resistance are particularly intriguing. This linkage was not strain specific but is apparently Pdr1-mutation specific. Relatedly, Ferrari et al. (17) previously reported that distinct Pdr1 gain-of-function mutations yield distinct phenotypes. Finally, we note that while Pdr1 gain-of-function mutants are readily isolated both in the laboratory and clinically, the isolation of loss-of-function mutants is typically more challenging, e.g., requiring mutagenesis, labor-intensive screening, and genetic dissection (18). In contrast, as shown here, C. glabrata pdr1-P927L strains provide a convenient genetic tool for isolating loss-of-function mutants, facilitating further exploration of Pdr1 structure/function and its linkage to H₂O₂ resistance.