Acquired amphotericin B resistance attributed to a mutated ERG3 in Candidozyma auris

ABSTRACT

First identified in 2009, Candidozyma auris (formerly Candida auris) is an emerging multidrug-resistant fungus that can cause invasive infections with a crude mortality rate ranging from 30 to 60%. Currently, 30–50% of C. auris isolates are intrinsically resistant to amphotericin B. In this study, we characterized a clinical case of acquired amphotericin B resistance using whole-genome sequencing, a large-scale phenotypic screen, comprehensive sterol profiling, and genotypic reversion using CRISPR. Data obtained in this study provide evidence that a deletion resulting in a frameshift in ERG3 significantly contributes to the observed resistant phenotype, and a nonsense mutation in ERG4 may more modestly contribute to resistance. Characterization of this isolate also revealed that a fitness cost is associated with the abrogation of ergosterol production and its replacement with other late-stage sterols. This article presents a clinical case description of amphotericin B resistance from a frameshift mutation in ERG3 in C. auris and marks an advancement in the understanding of antifungal resistance in this fungal pathogen.

INTRODUCTION

In 2019, the Centers for Disease Control and Prevention (CDC) designated Candidozyma auris (previously known as Candida auris) as an urgent antimicrobial threat, making it the first fungal pathogen to be raised to this level of concern (1, 2). This is attributed to the multidrug-resistant characteristics of C. auris, its capacity to spread easily in health care facilities, and its potential to cause invasive candidiasis, especially among patients with weakened immune systems (3–6). Out of the three primary antifungal classes approved for the treatment of Candida infections, approximately 80–93% of C. auris isolates are resistant to fluconazole (triazole drug class), 35–50% display resistance to amphotericin B (polyene drug class), and 5–7% show resistance to the echinocandin antifungals (7–9).

Amphotericin B is frequently utilized due to its broad spectrum of activity and success in treating systemic fungal infections (10). First isolated from Streptomyces nodosus in 1953, amphotericin B disrupts fungal membrane permeability via ergosterol binding and has been observed to additionally enact oxidative damage (11, 12).

At the time of manuscript preparation, there were no clinical amphotericin B susceptibility breakpoints set forth by the Clinical Laboratory Standards Institute (CLSI) for C. auris (13), largely due to limited clinical outcome data and variability in in vitro minimum inhibitory concentration (MIC) determinations across testing methodologies. However, the CDC has established a tentative amphotericin B MIC breakpoint for C. auris at 2 mg/L. Though up to a third of C. auris isolates from the United States have MICs of 1 mg/L, caution is urged for C. auris amphotericin B MIC interpretation (6, 13–15).

Even though amphotericin B has been used in clinical practice since the 1950s, very few instances of drug resistance have been documented. Previously, acquired cases of amphotericin B resistance in Candida spp. have been shown to be associated with mutations in the genes encoding the ergosterol biosynthesis pathway (8). In C. auris, the only established mechanism of clinically acquired amphotericin B resistance is attributed to an indel in ERG6 (16). Additionally, a nonsense mutation in ERG3 has once been associated with amphotericin B resistance in a single clinical isolate but has yet to be tested/confirmed (17). Other mechanisms of amphotericin B resistance have been identified in vitro. Amphotericin B resistance can be induced by culturing C. auris in the presence of sub-lethal doses of amphotericin B (18). In other Candida spp., mutations in ERG2, ERG3, ERG4, ERG5, ERG6, and ERG11 have been found to be associated with reduced susceptibility to amphotericin B (8, 10, 19–21).

Here, we present a case of acquired clinical amphotericin B resistance, which is attributed to mutations in the C. auris ERG3 and ERG4 genes, causing a premature stop codon in the C-5 sterol desaturase and delta (24 (24 (1)))-sterol reductase, respectively. We further investigated this isolate through a phenotypic screen, Cas9-mediated genetic repair, and sterol profiling. Our results support a causal link between the ERG3 mutation and much of the amphotericin B resistance phenotype.

MATERIALS AND METHODS

Collection, Culturing, and Confirmation of Specimens

The first isolate (LNV001) from a patient was collected from the bronchial lavage in July 2022, while the second isolate (LNV002) from the same patient was collected from the urine in September 2022. Both samples were initially grown in Salt Sabouraud Dulcitol broth at 38°C at 250 rpm for 24–48 h. From culture, a 10 µL loop was used to streak culture on CHROMagar Candida (CHROMagar, Paris, France), where it was grown for 24–48 h at 36°C. Plates displayed a pinkish purplish growth and were confirmed as C. auris by matrix-assisted laser desorption ionization–time-of-flight mass spectrometry with reference library MALDI Biotyper CA library (version 2022) (Bruker, Billerica, MA).

DNA extraction

Clinical isolates underwent bead-beating (FastPrep-24, MP Biomedicals, Irvine, CA) for four cycles at 6.0 m/s for 30 s with 5-min pauses in between. Then, genomic DNA (gDNA) from isolates was isolated using the PureFood Pathogen Kit on the Maxwell RSC (Promega, Madison, WI) per manufacturer’s protocol.

Library prep and whole-genome sequencing

Extracted gDNA was library-prepped using DNA Prep Kit (Illumina, San Diego, CA) per manufacturer’s recommended protocol utilizing a STARlet automated liquid handler (Hamilton Company, Reno, NV). Paired-end sequencing (2 × 151 bp) was performed using NovaSeq 6000 (Illumina, San Diego, CA) with a read depth of 94× and 123× and a genome length of 12,288,577 and 12,267,689 for isolates LNV001 (SRR23958537) and LNV002 (SRR23109153), respectively.

Bioinformatic analysis

The open-source software TheiaEuk was used to perform the de novo assembly, quality assessment, and genomic characterization of fungal genomes (22). Using the generated FASTA files, species taxon identification and clade typing were performed by Genomic Approximation Method for Bacterial Identification and Tracking (GAMBIT) (23). kSNP3 identified core genome single nucleotide polymorphisms (SNPs) between the two isolates, with confirmation and gene identification performed by Snippy to the C. auris clade III reference strain B11221 (GCA_0022775015.1) (24, 25).

Phenotypic screens

Growth curve

C. auris isolates were grown on CHROMagar plates at 36°C for 48 h, and then, a single colony was isolated and regrown for purity for another 48 h at 36°C. The high-throughput phenotypic screening was performed utilizing Biolog PM1 and PM2a phenotypic microarray plates (Biolog, Inc., Hayward, CA) per manufacturer’s protocol for Saccharomyces cerevisiae. Absorbance was taken every 6–8 h for 72 h. Graphs were designed in R studio using the growthcurver package (26). Significance was determined by a student’s t-test of growth rate with the Bonferroni method to adjust for a large data set.

Dilution Spot

Single C. auris colonies from respective isolates were grown on CHROMagar plates for mass growth at 36°C for 48 h. Both isolates were standardized by absorbance in RPMI broth. A ten-fold dilution series was performed to gain a final dilution of 10⁻⁴. Dilution spots were plated on RPMI agar and checked every 24 h for 72 h. In Fig. 1, the lines in the background of each image are the overhead lights.

Fig 1.

Dilution spot assay of amphotericin B-susceptible and resistant C. auris isolates. A ten-fold dilution series of LNV001 and LNV002 were plated on RPMI agar supplemented with varying concentrations of amphotericin B, and reading was observed every 24 h.

Sterol analysis

Overnight cultures of C. auris strains were used to inoculate 20 mL MOPS-buffered RPMI to 0.1 OD₅₉₀. Cultures were grown for 18 h at 35°C, 170 rpm. Cells were then harvested and split into two samples to enable the determination of dry weight and extraction of sterols. An internal standard of 10 µg of cholestanol was added to each sample, and sterols were extracted as previously described (27). Briefly, lipids were saponified using alcoholic KOH, and non-saponifiable lipids were extracted with hexane. Dried samples were derivatized by the addition of 0.1 mL BSTFA TMCS (99:1, Sigma) and 0.3 mL anhydrous pyridine (Sigma) and heating at 80°C for 2 h. TMS-derivatized sterols were analyzed and identified using gas chromatography–mass spectrometry (GC/MS) (Thermo 1300 GC coupled to a Thermo ISQ mass spectrometer, Thermo Scientific) and Xcalibur software (Thermo Scientific). The retention times and fragmentation spectra for known standards were used to identify sterols. Sterol composition was calculated from peak areas, as a mean of three replicates, and the relative quantity of sterols present was determined using a standard curve of the internal standard (cholestanol) and lanosterol and the dry weight of the samples.

Episomal Plasmid-Induced Cas9 (EPIC)-mediated ERG3 and ERG4 modification in C. auris isolate LNV002

EPIC components

EPIC-mediated transformation of C. auris was performed as previously described (28). Briefly, guide sequence primers targeting ERG3 and ERG4 were ligated into pJMR19 following LguI (Thermo Scientific) digestion as previously described. Transformation repair templates were amplified from gBlock sequences (Integrated DNA Technologies) by PCR using Phusion Green Master Mix per manufacturer’s instructions (Thermo Scientific, Waltham, MA, USA), followed by subsequent QIAquick PCR purification (Qiagen). All strains, primers, and templates are listed in Table S1.

C. auris transformation

C. auris was cultured overnight to an OD₆₀₀ of 1.8–2.2, and transformation reactions were assembled with salmon sperm DNA (Invitrogen), pJMR19, repair template DNA, and TE-LiAC + 55% PEG. EPIC-positive transformants were selected by growth on nourseothricin (200 mg/L)-supplemented YPD agar plates (28). Single colonies were patched onto YPD for screening. DNA isolation was achieved via treatment with a DNA extraction buffer consisting of 10 mM Tris pH 8.0, 2 mM EDTA, 0.2% Triton X-100, and 200 µg/mL Proteinase K. ERG3 and ERG4 PCR amplification and subsequent Sanger sequencing (Hartwell Center, St. Jude Children’s Research Hospital) were used to screen positive transformants. Sequenced transformants were then replica plated for plasmid ejection as previously described, and proper ERG3 and ERG4 sequence retention was confirmed with a second Sanger sequencing run (16, 29).

Antifungal susceptibility testing

MICs for amphotericin B (Sigma-Aldrich) were determined by broth microdilution (BMD) in accordance with the M27-A4 from the CLSI, with considerations recommended by the CDC (15, 30). Amphotericin B MIC by Etest (bioMérieux, Marcy-l'Étoile, France) was determined per manufacturer’s instructions with CDC recommendations implemented. All susceptibility testing was performed in biological triplicate and determined visually for growth inhibition at 24 h.

RESULTS

Patient case description

Consecutive C. auris isolates were collected over a 2-month span in the summer of 2022 from a 76-year-old female. These two isolates were collected from the bronchial lavage and urine samples of the patient, respectively (Table 1). The initial isolate, LNV001, presented with an MIC value of 0.19 mg/L, yielding susceptibility to amphotericin B, while the latter isolate, LNV002, presented with a >32 mg/L MIC value, yielding high resistance to amphotericin B (Fig. 1 and 2). MIC values were determined by the Etest.

TABLE 1.

Case study of acquired amphotericin B resistance in C. auris^a

Isolate ID	Collection date	Isolation site	Amphotericin B MIC (mg/L)	Number of coding mutations
LNV001	July 2022	Bronchial lavage	0.19	9
LNV002	September 2022	Urine	>32

^a C. auris isolates were collected over a 2-month span. Whole-genome sequencing was performed on both isolates, revealing nine sequence differences that result in coding changes between the genomes. MIC values determined by Etest.

Fig 2.

Amphotericin B MICs following ERG3 and ERG4 reversion to wild type. (A) Representative images of amphotericin B MICs at 24 h as determined by Etest (bioMérieux). (B) MICs determined by BMD in accordance with CLSI susceptibility testing. MICs were read visually for 100% growth inhibition at 24 h. Bars represent the modal MIC with points plotted for three biological replicates for each isolate and independent strain with MIC values for two independently derived LNV002_ERG3^WT and LNV002_ERG3^WT, ERG4^WT strains shown.

Whole-genome sequencing

The two isolates underwent paired-end sequencing on the NovaSeq 6000. Sequences were de novo assembled and quality checked by the bioinformatic workflow TheiaEuk (22). Confirmation of the respective isolates as C. auris was performed using GAMBIT, which further described each isolate as a member of Clade III (23). The sequences of the two isolates were genetically discordant from each other by nine sequence differences in coding regions determined by the bioinformatic workflow Snippy (Table 2) (25). Both isolates were resistant to fluconazole with an MIC value of >256 mg/L, and each was found to possess a mutation encoding the VF125AL amino acid substitution in sterol 14α-demethylase encoded by ERG11, which has been previously established to confer clinical fluconazole resistance and has exclusively been observed in Clade III isolates (7, 8). When these two isolates were compared with the Clade III reference isolate, LNV001 and LNV002 were 35 and 41 SNPs different in the core genome. Between the amphotericin B-susceptible and resistant isolates, the mutations of greatest interest to public health are the 10 bp deletion causing a frameshift at amino acid 71, leading to a premature stop codon at amino acid 132 in the C-5 sterol desaturase gene, ERG3, and the nonsense mutation causing a premature stop codon at amino acid 106 in the delta (24 (24 (1)))-sterol reductase gene, ERG4. Since amphotericin B is hypothesized to exert antifungal activity through direct interaction with ergosterol, we focused on ERG3 and ERG4 of the ergosterol biosynthesis pathway. When comparing the ERG3 allele to a group of 44 isolates across Clades I–VI, with 25 of the isolates presenting with MIC values at or slightly above the amphotericin B breakpoint, we found that none of these isolates had an ERG3 allele different from the respective clade reference strain (31).

TABLE 2.

SNP differences between LNV001 and LNV002 in coding regions^a

Protein (gene)	Predicted ortholog	Gene locus	Mutation	Mutation type
C-5 sterol desaturase (ERG3)	C-5 sterol desaturase (ERG3)	CJI97_003811	G71fs - > P132*	Frameshift/stop
Delta (24 (24(1)))-sterol reductase (ERG4)	Delta (24 (24(1)))-sterol reductase (ERG4)	CJI97_002908	L106*	Stop
1-phosphatidylinositol 4-kinase	1-phosphatidylinositol 4-kinase	CJI97_001150	N126D	Missense
Retromer subunit VPS35	Retromer subunit VPS35	CJI97_003028	I175V	Missense
Uncharacterized protein	Major Facilitator Superfamily	CJI97_000798	F232S	Missense
Uncharacterized protein	Mediator of RNA polymerase II transcription subunit 5	CJI97_002146	F275fs	Frameshift
Uncharacterized protein	Selenoprotein O	CJI97_002594	R243fs	Frameshift
Uncharacterized protein	STE3	CJI97_005168	A178V	Missense
Uncharacterized protein	Transcriptional regulator ADR1	CJI97_002218	E41Q	Missense

^a Nine coding regions have been identified to possess mutations from the former susceptible isolate. ERG3 and ERG4, which are hypothesized to be responsible for the resistant phenotype, are bolded. Gene loci are from the clade III reference strain B11221. The “*” symbol designates a stop codon. SNP differences selected for further evaluation in this work are shown in shaded rows.

EPIC-mediated reversion of ERG3 and ERG4 to wild type

To identify the influence of each mutated gene on the amphotericin B resistance, we utilized the EPIC genetic manipulation system to revert the Erg3 and Erg4 sequences to wild type (matching LNV001 and the B11221 reference) in LNV002 (28). While we were able to generate two independent ERG3^WT single reversions in LNV002, we were unable to generate ERG4^WT single reversions in LNV002 after multiple transformations. However, once ERG3 had been reverted to wild type, we were able to generate two independent ERG3^WT and ERG4^WT double correction strains. When testing amphotericin B susceptibility using the diffusion test strip (Etest, bioMérieux) method, as recommended by the CDC, the ERG3^WT reversion in LNV002 resulted in dramatically increased amphotericin B susceptibility with an MIC shift from >32 mg/L to 0.38 mg/L (Fig. 2A). However, the additional reversion of ERG4 to the wild-type allele (in strain LNV002_ERG3^WT, ERG4^WT) did not further increase amphotericin B susceptibility. By comparison, when testing amphotericin B MIC using the BMD method, reversion of the ERG3 gene to the wild-type sequence resulted in a more modest decrease in amphotericin B MIC, and correction of both ERG3 and ERG4 to the wild-type allele resulted in a further one-dilution decrease in amphotericin B MIC (Fig. 2B).

Absence of Ergosterol

We sought to determine if LNV002 produced ergosterol via sterol profiling because of the predicted interaction between amphotericin B and ergosterol. TMS-derivatized sterols in biological triplicate were analyzed and identified using GC/MS and Xcalibur software. The sterol profile of the susceptible LNV001 isolate largely consists of ergosterol and lanosterol (Table 3; Fig. S1). The sterol profile of the resistant isolate (LNV002) is mainly composed of ergosta-7,22,24 (28)-trienol, episterol, and lanosterol. Ergosterol was not detectable in the resistant isolate. LNV002_ERG3^WT (LNV002 ERG3c 39B; CRISPR repaired to ERG3^WT) has a sterol profile that largely consisted of ergosta-5,7,22,24 (28)-tetraenol, lanosterol, and 14-methyl fecosterol, with a notable absence of detectable ergosterol. However, LNV002_ERG3^WT, ERG4^WT (LNV002 ERG3c 39B ERG4c 7A; CRISPR repaired to ERG3^WT and ERG4^WT) is mainly composed of ergosterol and lanosterol.

TABLE 3.

Sterol composition of clinical and CRISPR-repaired C. auris isolates^a

	Percentage of total sterol
	LNV001(S)		LNV002(R)		LNV002 ERG3c 39B		LNV002 ERG3c 39B ERG4c 7A
Sterol	Mean	±st.dev.	Mean	±st.dev.	Mean	±st.dev.	Mean	±st.dev.
Ergosta-5,8,22,24(28)-tetraenol	0.3	0.1					0.8	0.6
Ergosta-5,8,22-trienol	0.1	0.0			0.6	0.2	0.4	0.1
Ergosta-8,22,24(28)-trienol			1.5	0.0
Zymosterol	0.3	0.0					0.3	0.2
Ergosterol	45.2	6.4					62.8	2.8
Ergosta-7,22-dienol	0.6	0.1					0.6	0.1
Unidentified			1.1	0.1
Ergosta-5,7,22,24(28)-tetraenol					49.2	3.1	1.6	0.2
4,14-Dimethyl zymosterol	0.3	0.1	2.8	1.0			1.1	0.1
Fecosterol (ergosta-8,24(28)-dienol)			0.4	0.3
Ergosta-7,22,24(28)-trienol			53.7	8.6
14-Methyl fecosterol	2.0	0.5	3.2	1.0	11.1	0.5	4.9	0.4
Ergosta-5,7-dienol	5.7	7.6					3.8	0.4
Unidentified			0.6	0.1
Episterol (ergosta-7,24(28)-dienol)			17.3	4.8	0.5	0.1	0.4	0.1
Ergosta-7-enol	0.3	0.2					0.3	0.1
4,4-Dimethyl-ergosta 8,14,24(28)-trienol	0.2	0.0	0.1	0.0
14-Methyl ergosta-8,24(28)-dien-3-6-diol	0.2	0.0			1.0	0.1	1.1	0.1
Lanosterol	39.7	1.3	17.3	2.1	34.2	3.5	19.9	2.0
4-Methyl ergosta-8,24(28)-dienol	0.7	0.5
4,4-Dimethyl cholesta-8,24-dienol	1.4	0.7	0.8	0.2
4,4-Dimethyl cholesta-8,14,24-trienol			0.3	0.1
Eburicol	3.0	0.5	0.9	0.3	3.4	0.3	1.9	0.3
Total	100.0		100.0		100.0		100.0

^a TMS-derivatized sterols were analyzed and identified using GC/MS and XCALIBUR software. Sterols that mainly make up each isolate are bolded. Mean values are shown in shaded columns. Sterol compositions are out of 100%. Dry weight content of each strain is directly below the percentage of sterol table.

Fitness cost associated with amphotericin B non-susceptibility

Amphotericin B resistance has been observed rarely for most Candida species (32). Evidence suggests that this is due to severe fitness defects caused by mutations in the ergosterol biosynthesis pathway (33). Since LNV002 had mutations in the ergosterol biosynthesis pathway, we hypothesized that there could be a fitness cost associated with ERG3 and ERG4 mutations in C. auris. To test this, we performed growth assays over 72 h on 190 different carbon sources. We determined fitness differences between LNV001 and LNV002, as described in the materials and methods (Table S2). Fourteen carbon sources were determined to be statistically significant for fitness defects in isolate LNV002. Carbon sources β-D-allose, phenylethylamine, xylitol, α-keto-glutaric acid, glycyl-L-glutamic acid, and glycyl-L-aspartic acid are associated with the highest fitness defects (Fig. 3). In general, this experiment concluded that there is a fitness cost associated with the nine coding mutations present in the resistant isolate when utilizing certain carbon sources for energy.

Fig 3.

Significant growth of amphotericin B-susceptible isolate compared with resistant isolate Biolog Phenotypic plates PM1 and PM2a were inoculated with LNV001 and LNV002. Optical density was measured every 6–8 h for 72 h. The blue line represents LNV001, and the red line represents LNV002. Standard deviation is denoted by the blurring surrounding the lines. Significance was determined by a student’s t-test with the Bonferroni correction applied (α = 0.000263).

DISCUSSION

Amphotericin B is critical to managing this public health threat

Over 90% of C. auris clinical isolates demonstrate resistance to fluconazole, increasing our dependence on echinocandins and amphotericin B as the recommended frontline therapy for the treatment of invasive candidiasis and infections caused by C. auris specifically (15, 34). In 2021, the CDC noted that the number of echinocandin-resistant C. auris infections has tripled since 2020, leaving amphotericin B as a last resort to potentially treat these cases (3). Even though it has been observed that approximately 30% of C. auris isolates are resistant to amphotericin B, this agent remains a crucial drug in the antifungal armamentarium due to its broad-spectrum antifungal activity (10). C. auris is an emerging public health threat, and understanding its genetic basis of resistance is crucial for effective infection prevention and control.

Interpretation of CRISPR results

From a clinical case, we isolated a pair of patient isolates, one susceptible and one resistant to amphotericin B collected within 2 months of each other. Between both isolates, a narrow list of non-synonymous mutations was revealed. Due to the involvement of ergosterol in the mechanism of action of amphotericin B, we focused our efforts on mutations present in genes, which are predicted to play a role in the ergosterol biosynthesis pathway, ERG3 and ERG4.

Correction of the ERG3 frameshift mutation (ERG3^G71fs..P132*) in LNV002 restored clinical amphotericin B susceptibility, thus demonstrating a documented case of acquired amphotericin B resistance resulting from a deleterious ERG3 mutation in a clinical C. auris isolate. The restored LNV002_ERG3^WT strain was found to possess an MIC value below the current determination of clinical resistance set by the CDC of 2 mg/L. When ERG3 and ERG4 were both restored to the wild-type sequence in LNV002, the MICs remained relatively unchanged from that of the individual LNV002_ERG3^WT strain, leaving only a small susceptibility difference persisting between LNV002_ERG3^WT, ERG4^WT and that of LNV001. It remains possible that one of the other identified mutations observed in LNV002 may modestly contribute to amphotericin B resistance and account for the 0.5–1-dilution difference in MIC not attributable to mutations in ERG3 or ERG4. One mutation of note that we speculate as a potential gene modifier candidate is gene locus CJI97_002218, an ortholog to ADR1, a transcriptional regulator in C. albicans. It has been found that ADR1 has undergone transcriptional rewiring from S. cerevisiae, where now it directs the regulation of the ergosterol biosynthesis pathway in C. albicans (35). It is plausible that ADR1 in C. auris could also contribute to the regulation of the ergosterol biosynthesis pathway, making this missense mutation a primary candidate for the difference in MIC between the LNV002_ERG3^WT, ERG4^WT corrected strains and the susceptible isolate LNV001.

Additionally, it is notable that amphotericin B MIC values for the isolates and strains in this study varied by the susceptibility testing method used, particularly in the amphotericin B-resistant isolate LNV002 (>32 vs 2 mg/L). Similar variation in C. auris amphotericin B MIC values by testing method has previously been reported, with higher MIC values obtained when using diffusion test strips (36). Microbroth dilution determines MIC by complete inhibition of visible growth in RPMI broth, while Etest diffusion strips determine MIC by an ellipse on solid RPMI media and where visible fungal colonies contact the testing strip. The Etest methodology can be difficult to read and may be discerned subjectively. For example, while the LNV002 growth shown in Fig. 2 is small, actual inspection reveals that there are colonies contacting the testing strip up the gradient length above the 32 mark, and thus, the MIC is reported as >32 mg/L. Specifically for amphotericin B, the dynamic testing range for microbroth dilutions has been observed to be highly condensed, which contributes to challenges in determining clinical resistance in C. auris, where a large proportion of amphotericin B MICs are within one dilution of the CDC tentative clinical breakpoint. Nevertheless, both microbroth dilution and Etest diffusion strip-derived amphotericin B MIC determine LNV002 to exhibit clinical amphotericin B resistance and further demonstrated a consistent restoration of amphotericin B susceptibility upon ERG3 mutation correction.

It is also notable that within the boundaries of our methods, the generation of an LNV002_ERG4^WT-independent strain was not found to be possible. Therefore, we hypothesize that the ERG3 mutation may have occurred prior to the ERG4 mutation, with the subsequent deleterious ERG4 mutation potentially compensating for fitness loss associated with amphotericin B resistance. Further elucidation of the role of ERG4 in this process will require additional study.

Replacement of ergosterol

In fungal cell membranes, ergosterol is a vital component responsible for membrane fluidity regulation, making the ergosterol biosynthesis pathway an attractive target for conventional antifungal use and drug development. In C. albicans and S. cerevisiae, the loss of function of lanosterol 14α-demethylase and C-5 sterol desaturase (ERG11 and ERG3) has been associated with acquired amphotericin B resistance and for the exchange of ergosterol for other sterols such as lanosterol, eburicol, and 4,14-dimethyl-zymosterol in the cell membrane (37, 38). The resistant isolate described here does not produce ergosterol and showed enrichment of lanosterol, episterol, and late-stage sterols such as ergosta-7,22,24(28)-trienol. This altered membrane composition, which likely leads to amphotericin B resistance, is consistent with a loss of function of ERG3 and ERG4 based on their function in ergosterol biosynthesis. The loss of detectable ergosterol, while still uncommon, has been observed before in C. auris. In another amphotericin B-resistant isolate that had a premature stop codon in ERG6, ergosterol was notably absent, with cholesta-type sterols and early sterols comprising the sterol profile (16). We hypothesize that LNV002 is able to survive because late-stage sterols such as ergosta-7,22,24(28)-trienol and episterol are proposed to be able to stabilize the cell membrane in the absence of ergosterol. Additionally, the VF125AL amino acid substitution in lanosterol 14α-demethylase may be compensatory in nature, in order to support a truncated, non-functional C-5 sterol desaturase. However, determining the validity of this hypothesis is beyond the scope of this paper.

In the CRISPR-repaired isolates, LNV002_ERG3^WT and LNV002_ERG3^WT, ERG4^WT, susceptibility to amphotericin B, was restored. LNV002_ERG3^WT was mainly composed of the late-stage sterol Ergosta-5,7,22,24(28)-tetraenol with no detectable ergosterol. Since susceptibility is restored even though ergosterol is absent, amphotericin B seems to be able to interact with this late-stage sterol and exert its antifungal properties.

Fitness cost identified with an amphotericin B-resistant isolate

Amphotericin B resistance is rarely observed for most Candida spp. This may be due to the fitness defects of mutations in the ergosterol biosynthesis pathway (33). In characterizing this resistant phenotype, a comparison between the susceptible and resistant isolates supplemented with 190 different carbon sources revealed a statistically significant fitness defect present on a few of these sources. These data may indicate a narrow fitness cost for combinatorial mutations in ERG3 and ERG4 in this isolate. Conversely, it is possible that among the other seven mutations detected by whole-genome sequencing, there are one or more modifiers that alleviate the fitness cost associated with perturbation of the ergosterol biosynthesis pathway. Furthermore, epigenetic mechanisms may contribute to amphotericin B resistance and potentially could compensate for this fitness cost.

Conclusion

C. auris is spreading broadly, and the initially identified four major clades are no longer endemic solely to the area in which they were initially associated (39). The spread of this agent is resulting in extensive morbidity and mortality to patients in health care facilities across the world. Antifungal drug resistance will continue to evolve and spread unless we can identify and implement effective infection control mechanisms. With only three primary antifungal drug classes available to treat fungal infections in the United States and the multidrug-resistant characteristics of C. auris, the emphasis on anti-microbial stewardship, along with the development of novel antifungals, is more essential than ever to combat this pathogen.

This article characterizes a clinical case description of C. auris amphotericin B resistance resulting from a frameshift mutation in ERG3. In some outbreaks, one-third of C. auris isolates demonstrates amphotericin B resistance that is often attributed to an unidentified mechanism (9). This finding represents a significant advancement in understanding antifungal resistance in C. auris, and the knowledge generated can be used to design diagnostic tests to combat antifungal drug failure.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aac.00601-25.

Fig. 1. aac.00601-25-s0001.tif.

End of ergosterol biosynthesis pathway.

Supplemental material. aac.00601-25-s0002.docx.

Captions for supplemental figure and tables.

Table S1. aac.00601-25-s0003.jpg.

Strains and primers from EPIC mediated ERG3 and ERG4 transformation.

Table S2. aac.00601-25-s0004.xlsx.

Growth curve assay across 190 carbon sources in LNV001 and LNV002.

Table S1 cont. aac.00601-25-s0005.jpg.

Strains and primers from EPIC mediated ERG3 and ERG4 transformation.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.