Isolation and Genomic Profiling of Multi-Drug-Resistant pathogen Candidozyma auris, Including the New Clade VI, Through Wastewater Surveillance in Abu Dhabi, United Arab Emirates

Abstract

Background

Candidozyma auris (C auris) is an opportunistic, multidrug-resistant fungus associated with a high mortality rate and sustained transmission in both healthcare and environment settings, representing significant public health concerns. Genomic studies have identified distinct clades of C auris, each with specific antifungal resistance traits. This study evaluated the potential of wastewater surveillance for the detection and clade-level characterization of C auris.

Methods

Between March 2023 and February 2024, a total of 23,212 wastewater samples were systematically collected from multiple sampling sites across Abu Dhabi and Al Ain. Isolates were confirmed by Matrix-Assisted Laser Desorption/Ionization–Time of Flight Mass Spectrometry and subsequently characterized through whole-genome sequencing (WGS) for taxonomic resolution. Antifungal susceptibility testing was performed by broth microdilution.

Results

Here we demonstrate direct culture-based isolation of ten C auris positives ( ~ 4 per 10,000 samples). Phylogenetic and mating-type analyses classify nine isolates as Clade I and one as Clade VI. WGS further reveals ERG11 gene mutation in the clade VI isolate, conferring azole resistance. Antifungal susceptibility testing by broth microdilution reveals elevated Minimum Inhibitory concentration (MIC) to at least one antifungal drug in all isolates studied, while two isolates exhibited elevated MICs to agents spanning more than two drug classes.

Conclusions

To our best of knowledge, these findings represent the first regional detection of Clade VI C auris, highlighting wastewater surveillance as a valuable complementary tool for healthcare-based monitoring. Such approaches can provide insights into the transmission dynamics and antifungal resistance of emerging C auris clades that may otherwise remain undetected until its clinical detection.

Plain language summary

Candidozyma auris (C. auris) is a harmful fungus that can cause serious infections, especially in people with weak immune systems. It spreads easily in hospitals, is hard to detect, and is often resistant to medicines, which makes outbreaks difficult to control. In this study, we looked at whether wastewater could be used to monitor the presence of C. auris. We found that C. auris can be detected in wastewater samples, importantly, one isolate belonged to a rare type that, to the best our knowledge has not been seen in this region before. Many of the isolates were resistant to one or more antifungal drugs. These results show that wastewater monitoring could help public health authorities to detect C. auris early and take action to prevent infections before they spread.

Kusuma, Almansoori, Shanmugam, et al. analyze wastewater samples from Abu Dhabi, UAE, to assess the presence of the pathogenic fungus Candidozyma auris. The study demonstrates that C. auris can be detected in wastewater, including a rare clade (VI) not previously reported in the region, highlighting wastewater surveillance as an effective early warning tool for outbreak prevention.

Introduction

C auris, a multidrug-resistant fungal pathogen, was first identified in 2009 in Japan¹. Since then, it has been reported in more than 40 countries and continues to spread at an alarming rate². Due to limited treatment options, high pathogenicity, and increased incidence, both the Centers for Disease Control and Prevention (CDC) and World Health Organization (WHO) have classified C auris as an urgent threat^1,3–6. The global surge in clinical cases of C auris infection is driven by multiple factors, including resistance to disinfectants, rapid transmission, multi-drug resistance, and frequent misidentification due to phenotypic switching^5–7. These features are strongly linked to the species’ genetic diversity, which is structured into six clades, defined by their geographic origin and genomic divergence. Clade I (South Asian), Clade II (East Asian), Clade III (South African), Clade IV (South American), Clade V (Iranian), and Clade VI (Singapore). Clade VI has been recently identified in patient samples from Singapore and Bangladesh, displaying substantial genetic divergence from the previously recognized clades^8,9. Comparative genomic analyses revealed that Clade VI isolates differ by 37,000–42,000 single-nucleotide polymorphisms (SNPs) from other clades and display a distinct antifungal resistance profile. Each clade is associated with specific antifungal resistance traits, mating profiles, clinical presentations, and outbreak potential⁸.

The mortality rates of C auris range between 20% and 70%, and primarily affect individuals with underlying health conditions, medical device users, and long-term care residents. Many infected individuals are asymptomatic, with the fungus potentially remaining dormant for years before symptoms appear^2,10, making identification of asymptomatic carriers challenging and costly¹¹. In addition, C auris poses significant treatment challenges due to its high minimum inhibitory concentration (MIC), and resistance to multiple antifungal agents, including azoles, echinocandins, and polyenes, that ultimately increase the risk of therapeutic failure^12,13.

Rapid and precise species level identification and characterization of C auris is crucial^14,15. Typical diagnostic samples for C auris include dermal and blood samples and may also be extended to urine and stool^2,16. Conventional diagnostic testing of C auris relies on routine phenotyping methods such as VITEK 2 YST, API 20 C, BD Phoenix yeast identification system, and MicroScan. However, these methods are often less accurate, leading to systematic misidentification with other common species such as C haemulonii, C duobushaemulonii, C sake, Saccharomyces kluyveri, and Rhodotorula glutinis^17–19. For accurate identification, all Candida isolates should undergo confirmatory tests using MALDI-TOF equipped with an updated library, or sequencing²⁰. The delayed diagnosis, combined with the pathogen’s ability to colonize both patient and hospital environments, contributes to the rapid spread of C auris, particularly in healthcare settings. The remarkable resilience of C auris, including its ability to form biofilms and survive under harsh conditions²¹, has been demonstrated by its potential to exist in marine water, coastal sediments, recreational water, and hospital surfaces^22–24. These characteristics allow C auris to persist in wastewater systems, especially in regions surrounding healthcare facilities²⁵; highlighting its potential as a key candidate for population level surveillance. Wastewater monitoring, a method popularized during the COVID-19 pandemic, can enhance traditional clinical surveillance. Several recent studies have documented the detection of C auris in community wastewater. In Southern Nevada, USA, both culture and molecular methods identified C auris during outbreak investigations, establishing the region as a pilot site for wastewater surveillance^10,24. This surveillance strategy has subsequently been expanded to other U.S. regions, including Miami, Florida²⁵, St. George, Utah and Baltimore^26,27. Moreover, the WastewaterSCAN initiative now monitors multiple sites across the United States, enabling broader epidemiological tracking of C auris and related pathogens²⁸. The epidemiology of C auris in the United Arab Emirates has been marked by a dramatic increase in the number of cases since its initial detected in 2018^12,29. A recent retrospective analysis done by Thomsen et al identified a total of 908 non-duplicate C auris, of which most (89%) were from hospitals, especially ICUs³⁰. The isolations were mainly isolated from urine, blood specimens, and skin/soft tissue specimens. With respect to antifungal resistance, C. auris isolates in UAE exhibited higher resistance across multiple classes of anti-fungal agents³⁰. It was also observed that patients with C auris infection had higher ICU admission rates and longer hospital stays³¹.

The present study aims to evaluate the effectiveness of wastewater surveillance as a method for early detection and monitoring of C auris, in the emirate of Abu Dhabi, United Arab Emirates. Given the pathogen’s persistence in healthcare settings and its significant threat to public health, the study sought to identify the presence of C auris in wastewater samples, confirming its species through advanced molecular techniques and assessing the genetic and phenotypic characteristics of the isolates. By exploring the feasibility of this non-invasive surveillance approach, the study aims to provide valuable insights into public health management and strategies to combat emerging drug-resistant pathogens.

Here, we demonstrate culture-based isolation of Candida auris from wastewater, identifying ten positive samples, with phylogenetic analysis classifying nine isolates as Clade I and one as Clade VI. Whole-genome sequencing detects an ERG11 mutation associated with azole resistance, and antifungal susceptibility testing shows elevated minimum inhibitory concentrations to at least one antifungal agent in all isolates. These findings highlight wastewater monitoring as a vital, non-invasive complement to healthcare surveillance for pre-empting outbreaks of high-threat fungal pathogens.

Methods

Study design

Composite wastewater samples were routinely collected using auto-samplers. Wastewater samples were inoculated on selectively enriched SDA and Candida CHROMagar³² to promote Candida species growth using the BD Kiestra™ InoqulA automated system for uniform inoculation. Confirmation of species was performed through MALDI-TOF and underwent DNA extraction followed by WGS using the MGI sequencing platform. A total of twenty-four C auris genomes were studied, including ten newly sequenced whole genomes and fourteen publicly available genomes with known clade information. Among the publicly available genomes, three recently published Clade VI genomes from Singapore were retrieved from the Sequence Read Archive (SRA), as summarized in Table 1. Bioinformatics analyses, including taxonomic classification with Kraken2³³, variant calling, and phylogenetic tree construction, were conducted to identify clade affiliations, anti-fungal resistance profiles, and mating type loci. The flowchart of study design is available in Supplementary Fig. 1.

Table 1.

Clade information on ten Candidozyma auris isolates found in United Arab Emirates and fourteen publicly available clades of Candidozyma auris

S.No	Isolate	NCBI Identifier	Existing/Identified Clade	Sequenced in the study	Location
1	B13916	GCA_016772235.1	Clade I	No	UAE
2	B14000	None	Clade I	No	UAE
3	B11205	GCA_016772135.1	Clade I	No	India
4	B11220	GCA_003013715.2	Clade II	No	Japan
5	B13463	GCA_016495665.1	Clade II	No	Canada
6	B12631	GCA_016772195.1	Clade III	No	United States
7	B12037	GCA_016772215.1	Clade III	No	Canada
8	B12342	GCA_016772155.1	Clade IV	No	Colombia
9	B11245	GCA_008275145.1	Clade IV	No	Venezuela
10	IRFC2087	GCA_016809505.1	Clade V	No	Iran
11	SRR10102336	None	Clade VI	No	Singapore
12	SRR24877249	None	Clade VI	No	Singapore
13	SRR25455199	None	Clade VI	No	Singapore
14	SRR25455198	None	Clade VI	No	Singapore
15	AD23_00237_017	None	Clade I	Yes	UAE
16	AD23_00253_014	None	Clade I	Yes	UAE
17	AD23_00279_024	None	Clade I	Yes	UAE
18	AD23_00281_038	None	Clade VI	Yes	UAE
19	AD24_00060_043	None	Clade I	Yes	UAE
20	AD24_00075_015	None	Clade I	Yes	UAE
21	AD24_00080_023	None	Clade I	Yes	UAE
22	AD24_00086_022	None	Clade I	Yes	UAE
23	AD23_00239_059	None	Clade I	Yes	UAE
24	AD23_00275_008	None	Clade I	Yes	UAE

Wastewater Sample Collection and Processing on Selective Media

Composite wastewater samples were collected in a sterile container using a temperature controlled autosampler (HACH, AS950 and Teledyne ISCO, 6712) programmed to collect 200 mL/h from individual manholes of the sewer network across various locations in Abu Dhabi and Al Ain city, UAE. The samples collected were aseptically transported to the lab at 4 to 8 °C and processed immediately on receipt. The wastewater samples were registered and aliquoted into 15 ml sterile falcon tube and barcoded. To isolate pure colonies for MALDI- TOF identification, a limited sample volume of about 10 µL of the collected samples was inoculated on two selective media, SDA supplemented with chloramphenicol and Candida CHROMagar³². BD Kiestra™ InoqulA (BD Kiestra Netherlands), an automated sample processing machine was used for inoculation, enabling uniform streak pattern through magnetic bead-based streaking, and isolation of distinct pure colonies on media plates. Prior to inoculation, samples were mixed by the Kiestra system to ensure homogeneity of the sample. The inoculated SDA and Candida CHROMagar plates were then incubated at 40 °C for 48 h.

Identification of C auris by MALDI -TOF MS

Post incubation, all pure colonies were subjected to MALDI-TOF-based identification (in replicates). A small amount of isolated pure colonies was directly smeared onto a MALDI-TOF target plate. Then, 1 µL of formic acid was added to break down the cell wall and to extract the protein from the isolate followed by allowing the target plate to dry. This was followed by the addition of 1 µL of the matrix solution (α-cyano-4-hydroxycinnamic acid) to each spot, which was further allowed to air dry completely. The target plate was then placed into the MALDI® Sirius System Bruker instrument (MBT Compass reference library 2023, Ver. 4.1), and analysis was performed³⁴. The reliability of each batch of testing was confirmed by the bacterial test standard (BTS) and with a reference organism strain (ATCC). Positive samples were reanalyzed twice for additional confirmation. Results were considered valid only when both the BTS and reference organism were identified appropriately. A MALDI-TOF log score > 2 was considered valid and low score samples were not included for further analysis.

Anti-fungal Susceptibility Testing (AFST)

AFST was conducted for confirmed C auris isolates using MICRONAUT-AM. Antifungal MIC plates, following the manufacturer’s instructions³⁵. Results were interpreted according to the European Committee for Antimicrobial Susceptibility Testing (EUCAST) clinical breakpoints wherever available, and Epidemiological cutoff values (ECOFF) values were applied in their absence³⁶. Notably, clinical breakpoints are currently defined only for a limited number of antifungal agents.

Nucleic acid extraction

The pure colonies that were identified as C auris by MALDI-TOF were transferred to a 5 ml sterile saline tube and centrifuged at 3200x g for 10 min. The supernatant was decanted, and the pellet was mixed with 500 µl of lysis buffer (Mag max lysis solution A42361) and 20 µl of proteinase k, vortexed and incubated at 65 ˚C for 20 min. Post incubation, the 500 µl solution was used as sample for extraction using Mag max wastewater ultra kit (Thermofisher Scientific, MA, United States) in kingfisher flex machine according to the manufacturer’s protocol. The concentration and purity of the extracted nucleic acid was estimated using Nanodrop and Qubit system (Qubit™ dsDNA Quantification Assay Kits). Samples above 400 ng concentration and purity ratios for A260/280 and A260/30 above 1.8 and 2 respectively were processed for WGS studies.

WGS using the MGI platform

The single stranded library was prepared manually with 900 ng gDNA concentration using MGIEasy PCR-Free DNA Library Prep kit (MGI, China). The library preparation steps included fragmentation, size selection, end repair, adapter ligation, denaturation, circularization, and exo-digestion. The double-size selection was performed on the fragmented product using 0.6× and 0.2× MGI DNA easy clean beads to get the size ranging from 450 to 600 nm. The library concentration was quantified using Qubit SS DNA kit. DNA nanoballs (DNB) were prepared for the qualified libraries and pooled according to the manufacturer’s protocol for MGI T7 sequencing (MGI, China). DNB concentration between 8 and 40 ng/µL was considered qualified for sequencing. Pooled samples were loaded onto T7 flow cell and sequenced using DNBSEQ-T7RS DNB Sequencing Set (FCL PE100) (MGI, China).

Bioinformatics analysis of WGS data

The obtained sequences were compared with the fourteen publicly available C auris genomes with known clade information, retrieved from SRA (Table 1). The genomes were analyzed using our in-house developed bioinformatics pipeline. Adapter sequences and low-quality reads were trimmed from all samples using fastp v0.23.4³⁷. The clean sequences were subjected to K-mer based taxonomic classification using Kraken2 v2.1.3³³ followed by Bracken v2.9³⁸ to verify the presence of C auris using PlusPF-8 database (published on 5th June 2024 on Index zone by BenLangmead). The trimmed reads were aligned to the reference genome using Burrows Wheeler aligner (BWA) v0.7.18³⁹ to GCA_C and C auris_B8441_V3, which comprises seven chromosomes amounting to 12.4 megabases (Mb) and was published on 22nd April 2024, retrieved from NCBI. The resulting BAM files were sorted, and duplicate reads marked using Picard v3.0⁴⁰. Variant calling was performed with FreeBayes v1.3.6⁴¹, a haplotype-based variant detector, and the resulting variant call format (VCF) files were annotated using SnpEff v5.1⁴². These VCFs were further used to generate reference consensus FASTA files with bcftools 1.10.2⁴³. Multiple sequence alignment of the consensus FASTA files and core SNP detection was carried out with Parsnp v2.0.5⁴⁴, and the resulting phylogenetic tree was visualized using the Interactive Tree of Life (iTOL) viewer⁴⁵. Anti-fungal resistance analysis was conducted using a custom script that utilized SnpEff annotated VCF to identify drug resistance-associated amino acid mutations leveraging data from FungAMR database⁴⁶. database had unique amino acid mutations for 25 genes including ERG11, FKS1, FUR1 etc for azole and echinocandin drugs.

Additionally, short-read de novo assembly was performed for all samples using SPAdes v3.15.5⁴⁷, and the resulting scaffold FASTA files were analyzed to identify MTL loci (MTLa1(TQR88727.1), MTLa2(TQR88726.1), MTLα1(TQB77621.1), MTLα2(KAF6060548.1)) using command-line blastx 2.15.0 (Altschul, 1990). MTLα2 (KAF6060548.1) protein sequence for Candida albicans was considered in MTL analysis as the sequence for Candidozyma auris is not available on NCBI. The Bioinformatics analysis is detailed in Supplementary method 1.

Statistics and reproducibility

Summary statistics were generated using custom Python scripts (v3.12) applied to output files from fastp (v0.23.4), Picard (v3.0.4), Kraken2 (v2.1.3), and SnpEff (v5.1). No formal hypothesis testing was conducted, and no statistical methods were used to predetermine sample size. Wastewater samples and isolates were treated as biological replicates and were collected independently over the study period. Where applicable, each sample was sequenced once; no technical sequencing replicates were performed. Bioinformatics pipeline performance and reproducibility were evaluated using previously published positive-control whole-genome sequencing datasets from the NCBI Sequence Read Archive. All bioinformatics analyses were implemented using reproducible, version-controlled workflow management systems to ensure consistent execution across samples. Replicates were defined as independent wastewater samples processed and sequenced separately. All sample meeting predefined quality-control criteria were included in the analyses. Results were reproducible across independent samples and were consistent with previously reported findings. For manual antifungal susceptibility testing and MALDI-TOF MS analyses, measurements were performed in triplicate and mean values were used for interpretation.

Results

Isolation of C auris from wastewater sample

Composite wastewater samples were systematically collected between March 2023 to February 2024 at the rate of 100 samples per day (500 samples per week) from 100 distinct locations across Abu Dhabi and Al Ain, UAE. A total of 23,212 samples were screened from which C auris was isolated from 10 samples, yielding an estimated incidence rate for C auris in wastewater samples of 0.043% (or ~4 in 10,000). Detailed sample information is available in Table 1 and in Supplementary data 1 and 2.

The direct inoculation of wastewater onto selective media for C auris detection has not been well documented before. Most reported studies have employed the solid phase of wastewater, obtained using either centrifugation or filtration^26,28 followed by selective enrichment in the presence of antifungal agents such as fluconazole to isolate C auris²⁴. In the present study, the chosen approach was designed to be selective for C auris while also accommodating the detection of other Candida species prevalent in wastewater; therefore, no concentration or selective enrichment steps were applied. In our study, CHROMagar™ Candida plates were primarily colonized by C krusei (13.69%) and C tropicalis (12.09%), with limited direct recovery of C auris. The direct inoculation method was validated by spiking wastewater matrices with known concentrations of the reference strain, C auris (CDC B11903) and comparing direct plating with a pre-concentration step via centrifugation. The pilot study demonstrated that samples concentrated by centrifugation prior to plating achieved substantially higher recovery, with approximately 65% of the spiked organisms recovered. In contrast, direct plating of spiked wastewater samples resulted in a markedly lower recovery, at around 15%. The limit of detection for the direct inoculation method was found to be 1 – 5 ×10^{^6} CFU/L (i.e. 10 to 50 CFU/10 µl sample) (data not included).

Sequencing and Taxonomic assignment of C auris

The ten unique C auris isolates were confirmed by MALDI-TOF analysis and were subsequently analyzed through WGS (Supplementary data 3). The quality assessment of the sequencing reads demonstrated that 96% of the total bases exceeded a Phred quality score of 20 (Q20), and 90% of the bases sequenced surpassed Q30, ensuring high quality data (Refer to Supplementary Fig. 2a). The sequencing yield reaches a mean value of 4.42 gigabases (Gb) sequenced per sample (ranging from 2.55 Gb to 7.13 Gb) – of these, on average 4.38 Gb (ranging from 2.54 to 7.05 Gb) remained after removing adapters and low-quality reads (Refer to Supplementary Fig. 2b). Taxonomic classification identified an average of 1.24 Gb of C auris per sample (range of 0.81 to 2.16 Gb), while alignment to the reference genome (GCA_C and_auris_B8441_V3) indicated a mean value of 2.85 Gb of aligned base (about 2.23 to 3.16 Gb). When expressed as a proportion of total reads, k-mer based taxonomy assignment identified 22.97% (mean value of 32.18%) of the total reads to C auris while the alignment with the reference genome increases the classification rate to 73.10% of the bases (mean value of 69.35%) (Refer to Supplementary Fig. 2c). The underlying sequencing data supporting Fig. 1 is summarized in Supplementary data 4.

Fig. 1. Phylogenetic analysis of the 10 C auris isolates.

Phylogeny tree built from MSA of 14 publicly available C auris genome sequences along with those generated for the 10 isolates sequenced in this study (highlighted by box enclosing the samples). The isolates are color-coded according to their respective clades.

Genomic alignment, variant calling, and clade assignment of C auris isolates

The presence of C auris was also confirmed during alignment, with substantial coverage across all samples, indicated by a median value of 295X and standard deviation coverage of ~80X (Supplementary Fig. 3a) to the reference. Genetic variation across samples was distributed with a median of ~1,000 variants consisting of SNPs, Indels and other variants, as well as ~50 MNPs (Supplementary Method 1 and, Supplementary Fig. 3b). To identify the clade origin of the sequences, Multiple sequence analysis (MSA) of consensus FASTA was performed. Results from MSA analysis show that all isolates matched previously reported clades^8,9. Out of the 10 sequenced isolates, 9 belonged to Clades I, while 1 isolate belonged to Clade VI (Fig. 1).

The phylogenetic tree shown in Fig. 1 illustrates the genetic relationships among the analyzed C auris isolates and are color-coded according to their respective clades. A distinct clustering of isolates into six clades was noted whereas most isolates (nine) clustered within Clade I (red). One of the isolates (AD23_00281_038) was associated to Clade VI (orange) and showed tight clustering with other isolates from Singapore, suggesting a close genetic correlation. To further confirm their genetic distances, an SNP distance matrix was created using 10 sequenced isolates and 14 publicly available isolates (Fig. 2 and in Supplementary data 5A). From the distance matrix, it was observed that the SNP distance between AD23_00281_038 and Clade VI Singapore isolates were similar and could be clustered together, while other sequenced isolates clustered with Clade I sequences. The hierarchical clustering also confirmed that Clade IV and Clade VI are closely related with SNPs in the range of approximately 144K-161K. Notably, the Clade I isolates from this study showed strong concordance with previously reported UAE isolates B13916 and B14000. In both the phylogenetic tree and SNP-based hierarchical clustering, these reference isolates are grouped closely with newly sequenced Clade I genomes. The observed pairwise SNP differences ranged from 800-1000, further supporting their genetic relatedness. The list of sequenced isolates and corresponding clades is summarized below in Table 1.

Fig. 2. Hierarchical clustering of SNP Pairwise Distance Matrix on Candidozyma auris isolates observed across different clades (in 1000 s).

The heatmap presents a hierarchical clustering analysis of 10 sequenced isolates and fourteen publicly available C auris isolates based on pairwise genetic distances, visualized through a grayscale distance matrix. The isolates are categorized into six distinct clades, as indicated by color-coded labels on the row and column dendrograms represented by Clade legend bar on left. Clade I (red), Clade II (green), Clade III (blue), Clade IV (purple), Clade V (magenta), and Clade VI (orange). The scale bar on the right-hand side indicates the pairwise distances, with darker shades representing larger genetic distances and lighter shades indicating smaller distances.

Confirmation of Mating type in C auris clades

Certain clades exhibit one mating type due to clonal expansion, and the identified clades of C auris can be further confirmed by understanding their mating type⁴⁸. Mating type prediction indicates that the sequenced Clade I isolates possesses the MTLa locus, while the Clade VI isolate AD23_00281_038 exhibits the MTLα locus as expected in Clade VI samples previously reported from Singapore isolates⁸ (ref Supplementary data 8).

WGS-based Anti-fungal resistance analysis

C auris isolates were profiled using WGS data to evaluate potential resistance towards commonly used anti-fungal drugs. This involved examining specific amino acid mutations in 25 genes from FungAMR⁴⁶ database including ERG11, FKS1, FUR1, CYP51 and Tac1b genes, that are known to cause drug resistance in C auris infections⁴⁹. For validation purposes, we included Clade I isolate B13916 which was previously reported to carry AMR mutation in ERG11 gene at Y132F and FKS1 gene at S639F^50,51. The details of the sample with elevated MIC and gene mutation identified using sequencing are presented in Table 2 and in Supplementary data 9

Table 2.

Comparison of microdilution (AFST) results with mutations identified by sequence analysis

Antifungal agent	Sample Id	Antibiotic Concentration ((µg/mL)	Mutation	Gene
Amphotericin B	AD23_00239_059	4	-	-
5-Fluorocytosin	AD24_00075_015	1	-	-
Fluconazole	AD23_00275_008	8	-	-
	AD23_00239_059	8	-	-
	AD24_00075_015	16	-	-
	AD23_00253_014	16	-	-
	AD23_00281_038	8	E343D, K177R and N335S	ERG11
	AD23_00237_017	8	-	-
	AD24_00060_043	16	-	-
	AD24_00086_022	128	-	-
	AD23_00279_024	16	-	-
	AD24_00080_023	128	-	-
Voriconazole	No isolates	-	-	-
Posaconazole	No isolates	-	-	-
Itraconazole	No isolates	-	-	-
Micafungin	AD24_00075_015	8	-	-
Anidulafungin	AD24_00075_015	8	-	-
Caspofungin	AD24_00075_015	8	-	-

From the analysis, ERG11 mutations, namely E343D, K177R and N335S in Clade VI isolate AD23_0281_038 were identified. In some cases, the same mutations identified in Clade VI isolates were also detected in Clade IV isolate B12342 (Singapore) as well as Clade V isolate (Iran). Clade I isolates sequenced in the study exhibited wild type (WT) genotype at all evaluated amino acid positions. No amino acid mutations were detected in the genes involved in Anti-Fungal resistance. The underlying data for Fig. 3 is summarized in Supplementary data 5B.

Fig. 3. Amino Acid Mutations in Candidozyma auris Isolates Across Different Clades and Their Resistance Profiles.

The plot shows the presence of specific amino acid mutations associated with drug resistance in C auris isolates across different clades. Each row represents an isolate, color-coded by clade, while the columns represent five key mutations (E343D, K177R, N335S, Y132F, S639F). “WT” denotes the wild type (no mutation), and “MT” indicates a mutation. The mutations are linked to two genes: ERG11 (brown labels) and FKS1 (pink label).

Copy Number Variation analysis

Copy Number Variation (CNV) analysis was performed on ten C auris genomes using CNVnator v0.4.1⁵² to identify gene-level deletion or duplication with particular focus on ERG11, Mub1 and Flo8, genes previously implicated in antifungal resistance and pathogenicity and reported in FungAMR⁴⁶ database. No copy number of deletion or duplication event in the forementioned genes across any sequencing isolates (Refer Supplementary data 6 for more information).

Micro dilution based Anti-fungal resistance

The early surveillance and management of this organism were constrained by limited standardization, and absence of species-specific clinical breakpoints. Following EUCAST recommendation, fluconazole susceptibility was interpreted using non–species-specific Candida breakpoints (S ≤ 2, R ≥ 4), while for 5-fluorocytosine (no clinical breakpoint available), an ECOFF threshold of R ≥ 0.5 was applied. Based on these criteria, all 10 isolates exhibited elevated MICs to fluconazole, with AD24_00086.022 and AD24_00080.023 showing the highest values (≥128 µg/mL). Further, the isolates showed an increased MIC to other Azole drugs (Voriconazole, Posaconazole and Itraconazole) for which no recommended EUCAST breakpoint is available to date. Elevated MICs were also observed in AD23_00239.059 to amphotericin B ( ≥ 4 µg/mL) and in AD24_00075.015 for all tested echinocandins ( ≥ 8 µg/mL) (Table-2).

Multi-class non-susceptibility was observed in isolate AD24_00075.015, which exhibited elevated MIC values to 5-flucytosine (pyrimidine), fluconazole (azole), micafungin, anidulafungin, and Caspofungin (echinocandins). Similarly, isolate AD23_00239.059 demonstrated elevated MIC values to amphotericin B (polyene) and fluconazole (azole). Although these isolates can be categorized as multidrug-resistant⁵³, the breakpoints applied for interpretation were not derived from clinical data. This limitation underscores the need for standardized susceptibility testing methods and clinically validated breakpoints that reliably reflect the therapeutic potential of antifungals against this rapidly evolving pathogen. A detailed summary of AFST results is provided in Supplementary data 7.

Discussion

In this study, we report prevalence of C auris in Abu Dhabi’s wastewater. To the best of our knowledge, this represents the first wastewater investigation to utilize a high-throughput automation system (BD Kiestra™) for sample inoculation, coupled with direct culturing on selective media. Our study spans one of the longest durations and employs a more consistent sampling strategy compared to other reported studies. These findings underscore the potential of wastewater surveillance for detecting emerging fungal pathogens at the community level. We further hypothesize the possible route of C auris in the wastewater sample, by correlating detection patterns with the proximity of healthcare facilities. Notably, six of the C auris positive sampling sites receive wastewater from at least one hospital, whereas the remaining four sites were not linked to any healthcare facility in their catchment area.

The primary route of C auris entry into sewer system is urine and fecal shedding from infected individuals and skin desquamation from colonized hosts. While the exact shedding rates in humans are not fully characterized¹⁰, accurate correlation of wastewater detection with clinical cases, and performing source tracking is not feasible. Further, the species is reported to form biofilms on surfaces and are capable of withstanding high salinity and temperature, it is plausible that the climatic conditions of the study region may favor the persistence of C auris in wastewater. These phenomena further make it challenging to discern whether strains originate from community transmission or by proliferation within the sewer system. To the best of our knowledge, this study is the first to demonstrate direct plating of homogenized wastewater for detection of C auris. However, the prevalence may be underestimated, as greater recovery is expected using a solid fraction of the sample along with enrichment. Further optimization studies are warranted to address the knowledge gap.

To confirm the identity and characterize the genomic variation of the isolates, the in-house developed analysis pipeline was used to compare the obtained C auris sequences with reference sequences/libraries using two approaches: a k-mer method (Kraken2) and an alignment-based method (BWA). Both methods confirmed that the submitted sequence belonged to C auris, with a minor variation in results attributed to the inherent limitations of each method. Kraken 2 was used as initial rapid confirmation because of its ability to rapidly identify C auris presence in samples using PlusPF-8 database before performing alignment and variant calling. However, since Kraken 2 relies on exact k-mer matches, it may overlook regions with sequence variations, such as SNPs, which are not present in the reference database. Despite this limitation, Kraken 2 provides a reliable initial insight into the sample composition. Following the confirmation of the presence of C auris using Kraken 2, we applied the alignment-based method, BWA. Unlike Kraken 2, BWA accounts for both exact and inexact matches to the C auris reference genome. This approach enables more comprehensive analysis, as it allows for the detection of sequence variations, including SNPs, INDELs, and MNPs, through variant calling using Freebayes v1.3.10. We used Freebayes variant caller as it is natively built for haploid genomes. The integration of these methods ensured both rapid initial confirmation and detailed characterization of C auris clade within the samples. Importantly, the pipeline was benchmarked against the established MycoSNP⁵⁴ workflow v1.6.3 using B8441 C auris reference, with concordant results observed across both methods (Refer to Supplementary data 10). The validation supports the robustness of the integrated pipeline for accurate identification and clade-level resolution of C. auris genomes.

Among the identified C auris strains by short-read sequencing, isolate AD23_00281_038, was identified within the newly proposed sixth major clade (Clade VI). This classification was supported by phylogenetic analysis and SNP pairwise distance clustering of this isolate with publicly available sequences, showing a close relationship with Clade VI clinical samples from Singapore. Mating type locus (MTL) analysis determines genetic compatibility and is a crucial marker for distinguishing different C auris clades. The MTL analysis results supported clade assignment by phylogenetic analysis by revealing MTLα locus for AD23_00281_038 (Clade VI isolate) and MTLa2 locus for Clade I samples sequenced in the study as reported before. Additionally, three mutations were identified in the active site of ERG11 gene at positions K177R, E334D, and N335S in both isolated AD23_00281_038 and reported Clade VI reference isolates (SRR10102336, SRR24877249, SRR25455198 and SRR25455199). The amino acid mutations at K177R, E334D, and N335S in ERG11 gene are associated with resistance to azole-based anti-fungal drugs. These mutations are present in all previously known Clade VI isolates. Notably, The Clade I sequenced isolates in the study were all wild type with no known anti-fungal resistance mutation when compared to previously studied clinical isolates from UAE namely B14000 and B13916, which had known mutations at Y132F in ERG11 and S639F at FKS1. We hypothesize that sequenced isolates from this study differ from clinical settings as they were derived from wastewater, suggesting that wild type Clade I might be circulating in UAE. While ERG11 mutations are indeed the predominant mechanism of azole resistance in Clade I, Chow et al. ⁴⁸. analyzed over 300 global isolates and noted that although most resistant Clade I isolates carried ERG11 mutations (e.g., Y132F, K143R). Furthermore, a subset of ~27 isolates from Clades I–IV exhibited experimentally confirmed azole resistance despite lacking ERG11 or FKS1 mutations. These findings support the existence of alternative resistance mechanisms such as upregulation of efflux pumps or regulatory mutations and suggest the possibility that some wastewater-derived isolates may follow similar non-ERG11 and non-FKS1 based resistance mechanisms⁵⁵.

We additionally assessed copy-number variation (CNV) across all sequenced isolates and found no ERG11 duplications. This is consistent with population-scale evidence that ERG11 CNVs are mainly observed in Clade III which are associated with higher fluconazole MICs⁴⁸. The pipeline was able to detect an ERG11 duplication in Clade III - control isolate B12631, confirming that the method can identify ERG11 CNVs when present.

In this study, an attempt was made to correlate sequencing data with manual AFST results. Despite all 10 isolates demonstrating elevated MIC to fluconazole, a mutation in the ERG11 gene was detected exclusively in the isolate identified as clade VI. Similarly, elevated MIC to Echinocandins is typically associated with mutation at FKS1 gene, whereas no mutation was observed at S639F site in FKS1 gene⁴⁹. These mismatching results may be explained by the fact that there are multiple pathways for the strain to modify its resistance mechanism based on the exposed environmental conditions. However, it is important to highlight that the gene targets accessed in this study were mostly obtained from clinical isolates, whereas the isolated strains were from wastewater and might have been exposed to unique environmental conditions which differ markedly from clinical settings. A similar instance was reported, on C auris isolated from Turkey, which was phenotypically resistant to amphotericin B and caspofungin, yet no mutations were detected in the ERG2, ERG6, and FKS genes⁵⁶. These findings suggest that C auris in wastewater could develop unexplained mutations or resistance mechanisms that might contribute to the observed potential AMR. This complexity arises because resistance is often multifactorial, transient, and involves heterogeneous cellular responses rather than simple, predictable mutations⁵⁷. This highlights a critical knowledge gap where phenotypic resistance mechanisms are not fully elucidated at the genomic level. Furthermore, the lack of established clinical breakpoints and the reliance on tentative or provisional cutoff values emphasize key areas of discordance between phenotypic and genotypic resistance profiles.

We hypothesize that Clade VI C auris identified in the United Arab Emirates is phylogenetically related to strains identified and circulating in Bangladesh⁹, reflecting human migration as potential introduction routes from Asia diaspora. Although the UAE has experienced a significant upsurge in C auris cases, including isolates with considerable antifungal resistance, Clade VI has not yet been identified in clinical samples from the region³⁰.

This study has several limitations. While direct plating of raw wastewater represents an exploratory and less sensitive approach. The recovery efficiency is likely to be enhanced through sample concentration or enrichment strategies, including analysis of the solid fraction. The use of enrichment media (Sabouraud Salt Dulcitol Broth) may further improve detection sensitivity, indicating that the current methodology could underestimate the true prevalence of C. auris in wastewater. In the absence of established clinical breakpoints for all tested antifungals, we relied on the use of environmental cutoff values, which should be interpreted as tentative. In addition, antifungal resistance was inferred based on genomic analysis of known resistance-associated genes curated from the FungAMR database. While no canonical mutations were identified in key resistance target genes across the sequenced isolates, with the exception of AD23_00281_038 — several isolates, including AD24_00086_022, AD24_00080_023 (fluconazole), and AD24_00075_015 (echinocandins), exhibited elevated MICs without a corresponding genomic explanation. Resistance in Candida species may arise through alternative mechanisms not captured by the present analysis, including efflux pump upregulation, promoter-region or non-coding regulatory variants, chromosomal aneuploidy, and other structural genomic alterations. Although whole-genome sequencing was performed, the present analysis focused primarily on coding-region variants across curated resistance genes and targeted copy number changes in ERG11, Mub1, and Flo8.

Conclusion

Our findings emphasize the critical role of sentinel surveillance in monitoring the clonal proliferation of C auris or the introduction of new clades. Given the pathogen’s ability to rapidly adapt and develop resistance to multiple anti-fungal agents, it is imperative to detect and characterize its presence in various environments, such as wastewater, lakes, and city water reservoirs. Surveillance, particularly in regions with high international travel and dense populations, may provide invaluable data on the spread and evolution of C auris. Sustained genomic and phenotypic surveillance will be essential for enabling public health authorities to implement targeted interventions that mitigate outbreaks and curb the global spread of this emerging multidrug-resistant pathogen.

Author contributions

V.K., S.A., and B.K.S. contributed equally to this work. W.E.: methodology, supervision, formal analysis, and reviewing. V.K. performed the bioinformatic analyses. S.A. carried out the antifungal susceptibility testing (AFST). S.A., S.S., and M.Q. isolated strains from wastewater samples. B.K.S. and V.K. performed data analysis. B.K.S., S.M., S.P., S.B., T.C., and S.P. performed DNA extraction and sequencing. G.B., S.B.S., T.M., J.Q., T.C., and W.E. contributed to manuscript review and structural refinement. All authors reviewed the results and approved the final version of the manuscript.

Peer review

Peer review information

Communications Medicine thanks Joveria Farooqi, Elena de Carolis and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

The source data underlying the figures presented in this study are provided in the supplementary Information. Specifically, the source data for Fig. 1 is available in Supplementary Data 4; for Fig. 2 is Supplementary data 5, sheet A; for Fig. 3 in Supplementary Data 5, sheet B. Raw FASTQ files generated in this study have been deposited in the NCBI Sequence Read Archive (SRA) under BioProject accession PRJNA1447732. The corresponding BioSample accession number and SRA accession number are provided in Supplementary data 11. The data will be made publicly available upon publication. All supplementary data, materials, and analysis scripts used in this study are additionally archived on Zenodo (10.5281/zenodo.19568128).

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s43856-026-01642-2.