Candida auris isolates analyzed in this study were those recovered from patient specimens submitted for fungal investigation (prior to commencing patients’ therapy) at AUBMC Clinical Microbiology Laboratory (CML), accredited by the College of American Pathologists since 2004.

A total of 28 isolates were retrieved from 21 patients by plating the specimens on Sabouraud dextrose agar (SAB) medium and incubated them at 37°C. The specimen source of these isolates were as follows: 4 isolates from peripheral blood, 1 from central line catheter, 12 from deep tracheal aspirate (DTA), 8 from urine, 2 from skin screen, and 1 from bronchioalveolar lavage (BAL).

In six patients, isolates were simultaneously recovered from different specimen sources of the same patient: blood and central line catheter from 1 patient; DTA and urine from 2 patients; DTA and skin screening from 1 patient; DTA, skin screening, and urine from 1 patient; and DTA and blood from 1 patient.

The recovered Candida species on SAB medium was submitted directly for identification or subcultured on chocolate agar medium prior to identification. The colonies of these isolates were identified by MALDI-TOF system (Bruker Daltonik, GmbH, Bremen, Germany) and by the Vitek 2 system (BioMérieux, Marcy L’Etoile, France). Species identities were confirmed by PCR following by sequencing using ITS1F/ITS4R primers for the variable internal transcribed spacers ITS1 and ITS2 regions, located between universally conserved genes 18S, 5.8S, and 28S and NL1F/NL2R primers used to detect the D1–D2 region located at the 5′ end of the gene 28S encoding for the large nuclear ribosomal subunit (Schoch et al., 2012).

The Vitek 2 system, employing the antifungal susceptibility cards (AST-YS 08), was used to determine the minimum inhibitory concentrations (MICs) of the following antifungal agents: fluconazole, voriconazole, caspofungin, micafungin, amphotericin B, and flucytosine. The E-test (AB Biodisk, Solna, Sweden) was used to determine the MICs (μg/ml) of itraconazole (strip concentration range, 0.002–32 μg/ml), using Roswell Park Memorial Institute (RPMI) 1640 media (Sigma, St. Louis, MO, United States), according to what was reported earlier from our laboratory (Araj et al., 2015).

The interpretation of the minimum inhibitory concentrations (MICs) susceptibility breakpoints (μg/ml) for C. auris were based on CDC¹ (accessed July 7, 2021) and Clinical and Laboratory Standards Institute (CLSI) guidelines, essentially defined based on those established for closely related Candida species (Candida haemuloni) and on expert opinion. In this context, the designated resistant breakpoints are as follows: fluconazole, ≥32 μg/ml; anidulafungin, ≥4 μg/ml; caspofungin, ≥2 μg/ml; micafungin, ≥4 μg/ml; amphotericin B, ≥2 μg/ml. Voriconazole susceptibility breakpoints are not applicable and recommended to consider using fluconazole susceptibility as a surrogate susceptibility assessment. There is no published guidance about flucytosine’s breakpoint susceptibility.

The quality of test performance was controlled by including the reference strains C. albicans (ATCC 10231), C. parapsilosis (ATCC 22019), and C. kruseii (ATCC 6258).

The genomic DNA of 29 isolates (28 C. auris isolates and one C. haemuloni as outgroup) was extracted using the NucleoSpin microbial DNA kit (Macherey-Nagel, Duren, Germany). Subsequently, the extracted DNA was sheared to obtain 15-kb fragments using Hydropore long on the Megaruptor 2 (Diagnode). Express kit 2.0 (Pacific Biosciences, Menlo Park, CA, United States) was used for library preparation using the microbial multiplexing protocol according to the manufacturer’s recommendations. Library size selection was applied using the AMPure PB beads (Pacific Biosciences, Menlo Park, CA, United States) to select for fragments above 3 kb. Using the SMRT Link v9.0, HGAP4 and microbial assembly pipelines were used to assemble the sequences with a minimum seed coverage of 30, 40, and 50 depending on the coverage. The assemblies of CA3LBN and CA7LBN were annotated; gene prediction was achieved using BRAKER2 v2.1.6 pipeline on fungus mode, which combines GeneMarK-ES v4.65 and AUGUSTUS 3.4.0 for fungal gene prediction and identifying gene locations with the corresponding CDS and messenger RNA (mRNA) qualifiers (Altschul et al., 1990; Stanke et al., 2006, 2008; Ter-Hovhannisyan et al., 2008; Camacho et al., 2009; Hoff et al., 2016, 2019; Bruna et al., 2021). Interproscan 5.50-84.0 (Jones et al., 2014) was used on the COG database to create the xml file to further incorporate in the functional annotation pipeline created by the Funannotate 1.8.7 (Blachowicz et al., 2019; Vasquez-Gross et al., 2020; Smith, 2021). The pipeline starts by running HMMscan (HMMer v3.3) (hmmer.org) with default parameters on the PFAM database, then using emapper 2.1.2 based on eggnog orthology data (Huerta-Cepas et al., 2017, 2019). Sequence searches were performed using Diamond Blastp (Buchfink et al., 2015) on UniProt DB version 2021_02 and MEROPS v12.0; the resulting annotations were combined using Gene2Product v1.69, and later, Signalp 5.0 was used to predict secreted proteins (Almagro Armenteros et al., 2019). Furthermore, transfer RNA (tRNA) identification was done using ARAGORN v1.2.41 (Laslett and Canback, 2004); tRNAs identified and found to be overlapping with any CDS sequences were removed. The ribosomal RNA (rRNA) identification was done by downloading C. auris rRNA sequences from the National Center for Biotechnology Information (NCBI) and then by using BLAST + v2.11.0 (Camacho et al., 2009). Assemblies and annotations were assessed using BUSCO V5.2.2 (Manni et al., 2021).

The corresponding phylogenies of the 29 genomes of this study, along with all of the 80 C. auris sequences found in the NCBI assembly database, were performed. Briefly, the alignment of the core genome, detection of recombination events, and single nucleotide polymorphisms (SNPs) detection were performed using Parsnp v1.2, available in the Harvest suite (Treangen et al., 2014) using CA7LBN (index case), a reference genome for clustering and using CA3LBN (identified as C. haemuloni) as an outgroup as described elsewhere (Prakash et al., 2016). SNPs identified in local collinear blocks were subsequently used for reconstructing an approximate maximum-likelihood tree using FastTree2 (Price et al., 2010) while including the general time reversible (GTR) model of nucleotide substitution. The Shimodaira–Hasegawa test was used to assess the support for significant clustering in the observed phylogeny. The interactive tree of life or iTOL (Letunic and Bork, 2019) was used to annotate, modify, and edit the resulting phylogeny.

In this study, SNPs of the 27 C. auris genomes were compared to the SNPs of the first C. auris (CA7LBN detected in October 2020) by using snippy multicommand (snippy-base application v4.5.0) (Seemann, 2015) that generates a core genome multiple alignment against a common reference. The CA7LBN was used as a reference, since it was the index case. The pipeline detects the variants and generates single file for each isolate listing the different variations. The results were compared to detect any possible microevolution events among the genomes with respect to the time of detection.

The sequences of CDR1, CDR2, ERG1, ERG2, ERG3, ERG5, ERG6, ERG11, ERG24, MDR1, MRR1, TAC1, and UPC2 were extracted from the genome of C. auris B11221 and were used as reference. C. auris B11221 was chosen, since it is susceptible against fluconazole and amphotericin B. Then, the reference genes were blasted against the generated assemblies from the study, and the corresponding genes were detected using BLAST + v2.11.0 (Camacho et al., 2009). Protein products were then compared, and the corresponding amino acid substitutions were detected. All mutations have been confirmed by visually inspecting the alignment of the CCS reads to the assembly.

All genome assemblies have been deposited at GenBank under the following accession numbers: CP077052–CP077058, CP077045–CP077051, CP076661–CP076667, CP077038–CP077044, CP077031–CP077037, CP077024–CP077030, CP076749–CP076755, CP077017–CP077023, CP077010–CP077016, CP077003–CP077009, CP076996–CP077002, CP076989–CP076995, CP076982–CP076988, CP076975–CP076981, CP076968–CP076974, CP076961–CP076967, CP076954–CP076960, CP076947–CP076953, CP076940–CP076946, CP076933–CP076939, CP076926–CP076932, CP076919–CP076925, CP076912–CP076918, CP076905–CP076911, CP076898–CP076904, CP076891–CP076897, CP076884–CP076890, CP076877–CP076883, and CP076870–CP076876. We annotated two isolates (the index and the outgroup) with the following ascension codes CP076661–CP076667 and CP076749–CP076755 corresponding to CA3LBN and CA7LBN, respectively (Supplementary Table 1).

The gender distribution among the 21 patients showed 12 male (57%) and 9 female (43%). The average and range of age for the male and female patients were 71 years (range, 57–80 years) and 63 years (range, 34–82 years), respectively. The distribution of patients based on their clinical hospital location were nine in intensive care unit (ICU), two in respiratory care unit (RCU), four in coronary care unit (CCU), one in neuro ICU (NICU), one in emergency department (ED), and four in medical floor.

The MIC₅₀/MIC₉₀ and range of MICs (μg/ml) for each of the tested antifungal agents against the C. auris isolates are shown in Table 1 as follows: itraconazole, 0.25/1 (range, 0.19–1); fluconazole, ≥32/≥32 (16–32); voriconazole, 0.25/0.25 (range, 0.12–4); caspofungin, 0.25/0.25 (range, 0.25–0.25); micafungin 0.12/0.12 (range, 0.064–0.12); amphotericin B, 8/8 (range, 2–16); flucytosine 1/1 (range, 1–1).

Among the triazole class drugs, C. auris was considered to be uniformly resistant to fluconazole, although only 54% of the isolates showed clear resistance to fluconazole (MICs ≥ 32 μg/ml), while 46% of these isolates revealed a level close to the resistance values, an MIC of 16 μg/ml. The voriconazole rates of susceptible, intermediate, and resistant C. auris were 36, 61, and 3%, respectively, and that against itraconazole (n = 8 isolates tested) were 0, 75, and 25%, respectively.

Concerning the polyene class, the specified breakpoints of resistance (MICs ≥ 2 μg/ml) for amphotericin B was detected in 100% of the tested isolates. Extrapolating the breakpoint susceptibility reported for C. auris isolates by CDC, our results indicated uniform susceptibility (100%, MICs ≤ 4 μg/ml) of these isolates against the tested echinocandin class drugs (micafungin and caspofungin) (Table 1).

All strains expressed fluconazole and amphotericin B resistant phenotypes. Corresponding mutations in drug-resistance-associated genes, namely, CDR1, CDR2, ERG1, ERG2, ERG3, ERG5, ERG6, ERG11, ERG24, MDR1, MRR1, TAC1, and UPC2, were investigated in comparison with the reference genes in B11221 strain. Two mutations were detected: the first one in the lanosterol 14-α-demethylase-encoding gene ERG11 (Y132F) particularly in the first “hot-spot” region located between amino-acids (105–165) and another novel mutation D709E, found in CDR1 encoding for an ABC efflux pump.

The 29 genomes generated in our study, with CA3LBN (C. Haemulonii isolate) being the outgroup, were compared to all C. auris genomes available in the NCBI database (80 genomes). The results indicated that our isolates belonged to clade I (South Asian) among the five known global clades and closely clustered with the branch length showing zero differences between our isolates and the ones from United States, Germany, and India and located in close proximity to C. auris isolates recovered from Saudi Arabia, United Arab Emirates, and Hong Kong (Figure 1).

The genetic variations and microevolution among the 28 C. auris isolates were analyzed and compared to the first isolate detected in October 2020 (CA7LBN). Results revealed a range of one to six SNPs with one SNP in three isolates, two SNPs in five isolates, three SNPs in seven, four SNPs in four, five SNPs in six, and six SNPs in two isolates (Supplementary Table 1). SNPs were detected in coding and non-coding regions in the studied isolates. All isolates showed a mutation in THR1, a gene encoding for trihydroxynaphthalene reductase, with an amino acid substitution L188G. In addition, five other mutations were detected in different isolates as shown in Supplementary Table 2. The rest of the mutations were either in genes coding for uncharacterized proteins (hypothetical proteins) or in non-coding regions.