Introduction

The incidence of infections with fungi increased in the last years posing new challenges to health care professionals. Black yeast-like fungi are opportunistic and rarely found as pathogens in clinical specimen. The most prominent member of the group of black yeast-like fungi is Exophiala dermatitidis. Exophiala belongs to the Ascomycotina order Chaetothyriales. The taxonomic classification of E. dermatitidis has undergone revisions over time, and it is currently classified under the genera Exophiala, formerly known as Wangiella [1].

E. dermatitidis exhibits distinct characteristics, among others the melanized and thick cell walls, being responsible for high resistances to environmental stress, including temperature and salt concentrations, and the capability to switch between conidial and hyphal forms. The presence of melanin in the cell wall contributes to the fungus's virulence and enhances its resistance to host defense mechanisms and antifungal treatments [2, 3].

In nature, E. dermatitidis is ubiquitously distributed in various environments, including extreme natural habitats, hydrocarbon-rich artificial settings like steam baths and bathrooms, decaying organic matter, and even dishwashers [4]. Little is known about the natural habitat and the transmission routes of E. dermatitidis [2]. It is suggested that the natural habitat of E. dermatitidis is the warm and wet tropics, as E. dermatitidis is regularly found in environments with high temperatures, high humidity and pH changes. One potential transmission route for E. dermatitidis is through aerosol inhalation from household dishwashers, although household-acquired colonisations seems to be rare and predominantly observed in individuals with conditions such as cystic fibrosis (CF) or immunosuppression [4].

In the Western countries, E. dermatitidis is mostly isolated from the sputa of CF patients, only once in a while with clinical relevance. Furthermore, E. dermatitidis rarely acts as the primary causative agent of fungal infections affecting immunocompromised individuals [2]. Additionally, E. dermatitidis has been described to cause invasive infections of the central nervous system of otherwise healthy individuals with Asian origin. In systemic and invasive cases, it has been associated with mortality rates ranging from 25 to 80% [5]. However, E. dermatitidis is also described to cause cutaneous and superficial infections in both humans and animals [6].

Advancements in various methods, including molecular methods, have shown promising results in the identification of E. dermatitidis. Techniques such as matrix-assisted laser desorption-ionization time of flight mass spectrometry (MALDI-TOF MS), PCR, and ITS sequencing have been utilized for rapid and accurate identification [7,8,9]. The internal transcribed spacer region (ITS) has been particularly useful as a barcode marker for distinguishing black yeasts, including E. dermatitidis, and can even differentiate between different subtypes [10].

However, limitations include incomplete reference databases for black yeast species and challenges in distinguishing closely related species [7, 11]. This led us to conduct a thorough assessment of current typing techniques and select novel microsatellite markers to differentiate E. dermatitidis and to perform cluster analysis of various E. dermatitidis isolates. This is the first study on genotyping using short tandem repeat markers for discrimination of E. dermatitidis isolates from various sources.

Materials and Methods

Study Outline

After identifying short tandem repeat markers and establishing a microsatellite PCR, we applied the new method to a total of 82 clinical isolates of E. dermatitidis. A genotypic cluster analysis was performed.

Isolates

The study did not include patient’s details and did not result in additional constraints for the patients. All data (fungal strains) were anonymously analyzed without patients' consent due to the retrospective nature of the study. All procedures and methods were carried out in accordance with approved guidelines.

Isolates were predominantly collected from patient’s specimen, mainly CF sputa, as well as from environmental sources across different countries. All isolates are listed with their origin in Table 1. In case of the isolates obtained from CF sputa, for three patients, serially isolated E. dermatitidis were collected over time and included. These isolates are marked by patients' ID (a, b, c) in the table below.

Table 1 List of included Exophiala dermatitidis isolates
Full size table

All isolates were cultured on malt extract agar (Life Technologies GmbH, Darmstadt, Germany) and incubated at a temperature of 35°C for a duration of 48 h. Identification of the isolates relied on a combination of macroscopic and microscopic morphology evaluation. In case of uncertainty additionnally sequencing of the internal transcribed spacer region 1 (ITS1) was performed [12].

DNA Extraction

DNA was extracted using the Maxwell16 nucleic acid extraction instrument (Promega, Mannheim, Germany) and the Maxwell16 Tissue LEV Total RNA Purification Kit. Several colonies were inoculated in sterile water and vigorously shaken in a 2-mL innuSPEED Lysis Tube B (Analytic Jena, Jena, Germany) three times at 2000× g for 50 s each using a MagNA Lyser (Roche Diagnostics, Basel, Switzerland). After centrifugation at 8600× g for 30 s, the supernatant was transferred to the extraction cartridge. DNA was eluted in 50 μL of nuclease-free water. DNA concentrations were determined using the NanoDrop 1000 instrument (PeqLab Biotechnologie GmbH, Erlangen, Germany).

Identification of Microsatellite Primers and Microsatellite PCR

Next-generation sequencing and sequence filtering for microsatellites were conducted by ecogenics GmbH (Balgach, Switzerland) using the E. dermatitidis reference strain CBS 550.90. In total, 286 potential primer pairs were identified (data not shown).

The Illumina TruSeq nano library was analyzed on an Illumina MiSeq sequencing platform using a nano v2 500 cycles sequencing chip (Illumina, CAL, USA). The resulting paired-end reads which passed Illumina’s chastity filter were subject to de-multiplexing and trimming of Illumina adaptor residuals. Subsequently the quality of the surviving reads was checked with the software FastQC v0.117. In a next step the paired end reads were merged with the software USEARCH v10.0.240 to in-silico reform the sequenced molecule. The resulting merged reads were screened with the software Tandem Repeats Finder, v4.09. After this process, 7′409 merged reads contained a microsatellite insert with a tetra- or a trinucleotide of at least 6 repeat units or a dinucleotide of at least 10 repeat units. Primer design was performed with primer 3. Suitable primer design was possible in 5′848 microsatellite candidates. Primers were chosen according to the size of the amplification product in order to be able to perform multiplex pcr and according to the motif (motif variation).

Multiplex PCR was carried out using the Taq PCR Core Kit (Qiagen, Hilden, Germany) with three different primer sets. Each primer set consisted of three primer pairs, with one primer labeled with FAM, VIC/HEX, and NED markers, respectively (Table 2). The reaction mixture consisted of 10 μL of reaction buffer, 2 μL of dNTP mix, 2 μL of each primer per set, 0.8 μL of Taq polymerase, and 55.2 μL of nuclease-free water, resulting in a total volume of 80 μL. Subsequently, 20 μL of the reaction mix was combined with 5 μL of DNA at a concentration of 1 ng/μL. The microsatellite analysis took place on an abi3130 sequencing instrument (Life Technologies, Germany) in a total volume of 10 μL.

Table 2 Sequences of six primer pairs used for microsatellite PCR
Full size table

For analysis, 2 μL of each amplification product was mixed with 0.5 μL of GeneScan 1200 LIZ size standard and 7.5 μL of Hi-Di formamide (both supplied by Life Technologies). After denaturation at 92 °C for 2 min, the reaction was rapidly cooled on ice. Based on the fragment size, the DNA samples were subsequently classified into distinct genotypes.

Statistical Analysis

The analysis of microsatellite data was conducted using the R program version 4.2.2 (2022-10-31). The cluster dendrograms are created using the R program with the Pearson algorithm to visualize the similarities and differences among the samples, allowing for the identification of distinct clusters based on their geographic origin and sources (CF, invasive, environmental sources).

A robust approach was adopted to quantify the epidemiological cutoff values (ECVs/ECOFFs) for each average primer value, utilizing the 95th percentile method. This statistical approach, performed in R, allowed the calculation of upper thresholds for each primer value. This statistical procedure facilitated the derivation of ECVs, acting as upper thresholds for attribute values. Then, Isolates were classified based on primer values relative to the calculated ECVs. A classification system was implemented to categorize isolates as “related” or “unrelated,” which was integrated into the dataset. Isolates with primer values below the ECV were categorized as “related,” whereas those surpassing the ECV were classified as “unrelated.”

The discriminatory power of the STR typing method was mathematically defined by calculating the Simpson index of diversity (D): \(D=1-\frac{1}{N\left(N-1\right)}\sum_{j=1}^{s}{n}_{j}\left({n}_{j}-1\right)\), where N is the total number of isolates, s is the total number of clades, and nj is the number of isolates belonging to the jth type. A D value of 1 indicates good discriminatory power of the method whereas a D value of 0.0 indicates that all included isolates are defined as identical by this method.

Results

A cluster analysis with Pearson algorithm was performed. The analysis revealed a distinct cluster for the invasive isolates from patients with Asian origin, marked in Fig. 1 in green. In addition, the environmental isolates, marked in red, cluster with one exception: Isolate CBS109143, obtained from a shower in the Netherlands, clustered more likely with human isolates from CF than with the other isolates from environmental origin (Fig. 1).

Fig. 1
figure 1

Dendrogram of E. dermatitidis isolates from various sources with marked origin. Red: Environment; Blue: Human, CF; Green: Human, invasive; black: unknown

Full size image

Serial isolates from patients a and c were scattered across the dendrogram while the serial isolates from patient b were mainly of the same genotype. The dendrogram built for the cluster analysis of E. dermatitidis, in recognition of their origin, showed closely related isolates from the same or similar origin (Fig. 2). More than one cluster consists solely of isolates from Germany. In addition, the Greek isolates cluster together.

Fig. 2
figure 2

Dendrogram of E. dermatitidis isolates from various sources with marked country of origin

Full size image

Based on the calculated ECVs, a cut-off value of 95% was used to distinguish between related and unrelated strains. As a result, 77 isolates exhibited genetic relatedness and were classified as related strains. On the other hand, five isolates (F111, F88, F114, F05, F120) showed lower genetic relatedness and were classified as unrelated strains. In addition, strains were named related even though they were from distinct sources and geographical origins.

The calculated Simpson diversity index was 0.94, indicating that the typing method of microsatellite PCR for E. dermatitidis is suitable for the discrimination of unrelated isolates.

Discussion

We here performed for the first time a genetic cluster analysis via STR of 82 E. dermatitidis isolates from various origins. The included strain collection is diverse and the isolates were from three sources: 9 environmental isolates and 73 clinical isolates, among them 63 CF and 9 invasive isolates as well as one melanin-deficient mutant (mel−3 mutant). The data showed distinct cluster for isolates from different origins and sources. The CF isolates cluster together, although no person-to-person transfection is described yet and the suggested source of colonization is the environment. The invasive isolates from Asian and otherwise healthy patients formed distinct clusters. Additionally, the country of origin was influencing the genetic cluster. The distinct strains differ each in more than one STR from another, some even in all.

Fingerprinting of E. dermatitidis isolates has been performed with other methods before. However, these analyses were done solely for a set of strains from CF patients sputa [13, 14]. Rath et al. included eleven strains of E. dermatitidis from which ten were isolated from CF patients and one was a reference strain from an invasive infection from Japan. In this study from 1997, different methods were applied, none of them being microsatellite PCR. The authors recommend rather the application of fatty acid methyl ester (FAME) profiles and random amplification of polymorphic DNA (RAPD) analysis than assimilation tests. However, each of the results showed a separation of the included Japanese strain from the European CF isolates [13], which is comparable to the here obtained results.

Additionally, Packeu et al. performed molecular typing using RAPD of E-dermatitidis isolates from patients with CF. They included 71 isolates from 13 patients. They did not find genetic clustering of isolates according to their geographical origin, the date of isolation or their antifungal susceptibility [14]. Packeu et al. further included as well serially isolated strains in RAPD analysis. They detected for the majority of the sequential isolates a distribution in patient specific clusters only with a few exceptions [14]. In contrast to these findings, via the microsatellite approach we did see a clustering of isolates from geographical closeness. In this study, basing on the genotypes of the serial isolates, it can be hypothesized that patients a and c got recolonized and the isolates were replaced. All three patients were CF patients.

The here developed molecular typing method showed good discriminatory power with a Simpson index of diversity with 0.94, demonstrating the STR typing being capable to discriminate between most of the E. dermatitidis isolates. The method could find application in the genetic analysis of E. dermatitidis infection outbreaks, e.g. on clinical wards, as described to took place in the US in 2002 [15] and 2016 [16]. However, there are limitations of this method as relatedness has been detected here for isolates from different geographical origins as well as from differing sources e.g. patients, CF sputa and environment.

Conclusion

We here developed a novel short tandem repeat scheme for molecular typing of E. dermatitidis isolates from various origin, demonstrating geographical and source specific genetic clustering.