Case report: contamination of a drinking water distribution system by Exophiala-dominated biofilm in the Midwestern United States

Contamination of drinking water distribution systems by microorganisms has been recognized since the mid-1800s, and contamination events may result from introduction and/or regrowth of bacteria, viruses, protozoa, and fungi (Rochelle & Clancey 2006). For example, contamination with opportunistic pathogen bacteria such as Acinetobacter baumannii, Legionella pneumophila, and Mycobacterium avium is well-known (Falkinham 2011; Carvalheira et al. 2021; CDC 2021) with healthcare costs from these three species estimated at $600 million annually for the elderly in the United States (Naumova et al. 2016).

Fungal contamination of drinking water distribution systems is less frequently studied but is increasingly recognized (Mhlongo et al. 2019) with impacts upon water quality (e.g., color, odor, and taste), degradation of materials, and concerns about mycotoxin exposure and opportunistic infections (Nucci et al. 2002; Hageskal et al. 2009; Mesquita-Rocha et al. 2013; Mhlongo et al. 2020; Afonso et al. 2021). Available reports of fungal growth within distribution systems primarily implicate common, terrestrial, and filamentous genera, including Aspergillus, Cladosporium, and Penicillium (Afonso et al. 2021). These may co-occur with bacteria and protozoa in biofilm communities, and interkingdom interactions within such biofilms are poorly understood (Afonso et al. 2021).

Aside from common terrestrial fungi, members of the black yeast genus Exophiala are occasionally reported as distribution system contaminants in tap water and especially around outlets in bathrooms, kitchens, dishwashers, and laundry machines (Matos et al. 2002; Lian & De Hoog 2010; Biedunkiewicz & Schulz 2012; Adams et al. 2013; Isola et al. 2013; Babič et al. 2016, 2017; Moat et al. 2016; Zupančič et al. 2016; Wang et al. 2018; Kulesza et al. 2021). Within such environments, oligotrophy and tolerance of extreme conditions by certain Exophiala species enables their growth (Hamada & Abe 2010; Lian & De Hoog 2010; Heinrichs et al.2013b; Zupančič et al. 2016; Wang et al. 2018; Kulesza et al. 2021; Romsdahl et al. 2021). Moreover, many Exophiala spp. are opportunistic pathogens affecting both immune-competent and immune-compromised persons (Zeng et al. 2007; Sav et al. 2016; Singh et al. 2021; Usuda et al. 2021). Infections with Exophiala spp. are most often superficial but do include deep-tissue and systemic mycoses which most commonly affect the lungs (Zeng et al. 2007; Woo et al. 2013; Usuda et al. 2021). Dermal contact, ingestion, and inhalation may be relevant routes of exposure.

Recently, Heinrichs et al. (2013a, b) investigated black biofilms growing on aerators, shower heads, and toilet tanks in Germany. These biofilms were dominated by Exophiala lecanii-corni and smaller amounts of other Exophiala spp. and black yeast-like fungi. E. lecanii-corni may cause superficial mycoses effecting skin, nails, eyes, and sinuses in addition to deeper mycoses of the lungs, digestive system, and central nervous system (Zeng et al. 2007; Woo et al. 2013; Lee et al. 2016; Miyakubo et al. 2020; Hatta et al. 2021; Futatsuya et al. 2023). After further sampling of that distribution system, retrograde contamination with E. lecanii-corni was suggested (Heinrichs et al.2013b). However, it is unknown how frequently similar extensive E. lecanii-corni biofilms contaminate other distribution systems.

In this study, we report a series of Exophiala spp. biofilm contamination events similar to those reported by Heinrichs et al. (2013a), this time from a central Ohio (USA) distribution system. Our objective was to characterize these biofilms through DNA sequencing of the bacterial 16S and fungal ITS regions and to identify potentially pathogenic taxa of concern to water resource managers and for public health. This work highlights the potential importance of fungal biofilms in drinking water systems.

Fungal sequences were identified for samples S1 and S2, which yielded 36,342 and 26,873 sequences per sample, respectively, before denoising. Sample S3 failed to amplify during ITS sequencing. Both samples were dominated Order Chaetothryiales, and specifically by Exophiala spp. (Table 1). In sample S1, the putative species E. cancerae (85% of the reads) and Knufia epidermidis (11% of the reads) were dominant, whereas in S2, the putative species E. lecanii-corni was dominant (98% of the reads). E. lecanii-corni dominated the biofilm samples characterized by Heinrichs et al. (2013a). We view the identification of E. cancerae with caution because species-level identifications from next-generation DNA sequencing are tentative owing in part to sequencing and database shortcomings (Nilsson et al. 2006; Yamamoto et al. 2014). Moreover, E. cancerae is primarily reported from tropical locations. In South America, it is a causative agent of Lethargic Crab Disease (Orélis-Ribeiro et al. 2011) and we are aware of one report of gastrointestinal infection by E. cancerae from Hong Kong (Woo et al. 2013).

Several additional melanistic, black yeast-like fungi from orders Chaetothryiales and Venturiales that are commonly found in bathrooms (Lian & de Hoog 2010; Wang et al. 2018), and that are capable of human opportunism, were detected. First, E. oligosperma (0.6% of reads in S2) opportunistically infects cutaneous, subcutaneous, and various deep tissues including the lungs, heart, gastrointestinal tract, spleen, lymphatic system, blood, and brain (Tintelnot et al. 1991; de Hoog et al. 2003; al-Obaid et al. 2006; Zeng et al. 2007; Woo et al. 2013). Several additional species that opportunistically primarily infect human skin and nails were also detected, including Knufia epidermidis (11% of reads in S1; Li et al. 2008; Saunte et al. 2012; Martin-Gomez et al. 2019), Cyphellophora europaea (4% of reads in S2; de Hoog et al. 2000; Lian & de Hoog 2010; Saunte et al. 2012; Feng et al. 2014), Rhinocladiella similis (<0.001% of reads in S2; de Hoog et al. 2003; Lian & De Hoog 2010; Richarz et al. 2018), and Ochroconis mirabilis (0.1% of reads in S1; Giraldo et al. 2014; Shi et al. 2016; Yew et al. 2016).

Four phyla – Proteobacteria, Bacteroidetes, Acidobacteria, and Actinobacteria – were present in all samples, accounting for 70–97% of reads (Figure 2). The bacterial composition of samples was similar at the phylum and class levels, with more differentiation at the family and genus levels (Figure 2) as reported previously (Li et al. 2016). Across different geographic regions and distribution system designs, predominant phyla in distribution system biofilms are Proteobacteria, Actinobacteria, Acidobacteria, Cyanobacteria, Bacteroidota, Nitrospira, Firmicutes, and Planctomycetota (Proctor & Hammes 2015; Li et al. 2016; Stanish et al. 2016; Cruz et al. 2020; Ren et al. 2024). The most abundant classes identified in our samples – Alphaproteobacteria, Betaproteobacteria, Cytophagia, and Gammaproteobacteria – were also detected in a German distribution system, where biofilm samples also displayed high community variance (Henne et al. 2012). The possible opportunistic pathogens Legionella spp., Pseudomonas spp., Mycobacterium spp., and Acinetobacter spp. were all detected in at least one sample, as in previous studies (Douterelo et al. 2014; Li et al. 2016; Waak et al. 2018). Certain members of these genera are capable of growth within distribution system biofilms, resulting in illness (Falkinham 2011; Waak et al. 2018; Carvalheira et al. 2021). Moreover, emerging evidence suggests microbial communities in drinking water influence human health through the gut microbiome (Bowyer et al. 2020; Lugli et al. 2022; Vanhaecke et al. 2022). Microbiome impacts from ingesting the bacterial and fungal communities we describe are unknown.

Beyond health implications, the identification of ecological processes promoting growth of biofilms dominated by Exophiala and other black yeast-like fungi may assist control efforts. E. lecanii-corni is resistant to temperature, osmotic, and oxidative stresses (Romsdahl et al. 2021), is oligotrophic, exhibits extreme shear strength (Heinrichs et al.2013b), and thrives in environments laden with toxic hydrocarbons (Woertz et al. 2001; Pirnie-Fisker &Woertz 2007). For these reasons, Heinrichs et al. (2013b) proposed that volatile organic compounds (VOCs) from cosmetics or cleaning may contribute to biofilm contamination. Other considerations for future studies include depletion of chlorine residual, microbial regrowth and its promoting conditions, and water age. In the distribution system sampled, contamination events were somewhat clustered, especially in areas where construction activity necessitated reduction of flow for extended periods. Future studies of these biofilms could sample distribution systems more extensively and seek to understand the source and conditions that encourage growth.

We document the occurrence of Exophiala-dominant biofilm on distribution system taps following Heinrichs et al. (2013a, b), this time in the Midwestern USA. Additionally, we report on the bacterial composition of these biofilms. Biofilms samples contained potentially pathogenic bacteria and fungi including Acinetobacter spp., Legionella spp., Mycobacterium spp., Pseudomonas spp., Exophiala spp., and Knufia spp. Health implications of these biofilms are uncertain. Future studies might include more extensive sampling of drinking water distribution systems for fungal contamination and may seek to identify the environmental conditions that support growth to inform future control efforts.

This study was partially funded by the National Science Foundation (Grant 1942501). We thank the water distribution company and the homeowners that contributed samples. Alauren Lane created our graphical abstract. We also thank Mark Weir for consulting and Nick Nastasi for assistance with microscopy.

Raw sequences are available from GenBank (BioProject: PRJNA1072827).

The authors declare there is no conflict.