Three pathogenic Blastomyces species, B. dermatitidis, B. gilchristii, and B. helicus, have been identified in North America. In the United States, B. dermatitidis has been found throughout areas surrounding the Great Lakes, the Ohio and Mississippi River valleys, and the St. Lawrence River (1). In contrast, B. gilchristii has a smaller geographic range in Canada and the northern United States (2), and B. helicus has been found in the northwestern United States (3). No differences in clinical manifestations have been reported among these Blastomyces species.

In the United States, previous case reports have linked blastomycosis infections to outdoor activities, especially those involving moist soil and proximity to waterways (4,5). One of the largest reported outbreaks of blastomycosis occurred in 2015 among persons who had recreated along the Little Wolf River in Wisconsin (6). In Minnesota, blastomycosis is a reportable disease; epidemiologists at the Minnesota Department of Health (MDH) routinely collect demographic and clinical information for blastomycosis cases and attempt interviews to characterize illness and exposure history. The MDH Public Health Laboratory provides fungal identification services and stores isolates submitted by clinical laboratories.

Although whole-genome sequencing has been used to investigate outbreaks involving various fungal pathogens, such as Candida auris and Coccidioides spp. (7,8), this molecular technology has not been used to investigate Blastomyces spp. outbreaks in the United States. We performed whole-genome sequencing to determine the genetic diversity and phylogenetic relationships of 2 familial clusters of B. gilchristii infections identified in Minnesota.

In August 2020, five cases of blastomycosis were identified as cluster A, which comprised a family of 2 White Hispanic parents and 3 children (Table). Four of the 5 patients were hospitalized, of which 3 had sputum cultures that were positive for Blastomyces sp. All 5 patients recovered from illness. The mother reported that the family had visited rivers in St. Croix County, Wisconsin, numerous times during the summer. No other likely exposure locations or activities were reported.

In addition, 2 cases of blastomycosis were identified in White non-Hispanic sisters. Only 1 sister was hospitalized and had a positive culture for Blastomyces sp. from a bronchoalveolar lavage specimen. MDH learned that their father had blastomycosis in 2014, which was attributed to B. dermatitidis (9). The 2 patients with isolates (1 sister and the father) were classified as cluster B (Table). The family owned a cabin in Hubbard County, Minnesota, which is highly endemic for blastomycosis and was likely the exposure location for the three cases. All 3 patients recovered from illness.

Blastomyces identification is routinely performed by MDH only at the genus level. Therefore, the Centers for Disease Control and Prevention (CDC) determined the species in 4 isolates from the 2 blastomycosis clusters and performed Illumina (https://www.illumina.com) short-read sequencing (National Center for Biotechnology Information BioProject accession no. PRJNA786864). To investigate genetic diversity between strains, we performed whole-genome single-nucleotide polymorphism (SNP) analysis using the MycoSNP version 0.19 analytical workflow (https://github.com/CDCgov/mycosnp). We used publicly available sequences from B. dermatitidis isolates (NCBI run nos. SRR11849827, SRR11849828, SRR11849829) for comparison and genome assembly data for B. gilchristii strain SLH14081 from GenBank (accession no. GCA_000003855.2) as a reference. We constructed a neighbor-joining tree showing SNP differences and maximum-likelihood tree showing bootstrap values using MEGA software version 7.0, (https://www.megasoftware.net) and FastTree 2 (10).

All the isolates were B. gilchristii rather than B. dermatitidis. Phylogenetic tree analysis showed B. dermatitidis and B. gilchristii grouped into distinct clades, which were separated by 52,431 SNPs (Figure). Sequences from all 4 B. gilchristii isolates clustered with the reference genome SLH14081 and were separated by a minimum of 11,695 SNPs. Each familial cluster formed a subclade within the B. gilchristii clade; the subclades were separated by 5,214 SNPs. In cluster A, where all family members were infected at the same time and location, we found 63 SNPs separated the 2 cases. In cluster B, where exposures occurred in the same location but infections were 6 years apart, the cases differed by 120 SNPs (Figure).

Both B. dermatitidis and B. gilchristii have been reported in Minnesota (2). We used whole-genome sequencing and SNP analysis to evaluate clusters of blastomycosis infections caused by B. gilchristii in Minnesota. The genomic data showed that cases within cluster A or B were closely related genetically, whereas clusters A and B were genetically distinct. B. gilchristii is likely responsible for a higher proportion of blastomycosis clusters than is currently known. Therefore, pairing genomic data with clinical information and geographic location can be used to monitor blastomycosis infections and determine whether they are clusters, outbreaks, or sporadic occurrences. Our findings demonstrate the utility of genomic analyses for investigating blastomycosis outbreaks, determining genetic diversity of B. dermatitidis and B. gilchristii, and identifying common sources of environmental exposures among cases.