Biological diversity has been estimated for various phyla of life, such as insects and mammals, but in the microbe world is has been difficult to determine species richness and abundance. Here we describe a study of species diversity of fungi with a yeast-like colony morphology from the San Juan Islands, a group of islands that lies southeast of Vancouver Island, Canada. Our sampling revealed that the San Juan archipelago biosphere contains a diverse range of such fungi predominantly belonging to the Basidiomycota, particularly of the order Tremellales. One member of this group, Cryptococcus gattii, is the etiological agent of a current and ongoing outbreak of cryptococcosis on nearby Vancouver Island. Our sampling did not, however, reveal this species. While the lack of recovery of C. gattii does not preclude its presence on the San Juan Islands, our results suggest that the Strait of Juan de Fuca may be serving as a geographical barrier to restrict the dispersal of this primary human fungal pathogen into the United States.
A wide variety of ecological habitats have been investigated with regard to their richness and diversity of yeast species. Yeasts have been isolated from forest soils, agricultural soils, and marine environments (Slavikova & Vadkertiova, 2000, 2003; Gadanho, 2003; O'Brien, 2005). Most recently, the beetle gut has been shown to be inhabited by a significant number of yeast species (Suh & Blackwell, 2004; Boekhout, 2005). Yeasts have also been identified from extreme environments, such as hypersaline sea water (Butinar, 2005).
The aim of our study was to describe the species richness and diversity of the three major islands (Orcas Island, San Juan Island, and Lopez Island) of the San Juan Islands archipelago. The San Juan Islands is a group of c. 1970 islands located at the northern end of the Puget Trough and the eastern end of the Strait of Juan de Fuca. The archipelago lies within a ‘rain shadow’, receiving less rainfall and more sunshine than other parts of the Pacific-Northwest as a result of its close proximity to Canadian and United States mountain ranges. The average yearly rainfall of c. 710 mm occurs primarily during the mild winter months. The San Juan Islands climate is therefore dry and relatively moderate.
Our study focused on yeasts present in this locale, pursuing differences in species diversity between the islands and the mainland, as well as the role of the local ferry system on their dispersal. The motivation for this analysis is the occurrence of an outbreak of cryptococcosis on nearby Vancouver Island, where the etiological agent Cryptococcus gattii is causing infections in seemingly healthy individuals (Hoang, 2004; Fox & Muller, 2005). In contrast to previous known environmental sources of this species, on Vancouver Island C. gattii has been found in association with a number of native tree species (Douglas fir, alder, maple and Garry oak) rather than with eucalyptus trees, which are not native to this part of the world (Kidd, 2005). These native tree species reside on the east side of Vancouver Island in the Douglas Fir zone, a geographical feature that extends beyond the island to the continental mainland and the nearby San Juan Islands of the United States. The San Juan Islands therefore represent a temperate weather zone that is similar to but does not completely mimic the biogeoclimatic zone on Vancouver Island: they are, however, the closest United States landmass to the outbreak and an obvious region to explore in order to ascertain whether the C. gattii outbreak represents a public health risk in the United States.
Materials and methods
Environmental sampling in the San Juan Islands area of the Pacific Northwest was performed during September 2004. 397 Diverse environmental samples were collected, including swabs from various surfaces such as railings and car tyres (see Supplementary Table S1). Smaller collections were performed for the mainland control sites of Seattle, Mukilteo and Mount Vernon, Washington. The collection process was documented via the taking of notes, photography, maps, and using a Global Positioning System to record precise coordinates. Samples were shipped by express overnight delivery to our laboratory and processed immediately.
All media was made using ingredients from Becton, Dickinson and company (Le Pont de Claix, France). Antibiotics were purchased from Sigma-Aldrich (St. Louis, MO). All solid samples (5 g when available) were resuspended in 5 × weight in volume phosphate-buffered saline (PBS)+0.01% Tween 20, shaken, and allowed to settle for ∼15 min before 200 μL of undiluted or 1/100 diluted sample was plated on l-DOPA medium (Kwon-Chung & Bennett, 1992) supplemented with antibiotics (ampicillin 50 mg L−1, chloramphenicol 12.5 mg L−1, kanamycin 25 mg L−1). Swabs were soaked overnight in 1 mL PBS+0.01% Tween 20, vortexed, and plated as for solid samples. Samples were incubated for 7 days at 30°C. Test samples of avian excreta from Durham, NC processed at the same time by this protocol allowed rapid identification of serotype A isolates of Cryptococcus neoformans, and also of Cryptococcus gattii when this organism was added to environmental samples from Durham, NC. Test samples obtained later from Vancouver Island also yielded C. gattii, even following shipping and storage conditions similar to those used for samples from the San Juan Islands (data not shown).
All isolates that exhibited a yeast-like morphology ranging in colour from white to black in the l-DOPA primary screen were subcultured first on antibiotic-supplemented Niger seed agar (Kwon-Chung & Bennett, 1992), and then on YPD agar for further analysis.
D1/D2 sequencing and species identification
Deoxyribonucleic acid was prepared for 222 purified yeast-like isolates (Rose, 1990), and the 26S rRNA gene amplified using the primers ITS 5 (5′-GGA AGT AAA AGT CGT AAC AAG G-3′) and LR6 (5-CGC CAG TTC TGC TTA CC-3) (Fell, 2000). The ∼650-bp D1/D2 region was sequenced for each isolate using primers F63 (5-GCA TAT CAA TAA GCG GAG GAA AAG-3′) and LR3 (5-GGT CCG TGT TTC AAG ACG G-3). Novel D1/D2 sequences were submitted to GenBank, and accession numbers are in Fig. 2. Species designation was further verified by Analytical Profile Index (API) strip 20°C carbon source assimilation testing (bioMérieux, Marcy-l'Etoile, France) (data not shown).
Results and discussion
Environmental sampling of the San Juan Islands
We initiated a large-scale environmental sampling effort throughout the San Juan Islands in September 2004, and collected 397 diverse environmental samples (incorporating, but not restricted to, soil, avian excreta, vegetable material and water). Samples collected included swabs from various surfaces, particularly at sites pertaining to the ferry system that links the San Juan Islands with both Vancouver Island and the United States mainland (Fig. 1, Supplementary Table S1). Smaller collections were performed for three mainland control sites (Seattle, the regional transportation hub for the State Ferry System in Mukilteo, and Mount Vernon) of 6–9 samples of material similar to that obtained from the San Juan Islands (Supplementary Table S1).
Both solid samples and swabs were resuspended in PBS+0.01% Tween 20, and plated on antibiotic-supplemented medium containing the diphenolic substrate l-DOPA (Kwon-Chung & Bennett, 1992). The samples were incubated at 30°C, and examined over a period of 7 days. While abundant filamentous isolates were present, yeast-like colonies were readily visible for many samples (Supplementary Table 1). All isolates that exhibited a yeast-like morphology ranging in colour from white to black in this l-DOPA primary screen were subcultured and purified on antibiotic-supplemented Niger seed agar (Kwon-Chung & Bennett, 1992), and then on YPD agar.
Characterization of the biodiversity of yeast-like fungi on the San Juan Islands
The D1/D2 domain of the 26S rRNA gene has commonly been employed as a molecular marker to identify species and describe species diversity (Fell, 2000). We employed this approach for 222 yeast-like isolates, utilizing a BLAST comparison to known rDNA sequences in the GenBank database from previously characterized fungi. This analysis revealed 22 phylogenetically diverse species present on the San Juan Islands. These include many species that are commonly found in such studies, particularly the basidiomycetes Cryptococcus laurentii, Rhodotorula minuta, and Trichosporon moniliiforme (Slavikova & Vadkertiova, 2000, 2003; Randhawa, 2001; Gadanho, 2003; Wuczkowski & Prillinger, 2004) (Fig. 2, Supplementary Table S1).
Most abundant of the ascomycetes were members of the opportunistic genus Exophiala, which is not a true yeast but a dimorphic black mould (De Hoog, 2000). Six genotypes of these moulds were identified, representing 10% of isolates with a yeast-like colony morphology (Supplementary Table 1). 86% of the isolates were Basidiomycota, and the most abundant species was also an opportunistic pathogen, C. laurentii (20% of isolates) (Supplementary Table 1) (Custis, 1995; Bauters, 2002). All C. Laurentii isolates were of the same genotype by D1/D2 analysis; however, a novel isolate most closely related to C. laurentii and Cryptococcus cellulolyticus was also identified (GenBank entry AY953953) and may represent a new species that will be characterized further elsewhere.
An absence of Cryptococcus gattii
Of particular importance to these sampling studies is the ongoing outbreak of C. gattii on Vancouver Island, which has recently spread to the mainland in Canada near to Vancouver city proper. While the encapsulated, occasionally opportunistic, soil-borne yeast C. laurentii is closely related, no isolates with rDNA sequence corresponding to C. gattii itself (Diaz, 2000) were detected from any samples obtained from either the San Juan Islands or the United States mainland control sites. This rDNA-based observation is supported by the absence of any strongly melanized isolates on both the l-DOPA and Niger seed diphenolic compound-containing media. If C. gattii were present, melanized colonies would be observed.
This may indicate that the outbreak has not yet spread to the San Juan Islands. Alternatively, our approaches may have failed to yield C. gattii for technical reasons, because the sampling was not extensive enough, or because weather or temporal conditions precluded recovery of the organism. We note that a range of environments were sampled, including those from which C. gattii is typically recovered, including soil and tree debris (Ellis & Pfeiffer, 1990; Pfeiffer & Ellis, 1991; Kidd, 2003; Kidd, 2004). We also sampled extensively on Lopez Island, on the beach of which one of the expired C. gattii-infected porpoises was discovered. Given the migratory habits of these large marine mammals, this animal may have been infected in the environs of Vancouver Island and then expired in the San Juan archipelago.
In order to test whether our culture methods were sufficient to identify C. gattii, we artificially spiked soil samples recovered locally from Durham, NC, and found that C. gattii could be readily recovered by our culture methods, as could the closely related opportunistic pathogen Cryptococcus neoformans var. grubii from Durham pigeon guano samples. In addition, we found that the culture methods used to analyse samples from the San Juan Islands were sufficient to recover and identify C. gattii from positive samples from Vancouver Island (generously provided by Dr Karen Bartlett, University of British Columbia, Vancouver). We therefore submit that the organism is either not yet present on the San Juan Islands, or is present at an extremely low level such that it escaped detection in our sampling survey.
Implications for the Vancouver Island outbreak
Given the importance of the Vancouver Island outbreak of C. gattii, the fact that it has spread recently to the mainland of Canada, and the ongoing concern as to its eventual extent and its origins, further sampling efforts on the San Juan Islands and the Puget Sound area are clearly warranted in order to ascertain whether or not this emerging pathogenic microbe is a threat to human health extending to within the borders of the United States.
The wide sampling range yet negative result of this study suggests that, possibly in combination with other factors such as a requirement for active transport, the Strait of Juan de Fuca is serving as a geographical barrier to impede the spread of this primary pathogen from Vancouver Island to the nearby San Juan Islands of the United States. This therefore may provide insight into the origin of the Vancouver Island outbreak itself. However, we do not yet completely appreciate what biogeoclimatic factors have contributed to the appearance of this organism on Vancouver Island, and further studies will be essential to illuminate these features. It has been suggested that the abundance of C. gattii in the environment varies in response to seasonal conditions (K. Bartlett, unpublished), and our sampling has been restricted to a single period of the year. Future studies will be required not only to take this into consideration, but also to continue our surveillance of the possible spread of this outbreak to the United States.
Our data suggest that the organism may be unable to spread via wind dispersal, via the waterways, or via migrating animals or insects, yet C. gattii is now well established as part of the biota of Vancouver Island and has recently been observed on the proximate Canadian mainland. How, then, did C. gattii reach Vancouver Island to cause the outbreak? Recent studies of C. gattii population structure suggest that the fungus was actively transported to the island (Fraser, 2005; Kidd, 2005; Campbell, 2005a, b), as has been proposed for C. gattii transport from Australia to California in association with imported eucalyptus trees (Pfeiffer & Ellis, 1991). The fact that the major genotype present on Vancouver Island has also been isolated from a eucalyptus tree in the bay area of San Francisco (strain CBS 7750) lends further support to this hypothesis (Pfeiffer & Ellis, 1991; Fraser, 2005). Discovering the means by which this pathogen has been actively transported could therefore be crucial in ensuring its confinement.
We thank Wiley Schell, Ana Litvintseva, Karen Bartlett and Leona Campbell for advice on environmental sampling; Wiley Schell and Ana Litvintseva for comments on the manuscript; Ted White for assistance; Karen Bartlett for control samples from Vancouver Island; and Jim Kronstad for strains. This work was supported in part by NIAID R01 grant AI50113 to Joseph Heitman. Joseph Heitman was a Burroughs Wellcome Scholar in Molecular Pathogenic Mycology and an investigator of the Howard Hughes Medical Institute, while this work was in progress.
The following supplementary material is available for this article online:
Figure S1. GPS-based distribution of sampling sites in the San Juan Islands. During the sampling procedure, waypoint coordinates for each site were obtained when the GPS system could triangulate. Further details are provided in Table S1.
Table S1. Data summary of microbial sampling in the San Juan Islands, September 2004. GPS triangulation for each site is given, except when confounded by geographical restrictions (tree cover or building proximity). GenBank accession numbers are given for the rDNA D1/D2 region of each isolate, taken either from previously submitted sequences or from novel submissions from this study (in bold).
The material is available as part of the online article at http://www.blackwell-synergy.com