Unusual bacterioplankton community structure in ultra-oligotrophic Crater Lake

The bacterioplankton assemblage in Crater Lake, Oregon (U.S.A.), is different from communities found in other oxygenated lakes, as demonstrated by four small subunit ribosomal ribonucleic acid (SSU rRNA) gene clone libraries and oligonucleotide probe hybridization to RNA from lake water. Populations in the euphotic zone of this deep (589 m), oligotrophic caldera lake are dominated by two phylogenetic clusters of currently uncultivated bacteria: CL120-10, a newly identified cluster in the verrucomicrobiales, and ACK4 actinomycetes, known as a minor constituent of bacterioplankton in other lakes. Deep-water populations at 300 and 500 m are dominated by a different pair of uncultivated taxa: CL500-11, a novel cluster in the green nonsulfur bacteria, and group I marine crenarchaeota. b-Proteobacteria, dominant in most other freshwater environments, are relatively rare in Crater Lake (#16% of nonchloroplast bacterial rRNA at all depths). Other taxa identified in Crater Lake libraries include a newly identified candidate bacterial division, ABY1, and a newly identified subcluster, CL0-1, within candidate division OP10. Probe analyses confirmed vertical stratification of several microbial groups, similar to patterns observed in open-ocean systems. Additional similarities between Crater Lake and ocean microbial populations include aphotic zone dominance of group I marine crenarchaeota and green nonsulfur bacteria. Comparison of Crater Lake to other lakes studied by rRNA methods suggests that selective factors structuring Crater Lake bacterioplankton populations may include low concentrations of available trace metals and dissolved organic matter, chemistry of infiltrating hydrothermal waters, and irradiation by high levels of ultraviolet light. Molecular studies of bacterioplankton have identified a set of lake-water microorganisms that are surprisingly consistent, despite wide variation in trophic levels, geologic settings, and water chemistry. The b-proteobacterial division is dominant in all oxygenated lakes studied, including a suite of mostly small, oligotrophic lakes in the Adirondack Mountains (Hiorns et al. 1997; Methé et al. 1998), eutrophic Lake Loosdrecht (oxic due to constant, wind-induced turbulence; Zwart et al. 1998), immense Lake Baikal (Semenova and Kuznedelov 1998), an oligotrophic mountain lake (Alfreider et al. 1996), and lakes in the arctic (Bahr et al. 1996) and 1 Present address: Department of Microbiology, University of Tennessee, Knoxville, Tennessee 37996. Acknowledgments We thank M. Buktenica, A, Gibson, and S. Girdner of the National Park Service for sampling expeditions; R. Truitt, for phytoplankton and zooplankton data; R. Hoffman, S. Rumsey, and R. Collier for processing limnological data; J. Barrington for TOC analyses; N. Hamamura for N. europea RNA; M. Suzuki and E. DeLong for unpublished sequences; A. W. Groeger, B. W. Methé, and J. P. Zehr for unpublished data and probe sequences; N. Pace for bacterial divisions aligned sequence data; M. R. Fisk for geological data; A. Bernhard and M. Rappé for critical readings of the manuscript; the OSU Center for Gene Research and Biotechnology central services laboratory for DNA sequencing; and the OSU Biological Computing Consortium for computer facilities and software. This work was supported by NSF grant DEB-9709012 (S.J.G. and E.U.). Technical Paper 11729 from Oregon State University Extension and Experiment Station. antarctic (Priscu et al. 1999) regions (Table 1). The lakes studied to date cover nearly the entire spectrum of conditions encountered among oxygenated lakes. Crater Lake, located at the crest of the Cascade Mountains in southwestern Oregon, was chosen as a site for the present study of aquatic ecological processes because it is an isolated caldera lake with little input of allocthonous or anthropogenic materials. The oligotrophic lake is protected as a national park. About 78% of water entering the lake falls directly onto its surface as rain or snow, with the remainder entering as runoff from the lightly forested caldera walls or as a minor influx of hydrothermal fluid from the bottom (Collier et al. 1991). There is no surface outlet. Depletion of nitrogen compounds near the surface suggests that phytoplankton growth is N-limited, though bioassay studies imply colimitation with a trace metal (Groeger unpubl. data; Lane and Goldman 1984; Groeger and Teitjen 1993). The lake receives no anthropogenic nutrient inputs aside from low amounts in precipitation and waste from seasonal tour boats. A phylogenetic study of bacteria in Crater Lake was undertaken as a preliminary step in the characterization of bacterioplankton processes in the lake. The primary objective of this work was to identify dominant bacterial and archaeal taxa by using small subunit ribosomal ribonucleic acid (SSU rRNA) gene clone library analyses, with confirmation by oligonucleotide probe hybridization to lake-water rRNA. A second objective was to compare Crater Lake’s physical, chemical, and biological properties with those of other lakes