Effects of abiotic factors on the phylogenetic diversity of bacterial communities in acidic thermal springs

Abstract

Acidic thermal springs offer ideal environments for studying processes underlying extremophile microbial diversity. We used a carefully designed comparative analysis of acidic thermal springs in Yellowstone National Park to determine how abiotic factors (chemistry and temperature) shape acidophile microbial communities. Small-subunit rRNA gene sequences were PCR amplified, cloned, and sequenced, by using evolutionarily conserved bacterium-specific primers, directly from environmental DNA extracted from Amphitheater Springs and Roaring Mountain sediment samples. Energy-dispersive X-ray spectroscopy, X-ray diffraction, and colorimetric assays were used to analyze sediment chemistry, while an optical emission spectrometer was used to evaluate water chemistry and electronic probes were used to measure the pH, temperature, and E(h) of the spring waters. Phylogenetic-statistical analyses found exceptionally strong correlations between bacterial community composition and sediment mineral chemistry, followed by weaker but significant correlations with temperature gradients. For example, sulfur-rich sediment samples contained a high diversity of uncultured organisms related to Hydrogenobaculum spp., while iron-rich sediments were dominated by uncultured organisms related to a diverse array of gram-positive iron oxidizers. A detailed analysis of redox chemistry indicated that the available energy sources and electron acceptors were sufficient to support the metabolic potential of Hydrogenobaculum spp. and iron oxidizers, respectively. Principal-component analysis found that two factors explained 95% of the genetic diversity, with most of the variance attributable to mineral chemistry and a smaller fraction attributable to temperature.

FIG. 1.

(A) Study design diagram of the four springs. Ovals indicate the spring source (S = sulfur, Fe = iron-dominated spring), and open circles indicate sample collection points. (B) Flowing spring at RM. Origin and 70°C sample site (lower of two pools) are marked with temperature and pH. (C) Flowing springs at AS. (D) Contour map of AS and RM (1:100,000 scale) based on a 1983 U.S. Geological Survey 50-min interval map (Yellowstone National Park North, Wyoming-Montana).

FIG. 2.

Scanning electron micrograph of rod-shaped bacterial cells (arrows) attached to sulfur crystals (S) in a sulfur-dominated spring, AS102. Scale bar = 10 μm.

FIG. 3.

Comparison of matched spectra (S and Fe) from AS103 and RM. Peaks from Au and Pd result from metal coating during SEM sample preparation. (A) Spectrum of yellow crystals sampled near the spring origin in a zone of turbulence in AS103. S peak is prominent; C and Si are also present. (B) Reddish-rust-hardened AS103 streambed sampled in the laminar flow of the spring just below the yellow crystals shown in panel A. In addition to Fe peaks, Si, S, and C also appear. Carbon-coated scans of the same sample revealed minor peaks of B, K, and Na and peaks of thallium (Tl) and Mn. (C) AS103 midstream reddish-tinged sulfur crystals. Prominent peaks for S with minor peaks for C, Fe, Mn, and Ta are shown. The sample was carbon coated. (D) RM sediment with red-rust deposits showing major Fe peaks. The sample was carbon coated.

FIG. 4.

Results of phylogenetic analyses with rRNA gene sequences determined from AS that were related to Hydrogenobaculum spp. The tree shown is the ML tree (ln = −2,118.2). MP and NJ analyses had similar topologies, and the minor differences among the trees found by each algorithm were not supported by bootstrap analyses. Filled circles indicate branches with ML and NJ bootstrap support of greater than 80%, and open circles indicate bootstrap support exceeding 50%. A1 = AS101; A2 = AS102; A3S = AS103 containing sulfur. The sampling temperatures were 70, 65, and 60°C for AS101 and AS102 and 75°C for AS103.

FIG. 5.

Results of phylogenetic relationships of rRNA gene sequences determined from RM and AS103 Fe and AS103 Fe-S samples. The tree shown is the ML tree (ln = −2,882.6). MP and NJ analyses had similar topologies, and the minor differences among the trees found by each algorithm were not supported by bootstrap analyses. Filled circles indicate branches with ML and NJ bootstrap support of greater than 80%, and open circles indicate bootstrap support exceeding 50%. AFeS = AS103 containing iron and sulfur; A3Fe = AS103 containing iron.

FIG. 6.

Pairwise Φ_ST values between all samples were projected onto two dimensions by PCA, showing the genetic clustering of samples. Note that although the two factors are scaled equally, factor 2 only accounts for a minor part of the variance in Φ_ST values.