Spatial and temporal patterns in the microbial diversity of a meromictic soda lake in Washington State

Abstract

The microbial community diversity and composition of meromictic Soap Lake were studied using culture-dependent and culture-independent approaches. The water column and sediments were sampled monthly for a year. Denaturing gradient gel electrophoresis of bacterial and archaeal 16S rRNA genes showed an increase in diversity with depth for both groups. Late-summer samples harbored the highest prokaryotic diversity, and the bacteria exhibited less seasonal variability than the archaea. Most-probable-number assays targeting anaerobic microbial guilds were performed to compare summer and fall samples. In both seasons, the anoxic samples appeared to be dominated by lactate-oxidizing sulfate-reducing prokaryotes. High numbers of lactate- and acetate-oxidizing iron-reducing bacteria, as well as fermentative microorganisms, were also found, whereas the numbers of methanogens were low or methanogens were undetectable. The bacterial community composition of summer and fall samples was also assessed by constructing 16S rRNA gene clone libraries. A total of 508 sequences represented an estimated >1,100 unique operational taxonomic units, most of which were from the monimolimnion, and the summer samples were more diverse than the fall samples (Chao1 = 530 and Chao1 = 295, respectively). For both seasons, the mixolimnion sequences were dominated by Gammaproteobacteria, and the chemocline and monimolimnion libraries were dominated by members of the low-G+C-content group, followed by the Cytophaga-Flexibacter-Bacteroides (CFB) group; the mixolimnion sediments contained sequences related to uncultured members of the Chloroflexi and the CFB group. Community overlap and phylogenetic analyses, however, not only demonstrated that there was a high degree of spatial turnover but also suggested that there was a degree of temporal variability due to differences in the members and structures of the communities.

FIG. 1.

Bacterial (A) and archaeal (B) diversity obtained from DGGE profiles. The Shannon-Weaver indices (H) are values averaged across layers. The arrows indicate sampling dates for which clone libraries were constructed. The error bars in panel A indicate one standard deviation (n = 5) for illustrative purposes; archaeal diversity fluctuations exhibited similar trends (data not shown). mixo, mixolimnion; chemo, chemocline; moni, monimolimnion; mixo_sed, mixolimnion sediment; moni_sed, monimolimnion sediment.

FIG. 2.

Bacterial diversity in Soap Lake. (A) Collector's curve of predicted richness (Chao1) for pooled July and October sequences. The subscript numbers indicate the genetic distance (divergence), and the bold numbers indicate the endpoint estimates and are followed by 95% confidence limits. The vertical lines indicate 95% confidence intervals. (B) Observed and estimated richness (Chao1 and ACE) with divergence levels set at 0 to 80%.

FIG. 3.

Frequencies of sequences affiliated with major phylogenetic groups in libraries from the water column and sediments for July and October. Others*, Synechococcus, chloroplast DNA, Alphaproteobacteria, candidate taxon OD1, Actinobacteria, Planctomycetes, and Spirochaetes; Others*1, Planctomycetes, candidate taxon OP1, candidate taxon NKB19, and candidate taxon OD1.

FIG. 4.

Dendrogram (constructed by the unweighted-pair group method with arithmetic means) of the pairwise nonparametric estimates of community structure similarity (θ) between Soap Lake samples obtained from the mixolimnion (5m), the chemocline (20m), the monimolimnion (23m), and sediments (mixsed and monsed). The suffixes “j” and “o” indicate July and October, respectively. Distance = 1 − θ.