Population genetics of Synehococcus species inhabiting the Mushroom Spring microbial mat, Yellowstone National Park

Abstract

The species concept in microbiology is under considerable debate. Some scientists believe that traditional approaches are adequate, while others search for more natural concepts. The ecotype concept (ecological species concept) was evaluated in this work. Two temperature sites of a well-studied microbial mat system in Yellowstone National Park were investigated. Previous molecular analyses with 16S rRNA and the adjacent internal transcribed spacer (ITS) suggested the dominance of two putative ecotypes (PEs) of cyanobacteria in these sites, Synechococcus genotypes A and B'. Higher resolution molecular approaches were developed to address the hypotheses that (i) there are more Synechococcus PEs than those discerned by 16S rRNA/ITS sequence variation, (ii) these PEs exhibit distinct ecological distribution patterns and (iii) recombination has been less important than mutation in shaping the evolution of these Synechococcus populations. Analysis of single protein-encoding loci revealed more sample-specific PEs than previously detected, but didn't account for recombination. Bacterial artificial chromosome (BAC) libraries were constructed to sample multiple loci near 16S rRNA genes for multi-locus sequence analyses (MLSA). Analysis of BAC clone end sequences revealed that 16S rRNA regions of the genomes of Synechococcus A- and B'-like populations have undergone rearrangement. Multiple BAC loci were analyzed using two population genetics algorithms; Evolutionary Simulation (ES) and eBURST. ES of concatenated MLSA sequences, but not eBURST analysis, suggested a much greater number of PEs than were detected by 16S rRNA and ITS and provided stronger evidence of sample-specificity. Recombination, suggested by phylogenetic incongruency among loci, multiple recombination tests and polymorphism patterns, appears to have been more frequent than mutation, but not to have erased ecotype structure. Many PEs predicted by ES contained a dominant variant surrounded by rare variants. eBURST predicted some clonal complexes with the same dominant variant, but different rare variants. ES appears to miss phylogenetically distant variants that differ at one locus, whereas eBURST appears to miss phylogenetically similar variants that differ at >1 locus. True ecotype populations in nature may contain both types of variants, but this must be evaluated by examining the distribution of all variants relative to environmental gradients.