Microbial community structure across the tree of life in the extreme Río Tinto

Abstract

Understanding biotic versus abiotic forces that shape community structure is a fundamental aim of microbial ecology. The acidic and heavy metal extreme Río Tinto (RT) in southwestern Spain provides a rare opportunity to conduct an ecosystem-wide biodiversity inventory at the level of all three domains of life, because diversity there is low and almost exclusively microbial. Despite improvements in high-throughput DNA sequencing, environmental biodiversity studies that use molecular metrics and consider entire ecosystems are rare. These studies can be prohibitively expensive if domains are considered separately, and differences in copy number of eukaryotic ribosomal RNA genes can bias estimates of relative abundances of phylotypes recovered. In this study we have overcome these barriers (1) by targeting all three domains in a single polymerase chain reaction amplification and (2) by using a replicated sampling design that allows for incidence-based methods to extract measures of richness and carry out downstream analyses that address community structuring effects. Our work showed that combined bacterial and archaeal richness is an order of magnitude higher than eukaryotic richness. We also found that eukaryotic richness was highest at the most extreme sites, whereas combined bacterial and archaeal richness was highest at less extreme sites. Quantitative community phylogenetics showed abiotic forces to be primarily responsible for shaping the RT community structure. Canonical correspondence analysis revealed co-occurrence of obligate symbionts and their putative hosts that may contribute to biotic forces shaping community structure and may further provide a possible mechanism for persistence of certain low-abundance bacteria encountered in the RT.

Figure 1

The RT tree of life based on the V4–V8 portion of the small-subunit rRNA coding region and inferred using maximum likelihood as implemented in RAxML. Stars designate phyla not reported before in clone libraries from river water samples. The bar represents the number of substitutions per site.

Figure 2

NRI and NTI for the nine sites from our study. Solid symbols represent communities that are significantly structured by phylogenetic clustering (values >0) or by phylogenetic overdispersion (values <0) at the P=0.05 level.

Figure 3

Incidence-based (ICE) diversity estimates for bacteria and archaea (BAC+ARC) versus eukaryotes (EUK) for our nine sites. Error bars represent Bonferroni-corrected 95% confidence intervals. The y-axis is represented as a log scale.

Figure 4

Triplot of CCA axes 1 and 2 accounting for 54% of the variance of species (OTUs) with respect to environmental variables. The eigenvalue for axis 1 (horizontal) is 0.61 and for axis 2 (vertical) is 0.55, indicating that both are strong gradients. Presence or absence data were used for bacterial, archaeal and eukaryotic OTUs. Replicate samples are shown. Separate sites are indicated by different symbols, separate stations by different colors. Explanatory variables used in the analysis are shown in red arrows, whereas supplementary variables are shown in gray. The inset highlights eukaryotic OTUs and their relationship to putative symbiotic bacterial OTUs. See main text for explanation and interpretation and Supplementary Figure S2 for further details on the inset figure.