Streamlined and Abundant Bacterioplankton Thrive in Functional Cohorts

Abstract

While fastidious microbes can be abundant and ubiquitous in their natural communities, many fail to grow axenically in laboratories due to auxotrophies or other dependencies. To overcome auxotrophies, these microbes rely on their surrounding cohort. A cohort may consist of kin (ecotypes) or more distantly related organisms (community) with the cooperation being reciprocal or nonreciprocal and expensive (Black Queen hypothesis) or costless (by-product). These metabolic partnerships (whether at single species population or community level) enable dominance by and coexistence of these lineages in nature. Here we examine the relevance of these cooperation models to explain the abundance and ubiquity of the dominant fastidious bacterioplankton of a dimictic mesotrophic freshwater lake. Using both culture-dependent (dilution mixed cultures) and culture-independent (small subunit [SSU] rRNA gene time series and environmental metagenomics) methods, we independently identified the primary cohorts of actinobacterial genera "Candidatus Planktophila" (acI-A) and "Candidatus Nanopelagicus" (acI-B) and the proteobacterial genus "Candidatus Fonsibacter" (LD12). While "Ca Planktophila" and "Ca. Fonsibacter" had no correlation in their natural habitat, they have the potential to be complementary in laboratory settings. We also investigated the bifunctional catalase-peroxidase enzyme KatG (a common good which "Ca Planktophila" is dependent upon) and its most likely providers in the lake. Further, we found that while ecotype and community cooperation combined may explain "Ca Planktophila" population abundance, the success of "Ca. Nanopelagicus" and "Ca. Fonsibacter" is better explained as a community by-product. Ecotype differentiation of "Ca. Fonsibacter" as a means of escaping predation was supported but not for overcoming auxotrophies.IMPORTANCE This study examines evolutionary and ecological relationships of three of the most ubiquitous and abundant freshwater bacterial genera: "Ca Planktophila" (acI-A), "Ca. Nanopelagicus" (acI-B), and "Ca. Fonsibacter" (LD12). Due to high abundance, these genera might have a significant influence on nutrient cycling in freshwaters worldwide, and this study adds a layer of understanding to how seemingly competing clades of bacteria can coexist by having different cooperation strategies. Our synthesis ties together network and ecological theory with empirical evidence and lays out a framework for how the functioning of populations within complex microbial communities can be studied.

Keywords: Actinobacteria; alphaproteobacteria; aquatic; bacterioplankton; common goods; ecology; evolution; metagenomics; microbial communities; networks.

FIG 1

Cooperation models shown in a lake setting. The ecotype cooperation model describes anabolic variability within species as the main contributor to coexistence while allowing for recombination and swapping of metabolic modules across strains. The community cooperation model also encompasses anabolic complementarity but between different species and includes common goods ranging from expensive to costless. The community detrital model places complementarity both between and within species but describes catabolic variation.

FIG 2

(a) Location of Lake Erken and the sampling site within the lake. (b) Relative abundance of families from community time-series amplicon survey with metagenome-sequenced samples marked by blue asterisks. (c) Schematic for dilution mixed-cultures. Mixed-culture samples were taken in March 2016, the first spring thaw after the amplicon data set was sequenced. (d) Presence/absence of families detected in mixed-cultures via amplicon survey.

FIG 3

(a) Associations between lake cycle/season and dominant taxa in Lake Erken. The x axes show three seasons (spring, autumn, and winter) and three lake layers of summer stratification (epi-, hypo-, and metalimnion). The y axes show the relative abundance (ra) of the six most dominant taxa (acI, Stramenopiles, acIV, Comamonadaceae, LD12, and grouped phototrophs). Seasons/layers where ra was significantly different (KW P < 0.001, KWmc P < 0.001) to at least two others are denoted by blue asterisks. (b) Correlations between taxon relative abundance (percent ra) and environment parameters, only showing those with linear (Pearsons) correlation (r > |0.3|, P < 0.001). Data points with “×” are 2011 to 2015 MiSeq data, while circles show 454 data from 2008. Linear correlations are shown by blue lines.

FIG 4

OTU networks of the 1° cohorts from amplicons of time-series separated genera (a) and mixed-cultures (b). Nodes sized by average relative abundance in time-series and by percentage of cultures detected in for mixed-cultures. The pink dotted lines indicate negative correlations between OTUs, and gray solid lines show positive correlations.

FIG 5

Relationship between OTU phylogenetic distance (PD) and correlation of the 1° cohort phylotypes to selected genera: “Ca. Planktophila,” “Ca. Nanopelagicus,” “Ca. Fonsibacter,” and basal “Ca. Fonsibacter.”

FIG 6

RAXML phylogenetic tree of translated katG gene showing branches with >60% support. Bold taxon names are those with assay and/or crystallography data, high function in red and low function in pink. Lake Erken metagenomic contigs (names starting in GA) from Nanopelgicales and their 1° cohort are in bold black type with the assigned taxon listed at the far right.

FIG 7

Venn diagrams of the 1° cohorts with potential common goods encoded in genomes, genomic potential confirmed in lake metagenomes in bold. The plus or minus symbol in front of the genus indicates positive or negative correlation. All 1° cohorts were identified from time series (TS), both methods (TS&MC), or mixed cultures only (MC). The order of common goods listing is thiamine or vitamin B1, riboflavin or B2, niacin or B3, pantothenic acid or B5, pyridoxal-5P or B6, biotin or B7, folate or B9, cobalamin or B12, porphyrin ring or heme, bifunctional catalase/peroxidase or KatG, production of storage peptide cyanophycin or cphA, usage of cyanophycin or cphB. In this figure, “prototroph” refers to the coding potential to synthesize all B vitamins.