Influence of nutrients and currents on the genomic composition of microbes across an upwelling mosaic

Abstract

Metagenomic data sets were generated from samples collected along a coastal to open ocean transect between Southern California Bight and California Current waters during a seasonal upwelling event, providing an opportunity to examine the impact of episodic pulses of cold nutrient-rich water into surface ocean microbial communities. The data set consists of ~5.8 million predicted proteins across seven sites, from three different size classes: 0.1-0.8, 0.8-3.0 and 3.0-200.0 μm. Taxonomic and metabolic analyses suggest that sequences from the 0.1-0.8 μm size class correlated with their position along the upwelling mosaic. However, taxonomic profiles of bacteria from the larger size classes (0.8-200 μm) were less constrained by habitat and characterized by an increase in Cyanobacteria, Bacteroidetes, Flavobacteria and double-stranded DNA viral sequences. Functional annotation of transmembrane proteins indicate that sites comprised of organisms with small genomes have an enrichment of transporters with substrate specificities for amino acids, iron and cadmium, whereas organisms with larger genomes have a higher percentage of transporters for ammonium and potassium. Eukaryotic-type glutamine synthetase (GS) II proteins were identified and taxonomically classified as viral, most closely related to the GSII in Mimivirus, suggesting that marine Mimivirus-like particles may have played a role in the transfer of GSII gene functions. Additionally, a Planctomycete bloom was sampled from one upwelling site providing a rare opportunity to assess the genomic composition of a marine Planctomycete population. The significant correlations observed between genomic properties, community structure and nutrient availability provide insights into habitat-driven dynamics among oligotrophic versus upwelled marine waters adjoining each other spatially.

Figure 1

CalCOFI transect map. (a) Highlighted in blue are the sites sampled for GOS metagenomics sequencing. Two indicators for upwelling during the sampling period are shown, (b) chlorophyll-a and (c) sea surface temperature, both taken using satellite imagery. The GOS site identifications were superimposed onto maps.

Figure 2

GC composition of bacterial protein coding DNA sequences from the 0.1-μm filter of each site sampled. Percent GC is plotted on the y-axis and the % of bacterial sequences on the x-axis.

Figure 3

PCA using oceanographic metadata, including nutrients. The CalCOFI sample sites are given and a taxonomic representation of sequences from each site is shown in a pie chart.

Figure 4

Taxonomic classification using core bacterial HMMs with phylogeny inferred based on placement on the reference tree of core bacterial HMMs from all sequenced bacterial genomes. Taxonomic groups of interest are highlighted as follows: blue = Proteobacteria, red = Firmicutes, light blue = Actinobacteria, green = Cyanobacteria, white = Planctomycetes, yellow = Bacteroidetes. Heatmaps were generated based on the abundance of taxonomy at each node within the reference tree for every group of samples.

Figure 5

PhyML phylogenetic tree of 16S rDNA sequences of metagenomic sequences and SAR11 reference sequence. Each SAR11 is color-coded by subgroup: subgroup 1a (dark purple), subgroup 1b (light purple), subgroup 2 (blue), and subgroups 3 and 4 (grey). The numbers specify how many sequences were placed at each node. Red numbers indicate metagenomics sequences outside clades containing reference sequence(s).

Figure 6

Metabolic profiling of samples based on COG/KOG 20. Metagenomic samples are as follows: 0.1 μm (black), 0.8 μm (orange), 3.0 μm (purple) and oligotrophic (closed circle), aged–upwelled (closed triangle), upwelled (closed square). Reference genomes were binned into groups of varying diversity levels to show that the slope of the linear regression varies with taxonomic diversity: phyla (red) and phyla with fewer than six representatives and genera (blue).

Figure 7

Transport proteins binned based on substrate plotted with increasing genome size. (a) The mean transporter per genome for each substrate was computed and the difference from the mean (across all sites and size classes of each transporter family based on substrate) plotted for each sample, y-axis. Substrates were binned into three groups based on their proportion to the total number of proteins per genome; therefore (a1) denotes enriched in large genomes (proportion increases with genome size), (a2) enriched in small genomes and (a3) unchanged (proportion does not change with change in genome size), based on data obtained from the slope of the line in Figure S11. The x-axis shows the samples (sites/site classes) ordered by genome size (a4). (b) Average transporter per genome of nitrate/nitrite substrate transporters only belonging to the Major Facilitator Superfamily (MFS). (c) Taxonomic classification of nitrate/nitrite MFS transport proteins in B.