Linking activity and function to ecosystem dynamics in a coastal bacterioplankton community

Abstract

For bacterial communities containing hundreds to thousands of distinct populations, connecting functional processes and environmental dynamics at high taxonomic resolution has remained challenging. Here we use the expression of ribosomal proteins (%RP) as a proxy for in situ activity of 200 taxa within 20 metatranscriptomic samples in a coastal ocean time series encompassing both seasonal variability and diel dynamics. %RP patterns grouped the taxa into seven activity clusters with distinct profiles in functional gene expression and correlations with environmental gradients. Clusters 1-3 had their highest potential activity in the winter and fall, and included some of the most active taxa, while Clusters 4-7 had their highest potential activity in the spring and summer. Cluster 1 taxa were characterized by gene expression for motility and complex carbohydrate degradation (dominated by Gammaproteobacteria and Bacteroidetes), and Cluster 2 taxa by transcription of genes for amino acid and aromatic compound metabolism and aerobic anoxygenic phototrophy (Roseobacter). Other activity clusters were enriched in transcripts for proteorhodopsin and methylotrophy (Cluster 4; SAR11 and methylotrophs), photosynthesis and attachment (Clusters 5 and 7; Synechococcus, picoeukaryotes, Verucomicrobia, and Planctomycetes), and sulfur oxidation (Cluster 7; Gammaproteobacteria). The seasonal patterns in activity were overlain, and sometimes obscured, by large differences in %RP over shorter day-night timescales. Seventy-eight taxa, many of them heterotrophs, had a higher %RP activity index during the day than night, indicating a strong diel activity rhythm at this coastal site. Emerging from these taxonomically- and time-resolved estimates of in situ microbial activity are predictions of specific ecological groupings of microbial taxa in a dynamic coastal environment.

Keywords: activity; bacterioplankton; diel; marine; metatranscriptomics; seasonal.

Figure 1

Seasonal variation in %RP values for the top 200 transcript-recruiting genome bins. Error bars indicate the 95% confidence intervals determined by bootstrapping (1000 iterations). The activity cluster of each bin is given (see Figure 2 and Figure S1), and bins with significant variation in %RP either between seasons or between day and night samples are indicated.

Figure 2

Patterns in %RP expression among genome bins and in relation to environmental variables. (A) The top 200 transcript-recruiting taxa were hierarchically clustered based on pairwise Pearson correlations of %RP. To the right of the dendrogram, the 20 samples are arranged in rank order from lowest to highest %RP and colored by the sample's seasonal and day-night origins. See Figure S1 for the same dendrogram with taxon labels included. (B) Canonical Correspondence Analysis (CCA) of the 200 taxa ordinated by %RP and environmental variables. The taxa are colored according to their activity cluster as shown in part (A). (C) Taxonomic composition of cluster members.

Figure 3

Daytime enrichment in the transcriptome devoted to ribosomal protein synthesis. Example bins are from Activity Cluster 1 (gamma proteobacterium HTCC2080), Cluster 2 (Roseobacter Roseovarius sp. TM1035), and Cluster 4 (betaproteobacterium Methylophilales HTCC2181). Black bars = night samples, gold bars = day samples. For the %RP graphs of all top 200 transcript recruiting taxa see Figure S2.

Figure 4

Number of taxa with maximum %RP occurring in each season. Bars are colored based on activity cluster assignments.

Figure 5

Cluster 2 day and night indicator genes. (A) Orthologs (KEGG KOs) that were indicators for day or night samples. “genomes” indicates the number of Cluster 2 genomes expressing orthologs in the given category (out of 45 total genomes in the analysis). “KOs” indicates the number of unique KO classifications in that category that were expressed by Cluster 2 members. “Enriched day” or “Enriched night” indicates the number of those KOs significantly enriched in either the day or night samples. (B) Examples of individual KOs in the categories shown in (A).

Figure 6

Seasonal and day-night variability in phytoplankton concentrations at Sapelo Island, Georgia, USA. (A) Three year time-series of chlorophyll a (chl a) concentrations. Gold and black bars indicate day and night samples, respectively. Su, summer; Fa, fall; Wi, winter; Sp, spring. (B) Higher temporal resolution chlorophyll a measurements (black line) and phytoplankton concentrations during a four-day period in Summer 2010. Green line, centric diatoms; red line, pennate diatoms; blue line, dinoflagellates. Night hours are shaded in gray.