Shifts in bacterial community composition associated with increased carbon cycling in a mosaic of phytoplankton blooms

Abstract

Marine microbes have a pivotal role in the marine biogeochemical cycle of carbon, because they regulate the turnover of dissolved organic matter (DOM), one of the largest carbon reservoirs on Earth. Microbial communities and DOM are both highly diverse components of the ocean system, yet the role of microbial diversity for carbon processing remains thus far poorly understood. We report here results from an exploration of a mosaic of phytoplankton blooms induced by large-scale natural iron fertilization in the Southern Ocean. We show that in this unique ecosystem where concentrations of DOM are lowest in the global ocean, a patchwork of blooms is associated with diverse and distinct bacterial communities. By using on-board continuous cultures, we identify preferences in the degradation of DOM of different reactivity for taxa associated with contrasting blooms. We used the spatial and temporal variability provided by this natural laboratory to demonstrate that the magnitude of bacterial production is linked to the extent of compositional changes. Our results suggest that partitioning of the DOM resource could be a mechanism that structures bacterial communities with a positive feedback on carbon cycling. Our study, focused on bacterial carbon processing, highlights the potential role of diversity as a driving force for the cycling of biogeochemical elements.

Figure 1

Composite satellite images provided by GlobColour illustrating the temporal and spatial evolution of the phytoplankton blooms induced by natural iron fertilization in the region of Kerguelen Island. Surface ocean chlorophyll is derived from ocean color satellite observations. Stations are indicated on the satellite images that were available for corresponding sampling dates. See Table 1 for the exact dates of sampling of the stations.

Figure 2

Heatmap of the OTUs identified as enhanced in at least one bloom station compared with the reference Station R-2. Each of the OTUs is identified by a number and the most detailed phylogenetic assignment obtained. Blue-colored OTUs were not detected at the reference Station R-2, and red-colored OTUs were detected at station R-2. The different shades represent the relative abundance (Rel. Ab.) of the OTUs. White indicates absence of detection. The dashes indicate the OTU was detected in the sample but did not meet the criteria to be considered as enhanced (see Results and Discussion section). The stations are ordered according to a UPGMA clustering built from a weighted UNIFRAC distance matrix. The numbers at each node of the cluster indicate the percent of dissimilarity. Blue nodes correspond to a bootstrap support of 50% to 75% and red nodes correspond to a bootstrap support >75%. OTUs that were specifically associated to one of the two treatments in the continuous-culture experiment are noted with an asterisk (*). See Figure 3 and Results and Discussion section for details.

Figure 3

Relative abundances of the OTUs associated to the seawater cultures (blue) and of the OTUs associated to the diatom-DOM cultures (red) in the continuous-culture experiment after four bacterial-generation times (a) and relative abundance in surface waters of the visited stations of the OTUs identified in the continuous-culture experiment (b). The relative abundances in individual triplicate cultures are shown. OTUs that were also detected enhanced at a bloom site are noted with an asterisk (*). See Figure 3 and Results and Discussion section for details.

Figure 4

Regression analysis between bacterial heterotrophic production (mean value±s.d. in the mixed layer) and the number of enhanced OTUs identified at each bloom site (a) and between bacterial heterotrophic production and the extent of change, determined as the sum of the relative abundances of the enhanced OTUs identified at each bloom site (b). For bacterial heterotrophic production, the mean value±s.d. in the mixed layer is given. The gray data point corresponds to station E-3, which was excluded for the calculation of P and r² values.