Transitions in bacterial communities along the 2000 km salinity gradient of the Baltic Sea

Abstract

Salinity is a major factor controlling the distribution of biota in aquatic systems, and most aquatic multicellular organisms are either adapted to life in saltwater or freshwater conditions. Consequently, the saltwater-freshwater mixing zones in coastal or estuarine areas are characterized by limited faunal and floral diversity. Although changes in diversity and decline in species richness in brackish waters is well documented in aquatic ecology, it is unknown to what extent this applies to bacterial communities. Here, we report a first detailed bacterial inventory from vertical profiles of 60 sampling stations distributed along the salinity gradient of the Baltic Sea, one of world's largest brackish water environments, generated using 454 pyrosequencing of partial (400 bp) 16S rRNA genes. Within the salinity gradient, bacterial community composition altered at broad and finer-scale phylogenetic levels. Analogous to faunal communities within brackish conditions, we identified a bacterial brackish water community comprising a diverse combination of freshwater and marine groups, along with populations unique to this environment. As water residence times in the Baltic Sea exceed 3 years, the observed bacterial community cannot be the result of mixing of fresh water and saltwater, but our study represents the first detailed description of an autochthonous brackish microbiome. In contrast to the decline in the diversity of multicellular organisms, reduced bacterial diversity at brackish conditions could not be established. It is possible that the rapid adaptation rate of bacteria has enabled a variety of lineages to fill what for higher organisms remains a challenging and relatively unoccupied ecological niche.

Figure 1

Principle coordinates analysis of Baltic Sea bacterioplankton community composition. Pair-wise sample differences (β-diversity) were calculated by Spearman's rank-order correlations of OTU counts. Each dot is one sample and is colour-coded according to salinity and sized according to depth. PC1 (3.9% variation explained) correlated most strongly with salinity (Spearman's ρ=–0.92, P<10⁻¹⁶). Principal coordinate 2 (3.1% variation explained) correlated most strongly with depth (ρ=0.86, P<10⁻¹⁶), but also with depth-related factors, including oxygen, phosphate and light intensity (see also Supplementary Table S1).

Figure 2

Bacterial communities of surface samples along the Baltic Sea salinity gradient. (a) Map of the Baltic Sea with sampling stations colour-coded according to the measured surface water salinity. The connecting line indicates the stations displayed in Figure 3. (b) Hierarchical clustering based on bacterial community composition similarities. Nodes supported by high bootstrap values (>90%) are marked with red circles. The samples cluster in three salinity ranges: freshwater–brackish 0–3.2, brackish 4.6–7.7 and marine–brackish 10.5–30.9. Three samples do not show a clear affiliation to any cluster (8.1, 6.3 and 4.4). The samples having salinities of 6.3 and 4.4 were sampled outside Stockholm and St. Petersburg, respectively, the two largest cities in the region, and may potentially be influenced by anthropogenic emissions of nutrients and/or pollutants. (c) Relative abundance vs salinity for abundant phyla and proteobacterial classes.

Figure 3

Spatial distribution and abundance of selected OTUs along a Baltic Sea transect extending east of Gotland and into the Bay of Bothnia (see Figure 2). (a–e) Shifts in populations of (a) SAR86, (b) Acidimicrobiaceae, (c) SAR11, (d) Synechococcaceae and (e) Rhodobacteriaceae along the salinity gradient of the Baltic Sea. (f) The most abundant OTU in our study, which is related to Spartobacteriaceae, in the different depth and salinity regions.

Figure 4

Comparison of surface water OTU sequences obtained in our study to environmental sequences from freshwater and marine reference environments outside of the Baltic Sea. Blue and red areas indicate the number of OTUs ⩾97% identical to freshwater (blue) and marine sequences (red), respectively. Empty triangles are observed number of OTUs and black triangles are Shannon index. To account for sampling size differences, the OTUs corresponding to 800 randomly picked reads (without replacement) were used for each surface sample (samples with fewer than 800 reads were excluded from the analysis). Inset: The Remane curve (reconstructed from Remane (1934)) showing benthic fauna richness along a Baltic Sea salinity gradient. Blue area: number of freshwater species; red area: number of marine species; and yellow area: number of brackish specialist species.