Figure 1.
The Baltic Sea region and sampling locations.
A) Squares with red outlines indicate sites with metagenomic sequencing. 16S rRNA gene tag sequencing was conducted at all sites. B) Temperature, salinity, pH, oxygen, and chlorophyll a measurements across the Baltic Sea transect. Only stations shown on the section were used for section interpolation. The transition from limnic to marine conditions along the 1800 km axis of the Baltic Sea leads to an extensive area of slowly varying brackish conditions. Suboxic, low-pH conditions exist throughout the central basin below 50–60 m. All environmental metadata is in table S1. Sub-basin abbreviations are marked on the map in bold font (BB – Bothnian Bay, BS – Bothnian Sea, BP – Baltic Proper, BW – Baltic West).
Figure 2.
Salinity-driven shifts in bacterial community composition.
A) Bacterial communities characterized by AMPHORA2. The 12 most abundant groups are shown with “Other” defined as the remaining taxonomic groups. Sequences have been combined for all non-viral size fractions (0.1, 0.8 and 3.0 µm) with normalization for genome equivalents. Sampling depth (m) is indicated within circles below sample names, with interconnecting circles indicating samples from the same location. The corresponding basin for samples is shown in the top margin: TT = Torne Träsk, BB = Bothnian Bay, BS = Bothnian Sea, BP = Baltic Proper, BW = Baltic West. A merged set of metagenomes from the coastal eastern Pacific (CalCOFI) was included for comparison. B) RCCA analysis of community composition. The biplot shows correlations between environmental (red) and phylogenetic (color coded by superkingdom) variables with the corresponding first and second canonical variates. The first two canonical variates explain 65% of the environmental variables and 49% of the organismal variables. Biplots for a RCCA at the genus level is shown in figure S5. The outer circle (1) and inner circles (0.5) display the amount of variance explained by the linear combinations of variables. The variance explained for points within the inner circle is less than 25%, thus these have been dimmed.
Figure 3.
Salinity and functional potential.
A) KEGG module and metadata correlations with first and second canonical variates for the bacterial communities in combined size fractions. Text labels and associated arrows indicate the position of analogous modules shown in panel B. The outer circle (1) and inner circles (0.5) display the amount of variance explained by the linear combinations of variables. The variance explained for points within the inner circle is less than 25%, thus these have been dimmed. Biplots for each size fraction and the combined communities are shown in fig. S6. B) Plots showing the distribution of analogous central metabolic pathways highlighted in panel A along the salinity gradient. Note that the spermidine transport system (‘Sp.’) is not shown in panel A due to its non-normal distribution.
Figure 4.
Baltic Sea SAR11 phylogeny and function.
A) Metagenomic sequences classified as SAR11 placed onto a reference tree built from 31 concatenated core genes (see Methods). The Rickettsiales sister group (not shown for clarity) of Candidatus Pelagibacter was chosen as outgroup from AMPHORA2 reference trees. Terminal branches are colored according to the legend in the left margin. Sequences with likelihood weights<0.9 were pruned from the tree. Branches corresponding to sequences from the freshwater Lake Gatun and coastal California metagenomes are colored purple and dark blue, respectively. Numbers in the boxes indicate salinity (PSU). B) Functional differences between the low and high salinity SAR11 populations. Plot of differences in KEGG categories related to biosynthesis (“Synthesis”) and transport (“Transport”) of amino acids, nucleotides and cofactors. Bar plots show mean proportions for each category in the different populations and the confidence interval. Welch's t-test p-values (corrected using the Bonferroni method) are shown in the right margin of the plot.