Niche-directed evolution modulates genome architecture in freshwater Planctomycetes

Abstract

Freshwater environments teem with microbes that do not have counterparts in culture collections or genetic data available in genomic repositories. Currently, our apprehension of evolutionary ecology of freshwater bacteria is hampered by the difficulty to establish organism models for the most representative clades. To circumvent the bottlenecks inherent to the cultivation-based techniques, we applied ecogenomics approaches in order to unravel the evolutionary history and the processes that drive genome architecture in hallmark freshwater lineages from the phylum Planctomycetes. The evolutionary history inferences showed that sediment/soil Planctomycetes transitioned to aquatic environments, where they gave rise to new freshwater-specific clades. The most abundant lineage was found to have the most specialised lifestyle (increased regulatory genetic circuits, metabolism tuned for mineralization of proteinaceous sinking aggregates, psychrotrophic behaviour) within the analysed clades and to harbour the smallest freshwater Planctomycetes genomes, highlighting a genomic architecture shaped by niche-directed evolution (through loss of functions and pathways not needed in the newly acquired freshwater niche).

Fig. 1

Taxonomic milieu of Planctomycetes phylum in worldwide lacustrine habitats. The figure depicts the SILVA SSU (Ref NR 99 128) classification of 16S rRNA gene fragments (as unassembled shotgun reads) retrieved from 64 freshwater data sets. The X-axis shows the taxonomic ranks and the geographic distribution of the sample collection sites, while the Y-axis indicates the percentage of Planctomycetes within the prokaryotic communities (as assessed by 16S rRNA genes abundance). The sample collection time, following a four-seasons breakdown, is indicated by colored boxes arranged along the X-axis. The SRA identifier for each metagenome is indicated in the parentheses that follow the habitat name. The figure’s inset (upper right panel) shows the contribution of Planctomycetes (as assessed by 16S rRNA gene abundance in 298 metagenomic data sets) to the prokaryotic communities present in aquatic and freshwater sediments (64 lacustrine, 36 fluvial, 34 epipelagic, 46 deep chlorophyll maxima, 16 mesopelagic, 62 bathypelagic and 40 sediments). The colored circles highlight taxa that reached more than 1% abundance within prokaryotic communities. DCM: deep chlorophyll maxima

Fig. 2

Phylogenomics of Planctomycetes phyla. The left panel shows accurate whole-genome phylogenies through a maximum likelihood (phylogenomic) tree inferred from 138 genomes (complete and partial). The topology of the tree emphasizes the major phylogenomic groups found in lacustrine habitats (for details regarding tree inference see Methods). The names of the 60 metagenome-assembled genomes (MAGs), obtained in this study, are highlighted in boldface, while the culture-derived genomes (references) and other available MAGs are depicted in italic and roman type, respectively. The strength of support for internal nodes was assessed by performing bootstrap replicates, with the obtained values shown as colored circles (left legend). Ecological data (i.e., habitat of origin = H) and genomic characteristics (coding density = D, genome size = S, and completeness = C) are indicated by colored circles for each branch in the tree (top left legend). The relations between the genomic characteristics (i.e. estimated genome size, coding density, GC content, mean intergenic spacer length, genome completeness) of MAGs (Phycisphaerae and Planctomycetacia MAGs; see vertical taxonomic delineators) and reference Planctomycetes (31 culture-derived genomes) are shown by linear regressions in the 4 insets present in the right part of the figure. The lowermost insert (right side) shows the iRep values for Phycisphaerae (n = 4) and Planctomycetacia (n = 9) MAGs

Fig. 3

Phylogenomic subtrees a–d generated using maximum-likelihood methods and alignments of concatenated conserved proteins (54, 20, 206 and 315 proteins). The black colored branches designate aquatic groups, while the grey ones their closest relatives (found in soil/sediments). The circular symbols, situated at the tips of the branches, are proportional with genome size and depict gene densities (within genomes). The number of genomes present in the collapsed groups is specified in parenthesis. e Putative model of niche-directed genome evolution in freshwater Planctomycetes

Fig. 4

Spatio-temporal profiles of Planctomycetes relative abundance (horizontal bars), temperature (red line), chlorophyll a (green line) and oxygen (blue line) in Lake Zurich (a) and Rimov Reservoir (b) during 2015. The vertical axis shows the depth (m), within the water column, from which the samples were collected (9 for Lake Zurich and 6 for Rimov Reservoir). The upper X-axis shows the percentage of Phycisphaerae (red bars) and Planctomycetacia (dark cyan) within the prokaryotic communities (estimated as the total sum of DAPI-positive cells), while the lower one displays the values for temperature, chlorophyll a and oxygen. The sampling date is shown above the lower X-axis. c–e Display superimposed images of CARD-FISH-stained Planctomycetes (class Phycisphaerae, family Nemodlikiaceae) and DAPI-stained prokaryotes. The red arrows point towards free-living and particle-associated Planctomycetes, while the yellow ones designate unhybridized prokaryotic cells. The scale bar is 5 µm

Fig. 5

Hypothetical model of multiprotein complex (planctosome) involved in peptide degradation. The complex is tethered to extracellular membrane through a lamin A/C globular tail domain (LTD). The “anchoring” protein (2,210 aa) consists of a N-terminus signal peptide (26 aa) followed by the LTD, multiple proprotein convertase P-domains (PCD) divided by thrombospondin type 3 repeats (TSP), and a cohesin domain (CD). The “adaptor” protein contains a N-terminus signal peptide (26 aa), a dockerin domain (DD), a hyaline repeat domain (HYRD) and a cohesin (CD). The “adapter” binds Zn²⁺-dependent endo- (M12B Reprolysin4-like) and exopeptidases (M14 carboxypeptidase subfamily A) through Ca²⁺-dependent cohesin-dockerin interactions