Flexible genomic island conservation across freshwater and marine Methylophilaceae

Abstract

The evolutionary trajectory of Methylophilaceae includes habitat transitions from freshwater sediments to freshwater and marine pelagial that resulted in genome reduction (genome-streamlining) of the pelagic taxa. However, the extent of genetic similarities in the genomic structure and microdiversity of the two genome-streamlined pelagic lineages (freshwater "Ca. Methylopumilus" and the marine OM43 lineage) has so far never been compared. Here, we analyzed complete genomes of 91 "Ca. Methylopumilus" strains isolated from 14 lakes in Central Europe and 12 coastal marine OM43 strains. The two lineages showed a remarkable niche differentiation with clear species-specific differences in habitat preference and seasonal distribution. On the other hand, we observed a synteny preservation in their genomes by having similar locations and types of flexible genomic islands (fGIs). Three main fGIs were identified: a replacement fGI acting as phage defense, an additive fGI harboring metabolic and resistance-related functions, and a tycheposon containing nitrogen-, thiamine-, and heme-related functions. The fGIs differed in relative abundances in metagenomic datasets suggesting different levels of variability ranging from strain-specific to population-level adaptations. Moreover, variations in one gene seemed to be responsible for different growth at low substrate concentrations and a potential biogeographic separation within one species. Our study provides a first insight into genomic microdiversity of closely related taxa within the family Methylophilaceae and revealed remarkably similar dynamics involving mobile genetic elements and recombination between freshwater and marine family members.

Keywords: Methylophilaceae; cultivation; genome-streamlined bacteria; genomic islands; genomic microdiversity; genomics.

Figure 1

Phylogenomic and pangenome analyses of Methylophilaceae; (A) phylogenomic maximum likelihood tree of genomes used in this study; the tree was constructed with IQ-TREE with ultrafast bootstrapping using 851 common proteins; seven Methylotenera spp. genomes were used to root the tree; numbers in brackets indicate the number of genomes in each collapsed branch; a full version of this tree can be found as Fig. S1; (B) pangenome analysis of the three “Ca. Methylopumilus” and OM43 species. Symbols represent the core (diamonds) and pangenome (circles) on a species level; OM43-a was not included in this analysis due to the availability of only one genome.

Figure 2

Occurrence of “Ca. Methylopumilus” and OM43 in metagenomic samples; (A) Metagenomic read recruitment (coverage per Gb at 95% identity) of “Ca. M. rimovensis,” “Ca. M. planktonicus,” and “Ca. M. universalis" in epi- and hypolimnetic samples of 13 central European lakes of different trophic states and seasonal samples of the Římov reservoir from 2015 to 2019. Different seasons are color-coded at the top. Mean values for each species are shown, and the full dataset can be found in Table S3; E: Epilimnion (5 m depth); H: Hypolimnion (variable depths, see Table S1); (B) coverage per Gb boxplots of individual species in epi- and hypolimnetic samples. Boxes indicate the 25th and 75th quantiles, medians are displayed by central lines, whiskers indicate the 5th and 95th quantiles, outliers are displayed by open circles; significant differences are displayed as asterisks (*P < .05; ** P < .01; ***P < .001); Mrim: “Ca. M. rimovensis”; Mpla: “Ca. M. planktonicus”; Muni: “Ca. M. universalis"; (C) coverage per Gb boxplots of individual species in oligo-, meso-, and eutrophic lakes. Trophic states are color-coded as in (A) and significant differences are displayed as asterisks as in (B); (D) Spearman correlations between metagenomic recruitment values and environmental factors across the 13 central European lakes and the Římov reservoir; significant differences are displayed as asterisks as in (B); Temp: water temperature (°C); Chla: chlorophyll a concentrations (μg l⁻¹); Cond: conductivity (μS cm⁻¹); DOC: dissolved organic carbon concentrations (mg l⁻¹); NO₃: nitrate concentrations (mg l⁻¹); Green: chlorophyll a concentrations of green algae (μg l⁻¹); Bluegreen: chlorophyll a concentrations of cyanobacteria (μg l⁻¹); Diat: chlorophyll a concentrations of diatoms (μg l⁻¹); Crypt: chlorophyll a concentrations of cryptophytes (μg l⁻¹); (E) metagenomic read recruitment (coverage per Gb at 95% identity) of OM43-a, b, c, and d in samples of the Arctic Ocean and the Delaware Bay estuary.

Figure 3

Flexible genomic islands (fGIs) in Methylophilaceae. Genomic alignment (BLASTn) of genomes following the phylogenomic tree from Fig. 1A; dS profiles of each species are shown below each alignment group using one genome as reference; the location of tRNA and rRNA genes are shown as squares; the approximate location of the three major fGIs are outlined as dashed boxes; locations of other insertion sites and the hemolysin cluster are displayed as boxes; Methylotenera spp., OM43-a (HTTC2181), “Ca. M. rimovensis” RI-55 were omitted from the analyses due to their high divergence, and 22 clonal genomes of “Ca. M. universalis" are not shown due to redundancy in their alignment; a full version including all genomes can be found as Fig. S6.

Figure 4

Common patterns of the different flexible genomic islands (fGIs); (A) Venn diagram of shared and unique proteins in the replacement fGI (fGI-1); (B) COG categories assigned to proteins in the replacement fGI of each species; (C) the replacement fGI is flanked by a tRNA-Ser(gga) gene on one end and by a tRNA-Arg(cct) on the other with a tmRNA gene in the middle of the island; this pattern is common to all “Ca. Methylopumilus”, OM43, and some Methylotenera sp. genomes; (D) Venn diagram of shared and unique proteins in the additive fGI (fGI-2); (E) COG categories assigned to proteins in the additive fGI of each species; (F) the additive fGI is identified by recurring repeats of the last 20 base pairs (ACCAGCTGAGCTAATCCCCC) of the tRNA-Lys(ctt) gene present in all “Ca. M. universalis" and “Ca. M. planktonicus” genomes; (G) Venn diagram of shared and unique proteins in the tycheposon (fGI-3); (H) COG category assigned to proteins in the tycheposon of each species; (I) the tycheposon starts with a tRNA-met(cat), contains tyrosine recombinase xerD, and ends at the first conserved gene (glmS—glucosamine 6-phosphate synthetase). It is common to all “Ca. Methylopumilus” and OM43 genomes and can be separated in a conserved part and a part with potential genetic variability; (J) a tree of 86 common proteins (see Table S2) of the tycheposon of “Ca. Methylopumilus” separates “Ca. M. universalis" in strains isolated from lakes in Czechia and other countries, respectively.

Figure 5

Growth of “Ca. Methylopumilus” influenced by different methanol concentrations; (A) growth rates (day⁻¹) of seven strains of “Ca. M. planktonicus” and 10 strains of “Ca. M. universalis” (4 strains isolated from lakes in Czechia and 6 strains from other countries) under increasing methanol concentrations (0, 0.0001, 0.001, 0.02, 0.05, 0.1 and 1 mM); significant differences are indicated by asterisks (*P < .05; **P < .01); individual growth curves are shown in Fig. S10; (B) maximum likelihood tree of OstA (organic solvent tolerance protein) of Methylophilaceae; the tree was constructed with IQ-TREE with ultrafast bootstrapping; OM43-a HTCC2181 was excluded due to high sequence divergence.