Pangenomics reveals alternative environmental lifestyles among chlamydiae

Abstract

Chlamydiae are highly successful strictly intracellular bacteria associated with diverse eukaryotic hosts. Here we analyzed metagenome-assembled genomes of the "Genomes from Earth's Microbiomes" initiative from diverse environmental samples, which almost double the known phylogenetic diversity of the phylum and facilitate a highly resolved view at the chlamydial pangenome. Chlamydiae are defined by a relatively large core genome indicative of an intracellular lifestyle, and a highly dynamic accessory genome of environmental lineages. We observe chlamydial lineages that encode enzymes of the reductive tricarboxylic acid cycle and for light-driven ATP synthesis. We show a widespread potential for anaerobic energy generation through pyruvate fermentation or the arginine deiminase pathway, and we add lineages capable of molecular hydrogen production. Genome-informed analysis of environmental distribution revealed lineage-specific niches and a high abundance of chlamydiae in some habitats. Together, our data provide an extended perspective of the variability of chlamydial biology and the ecology of this phylum of intracellular microbes.

Fig. 1. MAGs from diverse environments expand known and add previously undescribed clades in the phylum Chlamydiae.

a Maximum likelihood phylogenetic tree based on a concatenated set of 43 conserved marker proteins (5704 sites) in which published genomes and 82 MAGs generated in the GEM initiative are shown in black and orange, respectively. Chlamydial monophyly was supported by optimized ultrafast bootstrap and SH likelihood ratio test support with 100% for both. Previously established chlamydial families are shaded in light gray, previously undescribed families are shaded in orange. The tree was inferred under the LG + C60 + G4 + F model with the IQ-TREE software. Nodes with an optimized ultrafast bootstrap support ≥ 95% are labelled with black circles. Tree annotations from inside to outside: (1) completeness, (2) contamination, (3) MAG with 16S rRNA gene, (4) high quality MAGs, (5) environmental origin (white stars indicate genomes from cultured isolates), (6) %GC content, (7) assembly size and estimated genome size (stacked white and gray bars, respectively), and (8) names of chlamydial families represented by more than ten genomes and added metagenomic clades indicated by orange segments. Scale bar indicates 0.1 substitutions per position in the alignment. b Number of MAGs retrieved per environmental category. c Completeness and contamination estimates of chlamydial MAGs from the GEM dataset. Shaded gradients behind the completeness and contamination boxplots represent the values in the heatmap boxes in the tree.

Fig. 2. MAGs from the GEM dataset broadly populate the taxonomy of the Chlamydiae at family-, genus-, and species rank.

MAGs from the GEM catalog significantly extend known chlamydial taxa, including 7 additional families, 34 genera, and 44 species, highlighting the taxonomic heterogeneity of the phylum. Packed circles represent chlamydiae taxonomic ranks and their higher level taxonomic structure. From the outermost to the innermost circle the family, genus, and species ranks are depicted. Violet indicates lineages with previously known genome representatives (family, genus, species rank in dark, medium, and light violet, respectively), while added lineages are shown in orange (family, genus, species in dark, medium, and light orange, respectively). The number in brackets next to the family names indicates the number of genome sequences available. a Known and previously undescribed chlamydial families containing MAGs from the GEM catalog. Bar charts represent the number of families, genera, and species recruited in this study. b Families without genome sequences from this study.

Fig. 3. Chlamydiae show conserved features of an intracellular lifestyle but versatility in oxygen adaptation.

The presence of selected genes and pathways across the chlamydial core and accessory genome is depicted. The phylogenetic tree includes 96 high quality genomes used for pangenome analysis and additional representatives (n = 109 in total). The tree is based on a concatenated set of 43 conserved marker proteins (6268 sites) and was inferred under the LG + C60 + G4 + F derived PMSF approximation by the IQ-TREE software. Branch support values are based on 100 non-parametric bootstraps, support ≥ 70% is indicated as black circles. MAGs from the GEM catalog are indicated by orange branch colors. Colored circles show full or partial presence of selected genes or metabolic pathways. Pyruvate fermentation refers to the presence of the full pathway for pyruvate fermentation to acetate and is differentiated based on the presence of the enzyme for acetyl-CoA generation from acetate, i.e., pyruvate dehydrogenase complex (PDC), or pyruvate ferredoxin oxidoreductase (PFO) together with phosphate acetyltransferase and acetate kinase, or acetate-CoA ligase. The arginine deiminase (ADI) pathway is only indicated if arginine deiminase, ornithine carbamoyltransferase, and carbamate kinase were found. Bar chart shows the completeness of the tricarboxylic acid cycle (TCA). Genes encoding nucleotide transport proteins (ntt); early upstream ORF (euo), transcriptional regulator of the chlamydial developmental cycle; histone-like developmental protein (hctA); serine/threonine protein kinase CopN (copN); pseudokinase Pkn5 (pkn5); glucose 6-phosphate transporter (uhpC); type IV secretion system (T4SS); type III secretion system (T3SS).

Fig. 4. Independent acquisition of light-driven ATP synthesis in two potentially amoeba-associated clades.

a Gene synteny plot of proteorhodopsin related gene clusters in Parachlamydiaceae MAGs. Comparisons are ordered according to the phylogenomic species tree in Fig. 1. Arrows colored in orange, yellow, and blue represent proteorhodopsin (prd), carotene biosynthesis, and circadian clock genes, respectively. Black arrows indicate genes with chlamydial homologs. Bands connect homologs and are colored according to their protein identity. All other proteins of contigs encoding proteorhodopsin gene clusters were blasted against the NCBI non-redundant (nr) database to confirm the chlamydial origin of the contig. b Maximum likelihood phylogenetic tree of proteorhodopsin (Prd) (ENOG4105CSB) with chlamydial sequences showing two distinct clades. Maximum likelihood tree was inferred under LG + C30 + G + F model with 1000 improved ultrafast bootstraps and 1000 replicates of the SH-like approximate likelihood ratio test. Filled circles at nodes indicate a bootstrap support ≥ 95%. Scale bar indicates the number of substitutions per site.

Fig. 5. Widespread fermentation pathways and molecular hydrogen production in chlamydiae.

a Representation of putative anaerobic pathways for ATP generation and molecular hydrogen metabolism in chlamydiae. Labels next to enzymatic reactions indicate the associated enzymes. Numbers in squares correspond to phylogenetic trees in (c). Colors indicate affiliation with different pathways—pyruvate fermentation (yellow), ADI pathway (violet), hydrogen metabolism (orange). b Species tree of chlamydial representative genomes as in Fig. 3 collapsed at the family rank. Branch support values are based on 100 non-parametric bootstraps, support ≥ 70% is indicated as black circles. Box next to family names indicates the number of non-redundant genomes in a family with the respective color coded metabolic pathway. Pyruvate fermentation to acetate was only counted if genes encoding the complete pathway were present, i.e., pyruvate dehydrogenase complex (PDC), or pyruvate ferredoxin oxidoreductase (PFO) together with phosphate acetyltransferase (Atp) and acetate kinase (AckA), or acetate-CoA ligase (Acs). Likewise, arginine deaminase (ADI) pathway was only counted if genes encoding arginine deaminase (ADI), ornithine carbamoyltransferase (OTC), and carbamate kinase (CK) were present in a genome. c Unrooted maximum likelihood phylogenetic trees with best fit models numbered and colored according to (a) with 1000 optimized ultrafast bootstrap and 1000 SH-like approximate likelihood ratio test support. Best fit models per gene are indicated under the gene name and clades are named by family.

Fig. 6. Members of novel chlamydial families predominantly occur in freshwater and marine environments.

a 16S rRNA gene maximum likelihood phylogenetic tree using near full-length sequences de-replicated at 99% under the SYM + R10 model inferred with IQTREE. Support was inferred from transfer bootstrap expectation (TBE) based on 100 non-parametric bootstraps. Circles at nodes indicate TBE support ≥70. The tree is pruned and does not include the outgroup. Chlamydial families are highlighted by gray background, 16S rRNA genes from novel MAGs in this study are indicated by orange shading. b All chlamydial full-length 16S rRNA genes in chlamydial genomes were used as a query against IMNGS with an identity cutoff of 1% to ensure species-specificity and summarized at the family level. Stars next to names indicate families with cultured representatives. Environmental categories “marine” and “freshwater” represent samples originating from the water column. The scatter plot on the left shows the relative abundance of chlamydial 16S rRNA gene amplicons. The bar plot in the middle shows the relative distribution of family members across diverse environments. The bar plot on the right indicates significant enrichment (adjusted p value ≤ 0.05) in environments based on one-sided Fisher’s exact test with false discovery rate adjusted p values expressed as z-scores. c Relationship of relative abundance of the anaerobic family MCF-E with oxygen concentration and depth in samples from Saanich Inlet. Y-axis depicts depth in meters below surface, top x-axis indicates molarity of oxygen and bottom x-axis indicates relative abundance in percent of total 16S rRNA amplicons. Dark blue and light blue areas depict mean oxygen concentration and standard deviation, respectively. Gray filled points, black line, and gray area represent relative abundance in a sample, mean relative abundance, and standard deviation, respectively.