Population Genomics Insights into Adaptive Evolution and Ecological Differentiation in Streptomycetes

Abstract

Deciphering the genomic variation that represents microevolutionary processes toward species divergence is key to understanding microbial speciation, which has long been under debate. Streptomycetes are filamentous bacteria that are ubiquitous in nature and the richest source of antibiotics; however, their speciation processes remain unknown. To tackle this issue, we performed a comprehensive population genomics analysis on Streptomyces albidoflavus residing in different habitats and with a worldwide distribution and identified and characterized the foundational changes within the species. We detected three well-defined phylogenomic clades, of which clades I and III mainly contained free-living (soil/marine) and insect-associated strains, respectively, and clade II had a mixed origin. By performing genome-wide association studies (GWAS), we identified a number of genetic variants associated with free-living or entomic (denoting or relating to insects) habitats in both the accessory and core genomes. These variants contributed collectively to the population structure and had annotated or confirmed functions that likely facilitate differential adaptation of the species. In addition, we detected higher levels of homologous recombination within each clade and in the free-living group than within the whole species and in the entomic group. A subset of the insect-associated strains (clade III) showed a relatively independent evolutionary trajectory with more symbiosis-favorable genes but little genetic interchange with the other lineages. Our results demonstrate that ecological adaptation promotes genetic differentiation in S. albidoflavus, suggesting a model of ecological speciation with gene flow in streptomycetes.IMPORTANCE Species are the fundamental units of ecology and evolution, and speciation leads to the astounding diversity of life on Earth. Studying speciation is thus of great significance to understand, protect, and exploit biodiversity, but it is a challenge in the microbial world. In this study, using population genomics, we placed Streptomyces albidoflavus strains in a spectrum of speciation and showed that the genetic differences between phylogenomic clusters evolved mainly by environmental selection and gene-specific sweeps. These findings highlight the role of ecology in structuring recombining bacterial species, making a step toward a deeper understanding of microbial speciation. Our results also raise concerns of an underrated microbial diversity at the intraspecies level, which can be utilized for mining of ecologically relevant natural products.

Keywords: adaptive evolution; ecological differentiation; population genomics; speciation; streptomycetes.

FIG 1

The pan-genome, core genome, and accessory genome profiles of S. albidoflavus. (A) The sizes of core and pan-genomes in relation to numbers of genomes added into the gene pool. Box plots show the 25th and 75th percentiles, with medians shown as horizontal lines, and whiskers indicate the lowest and highest values within 1.5 times the interquartile range (IQR) from the first and third quartiles, respectively. The curve for the pan-genome is fitted by the power-law regression model (y_pan = A_pan x^Bpan + C_pan), with r² = 0.999, A_pan = 1488.57 ± 1.67, B_pan = 0.37, and C_pan = 4,502.37 ± 4.34. B_pan is equivalent to the parameter γ, and α (= 1 − γ) < 1 indicates that the pan-genome does not approach a constant as more genomes are sampled. The curve for the core genome is fitted by the exponential curve fit model (y_core = A_core e^Bcore.x + C_core), with r² = 0.938, A_core = 1,060.74 ± 21.92, B_core = −0.13, and C_core = 4,838.38 ± 5.57. (B) Distribution of genes across strains.

FIG 2

Phylogeny and population structure of S. albidoflavus. Strains collected from different habitats are represented in different colors (red, soil or plants; green, insects; blue, sea; black, uncertain). “I,” “II,” or “III” indicates genetic clusters or clades. (A) Maximum-likelihood phylogeny generated from concatenated 4,116 single-copy core genes of the 33 strains, including the three outgroup strains. Bootstrap values less than 100% are shown in gray at the nodes. The scale bar indicates 1% sequence divergence. Ancestral genome contents of S. albidoflavus were reconstructed using COUNT (81). The numbers of gene gain (+) and loss (−) events are indicated next to the deep nodes, and the total numbers of ancestral orthologous genes present at each of the nodes are shown in parentheses. (B) Hierarchical cluster analysis based on the presence or absence of dispensable genes in the 33 strains. Numbers above branches are bootstrap support values from 1,000 replicates. Height indicates the dissimilarity between genomes. (C) Structure plot based on concatenated 4,663 single-copy core genes of the 30 S. albidoflavus strains, showing the contribution to each S. albidoflavus strain from each of the three hypothetical ancestral populations.

FIG 3

Distribution of the 226 accessory OGs significantly associated with free-living (soil/marine) (A) and entomic (denoting or relating to insects) (B) habitats. The OGs are ordered according to their relative locations in the genomes of strains NBRC 100770 (A) and CR33 (B). Dark gray and colored boxes indicate presence, with particularly described OGs in color (see key). Light gray boxes indicate absence. The tree on the left was derived from Fig. 2A; phylogenomic clades are indicated. Strains collected from different habitats are represented in different colors (red, soil or plants; green, insects; blue, sea; black, uncertain).

FIG 4

Genetic organizations of the regions containing habitat-associated gene clusters. Details of the habitat-associated accessory genes are shown in Data Set S1. Differential genes between the two ecological groups are marked with yellow background. (A) The nan gene cluster for sialic acid catabolism. (B) The gvp gene cluster for gas vesicle biosynthesis. Red lines between the two groups indicate corresponding homologous genes. (C) The gene clusters for griseobactin biosynthesis and the cobalt transport system. Genes griE, dhbA, dhbB, dhbG, and cbiO are habitat unassociated because some strains of clade I also contain corresponding homologous genes with low identity. DHBA, 2,3-dihydroxybenzoate; NRPS, nonribosomal peptide synthetase.

FIG 5

Functional verifications of sialic acid catabolic and griseobactin biosynthetic gene clusters. (A) The ability of representative free-living strains and insect-associated strains to utilize sialic acid as a sole carbon source. Strains were cultured for 5 days at 28°C on basal mineral salt medium agar plates supplemented with 1% N-acetylneuraminic acid or 1% glucose as a sole carbon source. (B) Arnow’s test of catechol-type siderophore production by representative insect-associated strains and free-living strains. The presence of a red color in the solution was recorded as a positive test for catechol.

FIG 6

Comparison of genetic organizations of the three genotypes in the versatile genomic region. The habitat-unassociated type exists in 8 strains (DSM 40233, D62, NBRC 13083, J1074, DSM 40455^T, NBRC 100770, “S. wadayamensis” A23, and CR13) from clade I and all strains from clade II, the free-living type exists in the remaining 6 strains from clade I and all outgroup strains, and the entomic type exists in all strains from clade III. Differential genes between the genotypes are marked with yellow background.

FIG 7

Density of habitat-associated SNPs along the core genome. The density of these SNPs was plotted along the core genome using a 5-kb sliding window (with an overlap of 2.5 kb between consecutive windows). The red line shows the average density of habitat-associated SNPs per 5 kb.

FIG 8

Map depicting the sampling locations of strains used in this study. The location symbol colors indicate different habitats (red, soil or plants; green, insects; blue, sea; black, uncertain), and the text shadow colors indicate the different phylogenomic clades detected in the species (red, clade I; gray, clade II; green, clade III). The geographic origin of strain J1074 is unavailable. The map is derived from Rawpixel Ltd.