Differential genome evolution between companion symbionts in an insect-bacterial symbiosis

Abstract

Obligate symbioses with bacteria allow insects to feed on otherwise unsuitable diets. Some symbionts have extremely reduced genomes and have lost many genes considered to be essential in other bacteria. To understand how symbiont genome degeneration proceeds, we compared the genomes of symbionts in two leafhopper species, Homalodisca vitripennis (glassy-winged sharpshooter [GWSS]) and Graphocephala atropunctata (blue-green sharpshooter [BGSS]) (Hemiptera: Cicadellidae). Each host species is associated with the anciently acquired "Candidatus Sulcia muelleri" (Bacteroidetes) and the more recently acquired "Candidatus Baumannia cicadellinicola" (Gammaproteobacteria). BGSS "Ca. Baumannia" retains 89 genes that are absent from GWSS "Ca. Baumannia"; these underlie central cellular functions, including cell envelope biogenesis, cellular replication, and stress response. In contrast, "Ca. Sulcia" strains differ by only a few genes. Although GWSS "Ca. Baumannia" cells are spherical or pleomorphic (a convergent trait of obligate symbionts), electron microscopy reveals that BGSS "Ca. Baumannia" maintains a rod shape, possibly due to its retention of genes involved in cell envelope biogenesis and integrity. Phylogenomic results suggest that "Ca. Baumannia" is derived from the clade consisting of Sodalis and relatives, a group that has evolved symbiotic associations with numerous insect hosts. Finally, the rates of synonymous and nonsynonymous substitutions are higher in "Ca. Baumannia" than in "Ca. Sulcia," which may be due to a lower mutation rate in the latter. Taken together, our results suggest that the two "Ca. Baumannia" genomes represent different stages of genome reduction in which many essential functions are being lost and likely compensated by hosts. "Ca. Sulcia" exhibits much greater genome stability and slower sequence evolution, although the mechanisms underlying these differences are poorly understood.

Importance: In obligate animal-bacterial symbioses, bacteria experience extreme patterns of genome evolution, including massive gene loss and rapid evolution. However, little is known about this process, particularly in systems with complementary bacterial partners. To understand whether genome evolution impacts symbiont types equally and whether lineages follow the same evolutionary path, we sequenced the genomes of two coresident symbiotic bacteria from a plant sap-feeding insect and compared them to the symbionts from a related host species. We found that the older symbiont has a highly reduced genome with low rates of mutation and gene loss. In contrast, the younger symbiont has a larger genome that exhibits higher mutation rates and varies dramatically in the retention of genes related to cell wall biogenesis, cellular replication, and stress response. We conclude that while symbiotic bacteria evolve toward tiny genomes, this process is shaped by different selection intensities that may reflect the different ages and metabolic roles of symbiont types.

FIG 1

Comparison of “Ca. Baumannia cicadellinicola” genomes from two different sharpshooter leafhopper host species, Graphocephala atropunctata (blue-green sharpshooter [BGSS]) and the previously sequenced Homalodisca vitripennis (glassy-winged sharpshooter [GWSS]) (11). Genes are color-coded as core shared genes or unique to BGSS or GWSS “Ca. Baumannia” as indicated in the key. The genomes are completely syntenic except for large contiguous gene deletions. The graph on the inner track shows genomewide GC skew, and the outer track shows predicted genes for each genome. Genes are color-coded according to whether they are unique to a genome (e.g., blue for unique BGSS genes) or whether they are shared between genomes (grey). Black tick marks around the outer track show inferred origins of replication.

FIG 2

Bar chart showing distribution of functional categories of clusters of orthologous groups (COGs) for predicted protein-coding genes for “Ca. Baumannia cicadellinicola” and “Ca. Sulcia muelleri” from BGSS and GWSS. Bars are color coded according to genome (see key). ORFS, open reading frames.

FIG 3

Schematics showing the presence and absence of genes involved in central cellular processes for the obligate symbionts of the Auchenorrhyncha. Colored boxes indicate gene presence, white boxes show gene absence, and hatched boxes show predicted pseudogenes. “Ca. Baumannia cicadellinicola” (B) and “Ca. Sulcia muelleri” (S) symbionts are abbreviated according to their hosts (BGSS and GWSS). Hodgkinia or Hodg. is “Candidatus Hodgkinia cicadicola,” and Nasuia is “Candidatus Nasuia deltocephalinicola.” A single “Ca. Sulcia muelleri” strain is shown as representative for the sharpshooter strains (S-BGSS), which have lost most of these major pathways in all auchenorrhynchan hosts. Genomes are arranged by size from smallest (inner ring) to largest (outer ring). The genomes sequenced in this study are highlighted in red.

FIG 4

Phylogenetic relationships and transmission electron micrographs (TEM) of “Ca. Baumannia cicadellinicola” strains. (A to C) Phylogenetic relationships of “Ca. Baumannia cicadellinicola” from across sharpshooter leafhoppers inferred with 16S rRNA sequences from previous studies (33, 34). Colored arrows show the focal “Ca. Baumannia cicadellinicola” strains and corresponding TEM images for B-BGSS (B) and B-GWSS (C). (D) Phylogenomic placement of the investigated “Ca. Baumannia cicadellinicola” genomes within the Enterobacteriaceae (Gammaproteobacteria) and as sister to the Sodalis-like clade (n = 4, see Fig. S3 for specific Sodalis species). Asterisks next to symbiont names show possible artifacts of long-branch attraction (66). Riesia pediculicola is “Candidatus Riesia pediculicola.”

FIG 5

Distribution of genomewide synonymous (dS) and nonsynonymous (dN) rates of molecular evolution in “Ca. Sulcia muelleri” (blue) and “Ca. Baumannia cicadellinicola” (orange). Rates are inferred from pairwise comparisons of all shared orthologs between the symbiont strains obtained from BGSS and GWSS. Adjacent box plots show median, quartiles, and maximum and minimum distributions for dS and dN values on the x- and y-axes, respectively. Box plots are color-coded according to genome.