Massive genome erosion and functional adaptations provide insights into the symbiotic lifestyle of Sodalis glossinidius in the tsetse host

Abstract

Sodalis glossinidius is a maternally transmitted endosymbiont of tsetse flies (Glossina spp.), an insect of medical and veterinary significance. Analysis of the complete sequence of Sodalis' chromosome (4,171,146 bp, encoding 2,432 protein coding sequences) indicates a reduced coding capacity of 51%. Furthermore, the chromosome contains 972 pseudogenes, an inordinately high number compared with that of other bacterial species. A high proportion of these pseudogenes are homologs of known proteins that function either in defense or in the transport and metabolism of carbohydrates and inorganic ions, suggesting Sodalis' degenerative adaptations to the immunity and restricted nutritional status of the host. Sodalis possesses three chromosomal symbiosis regions (SSR): SSR-1, SSR-2, and SSR-3, with gene inventories similar to the Type-III secretion system (TTSS) ysa from Yersinia enterolitica and SPI-1 and SPI-2 from Salmonella, respectively. While core components of the needle structure have been conserved, some of the effectors and regulators typically associated with these systems in pathogenic microbes are modified or eliminated in Sodalis. Analysis of SSR-specific invA transcript abundance in Sodalis during host development indicates that the individual symbiosis regions may exhibit different temporal expression profiles. In addition, the Sodalis chromosome encodes a complete flagella structure, key components of which are expressed in immature host developmental stages. These features may be important for the transmission and establishment of symbiont infections in the intra-uterine progeny. The data suggest that Sodalis represents an evolutionary intermediate transitioning from a free-living to a mutualistic lifestyle.

Figure 1.

Circular representation of the Sodalis glossinidius chromosome. The outer scale (indicated by numbers 1–4) is shown in megabases. From the outside in: circle 1, positive strand CDSs (red); circle 2, negative strand CDSs (blue); circle 3, pseudogenes (gray); circle 4, tRNA genes (green), rRNA genes (brown), genes for type III secretion systems (pink); circle 5, GC skew ([G-C]/[G+C]; khaki indicates values >0; purple indicates values <0); and circle 6, G+C content (higher values outward).

Figure 2.

CDSs in each size category, represented as a percentage of the total number of CDSs present (the legend indicates the size, in amino acids, of the corresponding encoded putative proteins). For each indicated bacteria, CDSs encoding putative proteins >100 amino acids were counted and used in this analysis. S. glossinidius CDS1 indicates the number of total CDSs in each category, while CDS2 excludes the pseudogenes assigned in this study.

Figure 3.

Comparative analysis of the number of genes present in each functional category as defined by the COG database for S. glossinidius, E. coli K-12, S. typhimurium, and Y. pestis. The number of genes in each COG category in Sodalis rendered nonfunctional (pseudogene) is denoted by the gray shaded area.

Figure 4.

Overview of Sodalis's general metabolic capabilities deduced from chromosomally located genes and their putative functions.

Figure 5.

Phylogenetic relationship shared between Sodalis' three TTSS structures and those characterized from pathogenic microbes. The tree was based on amino acid sequence alignments of the invA gene product. Scale bars representing branch lengths and bootstrap values are displayed at each internal node. Genes associated with each cluster are depicted as arrows that indicate the direction of transcription. The arrows with a cross denote pseudogenes. The light blue bars between loci indicate the regions of sequence similarity and gene order conservation. The different colors depict the functional roles of their putative products.

Figure 6.

Analysis of gene expression representative of SSR, flagella, and putative effector functions during host development. Transcript abundance levels were determined for invA1, invA2, fliC, motA, and hemolysin expressed at the following host developmental stages: 1, early larvae; 2, late larvae; 3, early pupae; 4, late pupae; 5, 6-d-old adults; and 6, 40-d-old fertile females. Results for each gene were normalized against the host developmental stage, during which their expression was least abundant. The data are presented as bar graphs, which represent fold difference in transcript abundance. No invA3 transcripts could be amplified from these cDNA pools tested (data not shown).