A Small Microbial Genome: The End of a Long Symbiotic Relationship?

See allHide authors and affiliations

Science  13 Oct 2006:
Vol. 314, Issue 5797, pp. 312-313
DOI: 10.1126/science.1130441


Intracellular bacteria are characterized by genome reduction. The 422,434–base pair genome of Buchnera aphidicola BCc, primary endosymbiont of the aphid Cinara cedri, is ∼200 kilobases smaller than the previously sequenced B. aphidicola genomes. B. aphidicola BCc has lost most metabolic functions, including the ability to synthesize the essential amino acid tryptophan and riboflavin. In addition, most retained genes are evolving rapidly. Possibly, B. aphidicola BCc is losing its symbiotic capacity and is being complemented (and might be replaced) by the highly abundant coexisting secondary symbiont.

Genome reduction in endosymbiotic bacteria is a continuous process derived from their adaptation to intracellular life. One unsolved question is whether reduction reaches a threshold, or if it is an ongoing process that inevitably leads to bacterial extinction and replacement by a new symbiont. Current debate centers on whether genomic streamlining is a result of deletion bias or natural selection, and has implications for the theory of genome complexity evolution (1). The obligate association between aphids and their maternally transmitted intracellular symbiont Buchnera aphidicola offers a model system to analyze genome reduction and its consequences. Genome sizes ranging from 450 to 641 kb have been reported in B. aphidicola strains from different aphid subfamilies, with the genome of B. aphidicola from the aphid Cinara cedri (B. aphidicola BCc) being the most dramatically reduced (2). A particular feature of C. cedri is the presence of large numbers of a secondary symbiont (3), “Candidatus Serratia symbiotica” (S. symbiotica) (4).

The genome comparison of three previously sequenced B. aphidicola strains (57) showed almost total conservation of genome architecture since their last common symbiotic ancestor. Selective gene losses in the extant lineages appear to be mainly related to host-specific properties (6, 7).

The B. aphidicola BCc genome is composed of a 416,380–base pair (bp) circular chromosome plus a 6045-bp plasmid for leucine biosynthesis (tables S1 to S3) (8, 9). Gene loss, scattered along the chromosome (fig. S1), is the main cause of genome shrinkage, because there is no reduction in the sizes of intergenic regions and open reading frames. With only 362 protein-coding genes, this genome represents the minimal known gene set able to support cellular life.

Gene loss affects all functional categories (figs. S1 and S2; table S4), although not evenly. Genes necessary for RNA metabolism (transcription and translation) are the most preserved, representing 35% of the genome's coding capacity. The DNA replication machinery is also complete, but the repair machinery is further reduced than in other strains. Chaperone systems and all essential components for protein translocation are also well preserved, ensuring proper folding and positioning of membrane protein components. These include a highly simplified flagellar apparatus, composed only of those elements homologous to the type III virulence secretion system required for the invasion of the host cells (10).

Gene losses affecting biosynthesis of nucleotides, cofactors, cell envelope, and transport are particularly acute. Hence, B. aphidicola BCc depends entirely on its host for nucleotide and cofactor provisioning. In addition, and in contrast to what has been described in other strains (11), B. aphidicola BCc is clearly unable to provide riboflavin to its host. Finally, it lacks most of the transporters encoded by other B. aphidicola genomes. Because it has also lost all the genes for aminosugar and peptidoglycan biosynthesis, it appears that B. aphidicola BCc must be close to a free-diffusing cell, in which most metabolites can be passively exchanged through a highly simplified cell envelope.

The putative B. aphidicola BCc metabolism inferred from the extant genes (fig. S3) is reduced simply to using glucose to obtain energy through substrate-level phosphorylation, plus the production of saturated fatty acids and all the essential amino acids, except tryptophan. All genes encoding the adenosine 5′-triphosphate synthase subunits have been lost, indicating that the retained components of the electron transport chain must be involved in the regeneration of nicotinamide adenine dinucleotide for glycolysis and acetyl–coenzyme A biosynthesis. In the absence of all genes necessary for the biosynthesis of phospholipids, the preservation of the complete saturated fatty acid pathway indicates that B. aphidicola BCc, and B. aphidicola in general, probably provide them to the host.

Aphids, like other animals, require adequate quantities of 10 essential amino acids that are lacking in their diet and must be provided by their endosymbionts. B. aphidicola BCc has retained the biosynthetic capacity for all essential amino acids except tryptophan. The importance of tryptophan production by the endosymbiont has been experimentally demonstrated (12), and the close relative B. aphidicola BCt, endosymbiont of the aphid Cinara tujafilina, possesses trpE and trpG (the two regulatory genes of the tryptophan pathway) on a plasmid (9). C. cedri and C. tujafilina are almost identical (13), and their plant hosts are also very similar. Yet, B. aphidicola BCt contains the genes for tryptophan biosynthesis whereas B. aphidicola BCc has lost the complete pathway, which suggests that B. aphidicola BCc is not only unable to provide tryptophan to its host, but must obtain it from another source. Although secondary symbionts are considered facultative in other aphids, S. symbiotica is present in all the C. cedri clones we have worked with. They are always contained within well-defined bacteriocytes, are present at a density similar to that of B. aphidicola (Fig. 1; fig. S5), are located in the central part of the bacteriome, and are surrounded by primary bacteriocytes (3). The polymerase chain reaction amplification and sequencing of a trpE fragment from S. symbiotica (8) indicate that this symbiont synthesizes tryptophan and supplies it to the whole symbiotic system.

Fig. 1.

Microscopic analysis of C. cedri bacteriocytes. Semi-thin section showing two types of bacteriocytes, identifiable by their different tonality with toluidin blue. P, primary symbiont (B. aphidicola); S, secondary symbiont (S. symbiotica); n, bacteriocyte nuclei.

The evolution of B. aphidicola BCc sequences appears to have been particularly rapid. In general, the ratio of synonymous to nonsynonymous substitutions, dN/dS, of B. aphidicola protein-coding genes is higher than those of free-living bacteria, owing to an accelerated rate of nonsynonymous substitution (14). This pattern is more marked in B. aphidicola BCc (8) (table S5). Tests of the relative accumulation of nucleotide substitutions performed for all possible B. aphidicola strain pairs (table S6) revealed that the B. aphidicola BCc branch accumulates a significantly higher number of substitutions in most of its genes (table S7). The genes with higher dN/dS ratios are not associated with any particular functional role (fig. S4). Finally, we analyzed the type of selection that operates on protein-coding genes in B. aphidicola (table S8). Most of the genes are under purifying selection (dS > dN), but about 12% of the genes are under neutral selection in B. aphidicola BCc (dSdN), as expected for pseudogenes.

Taking together all functional, evolutionary, and microscopic data, we postulate that B. aphidicola BCc is undergoing a process of genome degradation and functional replacement by the coexisting S. symbiotica (3). Natural symbiont replacement of B. aphidicola by a fungus was postulated to have occurred in aphids of the tribe Cerataphidini (15), whereas experimental evidence of secondary symbionts taking on the role of B. aphidicola has been demonstrated in infection experiments of B. aphidicola–cured aphids (16). All the analyses performed point to a more extreme gene-degradation effect occurring in the B. aphidicola BCc genes than in other B. aphidicola lineages. Indeed, the loss of most DNA-protecting and DNA-repair mechanisms in B. aphidicola BCc, more so than in the other B. aphidicola lineages, would enhance the mutation rate. Further, B. aphidicola BCc has apparently lost its role as a tryptophan and riboflavin supplier to its host and indeed cannot even supply its own needs, which must be provided by S. symbiotica, not only to the host but also to B. aphidicola BCc. Thus, the mutualistic relationship between B. aphidicola and its aphid host seems to have taken on a new, more complex role that includes a second endosymbiont and might end up in a replacement.

Supporting Online Material

Materials and Methods

SOM Text

Figs. S1 to S5

Tables S1 to S8


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article