Decoupling of Genome Size and Sequence Divergence in a Symbiotic Bacterium

Abstract

In contrast to genome size variation in most bacterial taxa, the small genome size of Buchnera sp. was shown to be highly conserved across genetically diverse isolates (630 to 643 kb). This exceptional size conservation may reflect the inability of this obligate mutualist to acquire foreign DNA and reduced selection for genetic novelty within a static intracellular environment.

Bacteria may undergo significant changes in genome size through the duplication or loss of existing chromosome fragments and through the acquisition of exogenous DNA, including plasmids or other accessory elements. Because the vast majority of a bacterial chromosome consists of coding sequences (4), changes in genome size reflect differences in gene content. The variation among bacterial genome sizes, ranging from 0.6 to 9 Mb (15), reflects differences in biochemical capabilities and, hence, in the range of environments available to particular microbial lineages.

Based on early comparisons of genetic maps, the size, organization, and gene content of bacteria were viewed as being evolutionarily conserved (20). However, the physical mapping of bacterial genomes by pulsed-field gel electrophoresis (PFGE) has provided overwhelming evidence that the size and structure of the bacterial chromosome varies both within and among species (6). Genome sizes of very closely related taxa can differ by as much as 100%, often by more than a megabase in total length (Table 1). Recent evidence from these studies and from analyses of full genome sequences indicates that genome size and organization are more evolutionarily labile than gene sequences (12). This lability extends both to those bacteria with relatively large genomes, such as the enteric bacteria and Bacillus species, and to those with small genomes, such as the mycoplasmas and rickettsiae (Table 1).

TABLE 1.

Genome size range and divergence in 16S rDNA for representative bacterial taxa

Organism	% 16S rDNA divergence	Genome size range (Mb)	% Difference in genome size	Source or reference (for genome size)
Free-living/opportunistic pathogens
Escherichia coli	<1	4.5–5.5	22	4
Salmonella typhi	<1	3.9–4.9	25	23
Bacillus cereus	1	2.4–5.3	121	5
Chronic pathogens
Spotted fever group of Rickettsia spp.	<3	1.2–2.1	75	21
Mycoplasma hominis	<1	0.70–0.83	19	14, 18
Mycoplasma genitalium-M. pneumoniae	2	0.58–0.82	41	9, 11
Chlamydia psittaci-C. trachomatis	5	1.0–2.65	165	10, 22
Vertically transmitted symbiont
Buchnera aphidicola	6–8	0.63–0.64	<3	7; this study

Bacterial genome size is determined by a balance between the introduction of new sequences, through duplication or horizontal transfer, and the loss of genes, through mutations that eliminate function and subsequent deletions. Whether or not a given gene persists within a lineage depends upon the selection coefficient associated with its functional role, the effective population size, and mutation rate (13). Thus, in bacteria with small effective population sizes and high mutation rates, the selection required to maintain a given gene increases. The expected dependence of genome size on population size and mutation rate is supported by the repeated observation of reduced genome sizes in intracellular pathogenic bacteria (2, 25). Chronic pathogens and symbionts may experience small population sizes due to bottlenecks during infection of hosts (1, 16) and higher per-site mutation rates (19). These factors would increase the selection coefficient required to maintain a gene, so that only the most essential genes persist, resulting in small genomes as observed for these organisms.

Intracellular bacteria might be expected to show conservation of genome size if they retain only essential genes for a specialized lifestyle. However, although they show consistently small genomes, intracellular pathogens display wide variation in chromosome length (Table 1). This variation suggests the incidence of lateral transfer of DNA, as supported by findings of gene transfer between distantly related intracellular pathogens (24).

The intracellular mutualistic symbiont of aphids, Buchnera aphidicola, has an extremely reduced genome (643 kb) based on PFGE of symbionts of the host species Acyrthosiphon pisum (7; H. Ishikawa, personal communication). The clade corresponding to Buchnera is over 100 million years old and has cospeciated with aphids during this period (17). Based on observations for pathogenic bacteria with small genomes, Buchnera spp. of different aphid lineages might be expected to differ in genome size. Applying methods developed by Charles and Ishikawa (7), genome sizes of Buchnera lineages, including isolates from A. pisum, Rhopalosiphum padi, Schizaphis graminum, and Melaphis rhois, were determined using PFGE. Symbiont DNA was prepared using filtration techniques described previously (7), modified by the addition of 100 mM EDTA and the omission of MgCl₂ in buffer A. Buchnera chromosomes were digested with several rare-cutting restriction enzymes (I-CeuI, NotI, SmaI, ApaI, and RsrII) and run on pulsed field gels. Fragment sizes were calculated by comparison to concatemers of lambda DNA (Fig. 1).

FIG. 1.

Conservation of genomes sizes in Buchnera. Buchnera chromosomes digested with rare-cutting restriction enzymes were run on pulsed field gels to estimate genome sizes. Lanes are labeled by the abbreviation for the host name (Ap, A. pisum; Rp, R. padi; Sg, S. graminum; Mr, M. rhois). Note that I-CeuI cuts once and linearizes the Buchnera chromosome. With this enzyme, genomes sizes of Buchnera from S. graminum, R. padi, and A. pisum are indistinguishable. The finding of a slightly smaller genome size of Buchnera from M. rhois is supported by several independent enzyme digests.

Genome sizes in Buchnera vary by <5%, even among strains that diverged 100 to 200 million years ago and that show substantial divergence in rRNA sequences (Fig. 2). In contrast, other bacterial clades with less sequence divergence consistently have more variation in genome size. The uniformity of genome size in Buchnera strongly suggests that genome shrinkage occurred early in the evolution of the symbiosis and that modern lineages retain the genome size of a common ancestor.

FIG. 2.

Relationships and genome sizes of Buchnera taxa. Host abbreviations are as given in Fig. 1. Divergence times are based on the aphid fossil record (8, 17).

Two factors may contribute to the conservation of genome size in Buchnera. First, unlike pathogens that move horizontally among hosts and experience coinfection, Buchnera is entirely vertically transmitted. As a result, bacteria within a single host are genetically homogeneous, preventing opportunity for the uptake of novel genes. Second, Buchnera is confined permanently to a single host lineage and to a mutualistic lifestyle that depends on a static set of physiological capabilities (3). In contrast, pathogens sometimes acquire new host taxa or new ways of exploiting hosts, providing a selective context for the stable incorporation of new genes.

The upcoming availability of full genome sequences of Buchnera strains (H. Ishikawa, personal communication) will show whether the conservation of genome size corresponds to a uniformity in gene inventories and genome organization. Furthermore, these full genome sequences will reveal whether specific genetic features, such as lack of translocatable elements or the loss of recombinase pathways, contribute to this unusual stability in genome size.