Intraspecific Variation in Symbiont Genomes: Bottlenecks and the Aphid-Buchnera Association

Corresponding author: Daniel J. Funk, Department of Biology, Vanderbilt University, Nashville, TN 37235., daniel.j.funk{at}vanderbilt.edu (E-mail)

Communicating editor: W. F. EANES

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Buchnera are maternally transmitted bacterial endosymbionts that synthesize amino acids that are limiting in the diet of their aphid hosts. Previous studies demonstrated accelerated sequence evolution in Buchnera compared to free-living bacteria, especially for nonsynonymous substitutions. Two mechanisms may explain this acceleration: relaxed purifying selection and increased fixation of slightly deleterious alleles under drift. Here, we test the divergent predictions of these hypotheses for intraspecific polymorphism using Buchnera associated with natural populations of the ragweed aphid, Uroleucon ambrosiae. Contrary to expectations under relaxed selection, U. ambrosiae from across the United States yielded strikingly low sequence diversity at three Buchnera loci (dnaN, trpBC, trpEG), revealing polymorphism three orders of magnitude lower than in enteric bacteria. An excess of nonsynonymous polymorphism and of rare alleles was also observed. Local sampling of additional dnaN sequences revealed similar patterns of polymorphism and no evidence of food plant-associated genetic structure. Aphid mitochondrial sequences further suggested that host bottlenecks and large-scale dispersal may contribute to genetic homogenization of aphids and symbionts. Together, our results support reduced N_e as a primary cause of accelerated sequence evolution in Buchnera. However, our study cannot rule out the possibility that mechanisms other than bottlenecks also contribute to reduced N_e at aphid and endosymbiont loci.

OVER the last decade, molecular and phylogenetic investigations of the endosymbiotic bacterium Buchnera aphidicola and its aphid hosts (BAUMANN et al. 1997 ; MORAN and TELANG 1998 ) have provided compelling evidence for two phenomena: (1) a strictly congruent phylogenetic association between aphid and Buchnera loci (MUNSON et al. 1991 ; MORAN et al. 1993 ; CLARK et al. 2000 ; FUNK et al. 2000 ); and (2) an elevated rate of DNA sequence evolution in Buchnera that can largely be attributed to increased fixation of nonsynonymous mutations (MORAN 1996 ; ROUHBAKHSH et al. 1997 ; BRYNNEL et al. 1998 ; LAMBERT and MORAN 1998 ; CLARK et al. 1999 ; WERNEGREEN and MORAN 1999 ).

Potential explanations for these patterns are suggested by the biology of this endosymbiosis (BUCHNER 1965 ; BAUMANN et al. 1997 ). Buchnera belongs to the gamma division of the Proteobacteria, which also contains Escherichia coli. These bacteria are usually housed in specialized cells (bacteriocytes) within the body cavity of aphids (BUCHNER 1965 ). However, a small number of Buchnera infect each developing aphid egg or embryo, thus ensuring maternal transmission into the next host generation. In combination with fossil evidence, the strict phylogenetic congruence of Buchnera and aphids suggests that their association dates to the origin of the Aphidoidea 200 mya (MORAN et al. 1993 ). This ancient mutualism is also obligate and neither aphid nor bacterium can survive and reproduce without its symbiont. Despite its small genome (CHARLES and ISHIKAWA 1999 ), Buchnera retains genes for biosynthesis of essential amino acids (BAUMANN et al. 1997 ) that are limiting in the phloem sap diet of their hosts (reviewed in DOUGLAS 1998 ; SANDSTROM and MORAN 1999 ). And in many Buchnera, genes encoding enzymes for leucine and tryptophan biosynthesis have been amplified on two types of plasmid, apparently as adaptations for increasing amino acid production and thus improving host nutrition (LAI et al. 1994 ; BRACHO et al. 1995 ; VAN HAM et al. 1997 ; BAUMANN et al. 1999 ).

Reduced N_e offers a potential explanation for the increased molecular evolutionary rates observed in Buchnera and other maternally inherited endosymbiotic bacteria relative to those of their free-living relatives (MORAN 1996 ; LAMBERT and MORAN 1998 ; PEEK et al. 1998 ; SPAULDING and VON DOHLEN 1998 ; CLARK et al. 1999 ; WERNEGREEN and MORAN 1999 ). If a substantial number of new mutations are slightly deleterious, reduced N_e will cause more mutations to drift to fixation, resulting in an overall increase in rates of substitution (OHTA 1992 ). For symbiont mutations deleterious at both the level of the host and that of the symbiont, fixation is dependent both on the number of infective cells transmitted between generations and on the N_e of female hosts in the population (BERGSTROM and PRITCHARD 1998 ; RISPE and MORAN 2000 ). This can be understood by noting that if few bacterial cells are transmitted to progeny, bacterial polymorphisms will be lost within individual aphids, whereas if hosts have small N_e due to strong selection, small census population sizes, or population bottlenecks, polymorphisms will also be lost from host populations. Thus, the N_e of symbionts will be dependent on the size, dynamics, and genetic structure of host populations. For these reasons, the N_e of animal symbionts will likely be smaller than that of free-living bacterial taxa (RISPE and MORAN 2000 ). The likely absence of recombination in vertically transmitted endosymbionts is also a contributing factor, since asexuality increases the rate of fixation of deleterious mutations (MULLER 1964 ; LYNCH 1996 , LYNCH 1997 ).

An alternative explanation for elevated rates of nonsynonymous substitution is positive selection for amino acid changes in endosymbionts. Positive selection, however, is not expected to show the consistent effects across all loci observed in Buchnera (MORAN 1996 ; WERNEGREEN and MORAN 1999 ) or to result in the decreased stability of gene products observed for rRNA genes in Buchnera and other endosymbionts (LAMBERT and MORAN 1998 ). By contrast, demographic history and population structure are expected to affect all loci and to promote slightly deleterious substitution.

One other viable hypothesis, however, is consistent with the genome-wide effects observed in various intracellular bacteria: relaxation of purifying selection across loci. Bacterial symbionts and chronic pathogens show low maximum replication rates and may inhabit relatively stable habitats compared to most free-living bacteria, which undergo occasional periods of exponential growth and may experience more variable environments. Any reduced selection intensity due to the symbiotic lifestyle might eliminate the fitness benefits associated with maximizing replication rate or maintaining resistance to extreme conditions and thus permit the fixation of nonsynonymous mutations that would be removed by selection in other bacteria.

These two competing hypotheses—decreased N_e vs. relaxed selection—cannot be distinguished on the basis of interspecific sequence divergence patterns, since a decrease in either N_e or s affects substitution rates similarly. However, the two hypotheses have different predictions for intraspecific polymorphism patterns. Under relaxed selection, a greater proportion of sites is effectively neutral, resulting in elevated polymorphism. By contrast, polymorphism is expected to be low under decreased N_e due to increased loss of alleles under drift. The distribution of allele frequencies is also expected to differ under the two hypotheses, with an excess of rare alleles expected only under decreased N_e. However, because certain patterns are expected whether N_e is reduced due to bottlenecks or to other mechanisms such as selective sweeps, quantifying their relative contribution to these patterns is difficult.

In this article we evaluate the "bottlenecks" hypothesis of increased genetic drift in the first study of DNA sequence variation among Buchnera [hereafter, referred to as "Buchnera (Ua)"] within a species of aphid host. To this end, we sequenced regions from aphid mitochondrial DNA, the Buchnera chromosome, and two plasmids from individuals of the ragweed aphid, Uroleucon ambrosiae, from 20 widely distributed U.S. localities. If reduced N_e has played a role in symbiont evolution, expected patterns include: (1) low levels of overall polymorphism compared to other bacteria and insects, due to the loss of neutral alleles under drift (KIMURA 1983 ; MCDONALD and KREITMAN 1991 ); (2) higher ratios of nonsynonymous to synonymous variation for within-species vs. between-species comparisons, as predicted if (slightly deleterious) mutations are sometimes eliminated by selection before fixation (e.g., HASEGAWA et al. 1998 ); and (3) an excess of young or rare mutations compared to the strictly neutral expectation, reflecting the recent origin of mutations following a bottleneck and the low frequencies of slightly deleterious alleles (e.g., TAJIMA 1989 ).

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Samples and loci:
U. ambrosiae (Thomas) (redescribed in OLIVE 1963 ) (Aphididae) is a cyclically parthenogenetic aphid with a broad North American distribution. It uses native food plants from the Asteraceae and especially the giant ragweed, Ambrosia trifida, on which it specializes over much of its range (MORAN 1984 ). Our U. ambrosiae samples consisted of single individuals collected from food plants in the field at different localities across the United States, with two individuals from a Minnesota site in different years (Table 1, Fig 1). Most collections for the principal analyses of this study were made from A. trifida, but New York and Utah samples were collected from unidentified Ambrosia species and the Sapelo Island, Georgia, sample was taken on Iva frutescens, a close relative of Ambrosia and an alternative food plant of U. ambrosiae along the U.S. East Coast (OLIVE 1963 ).

View larger version (21K): In this window In a new window Download PPT slide   Figure 1. Collection localities for U. ambrosiae aphids. Dark sites yielded haplotypes belonging to the "common" lineage and pale sites yielded the "divergent" alleles (see RESULTS).

All aphid samples were frozen alive and stored at -80° with vouchers from the same aphid colony stored in 70% ethanol. From each aphid, DNA sequences were collected from a segment spanning portions of the cytochrome oxidase I (COI) and cytochrome oxidase II (COII) genes of aphid mitochondrial DNA [1354 nucleotides (nt)], the entire dnaN gene on the Buchnera chromosome (1107 nt), and segments spanning parts of leuBC (1674 nt) and trpEG (1200 nt) from the leucine and tryptophan plasmids, respectively. Results reported below refer to coding sequences at these four loci, totaling 5325 nucleotides per haplotype.

DNA extraction, amplification, and sequencing:
For each sample, total genomic DNA was extracted from a single aphid and suspended in TE following a protocol modified from BENDER et al. 1983 . Sequencing templates were amplified in a single polymerase chain reaction (PCR) experiment by submitting 50-µl reactions (cf. MORAN et al. 1998 ) to 30–35 cycles of 1 min at 94°, 1 min at 54–60°, and 1.5–2.5 min at 72°. PCR products were filtered using the Concert Rapid PCR purification system (Life Technologies), resuspended in water, and submitted to the Laboratory of Molecular Systematics and Evolution Automated Sequencing Facility at the University of Arizona. For COI/II, dnaN, and leuBC, sequencing was performed in both directions using PCR primers plus an internal sequencing primer, yielding a continuous sequence fragment. For trpEG, analyzed sequences represent concatenated 600 nt 5' and 3' segments of the amplified fragment, with several hundred nucleotides missing between them.

Primers 8 and 18 from the Insect Mitochondrial DNA Primer Oligonucleotide Set (obtained from the University of British Columbia Nucleic Acid-Protein Service Unit; described in SIMON et al. 1994 ) were used to amplify COI/II sequences and an internal sequencing primer in COI was designed based on initial U. ambrosiae sequences. Primers designed and used to amplify the other study loci are the following: dnaN (dnaN.430seqF 5'-TCT ATG GGM AAA CAA G-3'), leuBC (leuB.689F 5'-TGT ATG CAA ATT ATT AAA RAT CC-3'; leuC.1337R 5'-GCW GCC ATA ATA GGA CTA ACT A-3'; leuC.180F 5'-AAT ACA TTT GCH ACT ATG GAT C-3'), and trpEG (trpEUrol.151F 5'-AAG YAT YAT GAT YAT YGA YAG TGC-3'; trpGUrol.507R 5'-ATA GAT TCR GGR TGA AATTG-3').

Sequence alignment and data analysis:
Sequences from two Minnesota aphids were analyzed, and trpEG sequences were not collected for the IL.1 and TN.2 samples (Table 1) due to technical difficulties. With these exceptions, the principal analyses of this study treated mitochondrial, bacterial, leucine plasmid, and tryptophan plasmid DNA sequences obtained from a single aphid from each of the 20 sampling localities. These sequences can be found under GenBank accession nos. AF196354, AF196467.

Sequences were inspected and nucleotides at all variable sites confirmed using Sequence Navigator (Applied Biosystems, Foster City, CA). Alignment of sequences using Clustal, as implemented in Sequence Navigator, was straightforward as no insertions or deletions were inferred within U. ambrosiae and overall variation was low. Using MacClade (MADDISON and MADDISON 1992 ), coding regions were identified and translated by comparison with previously published sequences and then aligned with sequences from U. rudbeckiae and U. aenum, two congeners sequenced for another study (J. J. WERNEGREEN and N. A. MORAN, personal communication). U. rudbeckiae is the most closely related Uroleucon species for which sequences were available (MORAN et al. 1998 ).

Using the four-gamete test (HUDSON and KAPLAN 1985 ) in the computer program DnaSP (ROZAS and ROZAS 1999 ), we obtained the minimum number of historical recombination events implied by a combined data set including all four gene regions. The degree of association between nucleotides at different polymorphic sites was also tested in DnaSP using chi-square tests and applying a Bonferroni adjustment to maintain a constant experiment-wide critical despite multiple comparisons (SOKAL and ROHLF 1991 ).

We used DnaSP and another program, MEA (MORIYAMA 1997 ), to calculate synonymous and nonsynonymous sites and two common measures of nucleotide variation, (NEI and LI 1979 ) and _w (WATTERSON 1975 ). Nucleotide diversity, , was calculated as the average number of nucleotide differences between all pairs of sequences, and _w is based on the proportion of segregating nucleotide sites in our sample. Both and _w are estimators of the neutral parameter = 2N_eµ for maternally inherited haploid genomes, where N_e is female effective population size. We used _w and a published rate of synonymous mutation in these bacteria (CLARK et al. 1999 ) to calculate effective population size for Buchnera (Ua) using the Buchnera chromosomal gene dnaN. Our mutation rate estimate was an average derived from 14 Buchnera loci with similar synonymous rates and calibrated using the host fossil record and estimates of the number of Buchnera cell divisions per generation (CLARK et al. 1999 ).

Phylogenetic relationships among haplotypes were determined by branch-and-bound parsimony analyses of the equally weighted data set as implemented in PAUP* (SWOFFORD 1999 ) using different combinations of U. rudbeckiae and U. aenum as outgroups. Statistical support for each node was obtained by bootstrap resampling based on 1000 branch-and-bound replicate trees. Evolutionary reconstructions of polymorphic nucleotide positions and amino acid residues over the topology were inferred with MacClade (MADDISON and MADDISON 1992 ).

We compared observed sequence variation with neutral expectations for each of the four gene regions of our study. In an equilibrium population undergoing neutral mutation, the estimate of based on the number of segregating sites will be the same as the estimate derived from . We used Tajima's D (TAJIMA 1989 ) to evaluate whether our estimates of and _w were significantly different. Tajima's D depends on the frequency distribution of segregating sites, with negative values ( < _w) indicating an excess of rare alleles and positive values ( > _w) indicating an excess of intermediate frequencies. Fu and Li's D (FU and LI 1993 ) evaluates whether the numbers of mutations on external branches (singletons) and internal branches in a genealogy are equal, as expected under neutrality. Fu's F's is a statistic that compares the observed to the expected number of alleles from a sample for a given and may offer a more powerful means of rejecting neutrality than Tajima's D and Fu and Li's D (FU 1997 ). Several other statistical comparisons of estimators (Tajima's T, Fu and Li's F, Fu and Li's D*, Fu and Li's F*; FU and LI 1993 ; FU 1997 ) were also applied.

The previous tests evaluate the hypothesis that all mutations in a DNA region are neutral in a population at stationarity. We also employed the method of MCDONALD and KREITMAN 1991 , which asks whether the ratio of replacement to synonymous polymorphisms within species is equal to the ratio of replacement to synonymous substitutions between species, as per the neutral prediction. For the interspecific comparisons in these tests, we used sequences from U. rudbeckiae.

Local analyses of U. ambrosiae from southeastern Arizona:
To complement the principal analyses of this study, we also sequenced aphids collected from several localities and several food plants across southeastern Arizona, where U. ambrosiae is rather generalized in its food plant preferences (FUNK and BERNAYS 2001 ). Along with the A. trifida Arizona specimen from the principal analyses, all 5325 nucleotides were also sequenced for three aphids from three different non-Ambrosia food plants [Viguiera dentata (Asteraceae), Heterotheca subaxillaris (Asteraceae), and Sphaeralcea sp. (Malvaceae)] (Table 2). For 18 additional specimens, the first 534 nucleotides of dnaN were sequenced. This data set was compiled to evaluate the possibility that different food plants support genetically differentiated aphid populations and to compare genetic diversity at different geographic scales.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Sequence and haplotype variation:
Polymorphic nucleotide sites and amino acid residues are illustrated in Table 3 and summarized by gene region in Table 4. No variable positions were detected in noncoding regions (66 nt of tRNA between COI and COII, a 1-nt spacer between leuB and leuC, and a 2-nt spacer between trpE and trpG). A total of 59 nt sites varied, of which 25 (42%) were nonsynonymous polymorphisms. No insertions or deletions were detected among U. ambrosiae sequences, but U. ambrosiae exhibited a five-codon insertion in COI and a one-codon insertion in trpEG compared to U. rudbeckiae and U. aenum sequences.

The 21 individuals sampled for the principal analyses yielded eight different haplotypes (Table 3). Sequences collected from 13 individuals were identical and represent the "common" haplotype. Together with the common haplotypes, distinct haplotypes from IL.1 and IN and a third shared by KY and VA samples constitute the highly homogeneous "common lineage" and differ from each other at a maximum of two of 5325 sites (0.04%) (Table 5 and Table 6). The four remaining individuals showed distinct haplotypes differing from the common haplotype and one another at 17–30 sites (0.32–0.56% total sequence divergence). They are collectively referred to here as the "divergent" haplotypes (Table 3 and Table 6). This combination of extremely closely and comparatively distantly related alleles yielded a distinctly bimodal frequency distribution of pairwise sequence divergences (Fig 2A). For each gene region, 75% of polymorphic sites were singletons, and the minority nucleotide was observed in no more than four alleles at any site (Table 3 and Table 4), yielding a skewed frequency distribution (Fig 2B).

View larger version (15K): In this window In a new window Download PPT slide   Figure 2. (a) Frequency distribution of uncorrected sequence divergences derived from all pairwise comparisons among the 21 U. ambrosiae multi-locus haplotypes. (b) Frequency distribution of the number of synonymous and nonsynonymous polymorphic sites at which the minority nucleotide was observed in a given number of the 21 multi-locus haplotypes.

Estimates of and N_e:
Overall and _w values were low for each gene region compared to estimates from other bacteria and insect mitochondria (see DISCUSSION). Values are miniscule when calculated for the common lineage alone (Table 5). On the basis of our estimate of _w from synonymous sites in the Buchnera chromosomal gene dnaN (Table 5, 0.36%) and the per generation mutation rate from CLARK et al. 1999 (1.4–1.9 x 10^-10 mutations/site/generation), effective population sizes for Buchnera (Ua) are about 1 x 10⁷.

Phylogenetic reconstruction:
For each gene region, 25% of polymorphic sites, a total of 15 sites, were parsimony informative. The strong bootstrap support provided by these 15 characters (Fig 3) reflects a lack of homoplasy. Among the 59 nt sites that are polymorphic among U. ambrosiae haplotypes, only 1 (leuC site 1539) appears to have experienced multiple (two) substitutions. Using other Uroleucon species to root the tree introduced some homoplasy, with the result that the basal MN.b and GA alleles (Fig 3) switch positions when U. aenum is used as an outgroup. The rooted topology clearly demonstrates the common lineage to be cladistically derived relative to the basal divergent alleles. Evidence for recombination or the horizontal transfer of bacterial or plasmid symbionts among hosts is absent, and significant linkage disequilibrium was detected among gene regions. These phenomena are further illustrated by the complete congruence among trees separately reconstructed for each of the assayed loci, using the data described here (FUNK et al. 2000 ). The data exhibit no obvious phylogeographic structure as alleles from the common lineage are intermingled geographically with the divergent alleles, and both are widely scattered among U.S. collecting localities (Fig 1).

View larger version (29K): In this window In a new window Download PPT slide   Figure 3. Phylogenetic relationships among multi-locus haplotypes with bootstrap support. Inferred synonymous (open bars) and nonsynonymous (solid bars) changes are mapped onto the tree. Changes are labeled by nucleotides (below: transversions in boldface) and by gene region (above: C, COI/II; D, dnaN; L, leuBC; T, trpEG). Diagonally striped bars reflect the observation of two substitutions at one site; a horizontally striped bar indicates the instance of homoplasy introduced by outgroup comparison. Consistency index, 0.98.

Tests of neutrality:
The absence of multiple hits, recombination (or horizontal transfer), and population structure evident from our data is consistent with three assumptions of many population genetic models, those of infinite sites, no recombination, and panmixia. Multiple changes may separate the U. rudbeckiae sequences used in our McDonald-Kreitman tests from U. ambrosiae at some sites. However, per site rates of synonymous divergence between these species are <0.3 for each gene region (Table 5), below the onset of saturation.

Every gene region showed departures from a neutral frequency distribution for a majority of the tests of neutrality applied, as summarized in Table 7. In most cases, significantly negative skews were observed, implying an excess of rare polymorphisms or alleles. This is reflected in the prevalence of singletons (Table 4, Fig 2B) and the concentration of changes on external branches (31) relative to internal branches (21; Fig 3). Deviation from neutrality was also supported by a MacDonald-Kreitman test for dnaN, which showed an excess of nonsynonymous intraspecific polymorphism (Table 8). In contrast, the other gene regions showed similar K_a/K_s values for polymorphism and divergence (Table 4).

Local analyses:
Sequencing 22 Arizona aphid specimens for the first 534 nucleotides of dnaN yielded levels and patterns of polymorphism very similar to those observed at the larger geographic scale of the principal analyses. Of these sequences, 18 proved identical to the homologous segment of the common allele from the principal analyses, 3 specimens from non-Ambrosia food plants shared a second haplotype that differed from the common allele at two nucleotide sites (Table 9), and 1 specimen from A. confertaflora possessed a third haplotype. One of the three polymorphic sites in the local data set (site 111) was also polymorphic at the larger geographic scale. For the local data, = 0.00110 and _w = 0.00155. The common allele was observed for each of the three individuals from non-Ambrosia hosts for which the entire 5325-nucleotide data set was collected (Table 9).

In any study of intraspecific variation, the presence of morphologically cryptic species may compromise interpretations if unrecognized. In herbivorous insects, such cryptic species can sometimes be distinguished by their use of different food plants (see references in FUNK 1998 ). Our data provide no evidence that food plant association likewise structures genetic variation in U. ambrosiae. The common allele was observed in aphids from a variety of food plants in the southwestern United States (Table 2), and three other specimens collected on three different plant genera shared a single uncommon allele (Table 9). These findings corroborate the supposition that southwestern U. ambrosiae are food plant generalists rather than genetically differentiated populations that specialize on different food plants.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Bottlenecks and Buchnera:
Our findings support the hypothesis that reduced N_e in Buchnera promotes the fixation of nonsynonymous substitutions and contributes to the accelerated molecular evolution of Buchnera relative to free-living bacteria. This support is most clearly manifest in the very low levels of synonymous polymorphism exhibited by Buchnera (Ua) relative to those of nonsymbiotic bacteria, as would be expected if increased exposure to genetic drift had promoted the loss of neutral alleles in these symbionts. In another recent study, SREEVATSAN et al. 1997 similarly reported strikingly low levels of synonymous polymorphism when comparing their global samples of the human pathogen Mycobacterium tuberculosis with previously reported data from other bacterial species (Vibrio cholerae, Streptococcus pyogenes, Neisseria meningtidis, M. avium-intracellulare, Escherichia coli, Borrelia burgdorferi, and Salmonella enterica, in order of increasing polymorphism). The synonymous nucleotide diversity we observed for dnaN was also lower than for each of these seven bacterial taxa and was about three orders of magnitude lower than for E. coli or S. enterica, the closest relatives to Buchnera in this set, although it was still 20 times higher than in M. tuberculosis. Since Buchnera has a higher per site mutation rate (OCHMAN et al. 1999 ), the low observed must result from a substantial decrease in N_e.

ACHTMAN et al. 1999 describe yet greater genetic homogeneity from global assays of the plague-causing Yersinia pestis, finding zero synonymous polymorphisms in sequences from several gene regions. These authors propose that Y. pestis represents a recent clone descending from within a more diversified species, much as the common lineage of Buchnera (Ua) is nested within a set of more divergent lineages (Fig 3). Similarly analyzing this geographically widespread common lineage as a distinct evolutionary entity yields estimates of synonymous polymorphism on par with those for Y. pestis and M. tuberculosis ( averaged across dnaN, leuBC, and trpEG = 0.01%, due to a single synonymous change; Table 5, Fig 3). However, there is no evidence of such ecological differentiation from the samples yielding the common lineage of Buchnera (Ua), which were collected over a wide geographic range and principally on a single host plant.

Another point is raised by the Y. pestis example, however. If cryptic bacterial species or ecological types such as Y. pestis go unrecognized, estimates of polymorphism in the bacterial species that include them will be artificially inflated, with the result that other species [such as Buchnera (Ua)] might be inappropriately viewed as depauperate in polymorphism by the kinds of comparisons made here. Nonetheless, because Buchnera (Ua) are consistently and appreciably less variable than those from an unbiased and fairly comprehensive set of published bacterial studies, it can safely be said that its overall levels of nucleotide polymorphism are exceedingly small.

In addition to reduced overall neutral variation, Buchnera (Ua) exhibited an excess of rare alleles relative to neutral expectations in both chromosomal (dnaN ) and plasmid loci (leuBC and trpEG; Table 7). This excess may reflect the persistence of slightly deleterious mutations at low frequency under reduced N_e, as per the bottlenecks hypothesis. However, it is important to recognize that such patterns can also arise due to other deviations from the null model. Indeed, the observed concentration of rare alleles within a few unusual haplotypes (Fig 3) is not readily explained as the result of slightly deleterious mutations alone. This pattern is suggestive of a history of population subdivision resulting in divergence among haplotypes or of a selective sweep involving positive selection (ATWOOD et al. 1951 ; BEGUN and AQUADRO 1992 ) on haplotypes from the common lineage.

Applying the McDonald-Kreitman test to dnaN revealed a significant excess of nonsynonymous polymorphism for Buchnera sequences within U. ambrosiae compared to nonsynonymous divergence between U. ambrosiae and U. rudbeckiae. This pattern is as expected if some of the slightly deleterious mutations maintained by low N_e and drift have been purged by selection prior to fixation (Table 8). Similar excesses of intraspecific amino acid polymorphism occur in the mitochondrial DNA of many animals, where they have likewise been interpreted as evidence for slightly deleterious mutations (NACHMAN 1998 ; Rand AND KANN 1998 ).

Although the smaller number of polymorphic sites in our COI/II mitochondrial sequences provides more limited evidence, their patterns are also consistent with low N_e of the aphid hosts. The nucleotide diversities of U. ambrosiae (Table 5) are, for example, somewhat smaller than those reported for the ND5 mitochondrial gene in Drosophila melanogaster and D. simulans (Rand AND KANN 1996 ) and are generally low by insect standards. COI/II allele frequency distributions are also negatively skewed and sometimes diverge significantly from neutral expectations (Table 7). However, no excess of within-species nonsynonymous polymorphism was detected in COI/II, providing no direct evidence for slightly deleterious variation.

Selection on symbionts:
Increased rates of nonsynonymous substitution are a consistent pattern across Buchnera loci in previous studies, suggesting that positive selection is not a general cause of increased rates of molecular evolution in these bacteria (MORAN 1996 ). As the first assay of DNA sequence variation within Buchnera of a single aphid species, the present study is able to address an alternative hypothesis: that purifying selection is relaxed for endosymbiotic bacteria, perhaps due to the relative stability of their intracellular environment. As a sole explanation of the accelerated sequence evolution observed in Buchnera, the relaxed selection hypothesis does not predict, nor readily explain, our findings of lower levels of genetic diversity within species or the observed non-neutral distributions of allele frequencies. By contrast, these results are expected under the hypothesis of low N_e, as described above.

Our ability to infer the specific mechanisms responsible for low polymorphism in Buchnera (Ua) is, however, limited by the fact that multiple processes can contribute to small N_e. These processes include fluctuations in census population size (or bottlenecks), selective sweeps involving linkage and strong positive selection (ATWOOD et al. 1951 ; BEGUN and AQUADRO 1992 ), and background selection involving linkage and negative selection (CHARLESWORTH et al. 1993 ). Our analyses of the combined multi-locus data set indicate significant linkage and no evidence for recombination. These findings complement documented patterns of phylogenetic congruence (historical linkage) among aphid, bacterial, and plasmid genomes in this system. They also suggest ample opportunity for genetic hitchhiking since selection at a site in any of these regions will have consequences for sequence variation at the others. However, hitchhiking is not predicted to give an excess of nonsynonymous polymorphism as we observed at dnaN. Rather, if selective sweeps result in fixation of certain genotypes that experience favorable mutations, an excess of nonsynonymous differences between species might be expected, with a relative lack of nonsynonymous polymorphism. In contrast, our results suggest that small census population sizes have played an important role in the evolution of this endosymbiosis.

Our results do not, however, completely exclude relaxed selection, perhaps in combination with small N_e, as a partial basis for accelerated substitution of nonsynonymous alleles in Buchnera. It is plausible, for example, that the evolutionary transition to endosymbiosis may have simultaneously introduced a less stressful selective environment and bottleneck-induced population structure in the bacterial ancestor of Buchnera. Thus, relaxed selection may coincide with small N_e, making their relative contributions difficult to disentangle. Similar limitations apply to other studies that have argued for small N_e as the basis for accelerated evolution in mammals (OHTA 1993 , OHTA 1995 ) and in D. melanogaster (AKASHI 1996 ) since changes in population structure may often be accompanied by ecological shifts that affect selection intensity. However, in these cases and in Buchnera the acceleration in sequence evolution affects loci throughout the genome, a pattern unlikely to be caused by many ecological scenarios, but consistent with reduced N_e.

Insights from aphid biology:
Observations on natural U. ambrosiae populations suggest that they may experience repeated and severe bottlenecking. In southeastern Arizona, these aphids occur at very low densities for much of the year, but may be quite common during March and April and again in October and November. This fluctuation in census population sizes is due, in part, to changes in food plant abundance resulting from variable rainfall patterns between seasons and years. During the present study, for example, U. ambrosiae were relatively uncommon at Arizona collecting localities during the autumns of 1997 and 1999 but very abundant at the same localities in the autumn of 1998. This capacity of aphid populations to rapidly rebound from reduced densities owes to their rapid development and parthenogenetic life history, which allow a single female to produce thousands of descendants in a single season.

Our data also suggest that the genetic homogeneity of U. ambrosiae may be further promoted by considerable long-distance gene flow. The lack of geographic structure in our principal data set and the distribution of the common allele on opposite sides of North America indicate that dispersal of U. ambrosiae is far and frequent on an evolutionary time scale. These results, plus the observation of similar levels and patterns of nucleotide diversity at continental and regional scales, suggest that U. ambrosiae haplotypes are randomly mixed across the United States. The long-distance dispersal of aphids as aeroplankton has been documented in other aphid taxa and may provide a mechanism for such pronounced gene flow (LOXDALE et al. 1993 ).

Population genetic studies on other aphid species mostly indicate low levels of polymorphism compared to other insect species, although aphid taxa vary in the extent of both polymorphism and geographic structure (e.g., DE BARRO et al. 1995 ; SIMON et al. 1996 , SIMON et al. 1999 ; LOXDALE et al. 1998 ; WILSON et al. 1999 ). Some studies suggest that selection linked to host plants or other ecological factors sometimes erodes polymorphism or produces geographic or host-plant-related subdivision within aphid populations (e.g., SUNNUCKS et al. 1997 ). Such selection could decrease N_e below expectations on the basis of census population sizes alone. In the case of U. ambrosiae, behavioral studies have indicated genetic differentiation between eastern and southwestern populations in their degree of food plant specificity (FUNK and BERNAYS 2001 ). However, the lack of geographic or food plant-associated genetic structure observed in the present study indicates that this regional and possibly adaptive differentiation occurs in the face of gene flow (EHRLICH and RAVEN 1969 ).

Previous molecular genetic studies of aphids have focused on pest populations using introduced crops. By contrast, our study on U. ambrosiae represents the first extensive molecular population genetic assay of an aphid using wild host plants in its native range. A parallel study on Pemphigus obesinymphae, a distantly related, gall-dwelling aphid of cottonwoods with a very different life cycle, has revealed low levels of polymorphism comparable to those reported here (P. ABBOT, personal communication). Such natural systems may provide unique insights into the population genetics underlying the aphid-Buchnera symbiosis.

	ACKNOWLEDGMENTS

We thank various colleagues (Table 1) for contributing aphid specimens and Michael Nachman and Howard Ochman for comments on the analyses and text. Louise Helbling and Michael Gelsey helped to collect sequence data. D.J.F. and J.J.W. were funded by a National Institutes of Health (NIH) training grant in Molecular Insect Science administered through the Center for Insect Science at the University of Arizona. This work was supported by NIH and by National Science Foundation grants DEB9815413 and DEB9978518 to N.A.M.

Manuscript received April 17, 2000; Accepted for publication October 10, 2000.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ACHTMAN, M., K. ZURTH, G. MORELLI, G. TORREA, and A. GUIYOULE et al., 1999  Yersinia pestis, the cause of plague, is a recently emerged clone of Yersinia pseudotuberculosis.. Proc. Natl. Acad. Sci. USA 96:14043-14048[Abstract/Free Full Text].

AKASHI, H., 1996  Molecular evolution between Drosophila melanogaster and D. simulans: reduced codon bias, faster rates of amino acid substitution, and larger proteins in D. melanogaster.. Genetics 144:1297-1307[Abstract].

ATWOOD, K. C., L. K. SCHNEIDER, and F. J. RYAN, 1951  Selective mechanisms in bacteria. Cold Spring Harbor Symp. Quant. Biol. 16:345-355[Medline].

BAUMANN, P., N. A. MORAN, and L. BAUMANN, 1997  The evolution and genetics of aphid endosymbionts. Bioscience 47:12-20.

BAUMANN, L., P. BAUMANN, N. A. MORAN, J. SANDSTRÖM, and M. L. THAO, 1999  Genetic characterization of plasmids containing genes encoding enzymes of leucine biosynthesis in endosymbionts (Buchnera) of aphids. J. Mol. Evol. 48:77-85[Medline].

BEGUN, D. J. and C. F. AQUADRO, 1992  Levels of naturally-occurring DNA polymorphism correlate with recombination rates in D. melanogaster.. Nature 356:519-520[Medline].

BENDER, W., P. SPIERER, and D. S. HOGNESS, 1983  Chromosomal walking and jumping to isolate DNA from the ace and rosy loci and the bithorax complex in Drosophila melanogaster.. J. Mol. Biol. 168:17-33[Medline].

BERGSTROM, C. T. and J. PRITCHARD, 1998  Germline bottlenecks and the evolutionary maintenance of mitochondrial genomes. Genetics 149:2135-2146[Abstract/Free Full Text].

BRACHO, A. M., D. MARTÍNEZ-TORRES, A. MOYA, and A. LATORRE, 1995  Discovery and molecular characterization of a plasmid localized in Buchnera sp., a bacterial endosymbiont of the aphid Rhopalosiphum padi.. J. Mol. Evol. 41:67-73[Medline].

BRYNNEL, E. U., C. G. KURLAND, N. A. MORAN, and S. G. E. ANDERSSON, 1998  Evolutionary rates for tuf genes in endosymbionts of aphids. Mol. Biol. Evol. 15:574-582[Abstract].

BUCHNER, P., 1965 Endosymbiosis of Animals with Plant Microorganisms. John Wiley & Sons, New York.

CHARLES, H. and I. ISHIKAWA, 1999  Physical and genetic map of the genome of Buchnera, the primary endosymbiont of the pea aphid, Acyrthosiphon pisum.. J. Mol. Evol. 48:142-150[Medline].

CHARLESWORTH, B., M. T. MORGAN, and D. CHARLESWORTH, 1993  The effect of deleterious mutations on neutral molecular variation. Genetics 134:1289-1303[Abstract].

CLARK, M. A., N. A. MORAN, and P. BAUMANN, 1999  Sequence evolution in bacterial endosymbionts having extreme base compositions. Mol. Biol. Evol. 16:1586-1598[Abstract].

CLARK, M. A., N. A. MORAN, P. BAUMANN, and J. J. WERNEGREEN, 2000  Cospeciation between bacterial endosymbionts and a recent radiation of aphids (Uroleucon) and pitfalls of testing for phylogenetic congruence. Evolution 54:517-525[Medline].

DE BARRO, P. J., T. N. SHERRATT, G. R. CARVALHO, D. NICOL, and A. IYENGAR et al., 1995  Geographic and microgeographic genetic differentiation in 2 aphid species over southern England using the multilocus (GATA)(4) probe. Mol. Ecol. 4:375-382[Medline].

DOUGLAS, A. E., 1998  Nutritional interactions in insect-microbe symbioses: aphids and their symbiotic bacteria Buchnera.. Annu. Rev. Entomol. 43:17-37[Medline].

EHRLICH, P. A. and P. H. RAVEN, 1969  Differentiation of populations. Science 165:1228-1232[Free Full Text].

FU, Y. X., 1997  Statistical tests of neutrality of mutations against population growth, hitchhiking, and background selection. Genetics 147:915-925[Abstract].

FU, Y. X. and W. H. LI, 1993  Statistical tests of neutrality of mutations. Genetics 133:693-709[Abstract].

FUNK, D. J., 1998  Isolating a role for selection in speciation: host adaptation and sexual isolation in Neochlamisus bebbianae leaf beetles. Evolution 52:1744-1759.

FUNK, D. J., and E. A. BERNAYS, 2001 Geographic variation in host specificity reveals host range evolution in Uroleucon ambrosiae aphids. Ecology (in press).

FUNK, D. J., L. HELBLING, J. J. WERNEGREEN, and N. A. MORAN, 2000  Intraspecific phylogenetic congruence among multiple symbiont genomes. Proc. R. Soc. Lond. Ser. B 267:2517-2521[Medline].

HASEGAWA, M., Y. CAO, and Z. H. YANG, 1998  Preponderance of slightly deleterious polymorphisms in mitochondrial DNA: nonsynonymous/synonymous rate ratio is much higher within species than between species. Mol. Biol. Evol. 15:1499-1505[Free Full Text].

HUDSON, R. R. and N. L. KAPLAN, 1985  Statistical properties of the number of recombination events in the history of a sample of DNA-sequences. Genetics 111:147-164[Abstract/Free Full Text].

KIMURA, M., 1983 The Neutral Theory of Molecular Evolution. Cambridge University Press, New York.

LAI, C. Y., L. BAUMANN, and P. BAUMANN, 1994  Amplification of trpEG—adaptation of Buchnera aphidicola to an endosymbiotic association with aphids. Proc. Natl. Acad. Sci. USA 91:3819-3823[Abstract/Free Full Text].

LAMBERT, J. D. and N. A. MORAN, 1998  Deleterious mutations destabilize ribosomal RNA in endosymbiotic bacteria. Proc. Natl. Acad. Sci. USA 95:4458-4462[Abstract/Free Full Text].

LOXDALE, H. D., J. HARDIE, S. HALBERT, R. FOOTTIT, and N. A. C. KIDD et al., 1993  The relative importance of short- and long-range movement of flying aphids. Biol. Rev. 68:291-311.

LOXDALE, H. D., C. P. BROOKES, I. R. WYNNE, and S. J. CLARK, 1998  Genetic variability within and between English populations of the damson-hop aphid, Phorodon humuli (Hemiptera: Aphididae), with special reference to esterases associated with insecticide resistance. Bull. Entomol. Res. 88:513-526.

LYNCH, M., 1996  Mutation accumulation in transfer RNAs: molecular evidence for Muller's ratchet in mitochondrial genomes. Mol. Biol. Evol. 13:209-220[Abstract].

LYNCH, M., 1997  Mutation accumulation in nuclear, organelle, and prokaryotic transfer RNA genes. Mol. Biol. Evol. 14:914-925[Abstract].

MADDISON, W. P., and D. R. MADDISON, 1992 MacClade: analysis of phylogeny and character evolution. Version 3.0. Sinauer Associates, Sunderland, MA.

MCDONALD, J. H. and M. KREITMAN, 1991  Adaptive protein evolution at the Adh locus in Drosophila.. Nature 351:652-654[Medline].

MORAN, N. A., 1984  The genus Uroleucon (Homoptera: Aphididae) in Michigan: key, host records, biological notes, and descriptions of three new species. J. Kansas Entomol. Soc. 57:596-616.

MORAN, N. A., 1996  Accelerated evolution and Muller's ratchet in endosymbiotic bacteria. Proc. Natl. Acad. Sci. USA 93:2873-2878[Abstract/Free Full Text].

MORAN, N. A. and A. TELANG, 1998  Bacteriocyte-associated symbionts of insects—a variety of insect groups harbor ancient prokaryotic endosymbionts. Bioscience 48:295-304.

MORAN, N. A., M. A. MUNSON, P. BAUMANN, and H. ISHIKAWA, 1993  A molecular clock in endosymbiotic bacteria is calibrated using the insect hosts. Proc. R. Soc. Lond. Ser. B 253:167-171.

MORAN, N. A., M. E. KAPLAN, M. J. GELSEY, T. G. MURPHY, and E. A. SCHOLES, 1998  Phylogenetics and evolution of the aphid genus Uroleucon based on mitochondrial and nuclear DNA sequences. Syst. Entomol. 24:1-9.

MORIYAMA, E. N., 1997 Molecular Evolutionary Analysis Package. Department of Ecology and Evolutionary Biology, Yale University, New Haven, CT.

MULLER, H. J., 1964  The relation of recombination to mutational advance. Mutat. Res. 1:2-9.

MUNSON, M. A., P. BAUMANN, M. A. CLARK, L. BAUMANN, and N. A. MORAN et al., 1991  Aphid-eubacterial endosymbiosis: evidence for its establishment in an ancestor of four aphid families. J. Bacteriol. 173:6321-6324[Abstract/Free Full Text].

NACHMAN, M. W., 1998  Deleterious mutations in animal mitochondrial DNA. Genetica 102(103):61-69.

NEI, M. and W.-H. LI, 1979  Mathematical model for studying genetic variation in terms of restriction endonucleases. Proc. Natl. Acad. Sci. USA 76:5269-5273[Abstract/Free Full Text].

OCHMAN, H., S. ELWYN, and N. A. MORAN, 1999  Calibrating bacterial evolution. Proc. Natl. Acad. Sci. USA 96:12638-12643[Abstract/Free Full Text].

OHTA, T., 1992  The nearly neutral theory of molecular evolution. Annu. Rev. Ecol. Syst. 23:263-286.

OHTA, T., 1993  An examination of the generation-time effect on molecular evolution. Proc. Natl. Acad. Sci. USA 90:10676-10680[Abstract/Free Full Text].

OHTA, T., 1995  Synonymous and nonsynonymous substitutions in mammalian genes and the nearly neutral theory. J. Mol. Evol. 40:56-63[Medline].

OLIVE, A. T., 1963  The genus Dactynotus Rafinesque in North Carolina (Homoptera: Aphididae). Misc. Publ. Entomol. Soc. Am. 56:556-559.

PEEK, A. S., R. C. VRIJENHOEK, and B. S. GAUT, 1998  Accelerated evolutionary rate in sulfur-oxidizing endosymbiotic bacteria associated with the mode of symbiont transmission. Mol. Biol. Evol. 15:1514-1523[Free Full Text].

RAND, D. M. and L. M. KANN, 1996  Excess amino acid polymorphism in mitochondrial DNA: contrasts among genes from Drosophila, mice, and humans. Mol. Biol. Evol. 13:735-748[Abstract].

RAND, D. M. and L. M. KANN, 1998  Mutation and selection at silent and replacement sites in the evolution of animal mitochondrial DNA. Genetica 102(103):393-407.

RISPE, C. and N. A. MORAN, 2000  Accumulation of deleterious mutations in endosymbionts: Muller's ratchet with two levels of selection. Am. Nat. 156:425-441.

ROUHBAKHSH, D., M. A. CLARK, L. BAUMANN, N. A. MORAN, and P. BAUMANN, 1997  Evolution of the tryptophan biosynthetic pathway in Buchnera (aphid endosymbionts): studies of plasmid-associated trpEG within the genus Uroleucon.. Mol. Phylogenet. Evol. 8:167-176[Medline].

ROZAS, J. and R. ROZAS, 1999  DnaSP version 3: an integrated program for molecular population genetics and molecular evolutionary analysis. Bioinformatics 15:174-175[Abstract/Free Full Text].

SANDSTRÖM, J. and N. MORAN, 1999  How nutritionally imbalanced is phloem sap for aphids? Entomol. Exp. Appl. 91:203-210.

SIMON, C., F. FRATI, A. BECKENBACH, B. CRESPI, and H. LIU et al., 1994  Evolution, weighting, and phylogenetic utility of mitochondrial gene sequences and a compilation of conserved polymerase chain reaction primers. Ann. Entomol. Soc. Am. 87:651-701.

SIMON, J. C., D. MARTINEZTORRES, A. LATORRE, A. MOYA, and P. D. N. HEBERT, 1996  Molecular characterization of cyclic and obligate parthenogens in the aphid Rhopalosiphum padi (L). Proc. R. Soc. Lond. Ser. B 263:481-486[Medline].

SIMON, J. C., S. BAUMANN, P. SUNNUCKS, P. D. N. HEBERT, and J. S. PIERRE et al., 1999  Reproductive mode and population genetic structure of the cereal aphid Sitobion avenae studied using phenotypic and microsatellite markers. Mol. Ecol. 8:531-545[Medline].

SOKAL, R. R., and F. J. ROHLF, 1991 Biometry, Ed. 3. W. H. Freeman, New York.

SPAULDING, A. W. and C. D. VON DOHLEN, 1998  Phylogenetic characterization and molecular evolution of bacterial endosymbionts in psyllids (Hemiptera: Sternorrhyncha). Mol. Biol. Evol. 15:1506-1513[Free Full Text].

SREEVATSAN, S., X. PAN, K. E. STOCKBAUER, N. D. CONNELL, and B. N. KREISWIRTH et al., 1997  Restricted structural gene polymorphism in the Mycobacterium tuberculosis complex indicates evolutionarily recent global dissemination. Proc. Natl. Acad. Sci. USA 94:9869-9874[Abstract/Free Full Text].

SUNNUCKS, P., P. J. DE BARRO, G. LUSHAI, N. MACLEAN, and D. HALES, 1997  Genetic structure of an aphid studied using microsatellites: cyclic parthenogenesis, differentiated lineages and host specialization. Mol. Ecol. 6:1059-1073[Medline].

SWOFFORD, D. L., 1999 PAUP* Phylogenetic Analysis Using Parsimony (* and other methods). Version 4. Sinauer Associates, Sunderland, MA.

TAJIMA, F., 1989  Statistical method for testing the neutral mutation model in finite populations. Genetics 128:183-192[Abstract].

VAN HAM, R. C. H. J., A. MOYA, and A. LATORRE, 1997  Putative evolutionary origin of plasmids carrying the genes involved in leucine biosynthesis in Buchnera aphidicola (endosymbiont of aphids). J. Bacteriol. 179:4768-4777[Abstract/Free Full Text].

WATTERSON, G. A., 1975  On the number of segregating sites in genetic models without recombination. Theor. Popul. Biol. 7:256-276[Medline].

WERNEGREEN, J. J. and N. A. MORAN, 1999  Evidence for genetic drift in endosymbionts (Buchnera): analyses of protein-coding genes. Mol. Biol. Evol. 16:83-97[Abstract].

WILSON, A. C. C., P. SUNNUCKS, and D. F. HALES, 1999  Microevolution, low clonal diversity and genetic affinities of parthenogenetic Sitobion aphids in New Zealand. Mol. Ecol. 8:1655-1666[Medline].

This article has been cited by other articles:

				P. Roumagnac, F.-X. Weill, C. Dolecek, S. Baker, S. Brisse, N. T. Chinh, T. A. H. Le, C. J. Acosta, J. Farrar, G. Dougan, et al. Evolutionary history of Salmonella typhi. Science, November 24, 2006; 314(5803): 1301 - 1304. [Abstract] [Full Text] [PDF]

				N. A. Moran, H. E. Dunbar, and J. L. Wilcox Regulation of Transcription in a Reduced Bacterial Genome: Nutrient-Provisioning Genes of the Obligate Symbiont Buchnera aphidicola J. Bacteriol., June 15, 2005; 187(12): 4229 - 4237. [Abstract] [Full Text] [PDF]