© 2004 Society of Systematic Biologists
Host–Symbiont Stability and Fast Evolutionary Rates in an Ant–Bacterium Association: Cospeciation of Camponotus Species and Their Endosymbionts, Candidatus Blochmannia
Edited by Kevin Johnson: Associate Editor
1 Josephine Bay Paul Center for Comparative Molecular Biology and Evolution, Marine Biological Laboratory 7 MBL Street, Woods Hole, Massachusetts 02543, USA; E-mail: jwernegreen{at}mbl.edu (J.J.W.)
2 School of Biological Sciences, University of Nebraska Lincoln Nebraska 68588, USA
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionAppendix 1: Online ResourcesAppendix 2Appendix 3Appendix 4AcknowledgmentsReferences |
|---|
Bacterial endosymbionts are widespread across several insect orders and are involved in interactions ranging from obligate mutualism to reproductive parasitism. Candidatus Blochmannia gen. nov. (Blochmannia) is an obligate bacterial associate of Camponotus and related ant genera (Hymenoptera: Formicidae). The occurrence of Blochmannia in all Camponotus species sampled from field populations and its maternal transmission to host offspring suggest that this bacterium is engaged in a long-term, stable association with its ant hosts. However, evidence for cospeciation in this system is equivocal because previous phylogenetic studies were based on limited gene sampling, lacked statistical analysis of congruence, and have even suggested host switching. We compared phylogenies of host genes (the nuclear EF-1
F2 and mitochondrial COI/II) and Blochmannia genes (16S ribosomal DNA [rDNA], groEL, gidA, and rpsB), totaling more than 7 kilobases for each of 16 Camponotus species. Each data set was analyzed using maximum likelihood and Bayesian phylogenetic reconstruction methods. We found minimal conflict among host and symbiont phylogenies, and the few areas of discordance occurred at deep nodes that were poorly supported by individual data sets. Concatenated protein-coding genes produced a very well-resolved tree that, based on the Shimodaira–Hasegawa test, did not conflict with any host or symbiont data set. Correlated rates of synonymous substitution (dS) along corresponding branches of host and symbiont phylogenies further supported the hypothesis of cospeciation. These findings indicate that Blochmannia–Camponotus symbiosis has been evolutionarily stable throughout tens of millions of years. Based on inferred divergence times among the ant hosts, we estimated rates of sequence evolution of Blochmannia to be 
0.0024 substitutions per site per million years (s/s/MY) for the 16S rDNA gene and 0.1094 s/s/MY at synonymous positions of the genes sampled. These rates are several-fold higher than those for related bacteria Buchnera aphidicola and Escherichia coli. Phylogenetic congruence among Blochmannia genes indicates genome stability that typifies primary endosymbionts of insects.
Keywords: Blochmannia; Camponotus; cospeciation; endosymbiosis; molecular clock; phylogenetics
Accepted September 12, 2003
Bacteria have forged intracellular symbioses with insects several times independently, with significant effects on the fitness and diversification of their hosts. Primary (P) endosymbionts occur in several insect orders, including Coleoptera, Dictyoptera, Diptera, Hemiptera, and Hymenoptera (Buchner, 1965; Douglas, 1989). Common features of P endosymbionts include their need for host survival and reproduction and their restriction to specialized host cells (bacteriocytes) and female ovaries (or milk glands; Cheng and Aksoy, 1999) through which they are transmitted to the offspring (Buchner, 1965; Douglas, 1989; Baumann et al., 2000; Moran and Baumann, 2000; Thao et al., 2000b). In contrast to facultative symbionts that have transitory associations with hosts, P endosymbionts have been vertically inherited over long evolutionary timescales. Phylogenetic congruence suggests host–symbiont cospeciation over tens or hundreds of millions of years that can be traced back to an ancient infection event within each host group (Moran et al., 1993; Baumann et al., 1995; Chen et al., 1999; Clark et al., 2000; Thao et al., 2000b; Spaulding and von Dohlen, 2001; Lo et al., 2003). Genomes of P endosymbionts have signatures of extreme specialization resulting from this long-term intracellular lifestyle, including small genome size (
450–809 kilobases [kb]; Gil et al., 2002; Wernegreen et al., 2002) compared with the 4.5–5.5-megabase genomes of their close relative Escherichia coli (Bergthorsson and Ochman, 1998). Full genome sequence analysis illustrates that this reduction involves the loss of metabolic functions and DNA repair capabilities (e.g., Shigenobu et al., 2000; Tamas et al., 2001; Akman et al., 2002; van Ham et al., 2003), the latter of which may account for the extreme AT-richness of endosymbiont DNA sequences. Known functions of P endosymbionts include providing essential nutrients that are absent or in low quantities from the hosts' typically unbalanced diets (Nogge, 1981; Douglas and Prosser, 1992; Akman et al., 2002). Although these insect associates have originated independently in several bacterial groups, they are particularly common in the 
-Proteobacteria, which also includes the well-studied E. coli and related enterobacteria.
In contrast to P endosymbionts, certain bacteria form labile associations with insects and in addition to maternal transmission can spread horizontally to new hosts (Fukatsu et al., 2000; Fukatsu, 2001; Sandström et al., 2001; Russell et al., 2003). These secondary (S) endosymbionts are thought to have more subtle effects on host fitness, such as increased defense against parasitoids (Oliver et al., 2003). Phylogenetic analyses have revealed frequent horizontal transfer of these endosymbionts among insect species, often to entirely new host groups. Reconstructing the evolutionary history of insect–bacterium associations can help to distinguish whether a given bacterial associate better fits the model of P endosymbiont (cospeciation) or S endosymbiont (host switching). The aim of this study was to test for cospeciation between Camponotus, a diverse ant genus that represents > 10% of all ant species (Bolton, 1995), and its bacterial associate Blochmannia, a -3 proteobacterium that exhibits certain characteristics of both P and S endosymbionts.
Blochmannia of Camponotus
Blochmannia shows many hallmarks of a P endosymbiont that suggest long-term stability of its association with ants. Phylogenetic analysis of the 16S ribosomal DNA (rDNA) gene indicates a close relationship with Wigglesworthia and other endosymbionts in the -3 subdivision of the Proteobacteria (Schroder et al., 1996; Charles et al., 2001). Blochmannia is maintained with high fidelity and has been found in all field specimens of Camponotus, Polyrhachis, and Colobopsis screened thus far (Dasch, 1975; Dasch et al., 1984; Sameshima et al., 1999). This symbiosis is maintained in a variety of ecological contexts because the ant hosts have a cosmopolitan distribution across Old and New World environments ranging from deserts to tropical rainforests (Bolton, 1995). The ubiquity of Blochmannia in its host genera contrasts with the sporadic distribution of a related bacterium within the ant genus Formica, whose presence is variable among ant colonies of the same species (Dasch, 1975; Dasch et al., 1984). Like other P endosymbionts, Blochmannia is located within bacteriocytes, where the bacterial cells float freely in the host cytoplasm (although certain other P endosymbionts are contained within host-derived membranes) and within oocytes of queens, consistent with maternal transmission and integration into larvae during development (Sauer et al., 2002). The genome of Blochmannia, like that of many other P endosymbionts, has an extremely low GC content ( 27–30% GC; Dasch, 1975; Gil et al., 2003;) and is severely reduced ( 706–809 kb; Wernegreen et al., 2002; Gil et al., 2003).
Other aspects of Blochmannia biology suggest it might be horizontally transmitted among hosts. Although bacteriocytes of other insects are often clustered into an organlike bacteriome, those containing Blochmannia are intercalated among the ant midgut epithelial cells (Buchner, 1965; Schröder et al., 1996; Sauer et al., 2000, 2002). By virtue of this close proximity to the midgut, the bacteria might have greater opportunities for horizontal transfer among species (e.g., via body secretions). It is unclear whether Blochmannia provides an essential function for the host; these bacteria have been eliminated from lab-reared ants with no apparent detrimental effects on the host (Sauer et al., 2002). Certain arboreal Camponotus and Polyrhachis species have restricted diets composed mostly of extrafloral nectar and insect exudates (Davidson, 1997, 1998; Davidson et al., 2003), but most Camponotus species have diverse and complex diets that include plant material, insect secretions, and other arthropods (Dasch et al., 1984; Hölldobler and Wilson, 1990). Despite the complex diet of its host, the full genome sequence of Blochmannia from C. floridanus strongly suggests a nutritional basis for the symbiosis (Gil et al., 2003). The small Blochmannia genome retains genes for the biosynthesis of several amino acids essential to the host. It also retains genes for urease, which hydrolyzes urea to CO2 and ammonia, and genes for glutamine synthetase, which can recycle ammonia before it reaches toxic concentrations. The mutualism is truly reciprocal because Blochmannia relies on the host for most nonessential amino acids, vitamins, and cofactors. It has also been suggested that the energetic functions of Blochmannia might be most important at the colony founding stage (D. Hann, pers. comm.; Wernegreen et al., 2003), and that the bacterium might also be involved in the production of pheromones used for food recruitment (Sauer, 2000). If Blochmannia provides a more subtle or sporadic selective advantage than is typical of nutritional P endosymbionts, it may exhibit a more labile host association.
Reconstructing Host–Symbiont Associations
Phylogenetic approaches provide powerful tools for inferring the stability of host–symbiont associations. Even very rare symbiont transfer, if successful, is expected to disrupt phylogenetic congruence. An overall agreement of Camponotus and Blochmannia phylogenies suggests stability of symbiont transmission (Schröder et al., 1996; Sameshima et al., 1999; Sauer et al., 2000). Moreover, the monophyly of Camponotus, Polyrhachis, and Colobopsis and the co-occurrence of Blochmannia in all three genera (Sameshima et al., 1999) further support the hypothesis of a single infection of a common ancestor and subsequent codivergence. However, previous studies were based on a single gene of the host and symbiont, had limited power to resolve phylogenetic relationships, and suggested horizontal transfer of Blochmannia between two host species (from C. pennsylvanicus to C. castaneus; Sauer et al., 2000). Our aim in this study was to evaluate the stability of the Blochmannia–Camponotus association by analyzing several loci from the symbiont, host mitochondria, and host nucleus totaling > 7 kb for each of 16 host species. In addition, we used parameter-rich maximum likelihood (ML) and Bayesian analyses to improve the reliability of the estimated relationships (Felsenstein, 1981; Huelsenbeck and Ronquist, 2001). A variety of approaches can be employed to compare alternative evolutionary hypotheses, including ML (Goldman et al., 2000), Bayesian analysis (Huelsenbeck et al., 2000), and maximum parsimony (Brooks, 1981; Farris et al., 1994). We have selected a likelihood-based approach, the Shimodaira–Hasegawa (SH) test (Shimodaira and Hasegawa, 1999) based on several advantages reviewed previously (Goldman et al., 2000).
Demonstrating host–symbiont cospeciation offers a unique opportunity to calibrate rates of bacterial molecular evolution. Such rates are very difficult to estimate for free-living prokaryotes that lack a fossil record (Ochman et al., 1999) and experience lateral gene transfer (Lawrence and Ochman, 2002). The most reliable estimates of evolutionary rates in bacteria have been calculated for Buchnera, based on divergence times among the hosts as inferred from the fossil record of aphids or their preferred host plants (Moran et al., 1993; Clark et al., 1999). Here, we built upon evidence for cospeciation between Blochmannia and Camponotus by estimating absolute rates of sequence evolution in this endosymbiont for comparison with other bacterial groups.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionAppendix 1: Online ResourcesAppendix 2Appendix 3Appendix 4AcknowledgmentsReferences |
|---|
Specimens Examined
DNA sequence data were collected from 16 Camponotus species and two outgroup species, Formica fusca and Polyrhachis foreli. Most ant specimens were collected near the Southwestern Research Station (SWRS) in Portal, Arizona during the summer of 2001 and from Falmouth, Massachusetts during the summers of 2001 and 2002 (Table 1). Ants were identified in the field by J.J.W. or A.B.L., and vouchers were confirmed by Stefan Cover of the Museum of Comparative Zoology (MCZ, Cambridge, MA). Vouchers are available for inspection in the Wernegreen lab (MBL) or at the MCZ. All specimens used for DNA analysis were frozen and stored at –80°C. Specimens of C. floridanus, C. pennsylvanicus, and Polyrhachis foreli were obtained from colleagues (see Acknowledgments).
|
DNA Extraction, Amplification, and Sequencing
Genomic DNA was extracted from single ants using the DNeasy Tissue Kit (Qiagen) and diluted for polymerase chain reactions (PCRs). Six gene fragments were amplified: the Blochmannia genes for 16S rDNA, a heat shock protein (groEL), glucose inhibited cell division protein A (gidA), and ribosomal protein subunit B (rpsB), the Camponotus mitochondrial genes encoding cytochrome oxidase subunits I and II (COI/II), and the Camponotus nuclear gene encoding elongation factor 1 alpha F2 copy (EF-1
F2). Combined, these gene fragments totaled 7,342 base pairs (bp) of high-quality sequence data per Camponotus species sampled (Table 1). Specific PCR primers are listed in Table 2. PCRs contained 0.4 pmol/µl of primers, 1.0 mM dNTPs (Invitrogen), PCR buffer (Fisher and Promega), 0.02 U of Taq DNA polymerase (Fisher and Promega), 5 µ l of diluted genomic DNA, and MgCl2 (concentrations optimized for each primer pair) and were brought to a final volume of 50 µ l with double-distilled water. Reactions were performed on the PTC-0200 gradient cycler (MJ Research) under the following conditions: 2 min at 94°C followed by 35 cycles of 20 sec at 95°C, 50 sec at an annealing temperature specific to the primer pair, 1–2 min at 72°C. PCR products were cleaned using the Qiaquick PCR purification kit (Qiagen) and quantified by spectrophotometry and agarose gel analysis. Some products were cloned into plasmid vectors to obtain sufficient template for DNA sequencing. Appendix 1 lists detailed methods, including PCR conditions for each gene (Appendix 2). PCR products and clones were sequenced on an ABI 3700 automated sequencer using Big Dye 3.0 chemistry (Applied Biosystems). Internal sequencing primers were used to sequence products > 1 kb. Sequences were assembled and edited using PHRED, PHRAP, and CONSED. All DNA assemblies were checked by eye, and any ambiguous base assignments were changed to N.
|
Phylogenetic Analysis
Alignment and data sets considered
Protein-coding genes were aligned using translated amino acid sequences in ClustalX (Thompson et al., 1997) and manually edited using MacClade 4.05 (Maddison and Maddison, 2002). Non–protein-coding regions, including the tRNA and intergenic spacer between COI and COII and the introns of EF-1
F2, were aligned using ClustalX. Blochmannia 16S rDNA sequences were aligned using the Ribosomal Database Project II Sequence Aligner (Maidak et al., 2001), which considers the secondary structure of the rRNA molecule, and the resulting alignment was edited manually in MacClade. Regions with ambiguous alignment were few because the sequences were closely related, but any such regions were excluded from the analysis. After confirming that individual gene regions gave similar phylogenies, we concatenated the five protein-coding data sets (referred to hereinafter as the complete data set). We also analyzed combined data sets of the Blochmannia protein-coding genes gidA, groEL, and rpsB (the ALLBloch data set) and of the Camponotus protein-coding genes COI/II and EF-1F2 (the ALLCamp data set). Blochmannia 16S rDNA was not included in these concatenated protein-coding data sets because of the requirement for distinct models of sequence evolution for 16S rRNA (rate variation among sites following a gamma distribution) and protein-coding genes (varying rates among codon positions) in Bayesian analyses.
Outgroup selection
Phylogenetic analysis of 16S rDNA suggests that Wigglesworthia, the bacterial endosymbiont of tsetse flies, is closely related to Blochmannia (Charles et al., 2001). Thus, we used genes from the fully sequenced genome of Wigglesworthia brevipalpis (NC004344; Akman et al., 2002) as the outgroup for analyses of two of the endosymbiont protein-coding genes (gidA and groEL). As an outgroup for the 16S rDNA and rpsB analyses, we obtained these genes from Blochmannia of Polyrhachis foreli after confirming its position outside our Camponotus–Blochmannia sample at the host COI/II (data not shown). The related formicine ant Formica fusca was used as the outgroup for Camponotus COI/II and EF-1F2.
ML and Bayesian analyses
For ML analyses, substitution models for each data set were selected using Modeltest 3.06 (Posada and Crandall, 1998), and ML heuristic searches were performed using 100 random taxon addition replicates with tree bisection and reconnection (TBR) branch swapping. ML bootstrap support was determined by 100 bootstrap replicates, each using 10 random taxon addition replicates with TBR branch swapping. Searches were performed in parallel on a Beowulf cluster using the clusterpaup program (A.G. McArthur, jbpc.mbl.edu/mcarthur) and PAUP 4.0b10 (Swofford, 2002).
Bayesian analyses were performed using both informative and noninformative prior probabilities and a general time reversible (GTR) substitution model with unequal base frequencies and the portion of invariant sites estimated from the data. Rates were allowed to vary among sites according to a gamma distribution for 16S rDNA and introns. All data sets containing protein-coding genes were analyzed using a partitioned-likelihood Bayesian search where each codon position was assigned an independent substitution model similar to those listed previously. First and second coding positions were given a prior uniform shape parameter of the gamma distribution ( range of 0.1–1), and third coding positions and introns were given a range of 5–30. Concatenated data sets were further partitioned by gene origin; nuclear, mitochondrial, or endosymbiont. The parameters for all partitions were estimated independently: unlink shape = (all) revmat = (all) pinvar = (all). Four replicates of each independent and concatenated data set were run, each with four simultaneous and incrementally heated Markov chain Monte Carlo (MCMC) chains, starting from a random tree, for 10,000,000 generations (MrBayes 3.0b4; Ronquist and Huelsenbeck, 2003). The likelihood parameters and trees were sampled every 100 generations. The 50% majority rule consensus and posterior probabilities were calculated from the 50,001 trees sampled after the initial burn-in period and were averaged across the four replicates. To confirm that Bayesian runs had adequate convergence and mixing for a given data set, we tested for similarity of the means and variances of model likelihoods after burn-in across the four independent analyses.
Statistical tests of phylogenetic congruence
The above analyses generated the "best" ML and Bayesian phylogenies for each of nine data sets (six gene regions and three combined data sets). The ML tree topologies of the various data sets showed slight differences. We tested the significance of these topological differences using the SH test (1999), a nonparametric ML-based test of phylogenetic congruence. This test involved optimizing the ML parameters of each data set across its best ML tree and across alternative topologies (the best ML trees of other data sets). The significance of differences in negative log likelihoods (–lnL) was evaluated using the bootstrap (1,000 replicates) based on RELL sampling and full optimization (PAUP 4.0b10). For each tree topology, the summed likelihood scores of all data sets were calculated to identify the tree with the best cumulative support.
Calculating Rates of Sequence Evolution
Testing for correlated rates of synonymous divergence at host and symbiont genes
Correlated branch lengths on host and symbiont phylogenies provide further evidence for cospeciation (Hafner and Nadler, 1988). If hosts and symbionts each evolved at constant (but not necessarily identical) rates, then cospeciation would result in a correlation in the lengths of corresponding branches. We calculated rates of synonymous substitution (dS) along branches for each data set using codeML (PAML 3.13; Yang, 1997) across the complete topology. The codeML parameters allowed omega (ratio of the rates of nonsynonymous to synonymous substitution, or dN/dS) to vary across the topology (model = 1) and Ti/Tv was estimated from the data (fix_kappa = 0). We excluded any dS values along the root branches, values of zero, or values < 2, suggesting saturation of synonymous sites. The dS values at COI/II were plotted against the dS of corresponding branches of each symbiont protein-coding gene. Significance of the correlation was assessed using the Pearson product-moment correlation (r) (JMP 5.0, SAS Institute).
Evolutionary rates on an absolute timescale
Because Camponotus lacks a detailed fossil record, we necessarily took a less direct approach than previous studies of absolute evolutionary rates (Moran et al., 1993; Clark et al., 1999) (see Appendix 1). We used the program r8s (Sanderson, 2003) to estimate divergence dates among Camponotus species and then constrained the date of a basal Camponotus node to estimate evolutionary rates in Blochmannia. First, using codeML, we calculated dN along branches along a phylogeny of 828 bp of COI available for the relevant taxa. This phylogeny included Camponotus and other ants in the subfamily Formicinae, a representative of the subfamily Myrmicinae, and wasp outgroups (Appendix 1, Fig. A1). We constrained the ancestral node between the Formicinae and Myrmicinae to 90–110 million years ago (MYA) based on fossil evidence (Grimaldi and Agosti, 2000; Nascimbene and Silverstein, 2000). From this divergence time and dN at COI, we estimated more recent divergence times, including the basal node of the Camponotus species sampled (16.2–19.9 MYA). Next, using r8s we estimated absolute rates of Blochmannia sequence evolution based on this inferred date and branch lengths of Blochmannia gene trees. These branch lengths were estimated as dS of protein-coding genes (codeML) or as overall divergences at 16S rDNA using the model of Tamura and Nei (1993) and accounting for rate variation among sites using a gamma distribution with shape 0.5, as implemented by PAUP* (Swofford, 2002).
|
All estimates of dates and evolutionary rates above were calculated with the semiparametric penalized likelihood (PL) and nonparametric rate smoothing (NPRS) methods implemented by r8s. These smoothing methods are intended to correct for deviations from a molecular clock. PL outperforms NPRS when data are limited and some branches are very short (Sanderson, 2002), which is true of the COI data set. Thus, we placed greater emphasis on our results from the PL method.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionAppendix 1: Online ResourcesAppendix 2Appendix 3Appendix 4AcknowledgmentsReferences |
|---|
Sequence Characteristics
New sequence data for host nuclear, mitochondrial, and endosymbiont genomes totaled > 7 kb for each of 16 Camponotus species representing five subgenera. Protein-coding genes of Blochmannia showed a strong AT bias that is consistent with previous estimates of 30% genomic GC content (Dasch, 1975) (Table 3). The AT bias of the mitochondrial genes COI/II was consistent with previous analyses of Camponotus (30.4% GC at COI; Brady et al., 2000), and mitochondrial genes showed a pattern of GC1 = GC2 > GC3 observed previously in Camponotus and other ant genera (Chenuil and McKey, 1996; Wetterer et al., 1998; Brady et al., 2000). The exceptionally low GC content at silent sites of endosymbiont and mitochondrial genes would be expected if AT mutational pressure shaped base composition but was attenuated at replacement sites by selection on amino acid content (Muto and Osawa, 1987). In contrast to the AT-rich mitochondrial and endosymbiont genomes, the nuclear EF-1
F2 gene had a more moderate base composition (50.2%).
|
The COI/II sequence derived from C. vafer contained three unusual insertions and deletions that may shift the coding frame of both cytochrome oxidases I and II. These changes occurred at C. vafer alignment positions 511 and 512 (2-bp insertion), at positions 726–728 (3-bp insertion), and between positions 922 and 923 (4-bp deletion). These sequence results were confirmed by the independent PCR amplification and direct sequencing of an identical product from a separate C. vafer isolate. We introduced gaps in the COI alignment to account for these indels. The downstream region of COII, positions 1929–2117, contained several deletions, and this region was not included in subsequent analyses. Despite apparent frameshifts in C. vafer COI/II, the pairwise Ka/Ks between C. vafer and C. festinatus for the COI/II region included was 0.038, a value lower than most pairwise Ka/Ks between other Camponotus species at the same gene region (unpubl. data). Thus, if C. vafer COI is a pseudogene, it is quite young and has not accumulated mutations at nonsynonymous sites. This finding justifies our inclusion of C. vafer in the COI/II analysis and assignment of codon positions based on the full alignment.
Blochmannia 16S rDNA genes had moderate base composition (average of 48.7%) and contained putative intervening sequences (IVSs) observed previously (Sauer et al., 2000). Because of alignment ambiguity, we excluded IVS-like regions from phylogenetic analyses. In other bacterial species, IVSs are excised from the rRNA molecule during posttranscriptional modification (Rainey et al., 1996). We identified a region that aligns with an insert designated II (Sauer et al., 2000), which is inserted in E. coli 16S rDNA between bp 207–212 and 208–211 and varied from 6 to 48 bp among 15 of the taxa we sampled (data are incomplete for C. nearcticus). It was previously reported that both Wigglesworthia and Buchnera lack a similar insertion (Sauer et al., 2000). However, although Buchnera sp. APS lacks the insert (Shigenobu et al., 2000), both copies of 16S rDNA in W. brevipalpis contain a 14-bp insert.
Phylogenetic Reconstruction
ML tree topologies were generally congruent across datasets and with phylogenies generated from Bayesian approaches (Fig. 1). For each data set, the four independent Bayesian runs produced the same topology and very similar posterior probabilities, with 
3% variation among values of any such node. In addition, the four runs for a given data set had very similar means and variances of total likelihood. For all data sets, the mean of each run fell within the 95% credibility interval of the other three (MrBayes 3.0b4). For a given node, ML bootstrap values were typically lower than Bayesian posterior probability scores, as predicted from modeling approaches (Cummings et al., 2003). Most clades were highly congruent and well supported across trees, and each phylogeny clearly distinguished four major clades (Fig. 1). Any topological differences tended to occur at deep nodes, some of which had relatively low support in these single-gene analyses. ML nucleotide substitution parameters of each data set were reoptimized across the ML phylogeny of that data set. Variation in substitution patterns among data sets illustrated distinct modes of DNA sequence among the loci sampled (Table 4). Subgenera defined by morphological characters were not always monophyletic (Fig. 2), as shown previously (Brady et al., 2000).
|
EF-1
|
|
The combined data set of all protein-coding genes (complete data set) produced the same phylogeny using ML and Bayesian approaches, with bootstrap scores and posterior probabilities generally > 84% (Fig. 2). The one exception was the ML bootstrap support of 47% for the grouping of C. ulcerosus and C. floridanus, although this node had 96% posterior probability. Phylogenies estimated from the concatenated ALLCamp and ALLBloch data sets showed strong congruence at well-supported nodes (Fig. 3).
|
Statistical Comparison of Alternative Topologies
We compared the ML phylogenies of each data set using the SH test, which unlike the commonly implemented Kishino–Hasegawa test (Kishino and Hasegawa, 1989) accounts for the fact that alternative topologies were estimated from the data (Goldman et al., 2000). We first optimized each data set across its "best" ML tree and across the alternative topologies (i.e., the ML trees estimated from other data sets). The phylogeny estimated from all protein-coding genes (complete data set) had the best cumulative likelihood score, summed for all data sets across that tree (Lsum = 70341.29; Table 5). We next applied the SH test to evaluate the significance of difference between the –lnL of a given data set across its ML tree and the alternative topologies. We found few cases of significant conflict with alternative topologies, reflected in the scarcity of significant values in Table 6. The ML tree of EF-1
F2 showed the most conflict with other data sets, but this topology was poorly supported even by the EF-1 F2 data set because of low variation in this nuclear gene (Table 6; Fig. 1a). The topology of the tree generated by combined protein-coding genes (complete data set, Fig. 2) is identical to that of the tree estimated from all bacterial protein-coding genes (Fig. 3a) and was not rejected by any host or symbiont data set. We consider this tree our best estimate of the shared host and symbiont phylogeny. Although 16S rDNA was not included in the complete data set, the 16S rDNA phylogeny does not significantly disagree with the complete phylogeny.
|
|
Rates of Host and Symbiont Evolution
Correlated rates of synonymous substitution
Branch lengths of host and symbiont trees were compared by plotting dS values calculated along corresponding branches. The dS values of each bacterial gene showed a significant correlation with dS of COI/II (groEL, r = 0.685, P < 0.0002; gidA, r = 0.817, P < 0.0001; rpsB, r = 0.435, P < 0.038). EF-1
F2 was excluded because of extremely low dS values within Camponotus.
Absolute rates of sequence evolution
Absolute rates of evolution within the subfamily Formicinae were estimated based on fossil evidence that this subfamily diverged from the ant subfamily Myrmicinae approximately 90–110 MYA (Grimaldi and Agosti, 2000). Using the PL method in r8s, we estimated rates of nonsynonymous substitution at an 828-bp region of COI to be 0.000425 ± 6.5e-05 substitutions per site per million years (s/s/MY) (assuming the subfamilies diverged 110 MYA) or 0.000521 ± 8.3e-05 s/s/MY (assuming they diverged 90 MYA) (Appendices 1, 3). This rate is faster than rates previously calculated across several insect orders (0.00022 s/s/MY at second codon positions of COI; Gaunt and Miles, 2002) but is consistent with elevated rates of mitochondrial evolution in Hymenoptera (e.g., Crozier and Crozier, 1992). The same r8s analysis estimated divergence times within the Formicinae. The ancestral node of all ants known to have Blochmannia was dated at 29.3–35.9 MYA (Appendix 1 Fig. A1), suggesting the Blochmannia–ant association is at least 30 MY old and perhaps much older if the symbiont once infected other genera but was subsequently lost from them. The divergence time of Camponotus included here was estimated at 16.2–19.9 MYA (Fig. 2).
From this estimated divergence time, we used r8s to calculate absolute rates of synonymous substitution at Blochmannia gidA and groEL and total substitution rates at 16S rDNA (Table 7). Rates of synonymous substitution at Blochmannia gidA and groEL ranged from 0.0854 to 0.1372 s/s/MY, depending on the gene and the date at which the basal Camponotus node was constrained (16.2 or 19.9 MYA). Compared with Buchnera, rates in Blochmannia were about 14 times faster at 16S rDNA and 5–10 times faster at synonymous sites of gidA and groEL. Compared with E. coli, Blochmannia apparently evolves about 29 times faster at the 16S rDNA gene (Table 7, Appendix 4). Direct comparisons of dS values are difficult to interpret because of codon bias in E. coli at the high-expression genes sampled. Although the NPRS method gave slightly slower rate estimates than did PL (Table 7, Appendix 3), both methods suggested faster absolute rates in Blochmannia than in Buchnera and E. coli.
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionAppendix 1: Online ResourcesAppendix 2Appendix 3Appendix 4AcknowledgmentsReferences |
|---|
Obligate endosymbionts of insects, like most of the microbial world, cannot be cultured in the lab and have thus proved difficult to study with traditional genetic or physiological techniques. Full genome sequencing has illustrated the mechanisms and consequences of severe genome size reduction among obligate bacterial associates, including their extreme specialization for the host cellular environment and the retention of host-beneficial genes (e.g., Shigenobu et al., 2000; Tamas et al., 2001; Akman et al., 2002; Moran, 2002; Gil et al., 2003). Furthermore, molecular phylogenetic analysis has revealed that symbiont dynamics range from strict cospeciation of P endosymbionts and their hosts (Chen et al., 1999; Thao et al., 2000b), to the frequent host switching of S endosymbionts (Fukatsu et al., 2000; Thao et al., 2000a; Fukatsu, 2001; Sandström et al., 2001; Russell et al., 2003) and reproductive parasites such as Wolbachia (Werren et al., 1995). Distinguishing among these transmission patterns has important implications for understanding the evolutionary stability, approximate age, and potential for coevolution within a particular association.
The aim of this study was to test for cospeciation between the endosymbiont Blochmannia and its Camponotus hosts by expanding the available sequence data sets and by using statistically reliable approaches to estimate and compare phylogenies. The restriction of this bacterium to bacteriocytes and the female reproductive tissues suggests that speciation of the host and symbiont are linked (Schröder et al., 1996; Sauer et al., 2002). Although lineage sorting can disrupt phylogenetic congruence (e.g., Hardig et al., 2000), it is unlikely to operate in this system. Our PCR analysis of multiple host species has not detected more than a single Blochmannia strain per host. However, the ability to eliminate this symbiont from lab-reared ants without detectable host effects (Sauer et al., 2002) and previous suggestions of host switching (Sauer et al., 2000) suggest that Blochmannia may form a more labile association than is typical of other P endosymbionts.
Prior to this study, phylogenetic analyses had not accounted for rate variation across loci and were based on relatively small data sets. A molecular phylogenetic analysis of Camponotus based on 385 bp of COI showed that certain subgenera might not be monophyletic, suggesting homoplasy of the morphological characters on which subgenera are based (pilosity and head shape; Brady et al., 2000). We sampled a nuclear gene of Camponotus and an extended region of mitochondrial genes (COI/II) often used in insect systematics (e.g., Rokas et al., 2002). Our analyses also indicate that certain Camponotus subgenera are not monophyletic. Inconsistencies we observed include the placement of C. schaefferi (subgenus Camponotus) with species of the subgenus Myrmentoma and the close relationship of C. castaneus (subgenus Tanaemyrmex) to C. americanus (subgenus Camponotus). Most notably, the subgenus Tanaemyrmex apparently does not represent a monophyletic group, as observed previously (Brady et al., 2000).
Congruence between the phylogenies of host data sets and those of endosymbiont genes indicate that stable vertical transmission of symbionts to host offspring leads to bacterium–ant cospeciation. The phylogeny based on all protein-coding genes has the best cumulative likelihood score, is identical to the that estimated from bacterial protein-coding genes, and is consistent with the host data sets. In addition, correlations in dS values in host mitochondrial and endosymbiont phylogenies indicate similar levels of synonymous changes along branches for these two linked, maternally inherited genomes. These branch length correlations support the hypothesis of cospeciation but are not required by it. Even if cospeciation were true, variable evolutionary rates in either the host or symbiont lineages could disrupt the correlation in rates (Hafner and Nadler, 1988). Phylogenetic congruence among Blochmannia genes sampled here illustrates a lack of horizontal gene transfer and contrasts with frequent recombination among free-living bacterial strains (Dykhuizen and Green, 1991; Ochman et al., 2000). This apparent genome stability of Blochmannia is similar to that observed in Buchnera and Wigglesworthia, which show no evidence of lateral gene transfer (Funk et al., 2000; Wernegreen and Moran, 2001; Akman et al., 2002; Tamas et al., 2002).
Hypotheses other than cospeciation can explain topological congruence of host and symbiont phylogenies. For example, geographic effects may generate this pattern when hosts within a given geographic region tend to associate with symbionts in the same location. Our sample of Camponotus included both closely related species from distinct geographic regions and, conversely, phylogenetically diverse species from the same field site (often adjacent to the other species' nests). Thus, observed topological congruence in this study cannot be explained by a tendency for co-occurring host and symbionts to form associations. Moreover, geographic effects would not necessarily predict the observed correlation of host and symbiont branch lengths.
Establishing host–symbiont cospeciation opens the possibility of calibrating a symbiont clock based on host divergence times. Molecular clock estimates for Buchnera have been calibrated against a host fossil record and provide the most reliable clock available for bacteria (Moran et al., 1993; Clark et al., 1999). At synonymous sites of low codon bias genes, Buchnera evolves about twice as fast as E. coli and Salmonella typhimurium on an absolute timescale (Clark et al., 1999). This rate elevation might result from higher mutation rates in small endosymbiont genomes that lack many repair genes (Eisen and Hanawalt, 1999; Moran and Wernegreen, 2000), and/or accelerated fixation of slightly deleterious mutations due to increased genetic drift in small populations (Moran, 1996; Funk et al., 2001; Woolfit and Bromham, 2003). Here, we calculated rates of sequence evolution in Blochmannia from inferred divergence times of its Camponotus hosts. At homologous gene regions, Blochmannia evolves surprisingly faster than either Buchnera or the enterobacteria. Our results suggest that Blochmannia 16S rDNA evolves about 14 times faster than that of Buchnera and about 29 times faster than that of the enterics (represented here by E. coli and S. typhimurium). At synonymous sites of gidA and groEL, rates in Blochmannia are 5–10 times higher than those in Buchnera. Previous studies have interpreted high divergence among Blochmannia strains to mean that this ant–bacterium association is very old and perhaps an original attribute of ants that was subsequently lost in most ant lineages (Schröder et al., 1996; Boursaux-Eude and Gross, 2000). However, host–symbiont cospeciation means this high divergence occurred in the context of its association with Camponotus. Although our rate estimates may be error prone because of uncertain divergence times among Camponotus hosts, Blochmannia apparently evolves many times faster than Buchnera. For example, for the 16S rDNA, applying evolutionary rates for Buchnera (inferred here or by other investigators; Clark et al., 2000) to Blochmannia sequence divergences would predict that the Blochmannia–Camponotus association is > 100 MY old, which far predates the inferred origin of the subfamily Formicinae (Grimaldi and Agosti, 2000). This evidence for fast evolutionary rates in Blochmannia would be strengthened by comparing rates across other genes and by better fossil data to assign dates to nodes within the Camponotus phylogeny.
|
Appendix 1: Online Resources |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionAppendix 1: Online ResourcesAppendix 2Appendix 3Appendix 4AcknowledgmentsReferences |
|---|
PCR Amplification Strategies
A fragment of the 16S rDNA gene was amplified using the SL and SR universal eubacterial primers (Schroder et al., 1996) often in combination with a Blochmannia-specific 16S rDNA primer to generate two overlapping PCR fragments. A gene fragment spanning COI/II was generated using the C1-J-1754F and C2-N-3661R primers (Simon et al., 1994) for most isolates. The only exception was C. sansabeanus, for which a new forward COI primer was designed to create a slightly smaller fragment. The nuclear EF-1
genes occur in two copies (F1 and F2) in Hymenoptera (Danforth and Ji, 1998). We first amplified both copies using primers designed from the EF-1 genes of Apis mellifera (AF015267
[GenBank]
). Both EF-1 copies were TA cloned and sequenced from two divergent Camponotus species (C. pennsylvanicus and C. festinatus) to design PCR primers specific to the Camponotus F1 and F2 copies. EF-1F2 was selected for analysis because of its slightly larger size (reflecting a greater number of introns). Primers for Blochmannia genes rpsB, gidA, and groEL were designed from alignments of enterobacteria and Buchnera sequences available in GenBank. Combined, these gene fragments totaled 7,342 bp of high quality sequence data per isolate sampled (Appendix 2). When PCR failed to yield sufficient template for direct sequencing, fragments were cloned into plasmid vectors using the TOPO TA cloning kit (Invitrogen) and electrocompetent cells according to the manufacturer's instructions. Purified plasmid clones were prepared using the Qiaquick spin miniprep kit (Qiagen) or the RevOrbit alkaline lysis 96-well plate purification system (Genomic Solutions). Clones containing inserts were screened prior to sequencing, using both restriction digests with EcoRI and NotI (New England Biolabs) and PCR amplification with M13F and M13R primers.
Estimating Absolute Rates of Sequence Evolution: Details of the Approach and Calculations
The goal of this analysis was to estimate rates of nucleotide substitution in Blochmannia on an absolute timescale (as substitutions per site per million years: s/s/MY) for comparison with absolute rate estimates for Buchnera and the enterics (Moran et al., 1993; Clark et al., 1999; Ochman et al., 1999). To estimate the rates of substitution in Blochmannia, we would ideally assign the dates of divergence for their Camponotus host lineages. However, because Camponotus itself lacks a detailed fossil record, we necessarily took a less direct approach that involved the following steps.
Estimating rates for the subfamily Formicinae on an absolute timescale
We first estimated sequence divergence between the subfamilies Myrmicinae and Formicinae, which diverged 90–110 MYA according to fossil evidence (Grimaldi and Agosti, 2000). Divergences were calculated at nonsynonymous sites of the mitochondrial gene COI, the locus that provided the most robust molecular clock in a previous study of several insect orders (Gaunt and Miles, 2002). Two regions of COI were selected based on the availability of myrmicine sequences in GenBank: bp 1816–2214 (region A) and bp 2254–2652 (region B) (positions correspond to genome sequence of the Drosophila melanogaster mitochondrion, NC001709). Among myrmicine species, region A was available for the Crematogaster borneensis and region B was available for Leptothorax recedens. Relative rates tests (calculated by the method of Muse and Weir, 1992) showed no significant difference in rates of nonsynonymous substitution in formicines and myrmicines at either region of COI (A or B) when either Vespidae or Scoliidae were used as outgroups. We then calculated an evolutionary rate on an absolute timescale for each of the regions A and B by averaging the pairwise estimates of dN (PAML; Yang, 1997) between the single myrmicine representative and formicine COI sequences and then dividing by the inferred divergence time (90–110 MYA). The rates based on regions A and B both performed similarly; they both estimated comparable divergence times of Colobopsis and Camponotus. Given the lack of conflict between the dates estimated from regions A (Crematogaster) and B (Leptothorax), we combined these regions for the rates analysis below.
The r8s program (Sanderson, 2003) implements several methods and algorithms for estimation of absolute rates of molecular evolution and divergence times across a phylogeny. The penalized likelihood (PL) method was most appropriate for the COI data set, but the nonparametric rate smoothing (NPRS) method was also used for comparison of results (Appendix 3). All analyses assumed a divergence time of 90–110 MYA between Formicinae and Myrmicinae. These values were independently fixed (using the "fixage" command in r8s) to the shared node representing the most recent common ancestor of Formicinae and Mymicinae (Fig. A1). Cross-validation analyses of the PL method were used to optimize the smoothing parameters for each age, 90 and 110 MYA, using both the Powell and truncated Newton (TN) algorithms. Calculations were rerun using the optimal smoothing parameter (110 MYA: smoothing = 150; 90 MYA: smoothing = 100), which estimated rates at nonsynonymous sites of this COI region to be 0.000425 ± 6.5e-05 s/s/MY and 0.000521 ± 8.3e-05 s/s/MY, respectively. NPRS was run using the Powell algorithm and estimated a slower rates of 0.000251 ± 2.3e-04 and 0.000306 ± 2.8e-04 s/s/MY under the assumption of 110 and 90 MYA divergence times, respectively. The lower rate is generally consistent with a molecular clock previously calculated at the same gene across several insect orders (0.00022 s/s/MY at second codon positions of COI, which was calibrated using the Blattaria/Orthoptera divergence at 380–354 MYA; Gaunt and Miles, 2002).
Evolutionary rates generated by the PL and NPRS methods are quite different (Appendix 3). This result is not surprising, because simulation studies have shown that NPRS performs best when the data set is large and the number of substitutions per branch is high (Sanderson, 1997, 2002). The short length of this COI data set (828 bp) and zero values for dN along certain branches may compromise the performance of NPRS (Sanderson, 2002). For these reasons, we consider the PL results to be more reliable for this data set.
Based on the 90–110 MYA divergence and the estimated dN values at COI, we then used the PL method of r8s to estimate the divergence of Camponotus and Colobopsis at 29.3–35.9 MYA, a minimum age for the Blochmannia–ant association. The same r8s analysis calculated the basal Camponotus node to be 16.2–19.9 MYA. This dated node was used to estimate absolute rates of Blochmannia sequence evolution.
Calculating absolute rates of Blochmannia sequence evolution
We calculated dS values for Blochmannia gidA and groEL along branches of the complete topology, using codeML. Branch lengths for the 16S rDNA phylogeny were calculated in PAUP* using the Tamura and Nei (1993) model, accounting for rate variation among sites under a gamma distribution ( = 0.5). By constraining the basal Camponotus node to 16.2 or 19.9 MYA, Blochmannia branch lengths were then used to estimate absolute rates (r8s, PL method) (Table 7).
Estimating rates in Buchnera and the enterobacteria
For comparison, we calculated rates for Buchnera and the enterics at the same gene regions we analyzed for Blochmannia. Rates for Buchnera were based on pairwise divergences between Buchnera-Ap and Buchnera-Sg with an estimated divergence time of 50–70 MYA (Tamas et al., 2002). Rates for the enterics were calculated for E. coli and S. typhimurium, which diverged 100–150 MYA (Ochman and Wilson, 1987). See Appendix 4 for details of these calculations. Rates for Buchnera and E. coli were consistent with previous estimates based on many more genes (Clark et al., 1999). In the previous study, average rates of KS for Buchnera ranged from 0.0051 to 0.0095 s/s/MY depending on Buchnera lineages considered and assumptions about their divergence time (Clark et al.,1999). Our estimates of dS between Buchnera-Ap and Buchnera-Sg were similar but had a slightly higher maximum rate (0.0058–0.020344 s/s/MY at gidA and groEL). Rates at Buchnera 16S rDNA ranged from 0.00019 to 0.00045 s/s/MY (depending on the pair of Buchnera taxa and inferred divergence time) (Clark et al., 2000). Our slightly lower estimates (0.000148 to 0.000208 s/s/MY) may reflect different Buchnera taxa used or the restriction of our estimates to 1.36 kb of the of 16S rDNA gene (a region that is also available for Blochmannia). Previous estimates of 16S rDNA rates for the enterics were 0.00012–0.00018 s/s/MY (Clark et al., 1999); ours were about half that (0.000066–0.000100 s/s/MY) at the particular 16S rDNA region sampled.
|
Appendix 2 |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionAppendix 1: Online ResourcesAppendix 2Appendix 3Appendix 4AcknowledgmentsReferences |
|---|
|
|
Appendix 3 |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionAppendix 1: Online ResourcesAppendix 2Appendix 3Appendix 4AcknowledgmentsReferences |
|---|
|
|
Appendix 4 |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionAppendix 1: Online ResourcesAppendix 2Appendix 3Appendix 4AcknowledgmentsReferences |
|---|
|
|
Acknowledgments |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionAppendix 1: Online ResourcesAppendix 2Appendix 3Appendix 4AcknowledgmentsReferences |
|---|
We thank Stefan Cover and Bill MacKay for their assistance with ant identification and for sharing their insights, along with those of Gary Alpert, Brian Fisher, and Daniel Hahn, into Camponotus ecology. We thank Stefan Cover, Mark Seid, Chris Starkweather, and Simon Robson for contributing insect specimens. We are grateful to Josh Herbeck, Andrew McArthur, and Seth Bordenstein for critical insights and comments on the manuscript and to Seán Brady for helpful discussion of Camponotus systematics. This article benefited greatly from the constructive suggestions of Carol von Dohlen, Kevin Johnson, Chris Simon, and an anonymous reviewer. We also thank Kevin Johnson and Dale Clayton for organizing the 2002 SSB Symposium "Untangling Coevolutionary History" at the SSB/SSE annual meeting in Champaign, Illinois, where some of these results were presented. Funding for this research was provided by the NIH (R01 GM62626-01), the NSF (DEB 0089455 and associated REU supplement award), the NASA Astrobiology Institute (NCC2-1054), and the Josephine Bay Paul and C. Michael Paul Foundation.
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionAppendix 1: Online ResourcesAppendix 2Appendix 3Appendix 4AcknowledgmentsReferences |
|---|
-
Akman L., Yamashita A., Watanabe H., Oshima K., Shiba T., Hattori M., Aksoy S. Genome sequence of the endocellular obligate symbiont of tsetse flies, Wigglesworthia glossinidia. Nat. Genet. (2002) 32:402–407.[CrossRef][Web of Science][Medline]
Baumann P., Baumann L., Lai C. Y., Rouhbakhsh D., Moran N. A., Clark M. A. Genetics, physiology, and evolutionary relationships of the genus Buchnera: Intracellular symbionts of aphids. Annu. Rev. Microbiol. (1995) 49:55–94.[CrossRef][Web of Science][Medline]
Baumann P., Moran N., Baumann L. Bacteriocyte-associated endosymbionts of insects. In: The prokaryotes, a handbook on the biology of bacteria; ecophysiology, isolation, identification, applications—Dworkin M., ed. (2000) New York: Springer-Verlag. 3rd edition, release 3.1. http://link.springer-ny.com/link/service/books/10125.
Bergthorsson U., Ochman H. Distribution of chromosome length variation in natural isolates of Escherichia coli. Mol. Biol. Evol. (1998) 15:6–16.[Abstract]
Bolton B. Identification guide to the ant genera of the world (1995) Cambridge, Massachusetts: Harvard Univ. Press.
Boursaux-Eude C., Gross R. New insights into symbiotic associations between ants and bacteria. Res. Microbiol. (2000) 151:513–519.[Medline]
Brady S. G., Gadau J., Ward P. S. Systematics of the ant genus Camponotus (Hymenoptera: Formicidae): A preliminary analysis using data from the mitochondrial gene cytochrome oxidase I. In: Hymenoptera. Evolution, biodiversity and biological control—Austin A. D., Dowton M., eds. (2000) Collingwood Victoria, Australia: CSIRO. 131–139. Pages.
Brooks D. Hennig's parasitological method: A proposed solution. Syst. Zool. (1981) 30:229–249.
Buchner P. Endosymbiosis of animals with plant microorganisms (1965) New York: Interscience.
Charles H., Heddi A., Rahbe Y. A putative insect intracellular endosymbiont stem clade, within the Enterobacteriaceae, inferred from phylogenetic analysis based on a heterogeneous model of DNA evolution. C.R. Acad. Sci. III (2001) 324:489–494.
Chen X., Li S., Aksoy S. Concordant evolution of a symbiont with its host insect species: Molecular phylogeny of genus Glossina and its bacteriome-associated endosymbiont, Wigglesworthia glossinidia. J. Mol. Evol. (1999) 48:49–58.[CrossRef][Web of Science][Medline]
Cheng Q., Aksoy S. Tissue tropism, transmission and expression of foreign genes in vivo in midgut symbionts of tsetse flies. Insect Mol. Biol. (1999) 8:125–132.[CrossRef][Web of Science][Medline]
Chenuil A., McKey D. B. Molecular phylogenetic study of a myrmecophyte symbiosis: Did Leonardoxa/ ant associations diversify via cospeciation? Mol. Phylogenet. Evol. (1996) 6:270–286.[CrossRef]
Clark M. A., Moran N. A., Baumann P. Sequence evolution in bacterial endosymbionts having extreme base compositions. Mol. Biol. Evol. (1999) 16:1586–1598.[Abstract]
Clark M. A., Moran N. A., Baumann P., Wernegreen J. J. Cospeciation between bacterial endosymbionts (Buchnera) and a recent radiation of aphids (Uroleucon) and pitfalls of testing for phylogenetic congruence. Evolution (2000) 54:517–525.[CrossRef][Web of Science][Medline]
Crozier R. H., Crozier Y. C. The cytochrome b and ATPase genes of honeybee mitochondrial DNA. Mol. Biol. Evol. (1992) 9:474–482.[Abstract]
Cummings M. P., Handley S. A., Myers D. S., Reed D. L., Rokas A., Winka K. Comparing bootstrap and posterior probability values in the four-taxon case. Syst. Biol. (2003) 52:477–487.
Danforth B. N., Ji S. Elongation factor-1 alpha occurs as two copies in bees: Implications for phylogenetic analysis of EF-1 alpha sequences in insects. Mol. Biol. Evol. (1998) 15:225–235.[Abstract]
Dasch G. A. Morphological and molecular studies on intracellular bacterial symbiotes of insects (1975) New Haven, Connecticut: Yale Univ. Press.
Dasch G., Weiss E., Chang K. Endosymbionts of insects. In: Bergy's manual of systematic bacteriology—Holt J., Krieg N., eds. (1984) Baltimore: Williams & Wilkins. 811–833. Pages.
Davidson D. The role of resource imbalances in the evolutionary ecology of tropical arboreal ants. Biol. J. Linn. Soc. Lond. (1997) 61:153–181.[CrossRef][Web of Science]
Davidson D. Resource discovery versus resource domination in ants: A functional mechanism for breaking the trade-off. Ecol. Entomol. (1998) 23:484–490.[CrossRef][Web of Science]
Davidson D. W., Cook S. C., Snelling R. R., Chua T. H. Explaining the abundance of ants in lowland tropical rainforest canopies. Science (2003) 300:969–972.
Douglas A. E. Mycetocyte symbiosis in insects. Biol. Rev. Camb. Philos. Soc. (1989) 64:409–434.[Medline]
Douglas A., Prosser W. Synthesis of the essential amino acid tryptophan in the pea aphid (Acyrthosiphon pisum) symbiosis. J. Insect Physiol. (1992) 38:565–568.[CrossRef][Web of Science]
Dykhuizen D. E., Green L. Recombination in Escherichia coli and the definition of biological species. J. Bacteriol. (1991) 173:7257–7268.
Eisen J. A., Hanawalt P. C. A phylogenomic study of DNA repair genes, proteins, and processes. Mutat. Res. (1999) 435:171–213.[Web of Science][Medline]
Farris J. S., Kállersjö M., Kluge A. G., Bult C. Testing significance of congruence. Cladistics (1994) 10:315–320.[CrossRef][Web of Science]
Felsenstein J. Evolutionary trees from DNA sequences: A maximum likelihood approach. J. Mol. Evol. (1981) 17:368–376.[CrossRef][Web of Science][Medline]
Fukatsu T. Secondary intracellular symbiotic bacteria in aphids of the genus Yamatocallis (Homoptera: Aphididae: Drepanosiphinae). Appl. Environ. Microbiol. (2001) 67:5315–5320.
Fukatsu T., Nikoh N., Kawai R., Koga R. The secondary endosymbiotic bacterium of the pea aphid Acyrthosiphon pisum (Insecta: Homoptera). Appl. Environ. Microbiol. (2000) 66:2748–2758.
Funk D. J., Helbling L., Wernegreen J. J., Moran N. A. Intraspecific phylogenetic congruence among multiple symbiont genomes. Proc. R. Soc. Lond. B Biol. Sci. (2000) 267:2517–2521.
Funk D. J., Wernegreen J. J., Moran N. A. Intraspecific variation in symbiont genomes: Bottlenecks and the aphid–Buchnera association. Genetics (2001) 157:477–489.
Gaunt M. W., Miles M. A. An insect molecular clock dates the origin of the insects and accords with palaeontological and biogeographic landmarks. Mol. Biol. Evol. (2002) 19:748–761.
Gil R., Sabater-Munoz B., Latorre A., Silva F. J., Moya A. Extreme genome reduction in Buchnera spp.: Toward the minimal genome needed for symbiotic life. Proc. Natl. Acad. Sci. USA (2002) 99:4454–4458.
Gil R., Silva F. J., Zientz E., Delmotte F., Gonzalez-Candelas F., Latorre A., Rausell C., Kamerbeek J., Gadau J., Holldobler B., van Ham R. C., Gross R., Moya A. The genome sequence of Blochmannia floridanus: Comparative analysis of reduced genomes. Proc. Natl. Acad. Sci. USA (2003) 100:9388–9393.
Goldman N., Anderson J. P., Rodrigo A. G. Likelihood-based tests of topologies in phylogenetics. Syst. Biol. (2000) 49:652–670.
Grimaldi D., Agosti D. A formicine in New Jersey Cretaceous amber (Hymenoptera: Formicidae) and early evolution of the ants. Proc. Natl. Acad. Sci. USA (2000) 97:13678–13683.
Hafner M. S., Nadler S. A. Phylogenetic trees support the coevolution of parasites and their hosts. Nature (1988) 332:258–259.[CrossRef][Medline]
Hardig T. M., Soltis P. S., Soltis D. E. Diversification of the North American shrub genus Ceanothus (Rhamnaceae): Conflicting phylogenies from nuclear ribosomal DNA and chloroplast DNA. Am. J. Bot. (2000) 87:108–123.
Hölldobler B., Wilson E. O. The ants (1990) Cambridge, Massachusetts: Belknap Press.
Huelsenbeck J. P., Rannala B., Larget B. A Bayesian framework for the analysis of cospeciation. Evolution (2000) 54:353–364.
Huelsenbeck J. P., Ronquist F. MrBayes: Bayesian inference of phylogenetic trees. Bioinformatics (2001) 17:754–755.
Kishino H., Hasegawa M. Evaluation of the maximum likelihood estimate of the evolutionary tree topologies from DNA sequence data, and the branching order in Hominoidea. J. Mol. Evol. (1989) 29:170–179.[CrossRef][Web of Science][Medline]
Lawrence J. G., Ochman H. Reconciling the many faces of lateral gene transfer. Trends Microbiol. (2002) 10:1–4.[CrossRef][Web of Science][Medline]
Lo N., Bandi C., Watanabe H., Nalepa C., Beninati T. Evidence for cocladogenesis between diverse dictyopteran lineages and their intracellular endosymbionts. Mol. Biol. Evol. (2003) 20:907–913.
Maddison D. R., Maddison W. P. MacClade 4: Analysis of phylogeny and character evolution, version 4.05 (2002) Sunderland, Massachusetts: Sinauer.
Maidak B. L., Cole J. R., Lilburn T. G., Parker C. T. Jr., Saxman P. R., Farris R. J., Garrity G. M., Olsen G. J., Schmidt T. M., Tiedje J. M. The RDP-II (Ribosomal Database Project). Nucleic Acids Res. (2001) 29:173–174.
Moran N. A. Accelerated evolution and Muller's ratchet in endosymbiotic bacteria. Proc. Natl. Acad. Sci. USA (1996) 93:2873–2878.
Moran N. A. Microbial minimalism: Genome reduction in bacterial pathogens. Cell (2002) 108:583–586.[CrossRef][Web of Science][Medline]
Moran N. A., Baumann P. Bacterial endosymbionts in animals. Curr. Opin. Microbiol. (2000) 3:270–275.[CrossRef][Web of Science][Medline]
Moran N. A., Munson M. A., Baumann P., Ishikawa H. A molecular clock in endosymbiotic bacteria is calibrated using the insect hosts. Proc. R. Soc. Lond. B Biol. Sci. (1993) 253:167–171.
Moran N. A., Wernegreen J. J. Lifestyle evolution in symbiotic bacteria: Insights from genomics. Trends Ecol. Evol. (2000) 15:321–326.[CrossRef][Medline]
Muse S. V., Weir B. S. Testing for equality of evolutionary rates. Genetics (1992) 132:269–276.[Abstract]
Muto A., Osawa S. The guanine and cytosine content of genomic DNA and bacterial evolution. Proc. Natl. Acad. Sci. USA (1987) 84:166–169.
Nascimbene P., Silverstein H. The preparation of fragile Cretaceous ambers for conservation and study of organismal inclusions. In: Studies on fossils in amber, with particular reference to the Cretaceous of New Jersey—Grimaldi D., ed. (2000) Leiden: Backhuys. 93–102. Pages.
Nogge G. Significance of symbionts for the maintenance of an optimal nutritional state for successful reproduction in hematophagous arthropods. Parasitology (1981) 82:299–304.
Ochman H., Elwyn S., Moran N. A. Calibrating bacterial evolution. Proc. Natl. Acad. Sci. USA (1999) 96:12638–12643.
Ochman H., Lawrence J. G., Groisman E. A. Lateral gene transfer and the nature of bacterial innovation. Nature (2000) 405:299–304.[CrossRef][Medline]
Ochman H., Wilson A. C. Evolution in bacteria: Evidence for a universal substitution rate in cellular genomes. J. Mol. Evol. (1987) 26:74–86.[CrossRef][Web of Science][Medline]
Oliver K. M., Russell J. A., Moran N. A., Hunter M. S. Facultative bacterial symbionts in aphids confer resistance to parasitic wasps. Proc. Natl. Acad. Sci. USA (2003) 100:1803–1807.
Posada D., Crandall K. A. Modeltest: Testing the model of DNA substitution. Bioinformatics (1998) 14:817–818.
Rainey F. A., Ward-Rainey N. L., Janssen P. H., Hippe H., Stackebrandt E. Clostridium paradoxum DSM 7308T contains multiple 16S rRNA genes with heterogeneous intervening sequences. Microbiology (1996) 142:2087–2095.
Rokas A., Nylander J. A., Ronquist F., Stone G. N. A maximum-likelihood analysis of eight phylogenetic markers in gallwasps (Hymenoptera: Cynipidae): Implications for insect phylogenetic studies. Mol. Phylogenet. Evol. (2002) 22:206–219.[CrossRef][Web of Science][Medline]
Ronquist F., Huelsenbeck J. P. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics (2003) 19:1572–1574.
Russell J. A., Latorre A., Sabater-Munoz B., Moya A., Moran N. A. Side-stepping secondary symbionts: Widespread horizontal transfer across and beyond the Aphidoidea. Mol. Ecol. (2003) 12:1061–1075.[CrossRef][Medline]
Sameshima S., Hasegawa E., Kitade O., Minaka N., Matsumoto T. Phylogenetic comparison of endosymbionts with their host ants based on molecular evidence. Zool. Sci. (1999) 16:993–1000.[CrossRef][Web of Science]
Sanderson M. J. A nonparametric approach to estimating divergence times in the absence of rate constancy. Mol. Biol. Evol. (1997) 14:1218–1231.[Web of Science]
Sanderson M. J. Estimating absolute rates of molecular evolution and divergence times: A penalized likelihood approach. Mol. Biol. Evol. (2002) 19:101–109.
Sanderson M. J. r8s: Inferring absolute rates of molecular evolution and divergence times in the absence of a molecular clock. Bioinformatics (2003) 19:301–302.
Sandström J. P., Russell J. A., White J. P., Moran N. A. Independent origins and horizontal transfer of bacterial symbionts of aphids. Mol. Ecol. (2001) 10:217–228.[CrossRef][Medline]
Sauer C. Charakterisierung intrazellulärer, bakterieller Endosymbionten im Mitteldarm von Ameisen der Gattung Camponotus. In: Biozentrum Universität Würzburg (2000) WürzburgGermany. 112. Page.
Sauer C., Dudaczek D., Hölldobler B., Gross R. Tissue localization of the endosymbiotic bacterium "Candidatus Blochmannia floridanus" in adults and larvae of the carpenter ant Camponotus floridanus. Appl. Environ. Microbiol. (2002) 68:4187–4193.
Sauer C., Stackebrandt E., Gadau J., Hölldobler B., Gross R. Systematic relationships and cospeciation of bacterial endosymbionts and their carpenter ant host species: Proposal of the new taxon Candidatus Blochmannia gen. nov. Int. J. Syst. Evol. Microbiol. (2000) 50:1877–1886.
Schröder D., Deppisch H., Obermayer M., Krohne G., Stackebrandt E., Hölldobler B., Goebel W., Gross R. Intracellular endosymbiotic bacteria of Camponotus species (carpenter ants): Systematics, evolution and ultrastructural characterization. Mol. Microbiol. (1996) 21:479–489.[Web of Science][Medline]
Shigenobu S., Watanabe H., Hattori M., Sakaki Y., Ishikawa H. Genome sequence of the endocellular bacterial symbiont of aphids Buchnera sp. APS. Nature (2000) 407:81–86.
Shimodaira H., Hasegawa M. Multiple comparisons of log-likelihoods with applications to phylogenetic inference. Mol. Biol. Evol. (1999) 16:1114–1116.[Web of Science]
Simon C., Frati F., Beckenbach A., Crespi B., Liu H., Flook P. Evolution, weighting and phylogenetic utility of mitochondrial gene sequences and a compilation of conserved polymerase chain reaction primers. Ann. Entomol. Soc. Am. (1994) 87:651–701.[Web of Science]
Spaulding A. W., von Dohlen C. D. Psyllid endosymbionts exhibit patterns of co-speciation with hosts and destabilizing substitutions in ribosomal RNA. Insect Mol. Biol. (2001) 10:57–67.[CrossRef][Web of Science][Medline]
Swofford D. L. PAUP*: Phylogenetic analysis using parsimony (*and other methods), version 4 (2002) Sunderland, Massachusetts: Sinauer.
Tamas I., Klasson L., Canback B., Naslund A. K., Eriksson A. S., Wernegreen J. J., Sandström J. P., Moran N. A., Andersson S. G. 50 million years of genomic stasis in endosymbiotic bacteria. Science (2002) 296:2376–2379.
Tamas I., Klasson L. M., Sandström J. P., Andersson S. G. Mutualists and parasites: How to paint yourself into a (metabolic) corner. FEBS Lett. (2001) 498:135–139.[CrossRef][Web of Science][Medline]
Tamura K., Nei M. Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Mol. Biol. Evol. (1993) 10:512–526.[Abstract]
Thao M. L., Clark M. A., Baumann L., Brennan E. B., Moran N. A., Baumann P. Secondary endosymbionts of psyllids have been acquired multiple times. Curr. Microbiol. (2000a) 41:300–304.[CrossRef][Web of Science][Medline]
Thao M. L., Moran N. A., Abbot P., Brennan E. B., Burckhardt D. H., Baumann P. Cospeciation of psyllids and their primary prokaryotic endosymbionts. Appl. Environ. Microbiol. (2000b) 66:2898–2905.
Thompson J. D., Gibson T. J., Plewniak F., Jeanmougin F., Higgins D. G. The CLUSTAL_X windows interface: Flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. (1997) 25:4876–4882.
van Ham R. C., Kamerbeek J., Palacios C., Rausell C., Abascal F., Bastolla U., Fernandez J. M., Jimenez L., Postigo M., Silva F. J., Tamames J., Viguera E., Latorre A., Valencia A., Moran F., Moya A. Reductive genome evolution in Buchnera aphidicola. Proc. Natl. Acad. Sci. USA (2003) 100:581–586.
Wernegreen J. J., Degnan P. H., Lazarus A. B., Palacios C., Bordenstein S. R. Genome evolution in an insect cell: Distinct features of an ant–bacterial partnership. Biol. Bull. (2003) 204:221–232.
Wernegreen J. J., Lazarus A. B., Degnan P. H. Small genome of Candidatus Blochmannia, the bacterial endosymbiont of Camponotus, implies irreversible specialization to an intracellular lifestyle. Microbiology (2002) 148:2551–2556.
Wernegreen J. J., Moran N. A. Vertical transmission of biosynthetic plasmids in aphid endosymbionts (Buchnera). J. Bacteriol. (2001) 183:785–790.
Werren J. H., Zhang W., Guo L. R. Evolution and phylogeny of Wolbachia: Reproductive parasites of arthropods. Proc. R. Soc. Lond. B Biol. Sci. (1995) 261:55–63.
Wetterer J. K., Schultz T. R., Meier R. Phylogeny of fungus-growing ants (tribe Attini) based on mtDNA sequence and morphology. Mol. Phylogenet. Evol. (1998) 9:42–47.[CrossRef][Web of Science][Medline]
Woolfit M., Bromham L. Increased rates of sequence evolution in endosymbiotic bacteria and fungi with small effective population sizes. Mol. Biol. Evol. (2003) 20:1545–1555.
Yang Z. PAML: A program package for phylogenetic analysis by maximum likelihood. Comput. Appl. Biosci. (1997) 13:555–556.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  S. E. Massey The Proteomic Constraint and Its Role in Molecular Evolution Mol. Biol. Evol., December 1, 2008; 25(12): 2557 - 2565. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Ueda, S.-P. Quek, T. Itioka, K. Inamori, Y. Sato, K. Murase, and T. Itino An ancient tripartite symbiosis of plants, ants and scale insects Proc R Soc B, October 22, 2008; 275(1649): 2319 - 2326. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Conord, L. Despres, A. Vallier, S. Balmand, C. Miquel, S. Zundel, G. Lemperiere, and A. Heddi Long-Term Evolutionary Stability of Bacterial Endosymbiosis in Curculionoidea: Additional Evidence of Symbiont Replacement in the Dryophthoridae Family Mol. Biol. Evol., May 1, 2008; 25(5): 859 - 868. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. J. Thornhill, A. A. Wiley, A. L. Campbell, F. F. Bartol, A. Teske, and K. M. Halanych Endosymbionts of Siboglinum fiordicum and the Phylogeny of Bacterial Endosymbionts in Siboglinidae (Annelida) Biol. Bull., April 1, 2008; 214(2): 135 - 144. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. J. Stewart, C. R. Young, and C. M. Cavanaugh Lateral Symbiont Acquisition in a Maternally Transmitted Chemosynthetic Clam Endosymbiosis Mol. Biol. Evol., April 1, 2008; 25(4): 673 - 687. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Zientz, I. Beyaert, R. Gross, and H. Feldhaar Relevance of the Endosymbiosis of Blochmannia floridanus and Carpenter Ants at Different Stages of the Life Cycle of the Host Appl. Envir. Microbiol., September 1, 2006; 72(9): 6027 - 6033. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. A. Moran, P. Tran, and N. M. Gerardo Symbiosis and Insect Diversification: an Ancient Symbiont of Sap-Feeding Insects from the Bacterial Phylum Bacteroidetes Appl. Envir. Microbiol., December 1, 2005; 71(12): 8802 - 8810. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. H. Degnan, A. B. Lazarus, and J. J. Wernegreen Genome sequence of Blochmannia pennsylvanicus indicates parallel evolutionary trends among bacterial mutualists of insects Genome Res., August 1, 2005; 15(8): 1023 - 1033. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||