History of Infection With Different Male-Killing Bacteria in the Two-Spot Ladybird Beetle Adalia bipunctata Revealed Through Mitochondrial DNA Sequence Analysis

Corresponding author: J. Hinrich G. v. d. Schulenburg, Institut für Evolution und Ökologie der Tiere, Westfälische Wilhelms-Universität Münster, Hüfferstr. 1, 48149 Münster, Germany., hschulen{at}uni-muenster.de (E-mail)

Communicating editor: D. CHARLESWORTH

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The two-spot ladybird beetle Adalia bipunctata (Coleoptera: Coccinellidae) is host to four different intracellular maternally inherited bacteria that kill male hosts during embryogenesis: one each of the genus Rickettsia (-Proteobacteria) and Spiroplasma (Mollicutes) and two distinct strains of Wolbachia (-Proteobacteria). The history of infection with these male-killers was explored using host mitochondrial DNA, which is linked with the bacteria due to joint maternal inheritance. Two variable regions, 610 bp of cytochrome oxidase subunit I and 563 bp of NADH dehydrogenase subunit 5, were isolated from 52 A. bipunctata with known infection status and different geographic origin from across Eurasia. Two outgroup taxa were also considered. DNA sequence analysis revealed that the distribution of mitochondrial haplotypes is not associated with geography. Rather, it correlates with infection status, confirming linkage disequilibrium between mitochondria and bacteria. The data strongly suggest that the Rickettsia male-killer invaded the host earlier than the other taxa. Further, the male-killing Spiroplasma is indicated to have undergone a recent and extensive spread through host populations. In general, male-killing in A. bipunctata seems to represent a highly dynamic system, which should prove useful in future studies on the evolutionary dynamics of this peculiar type of symbiont-host association.

MALE-KILLING is known from an increasing number of arthropod host species and seems to be particularly common in ladybird beetles. It refers to the killing of male embryos or larvae by intracellular, maternally inherited microorganisms. Male-killing is used by these symbionts to enhance their spread through host populations, since infected female hosts gain in fitness by the death of their brothers due to inbreeding avoidance, reduction of antagonistic sibling interactions, and/or the availability of additional resources (HURST et al. 1997 ; HURST and JIGGINS 2000 ).

A unique situation was recently reported from the two-spot ladybird beetle, Adalia bipunctata L. (Coleoptera: Coccinellidae): It harbors at least four different male-killing bacteria, one each of the genus Rickettsia (-Proteobacteria) and Spiroplasma (Mollicutes), and two distinct strains of Wolbachia (-Proteobacteria; WERREN et al. 1994 ; HURST et al. 1999A , HURST et al. 1999B ). Two of these (Rickettsia, Spiroplasma) show a large overlap in their geographic distribution, whereas the Wolbachia could be isolated only from Central Russia. In Moscow (Russia), all four symbionts coexist (HURST et al. 1999A , HURST et al. 1999B ; MAJERUS et al. 2000 ; SCHULENBURG et al. 2000B ). The above finding is particularly unusual as minimal models of male-killing suggest the coexistence of several male-killers to be unstable at equilibrium (HURST et al. 1997 ; RANDERSON et al. 2000 ). Consequently, either male-killing in A. bipunctata is currently not at equilibrium or there are other factors such as resistance genes in the host species or continuously fluctuating environmental conditions that permit establishment of a stable equilibrium (see also HURST et al. 1999B ; MAJERUS et al. 2000 ; RANDERSON et al. 2000 ).

More detailed information on the unusual coexistence of several male-killers is of great importance for our general understanding of male-killing, as it may permit detection of the factors that shape the evolutionary dynamics of this type of symbiont-host association. This study focuses on characterizing the history of infection with male-killing symbionts. Such historical information can most efficiently be extracted from molecular markers. However, bacterial DNA sequences are unsuitable for this approach. Previous analyses highlighted that none of the DNA regions currently available from the three symbiont taxa permits inference of relative diversification dates because they do not evolve with constant rates, thus rendering application of a molecular clock invalid (SCHULENBURG et al. 2000A , SCHULENBURG et al. 2000B ). Therefore, we decided to use host mitochondrial DNA (mtDNA), which is in linkage disequilibrium with the intracellular bacteria, because both are maternally inherited. Consequently, mtDNA genealogies should permit reconstruction of symbiont evolution within the host.

An association between mitochondria and intracellular symbionts has already been documented for other symbiotic bacteria that manipulate host reproduction. The best-studied example refers to Drosophila simulans (Diptera), in which Wolbachia bacteria cause cytoplasmic incompatibility. Here, mtDNA variants were found to "hitchhike" with particular symbiont strains that spread through the host population, leading to congruence between mtDNA and Wolbachia gene genealogies (NIGRO and PROUT 1990 ; TURELLI et al. 1992 ; BALLARD et al. 1996 ; JAMES and BALLARD 2000 ). A similar relationship between the spread of Wolbachia symbionts and the evolution of the respective host mtDNA molecules was observed in the mosquito Culex pipiens (Diptera; GUILLEMAUD et al. 1997 ), the woodlouse Armadillidium vulgare (Crustacea: Isopoda; RIGAUD et al. 1999 ), the oak gallwasp Biorhiza pallida (Hymenoptera; ROKAS et al. 2001 ), fire ants of the genus Solenopsis (Hymenoptera; SHOEMAKER et al. 2001 ), and the Asian rice gall midge Orseolia oryzae (Diptera; BEHURA et al. 2001 ).

For male-killing symbionts, the dynamics of such an association has so far been investigated only by computer simulation, based on population genetic models of mtDNA evolution and spread of a single male-killing symbiont (JOHNSTONE and HURST 1996 ). Intriguingly, the results obtained suggest that the evolution of host mtDNA should be severely affected in this case. As killing of male hosts confers a fitness advantage to the surviving siblings, maternally inherited genes from male-killer-bearing individuals will gain increased population frequencies. Hence, an increase in the abundance of daughters of infected broods will produce a concomitant increase in the frequency of a mitochondrial variant originally associated with the symbiont. This mitochondrial variant can even be predicted to reach an equilibrium frequency in the host population higher than that of the male-killing symbiont because mitochondria must show 100% vertical transmission, whereas transmission of the bacteria is usually imperfect (e.g., HURST et al. 1997 ). Consequently, the distribution of mitotypes in both the infected and the uninfected part of the host population provides information as to the history of infection with male-killing symbionts.

Here, we present an analysis of host mtDNA for a combined inference of the history of the different male-killing symbionts in A. bipunctata. For this purpose, mtDNA sequence data from two variable regions, part of the cytochrome oxidase subunit I (COI) and part of the NADH dehydrogenase subunit 5 gene (ND5), were isolated from 52 A. bipunctata specimens with known infection status and from different locations across Eurasia. For reliable outgroup comparisons, data were also obtained from two additional ladybird beetle species, A. decempunctata L. and Coccinella septempunctata L. Both phylogenetic and population genetic analyses were subsequently used to infer mtDNA genealogy and population dynamics as estimates of the demography of male-killing symbionts in A. bipunctata.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Materials:
Ladybird beetle material was either derived from previous studies (HURST et al. 1993 , HURST et al. 1999A , HURST et al. 1999B ; WERREN et al. 1994 ; MAJERUS et al. 2000 ; SCHULENBURG et al. 2000B , SCHULENBURG et al. 2001A ) or obtained in the course of the current work, as indicated in Table 1. In the former case, infection status had been inferred using a combination of breeding experiments and molecular genetic techniques. In the latter case, ladybird specimens from different locations were first sexed using the method of RANDALL et al. 1992 . Female beetles were then analyzed for the presence of the previously identified male-killing bacteria. For this purpose, we employed PCR assays specific for three bacterial groups (the Spiroplasma ixodetis clade, the genus Rickettsia, and A + B group Wolbachia) as described previously (MAJERUS et al. 2000 ; SCHULENBURG et al. 2001A ). All infected specimens contained only one of the four different symbionts, consistent with previous studies that did not furnish evidence of double infections (HURST et al. 1993 ; HURST et al. 1999A , HURST et al. 1999B ; WERREN et al. 1994 ; MAJERUS et al. 2000 ; SCHULENBURG et al. 2000B , SCHULENBURG et al. 2001A ).

Strategy for choice of variable mtDNA regions:
The entire cytochrome oxidase subunits I and II (COI + II) gene region and almost the complete NADH dehydrogenase subunit 5 (ND5) gene region were first isolated from two A. bipunctata specimens and a single A. decempunctata individual. Sequence analysis of these regions served to identify mtDNA fragments that were sufficiently variable for population genetic analysis. The COI + II and ND5 gene regions were chosen for such an analysis because they have previously been indicated to contain some of the most variable parts of the mitochondrial genome in insects (e.g., BEARD et al. 1993 ; SIMON et al. 1994 ; LUNT et al. 1996 ; BALLARD 2000 ). Analysis of noncoding DNA such as the A + T-rich region was not attempted because in insects, this region has in various cases been found to be less variable than some of the coding DNA (ZHANG and HEWITT 1997 ). More importantly, interpretation of the variation observed in this region is hampered in this class of organisms by an extremely high preponderance of A and T nucleotides (usually >90%) and the abundance of indels and tandemly repeated elements, which all complicate reliable identification of homologies (ZHANG and HEWITT 1997 ). The most variable parts within the COI + II and ND5 gene regions were then sequenced from various A. bipunctata.

Molecular methods:
Genomic DNA was isolated from individual ladybirds, using the CTAB-based protocol outlined in SCHULENBURG et al. 2001B . The COI + II and ND5 gene regions were amplified via PCR from A. bipunctata from two different populations, Cambridge and Moscow (specimens BCA26 and MOS18, respectively), and A. decempunctata from Berlin. For this purpose, primers were designed in conserved regions adjacent to the genes, using a comparison with previously published complete mtDNA sequences from insects (CLARY and WOLSTENHOLME 1985 ; BEARD et al. 1993 ; CROZIER and CROZIER 1993 ; MITCHELL et al. 1993 ; FLOOK et al. 1995 ; LEWIS et al. 1995 ; see Table 2 for primer sequences). The COI + II gene region was isolated in two overlapping fragments, using primers TY-J-1462 x TL2-N-3014 and C1-J-2630 x TK-N-3795. The ND5 gene region was obtained in a continuous fragment from A. bipunctata with primers N3-J-5945 x N5-N-7849. These primers failed to amplify this region from A. decempunctata. On the basis of the sequences from A. bipunctata and published data, two additional primers, TN-J-6172 and N5-N-7789, were therefore constructed and subsequently employed for PCR isolation of the ND5 gene region from this ladybird species. Using PCR and partial sequencing of the amplification product, these two primers were also used on one of the A. bipunctata specimens (MOS18) to confirm that they produce results identical to those obtained with the more distal primers.

PCR was performed with the Expand High Fidelity PCR system (Boehringer Mannheim Ltd., Mannheim, Germany), following manufacturer's instructions. PCR cycling was controlled by the Progene thermal cycler (Techne Ltd.), using the following profile: 2 min at 94°, followed by 10 cycles of 15 sec at 94°, 1 min at 50°, 1 min and 30 sec at 70°, followed by 20 cycles of 15 sec at 94°, 1 min at 50°, and 1 min and 30 sec at 70° with an additional 10 sec for each cycle, and completed by a final extension step of 10 min at 70° (see instructions of the Expand High Fidelity PCR system; Boehringer Mannheim). PCR products were subsequently purified with Microcon-50 microconcentrators (Amicon, Beverly, MA) and directly sequenced for both strands, using a variety of internal sequencing primers (Table 2). DNA sequencing was performed with the ABI Prism BigDye Terminator cycle sequencing kit and the ABI Prism 377 DNA sequencer (Applied Biosystems, Foster City, CA). The resulting sequences were aligned by eye, taking into account the coding structure of the genes, and the variation between taxa and across sites was assessed.

Reliability of mtDNA data:
DNA sequencing reactions on the PCR fragments produced, in each case, distinct unambiguous signals for the identified nucleotides. Therefore, they did not indicate presence of intraindividual heteroplasmy, which may hamper reliable analysis of mtDNA (e.g., Rand 1993 ; VOGLER and DESALLE 1993 ; VOGLER et al. 1993 ). In addition, intraspecific variation between the protein-encoding genes was exclusively synonymous. The majority of differences between species were similarly due to silent substitutions. Since nuclear copies of mitochondrial genes lose their function and are thus expected to show a random distribution of substitutions (no bias toward synonymous substitutions; see ZHANG and HEWITT 1996 ), the sequences isolated are unlikely to represent nuclear integrations, which would show nonmaternal inheritance and would thus forbid inference of symbiont evolution.

Isolation of variable mtDNA regions:
On the basis of an analysis of variation across sites, the two most variable regions were chosen for demographic inferences: 610 bp in the middle of the COI gene and 563 bp toward the 5' end of the ND5 gene. These regions were isolated from A. bipunctata specimens with known infection status from a variety of locations (Table 1). They were also studied for two additional A. decempunctata specimens from Cambridge and Moscow and one C. septempunctata specimen from Berlin, to obtain reliable data for outgroup comparisons. For all specimens, the complete COI gene and most of the ND5 gene were amplified via PCR as above (primers TY-J-1462 x TL2-N-3014 for COI and TN-J-6172 x N5-N-7789 for ND5). The variable regions were then directly sequenced from PCR products with internal primers (C1-J-1951 and C1-N-2618 for COI; N5-J-7183 and N5-N-7789 for ND5).

Phylogenetic analysis:
DNA sequences obtained were aligned manually. The incongruence-length difference test (P = 0.42, FARRIS et al. 1995 ), as implemented in PAUP*, version 4.0b8 (SWOFFORD 1998 ), and the pattern of variation between sequences suggested homogeneity of character partitions between the two gene regions. Analysis was thus performed on combined COI and ND5 gene region sequences. For this purpose, four data sets were constructed as outlined in Table 3.

Phylogenetic tree estimation was performed with maximum likelihood as implemented in PAUP*, 4.0b8 (SWOFFORD 1998 ), using data set 1. Choice of substitution model was based on the approach of HUELSENBECK and CRANDALL 1997 . Likelihood scores were thus obtained for a variety of substitution models and thereafter compared using the program MODELTEST (POSADA and CRANDALL 1998 ). The general time reversible model with gamma-distributed rate heterogeneity across sites and a proportion of invariable positions (GTR--I; e.g., YANG et al. 1994 ; YANG 1996 ) was found to provide the most realistic representation of sequence evolution for our data set. This substitution model was subsequently used for tree estimation and nonparametric bootstrapping (FELSENSTEIN 1985 ). Both procedures were performed with the heuristic search modus using branch swapping by nearest-neighbor interchanges and random addition of sequences with 10 repetitions. Nonparametric bootstrapping was based on 100 replicate data sets.

Analysis of genetic differentiation:
MtDNA genetic differentiation within A. bipunctata was assessed using differently structured data sets (Table 3). Data set 2 included all sequences isolated from uninfected A. bipunctata and was subdivided into those derived from Eastern (St. Petersburg, Moscow, Tomsk, Tuva) and Western ladybird populations (Cambridge, Bielefeld, Berlin, Bayreuth, Ribe). Data set 3 was similarly split up into Western and Eastern populations but contained all A. bipunctata mtDNA sequences obtained from the respective locations. Data set 4 was subdivided into sequences isolated from uninfected specimens or from ladybirds that harbored Rickettsia, Spiroplasma, or the two Wolbachia male-killers. Note that the two Wolbachia strains were taken together to increase sample size of this data subset, and no additional Wolbachia-host lineages are as yet available. Combination of these two strains is justified because both form an exclusive monophyletic group (SCHULENBURG et al. 2000A ). The three structured data sets were then used to test whether mtDNA from A. bipunctata showed significant genetic differentiation in relation to bacterial infection or to geographic origin. In the former case, significant differentiation would indicate that bacteria were primarily maternally inherited and thus bias distribution of mitotypes among host specimens. In the latter case, geographic subdivision was analyzed for both uninfected and all studied specimens to assess whether the hypothesis of linkage disequilibrium between mitochondria and male-killing symbionts or the presence of additional bacteria in the Eastern populations (namely the Wolbachia symbionts) had any effect on the distribution of mitotypes between geographic regions. For each subset of data sets 2, 3, and 4, mitotype frequencies were recorded. In addition, the nucleotide diversity, (NEI 1987 ), was calculated with the program ARLEQUIN, version 2.0 (SCHNEIDER et al. 2000 ), assuming the TN- substitution model (TAMURA and NEI 1993 ). Genetic differentiation was investigated by an analysis of variance framework, as initially defined by COCKERHAM 1969 , COCKERHAM 1973 and WEIR and COCKERHAM 1984 , using the analysis of molecular variance (AMOVA) and the related fixation indices (F_ST) as implemented in the program ARLEQUIN (EXCOFFIER et al. 1992 ; SCHNEIDER et al. 2000 ). AMOVA and the F_ST measures allow consideration of both the frequency of haplotypes and also the evolutionary distances between these. In comparison to the original approach and most alternative methods, they therefore permit a more realistic assessment of the presence and extent of genetic differentiation (EXCOFFIER et al. 1992 ; SCHNEIDER et al. 2000 ). The evolutionary distances were estimated from pairwise compared mitotypes assuming the TN- model, which is the most sophisticated substitution model available in ARLEQUIN and thus most similar to the previously identified best-fitting model for this data set (GTR--I). The -shape parameter for gamma-distributed rate heterogeneity across sites, as required for the TN- model implementation in ARLEQUIN, was estimated a priori with PAUP*, version 4.0b8, using data set 1. In AMOVA, a significant deviation from a random distribution of haplotypes is inferred by permuting haplotypes among data subsets, using 1000 replications.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

General information:
COI and ND5 gene sequences were obtained from 52 A. bipunctata specimens from nine different populations. Of these, 13 were infected with Rickettsia, 10 with Spiroplasma, 3 each with Wolbachia strain Y and Z, and 23 were uninfected. The ladybirds included produced 10 different sequences for both COI and ND5. Linkage between COI and ND5 mitotypes differed between some specimens (Table 1). A combined gene data set therefore included 16 different mitochondrial variants. Data for 3 specimens of A. decempunctata from Cambridge, Berlin, and Moscow, and from a single ladybird of C. septempunctata were additionally considered for outgroup comparisons. As A. decempunctata from Cambridge and Bielefeld produced identical sequences in both regions, only one of them was included in subsequent analyses.The combined sequence alignment comprised 1173 positions (610 from the COI and 563 from the ND5 gene). Of these, 253 were variable between all sequences included (21.57% of the alignment sites) and 84 between those from A. bipunctata (7.16%). Pairwise compared sequences from this ladybird produced up to 68 nucleotide differences (5.80%), although the majority of unique sequences did not differ by more than six positions (0.51%). Variation between species was much higher, ranging between 115 and 185 differences between pairwise compared sequences (9.80–15.77%; Table 4).

Distribution of mitotypes among specimens:
Nine of the 16 different mitotypes were isolated from at least two specimens. Five were found in at least five beetles, including mitotypes 1A, 1B, 2A, 9I, and 10J. Specimens from Eastern and Western populations yielded about the same number of different mitotypes, irrespective of infected ladybird beetles being excluded from the data set or not (in the former case, 9 Western vs. 13 Eastern mitotypes; in the latter case, 7 vs. 9; see data sets 2 and 3, Table 5). Three mitotypes were found exclusively in the West (1F, 4A, 7A) and 7 exclusively in the East, including 1 of the more frequent mitotypes (1B; the others were: 1E, 1G, 3D, 5B, 6B, 8H). Nevertheless, 5 mitotypes were present in both the Eastern and Western populations, including 4 of the most frequent mitotypes (1A, 2A, 9I, 10J). Nucleotide diversity estimates were almost identical for the different geographic regions (data sets 2 and 3, Table 5). Beetles with male-killing symbionts produced a total of 10 mitotypes. One of these was associated with more than one male-killer taxon: mitotype 1B was present in three specimens with Wolbachia and two with Spiroplasma infections. The remaining mitotypes were never associated with more than one male-killing bacterium (data set 4, Table 5). In general, uninfected ladybird beetles and those that harbored Wolbachia symbionts produced a larger relative number of different mitotypes than did specimens with the other two male-killers. However, nucleotide diversity was clearly highest for the specimens infected with Rickettsia, followed by uninfected ladybirds, those with Wolbachia, and finally those with Spiroplasma male-killers (data set 4, Table 5). If the two Wolbachia strains were dealt with separately, then the three specimens with strain Y had 3 different mitotypes (3D, 6B, 8H). They produced a larger value for than did those associated with strain Z, which all bore identical mitotypes (mitotype 1B; = 0).

Tree estimation:
Maximum likelihood tree estimation identified mitotype 10J to be the representative of the most basal lineage of the A. bipunctata clade. Mitotypes 9I and 7A were respectively found to belong to the next two early diversifying lineages. The inferred tree also indicated a monophyletic origin of mitotypes 3B, 3D, 6B, and 8H, and also of this clade and 1B and 5B (Fig 1). However, the majority of A. bipunctata mitotypes, all of which share a high degree of sequence similarity (mitotypes 1A–8H), were separated by only very short branches. This suggests that the data may not contain sufficient information for reliable phylogenetic inferences at this level. This is confirmed by nonparametric bootstrapping. With the exception of an early diversification of mitotypes 10J and 9I, no other clades within the A. bipunctata assemblage were supported by bootstrap values >50. Uninfected specimens are found in the clade of the highly similar mitotypes and also in the lineage leading to mitotype 9I, but not 10J. Spiroplasma and Wolbachia male-killers are associated only with the clade of highly similar mitotypes, whereas Rickettsia-bearing hosts are found additionally among the members of the early diversifying lineages (Fig 1).

View larger version (12K): In this window In a new window Download PPT slide   Figure 1. Phylogeny of ladybird mtDNA sequences (left) and the number of specimens bearing one of the mitotypes from A. bipunctata (right). Tree topology and branch lengths were estimated with maximum likelihood, as implemented in PAUP*, version 4.0b8 (SWOFFORD 1998 ), using the general time reversible model with rate heterogeneity across sites and a proportion of invariable positions (e.g., YANG et al. 1994 ; YANG 1996 ). Tree search was based on branch swapping by nearest-neighbor interchanges and random addition of sequences with 10 repetitions. The tree shown produced a likelihood of ln . Values above branches refer to the results of nonparametric bootstrapping (FELSENSTEIN 1985 ), based on the same methods as above and 100 replicate data sets. Only values >50 are shown. Branch lengths are given in proportion to the estimated number of substitutions per site (see bar in bottom left corner). The branches leading to the outgroup CS were shortened to permit increasing the resolution of the A. bipunctata clade (see double slash). Note that no resolution is obtained for the relationships between (i) 3B, 8H, 6B, and 3D; (ii) this clade, 1B, and 5B; (iii) the latter group, 1A, 4A, 2A, 1G, 1F, 1E, 1C, and 7A. Nomenclature of A. bipunctata mitotypes is as in Table 1. AD1 and AD2 denote the sequences from A. decempunctata from Cambridge/Berlin and Moscow, respectively. CS refers to that from C. septempunctata. On the right, horizontal bars are given in proportion to the number of specimens with one of the mitotypes from A. bipunctata as indicated. Infection status of these specimens is illustrated such that different patterns denote that host specimens were uninfected (U) or harbored male-killing Rickettsia (R), Spiroplasma (S), Wolbachia strain Y (WY) or Wolbachia strain Z (WZ; see legend in the far right corner; see also Table 1 and Table 3 for details).

Analysis of genetic differentiation:
The results of AMOVA clearly showed absence of significant differentiation between mitotypes from the different geographic regions (data sets 2 and 3). The variance observed within data subsets was much higher than the variance between data subsets. This was reflected by a small value for the overall fixation index, which in this case was identical to the F_ST measure of the genetic distance between data subsets. This parameter was not significantly different from zero and, hence, supported genetic homogeneity between geographic regions (Table 6). Note that the value for (equivalent to F_ST) had a negative prefix for data set 2. This was due to the fact that it was computed from the variance components, one of which was similarly negative (that inferred from the distribution of mitotypes among subsets). Usually, negative variance components are produced only if their nominal value is close to zero, which, in the above case, indicated lack of genetic structure between data subsets. Such negative prefixes should not have any effect on the inferred statistical significance of population differentiation (see discussion on this topic on the ARLEQUIN webpage; http://lgb.unige.ch/arlequin/software/2.000/doc/faq/faqlist.htm). The above results contrasted with those obtained from data set 4, in which an association between mitotypes and bacterial infections was tested. Here, variances observed among and within data subsets were almost identical. In addition, the overall fixation index, , was higher than those previously inferred for data sets 2 and 3. Moreover, AMOVA clearly demonstrated that mitotypes cannot be assumed to be randomly distributed among data subsets (data set 4, Table 6). Therefore, homogeneity between these had to be rejected, suggesting that genetic differentiation of mtDNA is related to the presence of different male-killing symbionts. Interestingly, pairwise inferred F_ST values indicated significant genetic differentiation to be the result of differences between the Rickettsia and Spiroplasma, Wolbachia, or Uninfected subsets and also between those of Spiroplasma- and Wolbachia-bearing specimens. In contrast, the diversity of mitotypes from uninfected ladybirds did not appear to differ from that associated with either Wolbachia or Spiroplasma infections (Table 7).

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Our study included ladybird beetles from across almost the whole Eurasian continents. The majority of specimens were derived from a German and a Central Russian population. Despite the enormous geographical distances between collection sites, mtDNA diversity was not found to correlate with geographic origin, suggesting considerable genetic exchange between ladybird beetle populations. Significant mtDNA differentiation was instead associated exclusively with the presence of the different male-killing bacteria. This supports the basic assumption used in this study that the distribution of mitotypes is, in general, linked with the distribution of the different male-killing symbionts. Bacteria thus appear to be primarily maternally inherited without significant levels of horizontal transmission. Consequently, host mtDNA should represent a suitable marker to reconstruct the evolutionary history of male-killing symbionts in relation to each other. On the basis of the results of phylogenetic and population genetic analyses and in consideration of the theoretical work by JOHNSTONE and HURST 1996 , Rickettsia bacteria are inferred to have invaded the host species earlier than the other male-killing symbionts. Mitotypes associated with Rickettsia belonged to the three major lineages of the A. bipunctata clade, including the most basal branches, which clearly lack Spiroplasma- or Wolbachia-infected specimens. Uninfected ladybirds contained mitotypes that were identical or, at least, highly similar to those associated with the Rickettsia, in agreement with the model of JOHNSTONE and HURST 1996 . Furthermore, the Rickettsia data subset produced the highest nucleotide diversity index for data set 4 (data subdivided according to infection status). This is similarly consistent with an early association of Rickettsia male-killers and A. bipunctata. One peculiarity requires special consideration. Mitotype 10J shows considerable differences to the other mtDNA variants. It was found exclusively in specimens with Rickettsia but not in uninfected beetles or those with one of the other symbionts. There are four possible explanations. The finding of such a strict and apparently isolated association could be due to perfect vertical transmission efficiency of Rickettsia such that the associated mitotype cannot "escape" to the uninfected part of the host population. However, Rickettsia transmission rates were consistently found to be smaller than 100% in previous breeding experiments, which included host lineages associated with 10J (Rickettsia from A. bipunctata specimens BBI33, 110, and MOS3; HURST et al. 1993 ; MAJERUS et al. 2000 ). Another explanation is long-term coadaptation between Rickettsia and 10J mitochondria such that uninfected beetles with 10J mitotypes show reduced fitness. In this case, Rickettsia associated with 10J should clearly differ from those associated with other mitotypes. This is not confirmed by a previous analysis of Rickettsia gene sequences, where Rickettsia associated with different mitotypes (including 10J and 9I; host specimens BBI21, 110, MOS3) all bore identical sequences (SCHULENBURG et al. 2001A ). The observed peculiarity could also have resulted from a specific event in a subset of the populations such as a recent population bottleneck. In this case, mitotype 10J is expected to be confined to a small geographic region. However, as it is present in different Western and Eastern populations, this hypothesis also seems unlikely. The apparent absence of mitotype 10J in uninfected ladybirds is thus most likely due to incomplete sampling, in combination with a recent spread of one of the other male-killers that is associated with a different mitotype (see below).

Mitotypes isolated from ladybirds with either Spiroplasma or Wolbachia symbionts were highly similar. They were inferred by ML tree estimation to belong to the same clade and were consistently shown to produce the smallest values for data set 4. This clearly supports the above conclusion that Spiroplasma and Wolbachia male-killers have been present in A. bipunctata for a shorter period of time than the Rickettsia male-killer. In addition, almost all of the uninfected specimens bore mitotypes that were identical or highly similar to those associated with the Spiroplasma or Wolbachia male-killers. This indicates a high degree of genetic exchange between mitotype pools of Spiroplasma- or Wolbachia-bearing ladybird beetles and uninfected specimens. According to theoretical work (JOHNSTONE and HURST 1996 ), high prevalence of a particular class of mitochondrial variant among uninfected hosts should result from the recent spread of a male-killing symbiont through the host population. We here propose that it is most likely the Spiroplasma male-killer that has undergone a recent and extensive spread through host populations.

First, a recent spread of Rickettsia symbionts seems unlikely. If this were the case, then the Rickettsia-associated mitotypes 9I and 10J, which were clearly different from the remaining A. bipunctata mtDNA variants, would be expected to show a higher frequency among uninfected specimens.

Second, extensive screening of Eurasian host populations revealed that Wolbachia symbionts were present exclusively in Central Russia where they were found at comparatively low prevalence (HURST et al. 1993 , HURST et al. 1999A , HURST et al. 1999B ; MAJERUS et al. 2000 ; ZAKHAROV et al. 1996 , ZAKHAROV et al. 2000 ; SCHULENBURG et al. 2000B , SCHULENBURG et al. 2001A ). Such a finding is clearly inconsistent with a recent spread of this symbiont through the host species.

Third, Spiroplasma bacteria did show a high prevalence in, at least, the Eastern host populations (ZAKHAROV et al. 1996 ; HURST et al. 1999B ; MAJERUS et al. 2000 ; our unpublished data). In addition, they were found consistently in breeding experiments to show one of the highest transmission rates of the male-killers from A. bipunctata (ZAKHAROV et al. 1996 ; HURST et al. 1999B ; MAJERUS et al. 2000 ). High transmission efficiency can be assumed to be associated with increased competitiveness (HURST et al. 1997 ; RANDERSON et al. 2000 ).

In conclusion, analysis of host mtDNA indicated that the unusual coexistence of several male-killing symbionts in A. bipunctata is associated with subsequent invasion events and spread of particular symbionts through host populations. Male-killing in A. bipunctata thus appears to be a highly dynamic system. It is therefore expected to prove extremely valuable in future studies on the evolutionary dynamics of male-killing since it may aid in identifying the factors that determine invasion, spread, maintenance, and also loss of male-killing bacteria.

The history of infection in this host species clearly warrants further investigation. In particular, information content of the data should be increased in the future by the inclusion of additional variable mtDNA regions in order to define more precisely the evolutionary relationships of the highly similar mitotypes. Similarly, sample size of data subsets should be increased, particularly in those cases where the number of host lineages was small, e.g., the Wolbachia data subsets. This first requires isolation of additional Wolbachia-bearing host lineages since only six have been identified to date, all of which were included in this study. More detailed insights into symbiont demography in A. bipunctata would also be aided by an extension of the model of JOHNSTONE and HURST 1996 to explore the dynamics of mtDNA evolution in response to the spread of several male-killing symbionts.

	FOOTNOTES

Sequence data from this article have been deposited with the EMBL/GenBank Data Libraries under accession nos. AJ312060, AJ312061, and AJ313061, AJ313062, AJ313063, AJ313064, AJ313065, AJ313066, AJ313067, AJ313068, AJ313069, AJ313070, AJ313071, AJ313072, AJ313073, AJ313074, AJ313075, AJ313076, AJ313077, AJ313078, AJ313079, AJ313080, AJ313081, AJ313082, AJ313083, AJ313084.

	ACKNOWLEDGMENTS

We thank C. Maddren for technical assistance and J. W. O. Ballard, J. M. Koene, and two anonymous reviewers for comments on an earlier version of the manuscript. J. H. G. v. d. Schulenburg was funded by a Training and Mobility of Researchers (TMR) fellowship from the European Union and, for the collection of ladybird beetles, by travel grants from both Magdalene College (Cambridge, England) and the Cambridge Philosophical Society (Cambridge, England). G. D. D. Hurst was supported by a Biotechnology and Biological Sciences Research Council (BBSRC) D. Phillips fellowship and I. A. Zakharov was supported by grants of the Russian Foundation for Basic Research (grant nos. 99-04-48193 and 00-15-97777).

Manuscript received June 13, 2001; Accepted for publication December 18, 2001.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BALLARD, J. W. O., 2000  Comparative genomics of mitochondrial DNA in Drosophila simulans. J. Mol. Evol. 51:64-75[Medline].

BALLARD, J. W. O., J. HATZIDAKIS, T. L. KARR, and M. KREITMAN, 1996  Reduced variation in Drosophila simulans mitochondrial DNA. Genetics 144:1519-1528[Abstract].

BEARD, C. B., D. M. HAMM, and F. H. COLLINS, 1993  The mitochondrial genome of the mosquito Anopheles gambiae: DNA sequence, genome organization, and comparisons with mitochondrial sequences of other insects. Insect Mol. Biol. 2:103-124[Medline].

BEHURA, S. K., S. C. SAHU, M. MOHAN, and S. NAIR, 2001  Wolbachia in the Asian rice gall midge, Orseolia oryzae (Wood-Mason): correlation between host mitotypes and infection status. Insect Mol. Biol. 10:163-171[Medline].

CLARY, D. O. and D. R. WOLSTENHOLME, 1985  The mitochondrial DNA molecule of Drosophila yakuba—nucleotide sequence, gene organization, and genetic code. J. Mol. Evol. 22:252-271[Medline].

COCKERHAM, C. C., 1969  Variance of gene frequencies. Evolution 23:72-84.

COCKERHAM, C. C., 1973  Analyses of gene frequencies. Genetics 74:679-700[Abstract/Free Full Text].

CROZIER, R. H. and Y. C. CROZIER, 1993  The mitochondrial genome of the honeybee Apis mellifera —complete sequence and genome organization. Genetics 133:97-117[Abstract].

EXCOFFIER, L., P. SMOUSE, and J. QUATTRO, 1992  Analysis of molecular variance inferred from metric distances among DNA haplotypes: application to human mitochondrial DNA restriction data. Genetics 131:479-491[Abstract].

FARRIS, J. S., M. KALLERSJO, A. G. KLUGE, and C. BULT, 1995  Testing significance of incongruence. Cladistics 10:315-319.

FELSENSTEIN, J., 1985  Confidence limits on phylogenies—an approach using the bootstrap. Evolution 40:783-791.

FLOOK, P. K., C. H. F. ROWELL, and G. GELLISSEN, 1995  The sequence, organization, and evolution of the Locusta migratoria mitochondrial genome. J. Mol. Evol. 41:928-941[Medline].

GUILLEMAUD, T., N. PASTEUR, and F. ROUSSET, 1997  Contrasting levels of variability between cytoplasmic genomes and incompatibility types in the mosquito Culex pipiens.. Proc. R. Soc. Lond. Ser. B 264:245-251[Medline].

HUELSENBECK, J. P. and K. A. CRANDALL, 1997  Phylogeny estimation and hypothesis testing using maximum likelihood. Annu. Rev. Ecol. Syst. 28:437-466.

HURST, G. D. and F. M. JIGGINS, 2000  Male-killing bacteria in insects: mechanisms, incidence, and implications. Emerg. Infect. Dis. 6:329-336[Medline].

HURST, G. D. D., M. E. N. MAJERUS, and L. E. WALKER, 1993  The importance of cytoplasmic male killing elements in natural populations of the two spot ladybird, Adalia bipunctata (Linnaeus) (Coleoptera: Coccinellidae). Biol. J. Linn. Soc. 49:195-202.

HURST, G. D. D., L. D. HURST and M. E. N. MAJERUS, 1997 Cytoplasmic sex-ratio distorters, pp. 125–154 in Influential Passengers, edited by S. L. O'NEILL, A. A. HOFFMANN and J. H. WERREN. Oxford University Press, Oxford.

HURST, G. D. D., F. M. JIGGINS, J. H. G. V. D. SCHULENBURG, D. BERTRAND, and S. A. WEST et al., 1999a  Male-killing Wolbachia in two species of insect. Proc. R. Soc. Lond. Ser. B 266:735-740.

HURST, G. D. D., J. H. G. V. D. SCHULENBURG, T. M. O. MAJERUS, D. BERTRAND, and I. A. ZAKHAROV et al., 1999b  Invasion of one insect species, Adalia bipunctata, by two different male-killing bacteria. Insect Mol. Biol. 8:133-139[Medline].

JAMES, A. C. and J. W. O. BALLARD, 2000  The expression of cytoplasmic incompatibility and its impact on population frequencies and the distribution of Wolbachia strains in Drosophila simulans. Evolution 54:1661-1672[Medline].

JOHNSTONE, R. A. and G. D. D. HURST, 1996  Maternally inherited male-killing microorganisms may confound interpretation of mitochondrial DNA variability. Biol. J. Linn. Soc. 58:453-470.

LEWIS, D. L., C. L. FARR, and L. S. KAGUNI, 1995  Drosophila melanogaster mitochondrial DNA—completion of the nucleotide sequence and evolutionary comparisons. Insect Mol. Biol. 4:263-278[Medline].

LUNT, D. H., D. X. ZHANG, J. M. SZYMURA, and G. M. HEWITT, 1996  The insect cytochrome oxidase I gene: evolutionary patterns and conserved primers for phylogenetic studies. Insect Mol. Biol. 5:153-165[Medline].

MAJERUS, M. E. N., J. H. G. V. D. SCHULENBURG, and I. A. ZAKHAROV, 2000  Multiple causes of male-killing in a single sample of the two-spot ladybird, Adalia bipunctata (Coleoptera: coccinellidae) from Moscow. Heredity 84:605-609.

MITCHELL, S. E., A. F. COCKBURN, and J. A. SEAWRIGHT, 1993  The mitochondrial genome of Anopheles quadrimaculatus species a—complete nucleotide sequence and gene organization. Genome 36:1058-1073[Medline].

NEI, M., 1987 Molecular Evolutionary Genetics. Columbia University Press, New York.

NIGRO, L. and T. PROUT, 1990  Is there selection on RFLP differences in mitochondrial DNA? Genetics 125:551-555[Abstract].

POSADA, D. and K. A. CRANDALL, 1998  Modeltest: testing the model of DNA substitution. Bioinformatics 14:817-818[Abstract/Free Full Text].

RAND, D. M., 1993  Endotherms, Ectotherms, and mitochondrial genome size variation. J. Mol. Evol. 37:281-295[Medline].

RANDALL, K., M. E. N. MAJERUS, and H. FORGE, 1992  Characteristics for sex determination in British ladybirds (Coleoptera: Coccinellidae). Entomologist 111:109-122.

RANDERSON, J. P., N. G. C. SMITH, and L. D. HURST, 2000  The evolutionary dynamics of male-killers and their hosts. Heredity 84:152-160.

RIGAUD, T., D. BOUCHON, C. SOUTY-GROSSET, and R. RAIMOND, 1999  Mitochondrial DNA polymorphism, sex ratio distorters and population genetics in the isopod Armadillidium vulgare. Genetics 152:1669-1677[Abstract/Free Full Text].

ROKAS, A., R. J. ATKINSON, G. S. BROWN, S. A. WEST, and G. N. STONE, 2001  Understanding patterns of genetic diversity in the oak gallwasp Biorhiza pallida: demographic history of a Wolbachia selective sweep? Heredity 87:294-304[Medline].

SCHNEIDER, S., D. ROESSLI and L. EXCOFFIER, 2000 Arlequin: a software for population genetics data analysis. Version 2.0, Genetics and Biometry Lab, Dept. of Anthropology, University of Geneva, Geneva.

V. D. SCHULENBURG, J. H. G., G. D. D. HURST, T. M. E. HUIGENS, M. M. M. VAN MEER, and F. M. JIGGINS et al., 2000a  Molecular evolution and phylogenetic utility of Wolbachia ftsZ and wsp gene sequences with special reference to the origin of male-killing. Mol. Biol. Evol. 17:584-600[Abstract/Free Full Text].

V. D. SCHULENBURG, J. H. G., T. M. O. MAJERUS, C. M. DORZHU, I. A. ZAKHAROV, and G. D. D. HURST et al., 2000b  Evolution of male-killing Spiroplasma (Procaryotae: Mollicutes) inferred from ribosomal spacer sequences. J. Gen. Appl. Microbiol. 46:95-98.

V. D. SCHULENBURG, J. H. G., M. HABIG, J. J. SLOGGETT, K. M. WEBBERLEY, and D. BERTRAND et al., 2001a  Incidence of male-killing Rickettsia spp. (-proteobacteria) in the ten-spot ladybird beetle Adalia decempunctata L. (Coleoptera: Coccinellidae). Appl. Environ. Microbiol. 67:270-277[Abstract/Free Full Text].

V. D. SCHULENBURG, J. H. G., J. M. HANCOCK, A. PAGNAMENTA, J. J. SLOGGETT, and M. E. N. MAJERUS et al., 2001b  Extreme length and length variation in the first ribosomal internal transcribed spacer of ladybird beetles (Coleoptera: Coccinellidae). Mol. Biol. Evol. 18:648-660[Abstract/Free Full Text].

SHOEMAKER, D. D., K. G. ROSS, L. KELLER, E. L. VARGO, and J. H. WERREN, 2001  Wolbachia infections in native and introduced populations of fire ants (Solenopsis spp.). Insect Mol. Biol. 9:661-673.

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.

SWOFFORD, D. L., 1998 PAUP*—Phylogenetic Analysis Using Parsimony (* and Other Methods). Version 4. Sinauer Associates, Sunderland, MA.

TAMURA, K. and M. NEI, 1993  Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Mol. Biol. Evol. 10:512-526[Abstract].

TURELLI, M., A. A. HOFFMANN, and S. W. MCKECHNIE, 1992  Dynamics of cytoplasmic incompatibility and mtDNA variation in natural Drosophila simulans populations. Genetics 132:713-723[Abstract].

VOGLER, A. P. and R. DESALLE, 1993  Phylogeographic patterns in coastal North American tiger beetles, Cicindela dorsalis, inferred from mitochondrial DNA sequences. Evolution 47:1192-1202.

VOGLER, A. P., R. DESALLE, T. ASSMANN, C. B. KNISLEY, and T. D. SCHULTZ, 1993  Molecular population genetics of the endangered tiger beetle Cicindela dorsalis (Coleoptera: Cicindelidae). Ann. Entomol. Soc. Am. 86:142-152.

WEIR, B. S. and C. C. COCKERHAM, 1984  Estimating F-statistics for the analysis of population structure. Evolution 38:1358-1370.

WERREN, J. H., G. D. D. HURST, W. ZHANG, J. A. J. BREEUWER, and R. STOUTHAMER et al., 1994  Rickettsial relative associated with male killing in the ladybird beetle (Adalia bipunctata). J. Bacteriol. 176:388-394[Abstract/Free Full Text].

YANG, Z., 1996  Among-site variation and its impact on phylogenetic analyses. Trends Ecol. Evol. 11:367-371.

YANG, Z., N. GOLDMAN, and A. FRIDAY, 1994  Comparison of models for nucleotide substitution used in maximum-likelihood phylogenetic estimation. Mol. Biol. Evol. 11:316-324[Abstract].

ZAKHAROV, I. A., I. I. GORYACHEVA, E. V. SHAIKEVICH, J. H. G. V. D. SCHULENBURG, and M. E. N. MAJERUS, 2000  Wolbachia, a new bacteria causing sex ratio bias in the two-spot ladybird Adalia bipunctata L. Russ. J. Genet. 36:385-388.

ZAKHAROV, I. A., G. D. D. HURST, N. E. CHERNYSHEVA, and M. E. N. MAJERUS, 1996  Maternally inherited bacterium causing female bias in the St. Petersburg population of Adalia bipunctata does not belong to the genus Rickettsia. Russ. J. Genet. 32:1303-1306.

ZHANG, D. X. and G. M. HEWITT, 1996  Nuclear integrations—challenges for mitochondrial DNA markers. Trends Ecol. Evol. 11:247-251.

ZHANG, D. X. and G. M. HEWITT, 1997  Insect mitochondrial control region: a review of its structure, evolution and usefulness in evolutionary studies. Biochem. Syst. Ecol. 25:99-120.

This article has been cited by other articles:

				J. Engelstadter and G. D. D. Hurst The Impact of Male-Killing Bacteria on Host Evolutionary Processes Genetics, January 1, 2007; 175(1): 245 - 254. [Abstract] [Full Text] [PDF]

				F. M. Jiggins and M. C. Tinsley An Ancient Mitochondrial Polymorphism in Adalia bipunctata Linked to a Sex-Ratio-Distorting Bacterium Genetics, November 1, 2005; 171(3): 1115 - 1124. [Abstract] [Full Text] [PDF]

				K. A. Dyer and J. Jaenike Evolutionarily Stable Infection by a Male-Killing Endosymbiont in Drosophila innubila: Molecular Evidence From the Host and Parasite Genomes Genetics, November 1, 2004; 168(3): 1443 - 1455. [Abstract] [Full Text] [PDF]

				F. M. Jiggins Male-Killing Wolbachia and Mitochondrial DNA: Selective Sweeps, Hybrid Introgression and Parasite Population Dynamics Genetics, May 1, 2003; 164(1): 5 - 12. [Abstract] [Full Text] [PDF]

				F. M. Jiggins The Rate of Recombination in Wolbachia Bacteria Mol. Biol. Evol., September 1, 2002; 19(9): 1640 - 1643. [Full Text] [PDF]