INRA, UMR 1099 Biologie des Organismes et des Populations appliquée à la Protection des Plantes (BiO3P), BP Le Rheu, F-35653, France 2

Transcription

1 Biological Journal of the Linnean Society, 2009, 97, With 2 figures Complex trait differentiation between host-populations of the pea aphid Acyrthosiphon pisum (Harris): implications for the evolution of ecological specialisationbij_ ADRIEN FRANTZ 1,2,3,4 *, VINCENT CALCAGNO 2, LUCIE MIEUZET 1, MANUEL PLANTEGENEST 5 and JEAN-CHRISTOPHE SIMON 1 1 INRA, UMR 1099 Biologie des Organismes et des Populations appliquée à la Protection des Plantes (BiO3P), BP Le Rheu, F-35653, France 2 Institut des Sciences de l Evolution de Montpellier, UMR Université Montpellier II Centre National de la Recherche Scientifique (UMR 5554) F Montpellier Cedex 05 France 3 UPMC Univ Paris 06, UMR 7625, Ecologie & Evolution, F-75005, Paris, France 4 CNRS, UMR 7625, Ecologie & Evolution, F-75005, Paris, France 5 Agrocampus Ouest, UMR 1099 Biologie des Organismes et des Populations appliquée à la Protection des Plantes (BiO3P), Rennes, F-35042, France Received 29 October 2008; accepted for publication 9 January 2009 Variation in traits affecting preference for, and performance on, new habitats is a key factor in the initiation of ecological specialisation and adaptive speciation. However, habitat and resource use also involves other traits whose influence on ecological and genetic divergence remains poorly understood. In the present study, we investigated the extent of variation of life-history traits among sympatric populations of the pea aphid Acyrthosiphon pisum, which shows several host races that are specialised on various plants of the family Fabaceae plants and is an established model for ecological speciation. First, we assessed the community structure of microbial partners within host populations of the pea aphid. The effect of these microbes on host fitness is uncertain, although there is growing evidence that they may modulate various important adaptive traits of their host such as plant utilisation and resistance against natural enemies. Second, we performed a multivariate analysis on several ecologically relevant features of host populations recorded in the present and previous studies (including microbial composition, colour morph, reproductive mode, and male dispersal phenotype), enabling the identification of correlations between phenotypic traits. We discuss the ecological significance of these associations of traits in relation to the habitat characteristics of pea aphid populations, and their consequences for the evolution of ecological specialisation and sympatric speciation The Linnean Society of London, Biological Journal of the Linnean Society, 2009, 97, ADDITIONAL KEYWORDS: dispersal phenotype facultative symbionts host races reproductive mode sympatric speciation. INTRODUCTION The evolution of reproductive isolation between sympatric populations (i.e. living in the same microgeographic area) has long constituted one of the most puzzling issues in evolutionary biology (Bolnick & *Corresponding author. adrien.frantz@upmc.fr Fitzpatrick, 2007). For decades, sympatric speciation has been regarded as highly implausible, with the opposite scenario of allopatric speciation being considered as the only way to create new species (Coyne, 1992; Rice & Hostert, 1993). Subsequently, however, the possibility of divergence in the absence of geographic barriers has gained strong support owing to studies on organisms that exhibit genetically 718

2 LINKED PHENOTYPIC TRAITS IN THE PEA APHID 719 differentiated sympatric populations living in alternative habitats, such as phytophagous insects ( host races ) (Berlocher & Feder, 2002; Drès & Mallet, 2002). Under the assumption that adaptations on alternative hosts are negatively correlated ( tradeoff ), ecological specialisation can restrict gene flow through the selection of coadapted traits affecting habitat use and of mechanisms maintaining these associations of traits (e.g. mating between genotypes adapted to the same host or assortative mating ; Felsenstein, 1981; Kondrashov & Mina, 1986; Rice, 1987). Among traits influencing habitat and resource use by phytophagous insects, full attention has been given to performance on, and preference for, alternative hosts, whereby natural selection is thought to retain genotypes choosing the host on which they perform best (Drès & Mallet, 2002). Habitat use, however, relies on a broader array of phenotypic traits (including morphology, physiology, behaviour, dispersal, and reproductive mode) that may diverge in response to different selective pressures induced by contrasted environmental conditions between habitats. In addition, those traits may either facilitate or impede gene flow between host populations and adaptive evolutionary diversification. Both the level of divergence of such life-history traits between host races and their possible influence on gene flow and reproductive isolation have rarely been investigated; but see also Diegisser et al. (2004). In the present study, we investigated how phenotypic traits other than performance and preference can be affected by specialisation to alternative habitats and, in turn, how these traits can potentially influence the divergence between natural populations of the pea aphid Acyrthosiphon pisum (Harris) (Hemiptera: Aphididae). This species represents an ideal biological model to address this issue for three main reasons. First, A. pisum exhibits a large amount of phenotypic variability in diverse ecologically important traits (see below). Second, such populations are represented by clonal lineages for the main part of their life cycle, allowing the distinction between the components of phenotypic variance explained by plasticity and by genetic factors. Third, the pea aphid encompasses ecologically and genetically distinct host races living on perennial or annual, and cultivated or uncultivated Fabaceae species (Ferrari et al., 2006; Ferrari, Via & Godfray, 2008). Hence, the pea aphid has been the subject of intensive works to tackle ecological and genetic mechanisms responsible for incipient speciation (Via, 1999; Via & Hawthorne, 2002). In this aphid, phenotypic variability includes the following ecologically relevant traits: (1) several reproductive lineages may coexist in local populations, ranging from cyclically parthenogenetic lineages (also termed sexual lineages for the sake of simplicity) that alternate clonal generations in spring and summer with a sexual one in the autumn, to strictly parthenogenetic (i.e. asexual) lineages (Frantz, Plantegenest & Simon, 2006b); (2) this aphid exhibits a male dispersal polymorphism controlled by a biallelic locus on sexual chromosomes (Caillaud et al., 2002; Braendle, Caillaud & Stern, 2005); hence, pea aphid populations may include homozygous lineages producing only winged or only wingless males, and heterozygous lineages producing both male types in equal proportions (Smith & MacKay, 1989); (3) A. pisum presents a colour polymorphism, with red and green morphs, whose maintenance has been attributed to opposite selective pressures exerted by predation and parasitism (Losey et al., 1997); and (4) beside the primary symbiont Buchnera harboured by virtually all aphid species (Buchner, 1965), A. pisum populations may bear one to several additional bacterial taxa known as secondary, accessory or facultative symbionts. These include three g-proteobacterial lineages (Serratia symbiotica, Hamiltonella defensa, and Regiella insecticola), a Rickettsia-like a- Proteobacteria sp., and a Spiroplasma species (Darby et al., 2005; Moran et al., 2005). Some of these secondary symbionts influence important aphid traits, including plant utilisation (Chen, Montllor & Purcell, 2000; Wilkinson et al., 2001; Douglas et al., 2002; Haynes et al., 2003; Ferrari et al., 2004; Tsuchida, Koga & Fukatsu, 2004), thermal tolerance (Chen et al., 2000; Montllor, Maxmen & Purcell, 2002), and susceptibility to natural enemies (Ferrari et al., 2004; Oliver, Moran & Hunter, 2005), or even compensate for loss of Buchnera in laboratory conditions (Koga, Tsuchida & Fukatsu, 2003). Three host races have been intensively characterised in the pea aphid based on both biological and genetic criteria: one on pea (Pisum sativum L.) and broad bean (Vicia faba L.); a second on alfalfa (Medicago sativa L.); and a third on red clover (Trifolium pratense L.) (Caillaud & Via, 2000; Hawthorne & Via, 2001; Carré & Bournoville, 2003; Simon et al., 2003; Frantz et al., 2006a). Annual (e.g. pea and broad bean) and perennial (e.g. alfalfa and red clover) host plants present contrasted temporal dynamics that may induce different selective pressures and lead to phenotypic divergence between pea aphid host populations. Supporting this hypothesis, pea aphids from broad bean and pea exhibit much stronger levels of investment in sexual reproduction and male dispersal than those from red clover and alfalfa (Frantz et al., 2006b; Frantz, Plantegenest & Simon, 2009). In the light of the above factors, we first assessed the prevalence of all known facultative endosymbionts of pea aphid lineages characterised for their

3 720 A. FRANTZ ET AL. Table 1. Details on lineages used in the sequential analyses for the assessment of genotypic and phenotypic differentiation among host races of the pea aphid Plant origin Analysis Pea Broad bean Red clover Alfalfa Total Reference Symbiont composition Present study Multilocus genotype Frantz et al. (2006a) and present study Reproductive mode Frantz et al. (2006b) Male phenotype Frantz et al. (2009) genotype with microsatellites markers and for several sex-related history traits measured in controlled conditions. In a second step, we analysed these data together with data collected previously on genetic differentiation, body colour, reproductive mode, and male dispersal phenotype, and conducted a multivariate analysis to explore associations between genotypic, symbiotic, and phenotypic variation in A. pisum populations. Strong correlations were found between traits measured in pea aphids from various plant origins. As a result of these findings, we propose that this phenotypic differentiation results from complex interactions between the pea aphid and its variable biotic and abiotic environments, leading to the joint evolution of several biological traits. MATERIAL AND METHODS BIOLOGICAL MATERIAL The presence of the five facultative symbionts reported in A. pisum was investigated among French populations previously used for microsatellite scoring, reproductive mode characterisation, and male phenotype determination. Details of aphid collections and microsatellite scoring are provided elsewhere (Frantz et al., 2006a). One hundred and seventy-nine aphids collected in local fields (located near Rennes, Western France) of four crops (pea, broad bean, alfalfa, and red clover), together with 15 clonal lineages collected in Central France (Carré & Bournoville, 2003), had been previously used for the assessment of genetic differentiation between host populations (Frantz et al., 2006a), whereas the reproductive mode and the dispersal phenotype of the males have been characterised for 137 of these aphid lineages (Frantz et al., 2006b; Frantz et al., 2009). In the present study, we first included 107 additional individual aphids to the genotypic analysis, resulting in a total of 286 individuals genotyped (microsatellite scoring was performed at 14 loci as described in Frantz et al., 2006a). Second, we assessed the presence of facultative symbionts within Table 2. Distribution of phenotypic variation among host populations of pea aphids from Western-Central France Natal plant Frequency of green morph Frequency of sexual lineages* Frequency of api w allele Pea Broad bean Clover Alfalfa *Data from Frantz et al. (2006a). Data from Frantz et al. (2009). 325 individual aphids from the initial sampling. Lastly, we examined associations between genotypic and phenotypic data to detect global patterns of differentiation among pea aphid host races (Table 1). PHENOTYPIC MEASUREMENTS Body colour was recorded for each lineage used for phenotypic analysis. For the reproductive phenotype, we used data from Frantz et al. (2006a) corresponding to 137 genotypes analysed for their laboratory response to sex inducing conditions. Reproductive mode variation falls into three main categories: (1) sexual lineages, which produce both sexual females and males; (2) asexual lineages, which produce parthenogenetic females only; and (3) male-producing lineages, which produce males and parthenogenetic females but no sexual females. The proportion of winged versus wingless morphs was also recorded among the lineages able to produce males ( sexual and male-producing lineages), corresponding to 78 genotypes (Frantz et al., 2009). This allowed a direct inference to be made concerning the genotype at the locus aphicarus for each lineage. The distribution of phenotypic variation among host populations of pea aphids examined in the present study is summarised in Table 2.

4 LINKED PHENOTYPIC TRAITS IN THE PEA APHID 721 MOLECULAR DETECTION OF FACULTATIVE SYMBIONTS Detection of the five facultative symbionts (S. symbiotica, H. defensa, R. insecticola, Rickettsia and Spiroplasma) was based on a polymerase chain reaction (PCR) assay with specific nucleotide primers to amplify their 16S rdna. Two pairs of specific primers were used for each of the five facultative symbionts to insure reliable detection (Tsuchida et al., 2002). To prevent contamination errors, reactions included a negative control where sterile water was used instead of DNA. To prevent false negatives, the presence of obligatory symbiont Buchnera was also investigated in all aphid lineages using specific primers (Tsuchida et al., 2002). Among the 325 tested aphids, 12 lineages that were free of facultative symbionts and for which Buchnera detection failed were considered as ambiguous and eliminated in further analyses. Hence, the presence of facultative symbionts was successfully assessed among 313 aphids. Amplifications were performed in a final volume of 25 ml using 4 pmol of each primer, 50 mm of each dntp, 1.5 ml of 10 Mg 2+ free reaction buffer (500 mm of KCl, 100 mm of Tris-HCl ph 9.0, and 1.0% Triton X-100), 2.5 mm of MgCl 2, 0.5 U Taq DNA polymerase (Promega), and 2 ml of aphid genomic DNA (5 ng). PCRs were performed on a Hybaid thermocycler with the following programme: one denaturation step at 94 C for 5 min, followed by 40 cycles of a denaturation step at 94 C for 1 min, 30 s at 54 C (or 57 C depending on the primer), and an elongation step at 72 C for 1 min, and then a final cycle at 72 C for 10 min. Amplification products were resolved in 1.5% agarose gels in 0.5 TBE buffer ph 8.0 (89 mm of Tris, 89 mm of boric acid, 2 mm of ethylenediaminetetraacetic acid, ph 8.0) at 100 V, and visualised under ultraviolet light after staining with ethidium bromide. STATISTICAL ANALYSIS For each facultative symbiont, we tested the effect of plant origin on the proportion of lineages hosting the symbiont as a measure of its prevalence (generalised linear model; SAS Genmod procedure; assuming a binomial distribution with a logit link). All statistical analyses were executed with software package SAS, version 8.1 (SAS System). Within each host population, preferential associations or exclusions between each pair of symbionts were investigated using Pearson s correlations. The significance level of these correlations was tested with Fisher s two-tailed exact tests after Bonferroni correction for multiple comparisons. To visualise the association between genetic, symbiotic and phenotypic variation among A. pisum populations, a correspondence analysis was carried out with SPAD (2007) using allelic composition for each lineage as contributing factors, whereas plant origin and phenotypic traits (including symbiotic composition, body colour, reproductive mode, and male dispersal phenotype) were used as illustrative factors. RESULTS SYMBIOTIC AND GENOTYPIC COMPOSITION OF PEA APHID SAMPLES Among the 313 pea aphids analysed for symbiotic composition, 72 (23.0%) harboured only the primary symbiont Buchnera, 160 (51.1%) harboured one of the five facultative symbionts, 80 (25.6%) harboured two facultative symbionts, and a single individual collected on broad bean showed a triple infection with S. symbiotica, Rickettsia and H. defensa (Table 3). The prevalence of each facultative symbiont strongly differed between host populations (Fig. 1, Table 4). Serratia symbiotica and Rickettsia occurred more commonly among A. pisum individuals from pea and broad bean than among those from red clover and alfalfa. Aphids from red clover harboured R. insecticola in a greater proportion than those from other host plants, whereas those from alfalfa harboured H. defensa in a greater proportion than aphids from other hosts. Lastly, Spiroplasma was in higher proportions among aphids from both red clover and alfalfa. Aphids from pea and broad bean did not differ from each other whatever the symbiont considered. Seven out of the ten possible combinations of symbiont pairs were found in the sample, among which three represented 91% of dual infections. Specific associations between pair of facultative symbionts were detected for each host population with Serratia + Rickettsia being preferentially associated in aphids on pea (33.33%) and broad bean (61.90%), Regiella + Spiroplasma in the clover race (90.48%), and Hamiltonella + Spiroplasma in the alfalfa race (92.86%). Among the 286 lineages for which microsatellite profiles were obtained, 177 multilocus genotypes were detected, including 141 unique genotypes and 36 genotypes found more than once. The presence of copies of the same genotype is expected in clonal organisms such as aphids, and this phenomenon is exacerbated in populations where sexual reproduction has been partially or totally abandoned, as in A. pisum from Western France (Frantz et al., 2006a; Frantz et al. 2006b). Estimation of symbiotic prevalence carried out with or without repeated genotypes led to very similar proportions (data not shown), indicating no influence of potential over-represented genotypes due to clonal reproduction. Neither positive nor negative significant association between any pair of the five facultative symbionts was observed within host races among 25 combina-

5 722 A. FRANTZ ET AL. Table 3. Symbiotic status of French lineages of the pea aphid Infection status Symbionts Total No infection Buchnera only 72 Single infection Serratia symbiotica 54 N = 160 Rickettsia 6 Regiella insecticola 24 Spiroplasma 23 Hamiltonella defensa 53 Dual infection Serratia symbiotica + Rickettsia 21 N = 80 Serratia symbiotica + Hamiltonella defensa 1 Rickettsia + Spiroplasma 1 Rickettsia + Hamiltonella defensa 1 Regiella insecticola + Spiroplasma 24 Regiella insecticola + Hamiltonella defensa 4 Spiroplasma + Hamiltonella defensa 28 Triple infection Serratia symbiotica + Rickettsia + Hamiltonella defensa 1 Total 313 Figure 1. Prevalence of each facultative symbiont according to host plant determined over the whole sample (313 aphid lineages). Bars with the same letter indicate that the prevalence of a given symbiont is not statistically different between host populations of Acyrthosiphon pisum (P > 0.05). tions tested with Fisher s exact tests, indicating that symbionts were distributed independently from each other (P > 0.002) within each host race. RELATIONSHIPS BETWEEN GENOTYPIC AND PHENOTYPIC VARIABLES The correspondence analysis whose first two axes accounted for 29.35% of the total variance, allowed visualisation of the global pattern of genetic and phenotypic differentiation among A. pisum populations (Fig. 2). Axis 1 discriminated between host populations from red clover and alfalfa host compared to those from pea and broad bean, with the two latter being only slightly distinguishable from one another. Axis 2 tended to discriminate between red clover and alfalfa host populations. Some of the phenotypic traits appeared strongly associated with a given host population. Notably, lineages from pea exhibited a strong investment in

6 LINKED PHENOTYPIC TRAITS IN THE PEA APHID 723 sexual reproduction, produced only winged males, and were predominantly infected by S. symbiotica and Rickettsia. Lineages from broad bean only differed to those from pea by a lower investment in sexual form production. The red clover race was strongly associated with the red morph and R. insecticola infection and was typically composed of asexual Table 4. Results of the generalised linear model fitted for the analyses of the effect of host plant on the proportions of pea aphid lineages harbouring each endosymbiont (N = 313 lineages) Symbiont d.f. Chi squared P Serratia symbiotica < *** Rickettsia < *** Regiella insecticola < *** Spiroplasma < *** Hamiltonella defensa < *** Buchnera only NS ***P < d.f., degrees of freedom; NS, nonsignificant. or male-producing lineages with wingless males. Lastly, lineages from alfalfa resembled those from red clover in their reproductive mode, but specifically harboured H. defensa and Spiroplasma. DISCUSSION All five main facultative symbionts documented in the pea aphid world-wide were detected among sympatric French populations of the aphid collected on four host plants. Previous studies on the same populations also reported a high variability, both for the level of investment in the sexual phase of the aphid s life cycle (Frantz et al., 2006b) and for the dispersal phenotype of males (Frantz et al., 2009). This large phenotypic variability is clearly not distributed at random among populations, but rather is largely shaped by specific interactions with host plants. Previous studies have shown that French populations of A. pisum constitute three genetic clusters respectively specialising on: (1) on alfalfa; (2) on red clover; and (3) on pea and broad bean, with the latter being much more divergent from the two other ones (Via, 1999; Carré & Bournoville, 2003; Simon Figure 2. Correspondence analysis on allelic composition at each microsatellite locus for the 78 multilocus genotypes of Acyrthosiphon pisum found on pea (square), broad bean (triangle), alfalfa (black circle), and clover (grey diamond). Plant origin and phenotypic data (including colour morph, reproductive mode, wing male phenotype, and symbiotic status) of each lineage are plotted as illustrative variables.

7 724 A. FRANTZ ET AL. et al., 2003; Frantz et al., 2006a). In the present study, host-based genetic differentiation is strongly associated with phenotypic divergence (i.e. phenotypic differentiation mirrors genotypic partitioning). Interestingly, the two levels of population differentiation were also observed at the phenotypic level. Indeed, A. pisum genotypes from pea and broad bean showed greater phenotypic divergence with those from red clover and alfalfa, with these two nevertheless being distinguished at several phenotypic traits. Multivariate analysis revealed strong interrelations between genotypic profiles, facultative symbionts, colour morph, host plant use, reproductive mode, and male phenotype. A general pattern of phenotypic differentiation among pea aphids from Western-Central France emerged from this analysis as follows: A. pisum populations from pea and broad bean showed virtually no red morphs, had high investments in the sexual phase of their annual life cycle, produced almost exclusively winged males, and specifically harboured S. symbiotica and Rickettsia. Pea aphids from red clover, in contrast, displayed red morphs in high proportions, reproduced mainly asexually, produced wingless males, and specifically bore R. insecticola. Furthermore, pea aphids from alfalfa were characterised by high levels of asexual reproduction, produced wingless males, and were predominantly infected by H. defensa and Spiroplasma. These associations between traits within host populations of the pea aphid appear stable through time because they have been observed in the same populations sampled 5 years previously, at least for the relationships between four of the five facultative symbionts, body colour, and plant origin (Simon et al., 2003). The existence of such interrelations between various aphid population characteristics raises questions about their underlying causes and consequences. This pattern could first result from phenotypic traits associated by chance in the genotype(s) that initially switched host (i.e. through a founder effect). However, this is unlikely because it does not account either for the phenotypic variability observed within each host race, nor for the spread of this trait variation on the genetic tree relating French A. pisum genotypes (Frantz et al., 2006a; data not shown). Alternatively, this pattern may result from one given phenotypic trait, associated with a given host race either in response to selective pressures or by chance due to historical factors, having a direct influence on other traits (e.g. through pleiotropic effect). However, the direct effect of facultative symbionts on wing male morph, colour morph or on reproductive mode is unlikely because it has been shown that all these traits are under the control of the aphid genome (Müller, 1961, 1962; Smith & MacKay, 1989; Caillaud et al., 2002; Braendle et al., 2005; Simon et al., 2007). These associations between phenotypic traits and genotypic clusters are likely to result from selection exerted by: (1) the local environmental conditions of the habitat and (2) the resource characteristics of the host plant. Differences in environmental conditions between pea aphid habitats mainly depend on the contrasted spatio-temporal dynamics between annual crops (represented in the present study by pea and broad bean) and perennial crops (represented in the present study by red clover and alfalfa). This variation in environmental heterogeneity was proposed to explain the associations linking host races with both reproductive mode variation (Frantz et al., 2006b) and, independently, male phenotype (Frantz et al., 2009). In particular, higher temporal habitat variability could exert a selective pressure for the maintenance of sex in aphid populations living on annual crops. This would allow them to face variable biotic and abiotic constraints on different hosts and would explain why A. pisum on pea and broad bean showed greater investment in sex compared to those feeding on more temporally stable habitats (i.e. A. pisum lineages on clover and alfalfa). Differences in spatio-temporal dynamics between annual and perennial crops could also influence the selection of male dispersal phenotype (Frantz et al., 2009). The ephemeral nature of annual crops indeed represents a strong selective pressure that could favour lineages producing winged males in high frequencies (as observed for A. pisum lineages associated with pea and broad bean) as a means to prevent local extinction and to avoid the negative genetic consequences of local relatedness, such as local mate competition and inbreeding (Frantz et al., 2009). In the present study, pea aphid populations harboured facultative symbionts at high frequency (77%), confirming that a small fraction of French A. pisum had Buchnera only (Simon et al., 2003). In addition, pea aphids living on different plants strongly differed in their symbiotic composition, suggesting that either these bacteria infected some lineages of A. pisum before ecological specialisation, or that they may increase host fitness on specific plants. Supporting the second idea, R. insecticola appears to enhance the survival of pea aphids when feeding on clover (Tsuchida et al., 2004). Moreover, the specific association between R. insecticola and pea aphid populations from clover found in the present study was repeatedly observed world-wide (Japan: Tsuchida et al., 2002; California: Leonardo & Muiru, 2003; France: Simon et al., 2003; UK: Ferrari et al., 2004). Although the effects of the other accessory bacteria remain unclear, they may play a direct or indirect role in the utilisation of the aphid host in which the prevalence is particularly high.

8 LINKED PHENOTYPIC TRAITS IN THE PEA APHID 725 Lastly, host-specialised pea aphids exhibited different colour morphs. Aphids on pea and broad bean were found to be almost exclusively green, whereas those from red clover and alfalfa showed a mixture of red and green morphs. The balance between predation and parasitism possibly allows the maintenance of the red/green polymorphism in the pea aphid, with the red morph being preferentially consumed by predators (e.g. ladybirds) and the green morph being preferentially attacked by parasitic wasps such as the parasitoid, Aphidius ervi Haliday (Hymenoptera: Braconidae: Aphidiidae) (Losey et al., 1997; Libbrecht, Gwynn & Fellowes, 2007). Predation or parasitic pressures may also differ between habitats on which pea aphid populations are specialised. If true, this may account for the high frequency of H. defensa, a microbial mutualist conferring protection against parasitoids, in the alfalfa race of A. pisum (Oliver et al., 2005). More field studies are needed to link the biotic and abiotic characterisitics of pea aphid habitats to their phenotypic attributes. These associations between life-history traits and host-based differentiation complement our understanding of the role of phenotypic divergence in the initial phase(s) of ecological specialisation and gene flow restriction between incipient races. Phenotypic traits such as reproductive mode, dispersal, and microbial partners are of major importance for shaping the evolution of ecological specialisation in many phytophagous insects, but also in other organisms in which host ecology is influenced by mutualistic interactions (Leonardo & Mondor, 2006; Janson et al., 2008). These traits can affect gene flow and migration between host populations, and specific associations between these traits may strongly reduce or facilitate ecological divergence and genetic differentiation. For example, sexual reproduction and recombination, as well as the production of winged males at high frequencies, promote population mixing, which is consistent with the very low genetic differentiation observed between pea aphids living on annual crops (Frantz et al., 2006a). Conversely, asexual reproduction and philopatric males greatly diminish population mixing, which is also consistent with the high level of population differentiation between aphids living on perennial crops such as red clover and alfalfa. The present study thus sheds light on the importance and potential feedback effects of phenotypic traits submitted to disruptive selection in the process of ecological divergence and reveals ecological interactions to be more complicated and fascinating than hitherto imagined. It also highlights the need to take these characteristics into account in the growing research field of adaptive diversification. ACKNOWLEDGEMENTS We thank Paul Sunnucks, Hugh Loxdale, and an anonymous reviewer for their constructive comments on the manuscript. We also thank Marie-Thérèse Querrien and Bernard Chaubet for helping with the biological experiments; Lucie Mieuzet for helping in the microsatellite genotyping; and Jean-Pierre Gauthier for helping with correspondence analyses. This work was supported by a grant from Région Bretagne (Opération A1C701 Programme 691) and by program ECOGER. REFERENCES Berlocher SH, Feder JL Sympatric speciation in phytophagous insects: moving beyond controversy? Annual Review of Entomology 47: Bolnick DI, Fitzpatrick BM Sympatric speciation: models and empirical evidence. Annual Review of Ecology, Evolution and Systematics 38: Braendle C, Caillaud MC, Stern DL Genetic mapping of aphicarus a sex-linked locus controlling a wing polymorphism in the pea aphid (Acyrthosiphon pisum). Heredity 94: Buchner P Endosymbiosis of animals with plant microorganisms. New York, NY: Interscience Publishers. Caillaud MC, Boutin M, Braendle C, Simon JC A sex-linked locus controls wing polymorphism in males of the pea aphid, Acyrthosiphon pisum (Harris). Heredity 89: Caillaud MC, Via S Specialized feeding behavior influences both ecological specialization and assortative mating in sympatric host races of pea aphids. American Naturalist 156: Carré S, Bournoville R Specialization of spring sympatric populations of Acyrthosiphon pisum (Hemiptera: Aphididae) according to Fabaceae. Annales de la Société Entomologique de France 39: Chen DQ, Montllor B, Purcell AH Fitness effects of two facultative endosymbiotic bacteria on the pea aphid, Acyrthosiphon pisum, and the blue alfalfa aphid, A. kondoi. Entomologia Experimentalis et Applicata 95: Coyne JA Genetics and speciation. Nature 355: Darby AC, Chandler SM, Welburn SC, Douglas AE Aphid-symbiotic bacteria cultured in insect cells lines. Applied Environmental Microbiology 71: Diegisser T, Johannesen J, Lehr C, Seitz A Genetic and morphological differentiation in Tephritis bardanae (Diptera: Tephritidae): evidence for host-race formation. Journal of Evolutionary Biology 17: Douglas AE, Darby AC, Birkle LM, Walters KFA The ecological significance of symbiotic micro-organisms in animals perspectives from the microbiota of aphids. In: Hails RM, Beringer J, Godfray HCJ, eds. Genes in the environment. Oxford: Blackwell Scientific,

9 726 A. FRANTZ ET AL. Drès M, Mallet J Host races in plant-feeding insects and their importance in sympatric speciation. Philosophical Transactions of the Royal Society of London Series B, Biological Sciences 357: Felsenstein J Skepticism toward Santa Rosalia, or why are there so few kinds of animals? Evolution 35: Ferrari I, Darby AC, Daniell TJ, Godfray HCJ, Douglas AE Linking the bacterial community in pea aphids with host-plant use and natural enemy resistance. Ecological Entomology 29: Ferrari J, Godfray HCJ, Faulconbridge AS, Prior K, Via S Population differentiation and genetic variation in host choice among pea aphids from eight host plant genera. Evolution 60: Ferrari J, Via S, Godfray HCJ Population differentiation and genetic variation in performance on eight hosts in the pea aphid complex. Evolution 62: Frantz A, Plantegenest M, Mieuzet L, Simon JC. 2006a. Ecological specialization correlates with genotypic differentiation in sympatric host-populations of the pea aphid. Journal of Evolutionary Biology 19: Frantz A, Plantegenest M, Simon JC. 2006b. Temporal habitat variability and the maintenance of sex in host populations of the pea aphid. Proceedings of the Royal Society of London Series B, Biological Sciences 273: Frantz A, Plantegenest M, Simon JC Host races of the pea aphid Acyrthosiphon pisum differ in male wing phenotypes. Bulletin of Entomological Research, in press. doi: /S Hawthorne DJ, Via S Genetic linkage of ecological specialization and reproductive isolation in pea aphids. Nature 412: Haynes S, Darby AC, Daniell TJ, Webster G, van Veen FJF, Godfray HCJ, Prosser JI, Douglas AE Diversity of bacteria associated with natural aphid populations. Applied and Environmental Microbiology 69: Janson E, Stireman JO III, Singer MS, Abbot P Phytophagous insect-microbe mutualisms and adaptive evolutionary diversification. Evolution 62: Koga R, Tsuchida T, Fukatsu T Changing partners in an obligate symbiosis: a facultative endosymbiont can compensate for loss of the essential endosymbiont Buchnera in an aphid. Proceedings of the Royal Society of London Series B, Biological Sciences 270: Kondrashov AS, Mina MV Sympatric speciation, when is it possible? Biological Journal of the Linnean Society 27: Leonardo T, Muiru GT Facultative symbionts are associated with host plant specialization in pea aphid populations. Proceedings of the Royal Society of London Series B, Biological Sciences 270: S209 S212. Leonardo TE, Mondor EB Symbiont modifies host life-history traits that affect gene flow. Proceedings of the Royal Society of London Series B, Biological Sciences 273: Libbrecht R, Gwynn DM, Fellowes MDE Aphidius ervi preferentially attacks the green morph of the pea aphid, Acyrthosiphon pisum. Journal of Insect Behaviour 20: Losey JE, Ives AR, Harmon J, Ballantyne F, Brown C A polymorphism maintained by opposite patterns of parasitism and predation. Nature 388: Montllor CB, Maxmen A, Purcell AH Facultative baterial endosymbionts benefit pea aphids Acyrthosiphon pisum under heat stress. Ecological Entomology 27: Moran NA, Russell JA, Koga R, Fukatsu T Evolutionary relationships of three new species of Enterobacteriaceae living as symbionts of aphids and other insects. Applied and Environmental Microbiology 71: Müller FP Stabilität und Veränderlichkeit der Färbung bei Blattläusen. Archiv der Freunde der Naturgeshichte in Mecklenburg 7: Müller FP Biotypen un Unterarten der Erbsenlaus, Acyrthosiphon pisum Harris. Zeitschrift für Pflanzenkrankheiten 69: Oliver KM, Moran NA, Hunter MS Variation in resistance to parasitism in aphids is due to symbionts not host genotype. Proceedings of the National Academy of Science of the United States of America 36: Rice WR Selection via habitat specialization, the evolution of reproductive isolation as a correlated character. Evolutionary Ecology 1: Rice WR, Hostert EE Laboratory experiments on speciation what have we learned in 40 years. Evolution 47: Simon JC, Carré S, Boutin M, Prunier-Leterme N, Sabater-Muñoz B, Latorre A, Bournoville R Host-based divergence in populations of the pea aphid: insights from nuclear markers and the prevalence of facultative symbionts. Proceedings of the Royal Society of London Series B, Biological Sciences 270: Simon JC, Sakurai M, Bonhomme J, Tsuchida T, Koga R, Fukatsu T Elimination of a specialised facultative symbiont does not affect the reproductive mode of its aphid host. Ecological Entomology 32: Smith MAH, MacKay P Genetic variation in male alary polymorphism of pea aphid, Acyrthosiphon pisum. Entomologia Experimentalis et Applicata 51: SPAD SPAD-N: SystÈme Portable Pour L Analyse Des DonnÉes NumÉriques; CISIA, Saint-MandÉ (France). SPAD test and go 2007, Version Paris: SPAD. Tsuchida T, Koga R, Fukatsu T Host plant specialization governed by facultative symbiont. Science 303: Tsuchida T, Koga R, Shibao H, Matsumoto T, Fukatsu T Diversity and geographic distribution of secondary endosymbiotic bacteria in natural populations of the pea aphid, Acyrthosiphon pisum. Molecular Ecology 11: Via S Reproductive isolation between sympatric races of pea aphids. I. Gene flow restriction and habitat choice. Evolution 53:

10 LINKED PHENOTYPIC TRAITS IN THE PEA APHID 727 Via S, Hawthorne DJ The genetic architecture of ecological specialization: correlated gene effects on host use and habitat choice in pea aphids. The American Naturalist 159: S76 S88. Wilkinson TL, Adams D, Minto LB, Douglas AE The impact of host plant on the abundance and function of symbiotic bacteria in an aphid. Journal of Experimental Biology 204: