Summary
Virtually all aphids maintain an obligate mutualistic symbiosis with bacteria from the Buchnera genus, which produce essential nutrients for their aphid hosts. Most aphids from the Lachninae subfamily have been consistently found to house additional endosymbionts, mainly Serratia symbiotica. This apparent dependence on secondary endosymbionts was proposed to have been triggered by the loss of the riboflavin biosynthetic capability by Buchnera in the Lachninae last common ancestor. However, an integral large-scale analysis of secondary endosymbionts in the Lachninae is still missing, hampering the interpretation of the evolutionary and genomic analyses of these endosymbionts. Here, we analysed the endosymbionts of selected representatives from seven different Lachninae genera and nineteen species, spanning four tribes, both by FISH (exploring the symbionts’ morphology and tissue tropism) and 16S rRNA gene sequencing. We demonstrate that all analysed aphids possess dual symbiotic systems, and while most harbour S. symbiotica, some have undergone symbiont replacement by other phylogenetically-distinct bacterial taxa. We found that these secondary associates display contrasting cell shapes and tissue tropism, and some appear to be lineage-specific. a scenario for symbiont establishment in the Lachninae, followed by changes in the symbiont’s tissue tropism and symbiont replacement events, thereby highlighting the extraordinary versatility of host-symbiont interactions.
Originality-Significance Statement A key question in evolutionary biology is that of how mutualism evolves. One way to approach this problem is to investigate recently-established mutualistic associations, particularly by comparing various symbiotic systems in closely related hosts. Here, we present a most comprehensive study to investigate co-obligate symbioses in aphids, focusing in the Lachninae subfamily. While most aphids keep an obligate vertically-transmitted association with intracellular Buchnera bacteria, some, such as members of the Lachninae subfamily, host an additional putative co-obligate symbiont. Thus, the Lachninae dual symbiotic systems offer a unique opportunity to understand the evolutionary dynamics of host-symbiont associations, in particularly how secondary symbionts become obligate and eventually may be replaced. Through genome sequencing of three aphid species belonging to distantly related tribes within the subfamily, we have previously corroborated that they have indeed established co-obligate mutualistic associations with the S. symbiotica secondary endosymbiotic bacterium. This was putatively facilitated by an ancient pseudogenisation of the riboflavin biosynthetic pathway in Buchnera, rendering it unable to provide the essential vitamin to the host. However, not all Lachninae members harbour S. symbiotica, some species being associated to at least four different bacterial taxa. To correctly interpret the genomic data and to understand the evolutionary dynamics of these symbiotic associations, a wide-range analysis of both the phylogenetic relations as well as of the secondary symbionts’ localisation within the bacteriome is needed. To tackle this, we have combined phylogenetic analyses of the symbionts’ 16S rRNA gene sequences and FISH microscopy, to understand the symbiont’s identity as well as the morphological characteristics and tissue tropism. The phylogenetic affinities and patterns of co-divergence of the symbionts, in combination with previously published genomic data, have enabled us to build an evolutionary scenario for the establishment, changes in tissue tropism such as “stable” internalisation into bacteriocytes, and replacements of the putative “ancient” secondary endosymbiont from the Lachninae last common ancestor. Also, we were able to determine through phylogenetic analyses that some putative co-obligate endosymbionts may have evolved from once facultative ones. The evolutionary framework presented here reveals a dynamic pattern for the more recent evolutionary history of these symbioses, including replacement and novel acquisition of phylogenetically different co-obligate symbionts. This study opens new research avenues on this symbiont-diverse subfamily, providing insight into how mutualism in endosymbiotic associations can evolve, and the role these bacteria have played in the species’ adaptation and even in the speciation process.
Introduction
The increasing recognition that symbionts play an important role in the ecology and evolution of their hosts, as well as the rapid changes in the type and nature of these symbiotic associations, call for an evolutionary framework to understand these dynamics. Symbiotic relationships between aphids and their primary obligate bacterial endosymbiont Buchnera aphidicola, represent one of the most well studied cases of bacterial endosymbiosis within animals. The Buchnera symbiosis is found across all modern aphids (Aphididae Latreille, 1802 family) (Buchner, 1953), with the notable exception of members belonging to the monophyletic Cerataphidini tribe Baker, 1920, in which Buchnera has been replaced by an extracellular yeast-like symbiont (Buchner, 1953; Fukatsu and Ishikawa, 1992; Fukatsu et al., 1994). Buchnera cells have a round and pleomorphic shape (Michalik et al., 2014), and inhabit the cytoplasm of bacteriocytes (specialised cells evolved to house the endosymbiont), which make up a distinct organ-like structure called the bacteriome (Buchner, 1953; Fukatsu et al., 1998). The onset of the aphid-Buchnera symbiosis dates back to at least 80-150 million years ago (hereafter Mya) (von Dohlen and Moran, 2000). Buchnera, as other “ancient” obligate endosymbionts, underwent a rapid genome erosion early in its evolutionary history with aphids, resulting in a high degree of synteny among distantly related Buchnera (Tamas et al., 2002; van Ham et al., 2003), and since then, lineages of both partners have been co-diverging. This has been evidenced through phylogenetic reconstructions using Buchnera DNA or amino acid sequences, which parallel their aphid hosts’ evolutionary relationships (Munson et al., 1991; Jousselin et al., 2009; Liu et al., 2013). Besides Buchnera, aphids can also harbour secondary endosymbionts (in addition to the primary symbiont), these being of facultative or obligate nature in some lineages. Contrary to obligate symbionts, facultative ones are not required for the correct development, reproduction and survival of their host. Still, they can provide a benefit under certain environmental or ecological conditions (conditional mutualism) (reviewed in Oliver et al., 2010, 2014). To date, various secondary facultative bacterial endosymbionts have been identified, primarily in the pea aphid Acyrthosiphon pisum (Aphidinae subfamily) (Fukatsu et al., 2000, 2001; Sakurai et al., 2005; Degnan, Leonardo, et al., 2009; Degnan, Yu, et al., 2009; Guay et al., 2009; Tsuchida et al., 2014). These secondary symbionts have a very different tissue tropism than Buchnera, as they can be present in separate bacteriocytes (called secondary bacteriocytes), co-infecting the primary endosymbiont’s bacteriocytes, located in sheath cells (at the periphery of the bacteriome and found closely associated to bacteriocytes), and/or free in the haemocoel (Fukatsu et al., 2000; Sandström et al., 2001; Moran et al., 2005; Sakurai et al., 2005; Michalik et al., 2014). While many secondary symbionts are facultative for their hosts, some seem to have established co-obligate associations with their respective symbiotic partners. In this respect, the subfamily Lachninae Herrich-Schaeffer, 1854 of aphids is peculiar, in that all members analysed thus far by microscopy techniques have been found to house secondary endosymbionts (Buchner, 1953; Fukatsu and Ishikawa, 1998; Fukatsu et al., 1998; Lamelas et al., 2008; Pyka-Fościak and Szklarzewicz, 2008; Michalik et al., 2014).
The Lachninae can be divided into 5 monophyletic tribes: i) Lachnini Herrich-Schaeffer, 1854, ii) Stomaphidini Mordvilko, 1914, iii) Tramini Herrich-Schaeffer, 1854, iv) Tuberolachnini Mordvilko, 1942, and v) Eulachnini Baker, 1920 (Chen, Favret, et al., 2015) (Fig 1). The first four tribes comprise 112 known species organised into 13 genera, however, due to the lack of molecular data, the phylogenetic affiliation of three of these (Neonippolachnus, Sinolachnus, and Eotrama) remains unresolved, and thus are not included in the Figure. While the Lachnini, Tramini, and Tuberolachnini feed on angiosperms, the Stomaphidini feed on both angiosperm and gymnosperm trees (bark-trunk and root). The latter have evolved some of the longest mouthparts among aphids, particularly the trunk-feeding species (Blackman and Eastop, 1994). The Tramini are unique in that they solely feed on the roots of herbaceous plants, mostly composites (Blackman and Eastop, 2006). Finally, the Eulachnini, which exclusively feed on conifers, are classified into 4 genera: Essigella, Eulachnus, Pseudessigella (with no molecular data available), and Cinara. The latter is the largest genus within the Lachninae and has been traditionally taxonomically classified into three subgenera (Cinara, Cupressobium, and Cedrobium). However, recent extensive molecular work on members of the subgenus Cinara has found the subgenus Cinara (Cinara) polyphyletic, and thus has divided the genus into three major phylogenetic clades, termed simply A, B, and C (Meseguer et al., 2015).
Known secondary symbionts of Lachninae differ in tissue tropism and cell shape (Buchner, 1953; Fukatsu and Ishikawa, 1998; Fukatsu et al., 1998; Pyka-Fościak and Szklarzewicz, 2008; Michalik et al., 2014), as well as in phylogenetic origin (Russell et al., 2003; Lamelas et al., 2008; Burke et al., 2009; Jousselin et al., 2016). Although different bacterial taxa have been found associated to Lachninae aphids, many species of this subfamily have been systematically found associated with members of the bacterial genus Serratia, mainly Serratia symbiotica (Fukatsu et al., 1998; Russell et al. 2003; Lamelas et al., 2008; Burke et al., 2009; Chen et al. 2015; Jousselin et al., 2016) (for a summary see Table S1). Particularly, most Cinara species have been consistently found to house S. symbiotica strains, which form two phylogenetically distinct clusters (based on 16S rRNA gene sequences), one “facultative-like” and one “obligate-like” (hereafter FL and OL, respectively) (Lamelas et al., 2008; Burke et al., 2009). Whole genome sequencing and metabolic reconstruction of the Buchnera-S. symbiotica bacterial consortia of three Lachninae species revealed that S. symbiotica strains, belonging to FL and OL, had indeed established co-obligate associations along with Buchnera in these hosts (Gosalbes et al., 2008; Lamelas, Gosalbes, Manzano-Marín, et al., 2011; Manzano-Marín and Latorre, 2014; Manzano-Marín et al., 2016). Through comparative genomics of these di-(Buchnera-S. symbiotica) vs. mono-endosymbiotic (Buchnera-only) systems, it was postulated that the establishment of S. symbiotica as a co-obligate endosymbiont in the Lachninae last common ancestor (hereafter LLCA) was facilitated by a putative ancient pseudogenisation of the riboflavin biosynthetic pathway in Buchnera and the complementation of this loss-of-function by S. symbiotica (Manzano-Marín and Latorre, 2014; Manzano-Marín et al., 2016). Substantial differences regarding cell morphology and genomic characteristics observed among extant S. symbiotica strains suggest that these represent different stages of the genome reduction process towards a highly reduced obligate intracellular symbiont.
Given the overwhelming evidence pointing towards a dependency on secondary endosymbionts within the Lachninae, we sought to further understand the evolutionary succession of establishments, replacements, and changes in tissue tropism (e.g. “stable” internalisations into distinct bacteriocytes), of secondary endosymbionts among species of this subfamily. For this purpose, we have identified the secondary endosymbionts of distantly related Lachninae aphids belonging to 19 different species (8 different genera collected in three different countries) (Table S1) through 16S rRNA gene sequencing and phylogenetic analysis. In selected species, we have determined the location of the secondary endosymbionts using fluorescence in situ hybridisation (FISH) with 16S rRNA targeted specific oligonucleotide probes. We propose an evolutionary scenario for the establishment of an original secondary co-obligate endosymbiont in the LLCA, followed by symbiont replacements, “stable” internalisations of these into distinct bacteriocytes, and/or the putative establishment of tertiary obligate symbionts in different aphid lineages from this symbiont-diverse subfamily.
Results
S. symbiotica and S. marcescens-like secondary symbionts
Most Cinara spp. investigated so far have been found to be associated with different S. symbiotica strains (Lamelas et al., 2008; Burke et al., 2009; Jousselin et al., 2016). Here, we have collected 19 representatives (comprising 11 species) of Cinara clades A (n=5), B (n=7), and C (n=7), and have identified their endosymbionts through PCR, cloning, and sequencing of their 16S rRNA genes. We found that the secondary symbionts of all of the collected species – except for Cinara (Cinara) confinis (Cinara clade C) and Cinara (Schizolachnus) obscurus – were indeed affiliated with S. symbiotica (Table S1).
To test for co-speciation previously observed for Buchnera-Serratia symbiont pairs within the Lachninae in the light of our new data, we performed a Bayesian phylogenetic reconstruction using currently available 16S rRNA gene sequences of both Buchnera (Fig 2A) and Serratia (Fig 2B) from Lachninae aphids. Contrary to earlier studies (Lamelas et al., 2008; Burke et al., 2009), we failed to recover the previously described FL and OL S. symbiotica clusters. We found that all Lachninae S. symbiotica strains, but the one from Trama caudata, form a well-supported and unresolved monophyletic clade nested within a group composed mainly of facultative strains of S. symbiotica from Aphidinae aphids, a strain from Adelges tsugae (Hemiptera: Adelgidae), and one from the Lachninae aphid Trama troglodytes (Fig 2B). This unresolved clade contains the “early” co-obligate S. symbiotica strain from Cinara (Cupressobium) tujafilina (Manzano-Marín and Latorre, 2014) and a strain from the closely related Cinara (Cupressobium) cupressi. Within the Lachninae S. symbiotica clade, we recovered three well-supported monophyletic clades made up from: (i) Cinara (Cinara) ponderosae and Cinara (Cinara) terminalis (both from Cinara cluster A), (ii) some Stomaphis spp., and (iii) most Lachnus species. The latter belong to various closely related species, some suspected to be synonyms of Lachnus tropicalis (Blackman and Eastop, 1994), which would be consistent with the high sequence identity (>99%) of their S. symbiotica endosymbionts’ 16S rRNA gene. Interestingly, most S. symbiotica strains from Cinara clade A form a well-supported monophyletic clade, and within this, there is high congruency with the phylogenetic relationships of the Buchnera strains found in the respective hosts, particularly within two subclades (Fig 2A and B, vertical black lines). In contrast, most S. symbiotica strains from Cinara clade B are polyphyletic, and their phylogenetic relationships do not seem to mirror those of the respective Buchnera symbiont. Curiously, both the Buchnera and S. symbiotica from Cinara (Cupressobium) costata are recovered nested within strains from Cinara clade A. This contrasts with previously established Cinara phylogenetic relationships (Chen, Favret, et al., 2015; Meseguer et al., 2015), which show this species to be part of a basal group of Cinara clade C. On the other hand, S. symbiotica strains from Lachnus roboris, Lachnus quercihabitans, Tuberolachnus salignus, and Pterochloroides persicae are all recovered within a clade encompassing most S. symbiotica strains from Cinara spp., reflecting no congruency with neither their hosts’ nor their corresponding Buchnera relationships.
Serratia strains from Stomaphis spp. are recovered nested within both the free-living S. marcescens strains and the Lachninae S. symbiotica clade. The former constitutes what we denominate the S. marcescens-like secondary symbionts (hereafter SMLSS), all of which have been identified from aphids belonging to a single clade of Stomaphis spp. (Fig 2A and B). The latter are recovered as a monophyletic clade which is congruent with the Buchnera phylogeny, and as basal to the clade comprising most S. symbiotica strains from Cinara species.
Next, we investigated the diversity in tissue tropism and intracellular location of S. symbiotica within distantly related Lachninae by whole-mount FISH using aphid embryos. We found that all of the FISH-analysed specimens for S. symbiotica indeed housed this symbiont, meaning they were fixed in the population and suggesting that they represent obligate symbionts. Interestingly, we observed a great diversity in both cell-shape and tissue tropism of S. symbiotica among the selected Lachninae (Figs 3 and S1). In C. (Cu.) tujafilina and C. (Cu.) cupressi, S. symbiotica is present in the periphery of the Buchnera bacteriocytes, co-infecting them, and also occupying its own bacteriocytes (Figs 3A–B and S1A-D). On the other hand, in Cinara (Cinara) cedri (clade B); Tu. salignus; Pt. Persicae; and Tr. caudata, S. symbiotica is housed exclusively inside distinct bacteriocytes (Figs 3C, D, E, F, and S1E-L), and thus “stably” internalised in its own distinct host cells. The distribution of these secondary bacteriocytes is different between the aphid species. In C. (Ci.) cedri they are interspersed among Buchnera bacteriocytes (Figs 3C, and S1E-G), in Tu. salignus they are found forming a “bacteriome core” surrounded by primary bacteriocytes (Figs 3D and S1H-I), and in Pt. persicae they form a “layer” along the bacteriome (Figs 3E and S1J). S. symbiotica cells appear round-shaped in C. (Ci.) cedri, Tu. salignus, and Pt. Persicae, while in Tr. caudata the secondary symbiont retains a rod shape and cell size similar to free-living Serratia strains (Figs 3F and S1K-L). Moreover, the S. symbiotica strains of C. (Cu.) tujafilina (Figs 3A and S1A-C) and C. (Cu.) cupressi (Figs 3B and S1D) show an elongated filamentous cell shape, similarly to the facultative S. symbiotica symbiont of Ac. pisum (Moran et al., 2005).
Sodalis-like secondary symbionts
Sodalis-like 16S rRNA gene sequences have been previously amplified from some Lachninae aphids, including Eulachnus spp., Nippolachnus piri, and Cinara (Cinara) glabra (Cinara cluster C) (Burke et al., 2009). Here, we have confirmed, by 16S rRNA gene sequencing, the presence of a Sodalis-like secondary symbiont (hereafter SLSS) in different populations of Eulachnus mediterraneus (Spain) and Eulachnus rileyi (Austria, France, and Spain) (Table S1). Additionally, using a specific primer designed to target the 16S rRNA gene of the SLSS of Eulachnus spp. (see Materials and Methods), we detected and sequenced SLSS 16S rRNA gene amplification products from all the collected Cinara (Schizolachnus) obscurus (clade B) populations from Austria, France, and Spain, which were almost identical to each other, pointing towards this symbiont being fixed in this aphid species. Although we lacked enough specimens to perform FISH microscopy, we were able to amplify a sequence from a SLSS from a population of the closely related aphid species Cinara (Schizolachnus) pineti. In this species, we were able to corroborate that the SLSS was the sole secondary bacterium lineage found through a MiSeq amplicon sequencing of the V3-V4 region of the 16S (Figure 4F). A similar approach has been recently been used to successfully detect symbionts associated to Cinara species (Jousselin et. al., 2016). By means of a Bayesian phylogenetic reconstruction, we determined that SLSSs of Lachninae aphids constitute at least four different lineages, nested within an unresolved clade made up of Sodalis bacteria, and Sodalis-like symbionts from different insect species (Fig 4A). Interestingly, the SLSS from Eulachnus spp. form a well-supported monophyletic clade, reinforcing previous results (Burke et al., 2009) and pointing towards a common origin. Considering the close phylogenetic relationship of Eulachnus and Essigella (Fig 1; Chen, Favret, et al., 2015), and the fact that Essigella has not been found associated to neither S. symbiotica, Candidatus Hamiltonella defensa, nor Candidatus Regiella insecticola endosymbionts (Russell et al., 2003), we hypothesised that the SLSS detected in Eulachnus spp. could have been either fixed in the common ancestor of these two genera or right before the diversification of Eulachnus species. Regretfully, we were unable to recover any secondary symbiont’s 16S rRNA gene sequence from Essigella californica (collected in France), neither by specific PCR nor by molecular cloning (50 colonies analysed). In the case of the SLSSs from both analysed Cinara (Schizolachnus) species, we found they were also recovered as a well-supported monophyletic clade, providing evidence towards their common origin.
Given our failure to detect sequence belonging to a secondary symbiont in Es. californica using the aforementioned methods, we amplified the V3-V4 region of the 16S rRNA gene and performed massive sequencing in the MiSeq Illumina platform. Surprisingly, we were unable to detect an additional bacterial lineage in Es. californica (Figure S2). This could reflect either the very low quantity of DNA belonging to the secondary endosymbiotic bacteria (relative to Buchnera’s), or a strong bias of the “universal” PCR primers used for this protocol towards amplifying Buchnera’s 16S rRNA gene.
To localise the SLSS within the bacteriome, we used the Eulachnus SLSS specific reverse PCR primer as a probe for FISH on dissected aphid embryos (Table S2). We found that all individuals from both Eu. rileyi and Eu. mediterraneus harbour the SLSS inside specific bacteriocytes within the bacteriome (Figs 4B-C and S1M-O). Additionally, since we were unable to determine the 16S rRNA gene sequence of the putative secondary endosymbiont of Es. californica, we used a combination of a general bacterial and a Buchnera-specific probe for FISH in embryos of this aphid species. We observed that there was indeed a distinct secondary bacterial symbiont with a very similar morphology and location as that of Eulachnus spp. (Figs 4D and S1P-R). In both Eulachnus spp. and Es. californica, we found that the symbiont is somewhat underrepresented compared to Buchnera, similarly to what is observed for the SLSS of N. piri (Fukatsu et al., 1998). This could be the reason why we failed to detect it in Es. californica, even whith V3-V4 MiSeq amplicon sequencing. Regarding C. (Sc.) obscurus, we did not observe a staining when the SLSS probe (designed for the SLSS of Eulachnus spp.) was used for FISH in this aphid species. Therefore, we used the same approach as for Es. californica. Using a general bacterial probe in combination with a Buchnera- specific probe, we found that C. (Sc.) obscurus harbours two phylogenetically distinct spherical endosymbionts in separate bacteriocytes (Figs 4E and S1S). In contrast to Eulachnus, the SLSS bacteriocytes from C. (Sc.) obscurus are more abundant and located along the bacteriome surrounding Buchnera bacteriocytes.
“X-type” secondary symbionts
For the current study, we were able to collect two populations of Ma. submacula aphids. Since the an “X-type” (or PAXS) symbiont was suspected to be the secondary symbiont of this aphid species (Lamelas et al., 2008; Burke et al., 2009), we used a specific PCR assay to confirm the presence of this endosymbiont in both populations. Through this assay, we recovered 16S rRNA gene fragments sharing 100% sequence identity with each other and >99% with other X-type symbionts. To facilitate phylogenetic analysis, we additionally performed molecular cloning of the 16S rRNA using universal primers (Table S2). Additionally, through the same method, we found an “X-type” symbiont associated with the aphid C. (Ci.) confinis. A Bayesian phylogenetic analysis of the different Aphididae “X-type” symbionts revealed that these form a well-supported monophyletic cluster closely related to Candidatus Hamiltonella defensa and Candidatus Regiella insecticola, facultative endosymbionts from Ac. pisum (Fig 5A). Particularly, the sequences obtained from Ma. submacula populations from three different countries form a well-supported monophyletic clade (separate from that of C. [Cu.] confinis, C. [Cupressobium] juniperi [clade C], and Ac. pisum), and show a high sequence identity among each other (>99%). We then performed FISH analysis on Ma. submacula embryos using specific probes for Buchnera and X-type (see Materials and Methods). We found that all analysed individuals from both Ma. submacula populations contained X-type symbionts distributed along the bacteriome, both surrounding Buchnera bacteriocytes and in their own distinct ones (Figs 5B and S1T-U). Regarding C. (Cu.) confinis, we lacked enough individuals to perform FISH analysis, and therefore its localisation within the bacteriome remains undetermined.
“Candidatus Fukatsuia” gen. nov. and “ Candidatus Fukatsuia symbiotica” sp. nov
Given that X-type symbionts form a well-supported monophyletic clade with high sequence identity (>99%), we propose the specific name “Candidatus Fukatsuia symbiotica” for the lineage of enterobacterial symbionts found, so far, affiliated only to aphids (Hemiptera: Aphididae). Ca. Fukatsuia’s closest relative, by 16S rRNA gene sequence identity, would be Budvicia diplopodorum strain D9 (INSDC accession number HE574451.1), with which it shares 93% sequence identity. The generic name “Fukatsuia” is in honour of Dr. Takema Fukatsu (Prime Senior Researcher at the National Institute of Advanced Industrial Science and Technology, Japan), who has enormously contributed to the study of aphid biology and that of their endosymbionts, with particular emphasis on his early work on secondary endosymbionts form Lachninae aphids. The specific epithet ‘symbiotica’ alludes to the symbiotic habit of Ca. Fukatsuia bacteria.
In the Lachninae aphid Ma. submacula, “Ca. Fukatsuia” is found inhabiting the bacteriome tissue and its cell presents a filamentous shape of variable length. Its tissue localisation and cell shape in aphids other than Ma. submacula remains unknown. Similarly to S. symbiotica, different Ca. Fukatsuia lineages are of different dispensability to their hosts: being a defensive facultative symbiont (of variable degrees of protection, depending on the strain) in Ac. pisum (Guay et al., 2009; Heyworth and Ferrari, 2015), and being a putative co-obligate symbiont in Ma. submacula aphids. Co-obligate lineages of “Ca. Fukatsuia symbiotica” may well represent separately evolving units, and thus, some lineages may constitute separate species within the same genus. This could be the case for the well-supported specific lineage associated to Ma. submacula aphids, however further genome data from several “Ca. Fukatsuia” symbionts is needed to test this hypothesis. Currently available sequences that correspond to “Ca. Fukatsuia symbiotica” are deposited under INSDC accession numbers FJ821502.1, KP866544.1, KP866545.1, LT600381.1, EU348311.1, EU348312.1, FJ655539.1, LT600338.1, and LT600340.1.
Discussion
Many insects maintain intimate associations with obligate endosymbiotic bacteria harboured in specialised organs (bacteriomes). One crucial question within the field is how do these associations evolve. One way of approaching this question is through the study of “recently” acquired endosymbionts. In the Lachninae subfamily of aphids, it has previously been proposed that an ancient loss of the riboflavin biosynthetic capability by Buchnera promoted the settlement of a co-obligate secondary endosymbiont (putatively S. symbiotica) in the LLCA (Manzano-Marín and Latorre, 2014; Manzano-Marín et al., 2016). Yet, various extant members of this subfamily have been found to be associated to bacterial taxa phylogenetically distinct from S. symbiotica (Lamelas et al., 2008; Burke et al., 2009). Therefore, the study of secondary endosymbionts within this subfamily is expected to give important clues regarding symbiont establishment and transition from facultative to co-obligate relationships. However, the studies on the symbiotic systems of aphids belonging to the Lachninae have been hampered by the lack of an evolutionary framework to correctly interpret the genomic and metabolic changes, as well as their links with the different stages of the transformation process towards an obligate intracellular lifestyle (Pérez-Brocal et al., 2006; Lamelas, Gosalbes, Manzano-Marín, et al., 2011; Lamelas, Gosalbes, Moya, et al., 2011; Manzano-Marín and Latorre, 2014).
In this work, we have explored the diversity, phylogenetic relationships, and location of different secondary endosymbionts within key members of the Lachninae subfamily. This has enabled us to propose an evolutionary scenario for the settlement, “stable” internalisation into distinct bacteriocytes, and replacements of the original secondary co-obligate endosymbiont from the LLCA (Fig 6). Firstly, with a combination of specific PCR assays, 16S rRNA gene sequencing, and FISH microscopy, we determined that all analysed specimens indeed harbour fixed secondary endosymbionts at the population/species level. This fact, in combination with previously published microscopic (Buchner, 1953; Fukatsu and Ishikawa, 1998; Fukatsu et al., 1998; Lamelas et al., 2008; Pyka-Fościak and Szklarzewicz, 2008; Michalik et al., 2014) and molecular (Pérez-Brocal et al., 2006; Burke et al., 2009; Lamelas, Gosalbes, Manzano-Marín, et al., 2011; Lamelas, Gosalbes, Moya, et al., 2011; Manzano-Marín and Latorre, 2014; Manzano-Marín et al., 2016) data from Lachninae aphids, provides strong evidence for the dependence of members of this subfamily on co-obligatory secondary endosymbionts, putatively due to the ancient pseudogenisation of the riboflavin biosynthetic genes in the Buchnera harboured by the LLCA, which lived at least some 85-106 Mya. The detection of S. symbiotica in different aphid species from at least six Lachninae genera across all five tribes, along with the genomic data from three strains of this symbiont at different stages of the genome reduction process (Lamelas, Gosalbes, Manzano-Marín, et al., 2011; Manzano-Marín and Latorre, 2014; Manzano-Marín et al., 2016), point towards an early establishment of S. symbiotica as co-obligate in the LLCA. While spherical and found consistently inside bacteriocytes (with a highly reduced genome) in Tu. salignus and C. (Ci.) cedri, the filamentous and broadly distributed (with a mildly-reduced genome) S. symbiotica from C. (Cu.) tujafilina would preserve the traits of the putative “ancient” S. symbiotica from the LLCA. This hypothesis would be consistent with the high level of genomic, metabolic, and phenotypic similarity of the co-obligate S. symbiotica from C. (Cu.) tujafilina and the facultative S. symbiotica from Ac. pisum (Moran et al., 2005; Lamelas et al., 2008; Burke and Moran, 2011; Manzano-Marín and Latorre, 2014). We find this scenario to be most parsimonious, as it would require one single event of infection with a shared S. symbiotica ancestor in the LLCA followed by at least four “stable” internalisations of S. symbiotica into bacteriocytes. This ancient secondary symbiont would have then undergone at least six independent events of symbiont replacement. An alternative scenario would require additional events of symbiont replacement with distinct S. symbiotica strains in specific Lachninae lineages. This last scenario is suggested by the lack of general congruency between Buchnera and S. symbiotica 16S rRNA gene phylogenies and by the multiple S. symbiotica lineages recovered from the 16S rRNA gene phylogeny in Fig 2B. However, this pattern also suggests different rates of sequence evolution of the S. symbiotica symbionts, possibly driven by changes in the Buchnera-S. symbiotica relationship, such as metabolic pathway splits (e.g. tryptophan) and/or changes in the symbiont’s tissue tropism (Manzano-Marín et. al., 2016). We expect that further sequencing of complete genomes from several Lachninae aphids will help clarify this.
Within the Lachnini tribe, there could have been either one or two independent events of “stable” internalisation and confinement of S. symbiotica into distinct bacteriocytes. The latter hypothesis is supported by the lack of congruency between Buchnera and S. symbiotica lineages from Pt. persicae and Lachnus spp., suggesting separate events of genome reduction of S. symbiotica in these two aphid lineages. Regarding the Lachnus genus, microscopic observations of bacteriocytes from La. roboris have revealed that it keeps a vertically transmitted association with two spherical-shaped bacteria (presumably Buchnera and S. symbiotica) residing in separate bacteriocytes and a third filamentous bacterial symbiont residing in distinct bacteriocytes, whose identity remains unknown (Buchner, 1953). If this tertiary symbiotic bacterium was established in the common ancestor of La. roboris and La. quercihabitans, it could explain the longer branches, relative to La. tropicalis (Fig 2A and B), given that the presence of an additional symbiont could facilitate the process of genomic erosion in the S. symbiotica symbiont. Such a phenomenon was observed in those Buchnera strains from Lachninae species that have established co-obligate associations with S. symbiotica (Manzano-Marín et al., 2016).
Also, we have confirmed that Ma. submacula aphids indeed harbour Ca. Fukatsuia bacteria (which belong to the group of symbionts previously referred to as “X-type” or PAXS) and determined its location within aphid embryos. We found that Ca. Fukatsuia was present in all of the microscopy analysed individuals, which, in combination with previous analyses detecting the presence of these symbionts in a Spanish (Lamelas et al., 2008), and a UK (Burke et al., 2009) population, points towards its obligate status. The morphology and location of Ca. Fukatsuia (Fig 5B and S1T-U) resembles that observed for facultative endosymbionts of other aphids (Moran et al., 2005), similarly to what is observed for the co-obligate S. symbiotica from C. (Cu.) tujafilina (Lamelas et al., 2008; Manzano-Marín and Latorre, 2014). This suggests that Ca. Fukatsuia from Ma. submacula has not yet undergone a massive genome reduction, contrary to what is observed in the pleomorphic S. symbiotica of C. (Ci.) cedri (Lamelas, Gosalbes, Manzano-Marín, et al., 2011) and Tu. salignus (Manzano-Marín et al., 2016). This, in combination with the lack of a S. symbiotica endosymbiont, points toward a replacement of the “ancient” secondary co-obligate endosymbiont which occurred at least some 77-99 Mya in the branch leading to Ma. submacula. It is important to note that Buchnera strains from aphids identified as Ma. submacula form at least two phylogenetically distinct lineages (Fig 2A): one sister to Maculolachnus sijpkensis and one sister to this Ma. submacula+Ma. sijpkensis clade. Thus, in the current work, we refer to the latter as the one associated to Ca. Fukatsuia. Given that no cox1 gene sequence is provided for the Ma. submacula whose Buchnera strain is recovered as sister to that of Ma. sijpkensis, we are unable to judge if these the two Buchnera lineages have been indeed isolated form the same species or if the polyphyly of Buchnera strains from Ma. submacula is due to wrong taxonomic identification.
Regarding the Stomaphidini, we postulate that at least two events of symbiont replacement have occurred. Stomaphis pini and Stomaphis quercisucta (both belonging to the same clade: Stomaphis clade A) have been found to be associated with S. symbiotica (Burke et al., 2009; Chen, Wang, et al., 2015), and microscopic investigations into St. quercus have revealed that this species houses three vertically-transmitted endosymbiotic bacteria: Buchnera plus two secondary symbionts which apparently reside inside the same secondary bacteriocytes (Buchner, 1953; Pyka-Fościak and Szklarzewicz, 2008). These could be an Arsenophonus and/or a Gilliamella-like secondary symbiont (Figure S3), both of which have been found associated to different Polish populations of this aphid species (Burke et al., 2009). In addition, both Stomaphis aphananthae and St. yanonis are associated with SMLSSs (Burke et al., 2009), suggesting an establishment of this symbiont in the branch leading to the clade comprising these two species (the unresolved Stomaphis clade C). Furthermore, microscopic analyses of St. yanonis bacteriomes have revealed this species indeed houses a tubular secondary symbiont, putatively SMLSS, in separate bacteriocytes located on the surface of the bacteriome “core” formed by the Buchnera bacteriocytes (Fukatsu and Ishikawa, 1993, 1998). Consequently, we propose at least two events of acquisition of a new endosymbiont: one before the diversification of the clade comprising St. cupressi (Stomaphis clade A) and another one before the expansion of the large unresolved clade including St. yanonis (Stomaphis clade C) (Fig 2A).
With respect to the Tramini+Tuberolachnini clade, all currently analysed members have been found to be associated with S. symbiotica (Burke et al., 2009), except for N. piri, which contains a putatively pleomorphic SLSS housed inside separate bacteriocytes (Fukatsu and Ishikawa, 1993, 1998; Fukatsu et al., 1998; Burke et al., 2009). In the case of both Tu. salignus and Tr. caudata (which started diverging at least some 47-69 Mya), the S. symbiotica symbiont is found exclusively within bacteriocytes (Figs 3D, F, S1H-I, and K-L). However, the cell shape of the endosymbionts is strikingly different. While in Tu. salignus they occur as large spherical-shaped cells, in Tr. caudata the symbionts have a small rod-shaped morphology, resembling the cell shape of free-living Serratia strains. This could be indicative of this bacterium being in the very first stages of “stable” internalisation into bacteriocytes, while still preserving its rod shape and putatively having a genome resembling closely that of S. symbiotica from C. (Cu.) tujafilina, rather than that of Tu. salignus.
In regards to the Eulachnini, most Cinara spp. have been consistently found associated to S. symbiotica strains. Microscopic investigations of Cinara (Cinara) pini (clade A) and C. (Ci.) cedri have revealed they indeed harbour a pleomorphic secondary endosymbiotic bacterium obligatorily inside bacteriocytes (Fukatsu et al., 1998) (Figs 3C and S1E-G), and in the case of the latter, genomic-based metabolic inference has established that both Buchnera and S. symbiotica are required for the biosynthesis of various essential nutrients (Gosalbes et al., 2008; Lamelas, Gosalbes, Manzano-Marín, et al., 2011). Additionally, a high level of congruency between the phylogenetic relationships of Buchnera and S. symbiotica strains from clade A Cinara (Fig 2A and B) suggests a single event of drastic genome reduction followed by divergence, similar to what is observed for Buchnera. On the contrary, S. symbiotica from clade B Cinara do not show this congruent pattern, pointing possibly to independent events of drastic genome reduction. Within Cinara clade B, C. (Sc.) obscurus would represent a case of symbiont replacement by a SLSS, which is present obligatorily inside bacteriocytes (Fig 4E and S1S). This SLSS is also present as in the closely related C. (Sc.) pineti (in which it is the only other symbiont present), and according to a previous study, this species also presents a spherical secondary endosymbiotic bacterium (presumably the detected SLSS) which is vertically transmitted in both oviparous and viviparous generations (Michalik et al., 2014). Taken together, this suggests the replacement of S. symbiotica by a SLSS in the common ancestor of these two Cinara (Schizolachnus) species. Whether or not this symbiont is widespread within the Cinara (Schizolachnus) subgenus remains to be explored. As most Cinara, C. (Cu.) tujafilina and C. (Cu.) cupressi, are associated to S. symbiotica strains. However, both the location and 16S rRNA gene sequence of these more closely resemble the facultative strains from Aphidinae aphids (Figs 2B, 3A-B, S1A-D; Moran et al., 2005). Genome-based metabolic inference has provided evidence towards the obligate status of S. symbiotica in C. (Cu.) tujafilina, given the loss of the riboflavin biosynthetic capability of Buchnera, an essential co-factor now synthesised by S. symbiotica (Manzano-Marín and Latorre, 2014). This, in addition to the consistent association of these two Cinara (Cupressobium) species with S. symbiotica, led us to infer that these aphids do indeed keep an obligate association with closely related secondary endosymbiotic strains. Within Cinara clade C, evidence of at least one species being affiliated to a SLSS (Fig 4A, Cinara [Ci.] glabra) and two to Ca. Fukatsuia (Fig 5A, Cinara [Cu.] confinis and Cinara[Cu.]juniperi), rather than S. symbiotica, suggests that some events of symbiont replacement have occurred in this group of species. This could have been facilitated due to the niche occupied by S. symbiotica, being similar to that of facultative endosymbionts of Ac. pisum (Sandström et al., 2001; Moran et al., 2005; Sakurai et al., 2005; Tsuchida et al., 2005, 2010). Finally, we propose a symbiont replacement event by a SLSS in the branch leading to the Eulachnus species. Consistent with previous observations in Eu. rileyi (Michalik et al., 2014), we found that both Eu. rileyi and Eu. mediterraneus species harbour spherical SLSSs in separate bacteriocytes, spatially-arranged in a very similar fashion (Fig 4B, C, S1M, O). Also, even though we were unable to recover a 16S rRNA gene sequence belonging to a bacterial taxon other than Buchnera in Es. californica (in a greater abundance than 1%), we were able to detect by FISH microscopy the presence of spherical bacterial endosymbionts residing in distinct bacteriocytes, localised similarly to those inhabited by SLSSs in Eulachnus spp. (Fig 4D, SR). Therefore, pending further studies, it could be suggested that the secondary symbiont found in Es. californica could belong to the same lineage as the SLSSs of Eulachnus species.
A feature we observed repeatedly was the change in the endosymbionts’ tissue tropism along distantly related aphid species. This changes included both bacteriocyte arrangement within the bacteriome and “stable” internalisation of the secondary endosymbionts in distinct bacteriocytes. In the case of S. symbiotica and Buchnera, genome data is available for three species. Regarding the bacteriocyte arrangement within the bacteriome, a similar case has been previously reported within spittlebugs (Auchenorrhyncha: Cercopoidea). Within this superfamily, species belonging to the tribe Philaenini have undergone symbiont replacement of the ancient Candidatus Zinderia endosymbiont by a Sodalis-like symbiont (Koga et. al., 2013). This shift in co-obligate symbiont is also accompanied by a newly evolved type of bacteriocyte with a different arrangement. In Lachninae aphids, changes in the bacteriocyte arrangement within the bacteriome could also be linked to symbiont replacement events, and/or changes in the interdependent metabolic “wiring” of their symbionts which promotes “stable” internalisation into distinct bacteriocytes in a non-deterministic matter. Regarding the latter, we have previously postulated that these changes in the interdependent metabolic “wiring” could also be involved in the “stable” internalisation into distinct bacteriocytes (Manzano-Marín et. al., 2016), which speculatively could be triggered by the constraints on the exchange of certain intermediary metabolites. Developmentally, this shift in tissue tropism would involve a change in the development and colonisation of the bacteriome by the secondary symbionts. In Ac. pisum, S. symbiotica acquires a broad distribution in the bacteriome (e.g. occupying both sheath cells, secondary bacteriocytes, and co-infecting Buchnera’s) following the formation of sheath cells (Koga et. al., 2012). Before this, and after bacteriocyte cellularisation and symbiont sorting, the two “stably” internalised symbionts of the different Lachninae aphids would putatively remain confined to their own bacteriocytes until vertical symbiont transmission.
In summary, we propose an evolutionary framework which should assist in future studies on the Lachninae, a symbiont-diverse subfamily. Our findings reveal a dynamic pattern for the evolutionary history of “recently” established endosymbionts, thus contributing to a better understanding of how mutualism in endosymbiotic associations can evolve. We believe that further studies directed towards the bacteriome development and its colonisation by endosymbiotic bacteria in species from the Lachninae subfamily, could also provide hints towards the evolution of new spatial arrangements and even type of bacteriocytes. The role these recently-acquired bacteria have played in the adaptation of their aphid hosts to different niches/feeding sites/plants and their role in speciation in this peculiar subfamily remains to be explored.
Materials and Methods
Aphid collection and storage
All aphids used for this study were collected in various locations around Rennes (France), Vienna (Austria), and Valencia (Spain). Collection details can be found in Table S1. Aphids used for DNA extraction were stored at-20°C in absolute ethanol inside a 1.5 mL Eppendorf tube. Aphids used for fluorescence in situ hybridisation experiments were dissected in absolute ethanol to extract embryos. These were then directly transferred to modified Carnoy’s fixative (6 chloroform : 3 absolute ethanol : 1 glacial acetic acid) and left overnight, following (Koga et al., 2009) protocol to quench autofluorescence. Briefly, fixed embryos were washed with absolute ethanol and transferred into a 6% solution of H2O2 diluted in absolute ethanol and were then left in this solution for two to six weeks (changing the solution every three days). When bleached, they were washed twice with absolute ethanol and stored at-20°C.
Fluorescence in situ hybridisation
Hybridisation of aphid embryos was performed overnight at 28°C in standard hybridisation buffer (20mM Tris-HCl [pH 8.0], 0.9 M NaCl, 0.01% SDS, and 30% formamide) and then washed (20mM Tris-HCl [pH 8.0], 5mM EDTA, 0.1 M NaCl, and 0.01% SDS) before slide preparation. The slides were examined using a confocal laser scanning microscope (TCS SP5 X, Leica; and FV1000, Olympus). A list of specific probes used for each aphid species is available in Table S2. Buchnera, S. symbiotica, and Ca. Fukatsuia competitive probes were designed based on (Gómez-Valero et al., 2004) and adapted to match the target bacterial strain. The Eulachnus SLSS probe was designed based on (Attardo et al., 2008), adapted to match the target strain. The embryos from at least 10 individuals were analysed per sample.
16S rRNA gene PCR, cloning, and sequencing
Since all endosymbionts detected in Lachninae members so far are bacteria, we used the primers 16SA1 and 16SB1 (Fukatsu and Nikoh, 1998) to amplify partial 16S rRNA genes (circa 1.5 kbp) for cloning and sequencing. This strategy was adopted in selected cases to facilitate phylogenetic reconstruction. Resulting amplicons were cloned into the pGEM-T Easy Vector 18 (Promega) and SP6 (5’-ATTTAGGTGACACTATAG-3’) and T7 (5’-TAATACGACTCACTATAGGG-3’) primers were used for amplification and sequencing of the cloned DNA (at least 5 clones from each species). Specific primers for either Buchnera or S. symbiotica were designed based on the FISH probes. Specific PCR reactions and sequencing were done mainly to confirm the presence of the secondary endosymbionts. In the case of the SLSS and Ca. Fukatsuia, specific primers were designed based on (Attardo et al., 2008) (“ Sodalis specific”) and (Guay et al., 2009) (“PAXSF”), respectively. For a full list of primers pairs and PCR conditions see Table S2. All sequences have been uploaded to the European Nucleotide Archive and are pending accession (temporarily available at https://figshare.com/s/8eb9686e546394547fe4).
MiSeq sequencing of the V3-V4 region of the 16S rRNA gene from bacteria associated to Es. Californica and C. (Sc.) pineti
Using the same DNA extracted for PCR and cloning, amplification and sequencing of the V3-V4 region (using standard Illumina primers 5’-CCTACGGGNGGCWGCAG-3’ and 5’- GACTACHVGGGTATCTAATCC-3’) of the 16S rRNA gene was performed in an Illumina MiSeq machine (paired-end 2x300 bp) at the FISABIO Center (Generalitat Valenciana). Next, mothur v1.31.2 (Schloss et. al., 2009) was used for merging of the paired-ends reads and taxonomic assignment of the resulting contigs. Briefly, reads where first quality trimmed using fastx_toolkit v0.0.14 (http://hannonlab.cshl.edu/fastx_toolkit/, last accessed November 9, 2016). Next, we joined the overlapping paired ends reads using the make.contigs function of mothur and filtered out all contigs shorter than 420 bps. Then, contigs were aligned to the SILVA NR99 v128 database (Quast et. al., 2013) and those with less than 90% of their length aligned were filtered out. After another step of redundancy removal, rare sequences (possibly resulting from sequencing errors) were merged with frequent unique sequences with a mismatch no greater than 2 bp (precluster function). Resulting contigs were then screened for chimeric sequences using (chimera.uchime function). Remaining contigs were then taxonomically assigned to genus level using the classify.seqs function and the SILVA NR99 v128 database (cutoff=90). Lastly, the remaining unclassified sequences (namely unclassified Enterobacteriaceae) that exceeded 1% of the total contigs were used for a MegaBLAST(Zhang et. al., 2004) search against a database of representatives including all 16S aphid endosymbiont sequences available to date. The best hit for each read, if this surpassed 94.5% identity, was used for genus-level assignment.
Phylogenetic analyses
All phylogenetic analyses were performed as follows. First SSU-ALIGN v0.1 (Nawrocki, 2009) was used to align 16S rRNA sequences, followed by visual inspection of the alignments in AliView v1.17.1 (Larsson, 2014). Then, GBlocks v0.91b (Castresana, 2000) was used to eliminate poorly aligned positions and divergent regions with the option ’-b5=h’ to allow half of the positions with a gap. The final alignments were transformed into nexus format (available online at https://figshare.com/s/8eb9686e546394547fe4) for phylogenetic analysis in MrBayes v3.2.5 (Ronquist et al., 2012) under the GTR+I+G model. Two independent runs, each with four chains (three “heated”, one “cold”), were run for 5,000,000 generations discarding the first 25% as burn-in and checked for convergence. Visualisation and tree-editing was done in FigTree v1.4.1 (http://tree.bio.ed.ac.uk/software/figtree/, last accessed November 9, 2016) and Inkscape v0.91 (http://www.inkscape.org/en/, last accessed November 9, 2016), respectively. For a full list of accession numbers of sequences used for phylogenetic analyses see Table S3.
Supporting Information
S1 Table. Accession numbers, collection data, and taxonomic status of sampled aphids and their endosymbionts.
S2 Table. Primers and probes used to detect endosymbionts from the different aphid species. Primers and probes used in this study to detect endosymbionts and their specificities.
S3 Table. Accession numbers of sequences used for phylogenetic reconstruction.
S1 Fig. Location and morphology of secondary symbionts in selected Lachninae aphids. FISH microscopic images of aphid embryos from selected Lachninae aphids. Symbiont-specific probes were used for FISH, except for panels A-D and M-S in which either one of the symbionts was visualised by a general bacterial probe and the other. Colors are as in corresponding images of the same aphid species in the main text. (A) Lateral and (B-C) Dorsal views of a C. (Cu.) tujafilina bacteriome of an early and later embryos, respectively. (D) Lateral view of a C. (Cu.) cupressi bacteriome of an early embryo. (E) Lateral and (F-G) Dorsal views of a C. (Ci.) cedri bacteriome of an earlier and later embryos, respectively. (H) Lateral and (I) ventral view of a Tu. salignus bacteriome of an earlier and a late embryo. (J) Lateral-ventral view of a Pt. persicae bacteriome. Lateral and (L) Ventral view of a Tr. caudata bacteriome of an early and later embryo. (M) Ventral view of an Eu. mediterraneus bacteriome. (N) Lateral and (O) ventral view of an Eu. rileyi bacteriome of an early and later embryo. (P) Lateral and (Q-R) ventral-lateral views of an Es. californica bacteriome from an early and later embryos. (S) View of an early embryo of a C. (Sc.) obscurus bacteriome. (T) Dorso-lateral and (U) lateral views of early embryos of a Ma. submacula bacteriome. Thick white boxes indicate the magnified region, depicted in the top-right of each panel. The scientific name for each species along with the false colour code for each fluorescent probe and its target group are shown at the top-left of each panel. Scale bars from the unmagnified and magnified FISH images represent 20 and 5µm, respectively.
S2 Fig. 16S V3-V4 rRNA amplicon sequences taxonomic assignment. Stacked bar plot showing the relative abundance of contigs assigned to taxonomical units. On the right, pie chart showing the MegaBLAST assignment of V3-V4 contigs to reference 16S rRNA genes from aphid endosymbionts. “<1%” indicates the percentage relative to all the reads (N).
S3 Fig. 16S rRNA gene-based phylogenetic relationships of GLSS strains from the Aphididae. Bayesian phylogram depicting the relationships and placement of the currently available GLSS strains from Aphididae and selected Enterobacteriaceae, Pasteurellaceae, and Orbaceae. The superscript H at the end of the full species name indicates the symbiont’s host name was used. The accession numbers for each sequence used is indicated within parenthesis after the strain name.
Acknowledgements
The authors would like to acknowledge emeritus professor José Manuel Michelena Saval, professor Joaquin Baixeras, Nicolás Pérez Hidalgo, Christelle Buchard, and Evelyne Turpeau for their invaluable and expert help in collection and/or taxonomic identification of aphid samples. Also, we thank and acknowledge artist/scientist Jorge Mariano Collantes Alegre for the aphid cartoons in Fig 6. This work has been funded by the Ministerio de Economía y Competitividad (Spain) co-financed by FEDER funds [BFU2015-64322-C2-01-R to A.L.]; the European Commission [Marie Curie FP7 PITN-GA-2010-264774-SYMBIOMICS to A.M.M]; the Consejo Nacional de Ciencia y Tecnología (Mexico) [Doctoral scholarship CONACYT 327211/381508 to A.M.M.]; and the Austrian Science Fund [project no. P22533-B17 to M.H.]. The Plant Health and Environment department of INRA is also acknowledged for financial support to JC Simon. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. The authors declare no conflict of interest.