Evolution and Diversity of Facultative Symbionts from the Aphid Subfamily Lachninae

Extraction, PCR analysis, cloning, and sequencing of DNA from aphid samples.A panel of aphids belonging to the aphid subfamily Lachninae (with a focus on species of the genus Cinara) were screened for bacterial endosymbionts (Table 1; see Table S1 in the supplemental material). In most cases, screening was based on a single individual per aphid species or population, unless indicated otherwise (Table 1).

DNA extraction was performed with single aphids using a DNeasy kit (Qiagen). PCR was used to amplify 16S rRNA using primers 10F (5′-AGTTTGATCATGGCTCAGATTG-3′), 35R (5′-CCTTCATCGCCTCTGACTGC-3′), and 1507R (5′-TACCTTGTTACGACTTCACCCCAG-3′). Each reaction mixture (50 μl) contained 10 mM Tris-HCl (pH 9.0), 50 mM KCl, 0.1% Triton X-100, 1.5 mM MgCl₂, 1 mM deoxynucleoside triphosphates (Eppendorf), 250 pmol of each primer, and 2 U Taq polymerase (Eppendorf). PCR amplification was carried out with a Mastercycler Gradient (Eppendorf) as follows: 94°C for 2 min, followed by 35 cycles of 94°C for 1 min, 55°C for 1 min, and 72°C for 3 min and then a final incubation at 72°C for 10 min. Amplified products were examined using standard 1% agarose gel electrophoresis and staining with ethidium bromide. Single, discrete bands were cleaned using Agencourt AMPure PCR purification (Agencourt, Beverly, MA).

In order to separate Buchnera and facultative symbiont 16S rRNA gene sequences, cloning was performed using a pGEM-T Easy kit (Promega), and pGEM-T Easy with its insert was transformed into Escherichia coli competent cells (Promega). The insert size for eight colonies was checked using colony PCR performed with primers M13F (5′-GTAAAACGACGGCCAG-3′) and M13R (5′-CAGGAAACAGCTATGAC-3′), followed by digestion with SalI (Roche) as described by Sandström et al. to differentiate between Buchnera and facultative symbiont clones (30). The amplicon was sequenced from both ends using the M13F and M13R primers at the Genomic Analysis and Technology Core, University of Arizona. Base calls were verified and contigs were assembled using the Sequencher 4.7 DNA software (Gene Codes Corporation).

Symbiont classification and phylogenetic analyses.Symbiont types were classified based on best BLAST hits and were verified by placement in a strongly supported clade on phylogenetic trees with other bacterial taxa that are members of the lineages to which they belong. Sequences were aligned using Clustal X and were checked by hand using MacClade 4.06 (18). The best maximum likelihood (ML) trees were obtained using a heuristic search with the GTR+I+γ model and bootstrapped with 100 replicates using RAxML (34). A Bayesian analysis was completed using MrBayes 3.1.2 (27) with the settings nst = 6 and rates = invgamma and two independent runs with four chains for 1,250,000 generations, with sampling every 1,000 generations. The first 250 samples were discarded, and posterior probabilities were obtained using the remaining 1,000 samples.

A Shimodaira-Hasegawa (S-H) test was used in PAUP* (37) to compare the best ML tree to a tree in which subgenus Cinara aphids possessing “Ca. Serratia symbiotica” cluster B are monophyletic to the exclusion of aphid species possessing other symbiont lineages (except Cinara pinimaritimae, which belongs to a clade with “Ca. Serratia symbiotica” cluster B-containing aphid species with a high level of support). Using the parameters estimated for the best ML tree, the S-H test was performed with full optimization and 1,000 bootstrap replicates.

Branch lengths were estimated using the ML parameters employed for phylogenetic reconstruction of the ML tree.

Nucleotide sequence accession numbers.The GenBank accession numbers of the 16S rRNA gene sequences determined in this study are FJ655482 to FJ655545.

Phylogenetic reconstruction of symbiotic Serratia lineages in aphids.Partial 16S rRNA gene sequences were obtained for a range of Serratia symbionts and other bacterial facultative symbionts, as well as Buchnera (GenBank accession numbers FJ655482 to FJ655545 [see Table S1 in the supplemental material]). ML and Bayesian phylogenetic analyses of the 16S rRNA gene of these taxa resulted in a gene tree showing the evolutionary relationships among symbionts and free-living species in the genus Serratia (Fig. 1).

• Open in new tab
• Download powerpoint

FIG. 1.

ML tree for Serratia based on 16S rRNA genes. Symbiotic bacterial taxa are indicated by the name of the aphid species from which they were isolated and sequenced. ML bootstrap support values are indicated above nodes, and Bayesian posterior probabilities are indicated below nodes. Only support values greater than 75 or 0.75 are shown. There have been at least two independent acquisitions of aphid symbionts in the genus Serratia, one in the common ancestor of clusters A and B and the other in associates of the aphid genus Stomaphis (bold type).

Most aphid symbionts belonging to the genus Serratia fall into two sister clades corresponding to clusters A and B, as described by Lamelas et al. (17). These symbionts include almost all of the Serratia symbionts from the aphid family Aphididae, including the Lachninae.

Cluster A contains symbionts from members of both the Aphididae and the Lachninae, including Cinara. The majority of the symbionts belonging to cluster A show little diversity; the greatest pairwise distance is 0.064 substitution per site. Cluster B contains symbionts solely from the Lachninae. These symbionts show greater diversity than those in cluster A (the greatest pairwise distance is 0.096 substitution per site), and lineages from closely related aphid hosts are generally more closely related to each other than to other symbiont lineages in the clade (Fig. 1). “Ca. Serratia symbiotica” lineages from C. cedri and Pterochloroides persicae cluster with high levels of support.

A third group of Serratia strains from aphids does not branch with clusters A and B. The 16S rRNA gene tree indicates that these strains are more closely related to Serratia marcescens than to other symbiotic clades.

Phylogenetic reconstruction of Buchnera lineages.16S rRNA gene sequences were obtained from all bacterial strains from Lachninae aphids, including Buchnera. Phylogenetic reconstruction of the Buchnera sequences supports clustering of lineages into tribes and shows general congruence with expected relationships of the aphid hosts (Fig. 2) (6, 23), with the exception of the Cinarini. The Cinarini are split into two clades corresponding to the subgenus Cupressobium and the subgenus Cinara. The latter clade is grouped with Stomaphis aphids with high levels of support, excluding subgenus Cupressobium aphids. The evolutionary relationships of aphids in the subgenus Cinara clade are relatively unresolved, but the best ML tree and Bayesian posterior probabilities (Fig. 2) do not support monophyletic grouping of these aphids. However, constraining aphids infected with “Ca. Serratia symbiotica” cluster B into a monophyletic clade with respect to aphids with other symbiont types results in a significantly worse tree (P = 0.01, S-H one-tailed test).

• Open in new tab
• Download powerpoint

FIG. 2.

ML tree for Buchnera associated with Lachninae, based on 16S rRNA genes. For a description of the taxa and phylogenetic support see the legend to Fig. 1. Most species of Lachninae sampled were infected with a facultative symbiont (all lineages except Other Enterobacteriaceae). Circles separated by a comma indicate taxa that were from different individuals in the same population or belonging to the same species. Taxa infected with “Ca. Serratia symbiotica” cluster B are not monophyletic in the Cinara subgenus Cinara clade. Cluster B lineages have been lost from the Cinara subgenus Cinara clade at least five times, including a loss in C. cedri from Chile, as indicated by filled circles on branches. There is a strong geographic pattern for infection with facultative symbionts, with “Ca. Serratia symbiotica” cluster A found mainly in Asia and North America and cluster B found in Europe.

Prevalence of facultative symbionts and their distribution among aphid species.In aphids, facultative symbionts are generally defined as intracellular, vertically transmitted bacterial associates that are facultative from the perspective of the host. Bacteria identified in this study that are closely related to other insect-associated bacteria or are known to reside inside insect cells were assumed to be facultative symbionts, although this assumption remains to be verified experimentally. Almost all aphids screened had a facultative symbiont infection, but infection by more than one facultative symbiont in the same aphid was rare (Fig. 2 and Table 2). Based on data from a survey of aphid species belonging to the entire aphid family (Aphididae), the level of infection with “Ca. Serratia symbiotica” or any facultative symbiont is 13.9% or 45.8%, respectively (28). In contrast, for species of the subfamily Lachninae, the frequencies of infection with “Ca. Serratia symbiotica” and with any facultative symbionts were higher, 43.9% and 78.0%, respectively (P = 0.0006 and P = 0.0014, Fisher's exact test). In Cinara species, the frequency of “Ca. Serratia symbiotica” infection was quite high, 64.3% (P = 0.0002, Fisher's exact test). Only data from this study and the study of Russell et al. (28) were used in this analysis, as other studies may not have reported the absence of symbiont infections.

Aphids in Cinara subgenus Cinara were infected with “Ca. Serratia symbiotica” cluster B or cluster A, “Ca. Hamiltonella defensa,” Wolbachia, relatives of Sodalis, or a gammaproteobacterial lineage referred to in this study as “cluster L” (closely related to a gut-associated symbiont of honey bees [see Fig. S2 in the supplemental material]). Some species had no facultative symbionts. Stomaphis aphids were infected with “Ca. Serratia symbiotica” cluster A, Arsenophonus, relatives of free-living Serratia, and “cluster L” bacteria. The Buchnera sequences from Stomaphis aphids infected with bacteria similar to free-living Serratia form a monophyletic clade (Fig. 2), while the Serratia symbionts themselves appear to be polyphyletic (Fig. 1). Lachnus aphids had “Ca. Serratia symbiotica” cluster A and B infections. One sample from Nippolachnus piri was infected with Sodalis, and another sample from N. piri was infected with “Ca. Hamiltonella defensa.” Finally, Cinara (Cupressobium) aphids were infected only with “Ca. Serratia symbiotica” cluster A.

Closer examination of symbiont sequences shown in Fig. 2 shows that there is a geographical pattern of infection with “Ca. Serratia symbiotica.” All aphids with “Ca. Serratia symbiotica” cluster B were collected in Spain, Poland, and the Czech Republic, except for Cinara edulis collected in the United States. Excluding C. edulis, all Cinara species infected with the cluster B lineage are endemic in Europe, although several of them have been introduced into other parts of the world (Table 1) (1). All aphids infected with “Ca. Serratia symbiotica” cluster A were collected from Japan or the western United States, with the exception of the subgenus Cupressobium aphids found in Spain. Subgenus Cupressobium aphids are the only cosmopolitan aphids found to be infected with “Ca. Serratia symbiotica” cluster A in our samples; all other Lachninae aphids infected with this symbiont lineage were collected from western North America and Asia.

Reconstruction of aphid evolution using Buchnera sequences.As Buchnera evolves synchronously with aphid host lineages (21), the phylogenetic relationships of these lineages can be used as a proxy for the evolutionary relationships of aphids. Here, the phylogeny of Buchnera strains showed general congruence with the relationships of their aphid hosts except the Cinarini, which were split into clades corresponding to Cinara subgenus Cinara and subgenus Cupressobium. The Cinarini were not monophyletic in the analyses of Normark (23), but subgenus Cupressobium taxa were not included in those analyses. The proposed close relationship of the genera Nippolachnus and Tuberolachnus (23) is supported here.

Multiple acquisitions of Serratia symbionts in aphids.“Ca. Serratia symbiotica” was first identified as an organism closely related to free-living and other insect-associated Serratia lineages based on sequence similarity (20, 30, 41). Its vertical transmission and scattered presence in numerous aphid species around the world suggested that this bacterium is a facultative symbiont of aphids (3, 14, 20, 28, 30, 33, 39). Early phylogenetic reconstructions of symbiotic Serratia lineages showed that they arose once from a common ancestor (17, 28, 30, 40, 41). We have identified additional Serratia types associated with Stomaphis aphids in Japan; these types are more closely related to free-living S. marcescens than to “Ca. Serratia symbiotica,” indicating that free-living Serratia strains have been acquired more than once as symbionts in aphids. Although the additional Serratia types associated with Stomaphis species are themselves polyphyletic in our tree (Fig. 1), they are located in a region of the tree with little resolution. Although vertical transmission of these bacteria in Stomaphis aphids has not been demonstrated yet, these facultative symbionts have been visualized in the bacteriocytes of Stomaphis yanonsis using immunohistochemical staining techniques (7), so they are likely symbiotic associates rather than environmental contaminants.

“Ca. Serratia symbiotica” encompasses lineages belonging to both clusters A and B (20). While aphid symbionts in the “Ca. Serratia symbiotica” clade are closely related to insect-associated bacteria, including Serratia ficaria in fig wasps (12) and Serratia entomophila in beetles (36), there is no clear evidence that “Ca. Serratia symbiotica” evolved directly from insect-associated bacterial species. The Stomaphis symbionts seem to have evolved from a close relative of S. marcescens, a free-living bacterial entomopathogen that infects many insect orders and is also found as an opportunistic pathogen in humans (11). The independently derived symbiotic lineages of Serratia may have different functions within their hosts and may reflect a propensity for these bacteria to become successful long-term associates of aphids.

Functional roles of symbiont lineages in the “Ca. Serratia symbiotica” clade.The potential functions of several “Ca. Serratia symbiotica” lineages in particular host species have been evaluated in genetic and experimental studies. Phylogenetic reconstruction of the evolution of “Ca. Serratia symbiotica” lineages allows us to assign potential functional roles and to determine host specificity based on relatedness and evidence of codiversification.

Given that “Ca. Serratia symbiotica” cluster A has been demonstrated to confer heat tolerance in A. pisum (19, 29), it is likely that other symbiont lineages in this clade have similar effects on their aphid hosts, although this has not been tested experimentally.

It has been proposed that symbionts of Cinara aphids have transitioned from facultative symbionts to obligate symbionts for host survival and reproduction (9, 17, 26). A pattern observed for all obligate symbionts studied to date is their cocladogenesis with hosts as a result of long-term vertical transmission and rare horizontal transfer between lineages (13, 20, 21, 38). We have shown that “Ca. Serratia symbiotica” and other symbiont lineages from subgenus Cinara species do not form a clade and have not undergone the strict vertical transmission characteristic of obligate symbionts. The absence of cocladogenesis is not sufficient to deem a symbiotic lineage “not obligate”; however, the loss of symbiont lineages from taxa in a clade whose ancestor was infected with “Ca. Serratia symbiotica” cluster B demonstrates that these symbionts remain facultative in at least some lineages. Loss of cluster B symbionts has occurred five times in the subgenus Cinara according to the ML tree (Fig. 2). Two examples are C. cedri from Chile, which appears to lack any symbiont other than Buchnera, while in Europe “Ca. Serratia symbiotica” seems to be an obligate symbiont of this species (26), and Cinara pseudotaxifoliae from the United States, which is closely related to European aphids infected with “Ca. Serratia symbiotica” cluster B, yet also appears to lack any facultative symbiont. Although many European aphids are infected with “Ca. Serratia symbiotica” cluster B, close relatives or even aphids belonging to the same species in other geographic locations do not require the presence of this organism.

A more plausible scenario than ancient obligate symbiosis in lachnid aphids is that “Ca. Serratia symbiotica” cluster B was a relatively lachnid-specialized facultative symbiotic lineage that was horizontally transferred among clades during the diversification of the Lachninae.

Although “Ca. Serratia symbiotica” from C. cedri may be an obligate symbiont in European aphids, it seems to have been horizontally transferred from more distantly related host species in the Lachninae. Phylogenetic reconstruction using both 16S rRNA and atpD genes suggested that “Ca. Serratia symbiotica” from C. cedri is more closely related to symbionts from non-Cinara lachnids (Pterochloroides species for the 16S rRNA gene and Lachnus roboris and Tuberolachnus salignus for atpD) than to other lineages from the genus Cinara (17). Thus, characteristics of its genome may not be representative of characteristics of “Ca. Serratia symbiotica” lineages from other Cinara aphids.

High incidence of facultative symbionts in Lachninae aphids.Almost all aphids in the Lachninae are infected with a bacterial symbiont in addition to Buchnera (Table 2). Many of these symbionts are “Ca. Serratia symbiotica” or Serratia relatives, but bacteria belonging to other lineages are present in several of the lachnid species screened. The smallest known Buchnera genomes are found in this subfamily, and Buchnera has lost genes for tryptophan biosynthesis in at least one lineage (26). Due to their high incidence, we propose that facultative symbionts such as “Ca. Serratia symbiotica” may be beneficial in the Lachninae due to their ability to supplement nutrition and thus compensate for inadequate provisioning of nutrients by Buchnera. Alternatively, some lineages of facultative symbionts found in these aphids are known reproductive manipulators and could have negative effects on fitness in their hosts (8, 35).

Although facultative symbiont coinfections were rare, the presence of multiple bacterial lineages in an individual or host population could facilitate intersymbiont competition detrimental to host fitness (25). However, competition might be avoided by partitioning of bacterial cells into different host cells, a feature commonly observed for Buchnera and other facultative symbionts in aphids (7, 10, 20, 30). Additionally, the presence of so many bacterial types in the host populations may provide increased opportunity for gene exchange between facultative symbionts, despite the rarity of coinfections. Examination of APSE phage sequences implied that in the past there was phage-mediated horizontal transfer of genes between “Ca. Hamiltonella defensa” and Sodalis or its relatives (5), and our results document the cooccurrence of these two symbiont lineages in the same host species (N. piri) for the first time, showing one potential route by which transfer could have occurred.

In this phylogenetic analysis, we learned that diversity in Serratia symbiont lineages is due to multiple independent origins of symbiosis from free-living Serratia and also to evolution within distinct symbiont clades. We show here that codiversification of “Ca. Serratia symbiotica” cluster B lineages with all subgenus Cinara aphids is unlikely, as is an obligate role. Finally, we show that there is a high incidence of facultative symbiont infection across diverse lineages in the subfamily Lachninae, which may be advantageous due to impaired Buchnera function.