Endosymbionts facilitate rapid evolution in a polyphagous herbivore

Abstract

Maternally transmitted bacterial symbionts can be important mediators of the interactions between insect herbivores and their foodplants. These symbionts are often facultative (present in some host individuals but not others) and can have large effects on their host's phenotype, thus giving rise to heritable variation upon which selection can act. In the cowpea aphid (Aphis craccivora), it has been established that the facultative endosymbiont Arsenophonus improves aphid performance on black locust trees (Robinia pseudoacacia) but not on fava (Vicia faba). Here, we tested whether this fitness differential translated into contemporaneous evolution of aphid populations associated with the different plants. In a laboratory study lasting 16 weeks, we found that the frequency of Arsenophonus‐infected individuals significantly increased over time for aphid populations on black locust but declined for aphid populations on fava. By the end of the experiment, Arsenophonus infection was >3× more common on black locust than fava, which is comparable to previously described infection frequencies in natural field populations. Our results clearly demonstrate that aphid populations with mixed facultative symbiont infection status can rapidly evolve in response to the selective environments imposed by different host plants. This selection differential may be a sufficient explanation for the global association between Arsenophonus‐infected cowpea aphids and black locust trees, without invoking additional assortative mechanisms. Because the aphid and plant originate from different parts of the world, we further hypothesize that Arsenophonus infection may have acted as a preadaptation that has promoted functional specialization of infected aphids on a novel host plant.

Abstract

Aphid populations show rapid symbiont‐facilitated evolution in response to selection imposed by host plant. Aphids infected with the facultative symbiont Arsenophonus come to predominate on black locust over time, but become the minority on fava.

Open in new tabDownload slide

INTRODUCTION

Many arthropods are infected with heritable symbionts that have profoundly influenced their evolutionary trajectory (Moran, McCutcheon, & Nakabachi, 2008). For example, blood‐feeding and phloem‐feeding niches became accessible through ancient partnerships with microbes capable of providing nutrients that were lacking from these imbalanced diets (Baumann, 2005; Chen, Li, & Aksoy, 1999). Millions of years later, these obligate symbionts are intimately integrated into their host's biology and absolutely critical for host survival (Douglas, 1989; Hansen & Moran, 2011; Hosokawa, Koga, Kikuchi, Meng, & Fukatsu, 2010).

In the present day, facultative microbial symbionts can influence contemporaneous evolution (Oliver, Campos, Moran, & Hunter, 2008; Vorburger & Perlman, 2018; White, 2011). Facultative symbionts are often heritable, just like obligate symbionts, but are not strictly required for host survival and reproduction. Many facultative symbionts are reproductive parasites that persist in host populations due to manipulations such as cytoplasmic incompatibility (Werren, Baldo, & Clark, 2008). Alternatively (or additionally), facultative symbionts can confer large phenotypic effects that are only conditionally beneficial, such as defending against natural enemies (Oliver, Russell, Moran, & Hunter, 2003; Scarborough, Ferrari, & Godfray, 2005), changing coloration (Tsuchida et al., 2010), shifting thermal tolerance (Montllor, Maxmen, & Purcell, 2002) or affecting host plant range in herbivorous insects (Tsuchida, Koga, & Fukatsu, 2004; Wagner et al., 2015). These facultative symbionts therefore provide heritable variation upon which natural selection can act, and several experimental studies have demonstrated shifts in the composition of a host population in response to selective forces acting on symbiont‐conferred traits (Dykstra et al., 2014; Oliver et al., 2008). Such studies have primarily focused on interactions between hosts with facultative defensive symbionts and the enemies against which the symbionts defend (Jaenike, Unckless, Cockburn, Boelio, & Perlman, 2010; Parker, Hrček, McLean, & Godfray, 2017; Rouchet & Vorburger, 2014).

Here, we document symbiont‐facilitated evolutionary response to host plants in a polyphagous herbivore. The cowpea aphid, Aphis craccivora, can be infected by several different heritable facultative symbionts, one of which is a strain of bacteria in the genus Arsenophonus (Brady et al., 2014; Brady & White, 2013). Previously, it has been shown that Arsenophonus‐infected aphids enjoyed a substantial fitness benefit relative to uninfected aphids on black locust (Robinia pseudoacacia), exhibiting increased population growth through higher reproduction and/or survival (Wagner et al., 2015). In contrast, infected aphid fitness was reduced on other host plants such as fava (Vicia faba; Wagner et al., 2015). We hypothesized that aphid populations initiated with mixed infection status would evolve rapidly in response to foodplant‐imposed selection, with Arsenophonus‐infected individuals favoured on black locust and uninfected individuals favoured on fava. A sixteen‐week (~11 aphid generations; Rani & Remamony, 1998) laboratory experiment supported the hypothesis, resulting in populations with proportional infection rates comparable to those observed from field aphid populations collected from the respective host plants (Brady et al., 2014). We therefore speculate that differential selection may be a sufficient explanation for the global association between Arsenophonus‐infected A. craccivora and black locust. The aphid and plant are thought to originate from different parts of the world, and we further hypothesize that Arsenophonus infection is acting as a preadaptation that facilitates aphid specialization on a novel host plant.

METHODS AND MATERIALS

To assess host plant‐dependent demographic changes in Arsenophonus infection among cowpea aphids, we tracked the composition of replicate populations in a laboratory setting. This experiment used Arsenophonus‐infected (Ars+) and uninfected (Ars−) aphids from the same genetic clone (LW). The Ars− colony had initially been cured using antibiotics 3 years previously (Wagner et al., 2015), and both colonies were separately maintained on fava. We initiated each of 12 replicate aphid populations with 20 Ars+ and 20 Ars− aphid nymphs, that ranged from second to third instar. Each population was randomly assigned to either black locust or fava as a host plant and maintained on only that plant species for the duration of the experiment. Each population was restricted to potted plants enclosed in a plastic jar with fine mesh panels maintained in a laboratory under fluorescent grow lights at 16 L: 8 D and 22°C. We used 10 cm pots with two one‐month‐old black locust or two two‐week‐old fava seedlings raised in a greenhouse with PRO‐MIX® BX Mycorrhizae soil. Every two weeks, we moved a subset of 20 aphids onto a fresh plant and deposited a second subset of 20 aphids in 95% ethanol for storage at −20°C. We repeated this procedure 8 times for a total of 16 weeks. We extracted DNA from each individual aphid using a squish extraction protocol adapted from Gloor, Nassif, Johnson‐Schlitz, Preston, and Engels (1991). This methodology was the same as described by Dykstra et al. (2014), except the proteinase K incubation was conducted at 50°C for 60 min, followed by 10 min at 95°C to deactivate the proteinase K. Subsequent diagnostic PCR used Arsenophonus‐specific primers as described by Brady et al. (2014) and products were visualized on a 1% agarose gel stained with GelRed (Biotium). To avoid false negatives, we repeated the PCR for specimens that initially tested negative and also tested extraction quality via PCR that amplified a segment of the aphid COI gene (Brady et al., 2014). Specimens that were negative for COI were excluded from dataset, yielding a mean ± SE of 17.9 ± 0.3 aphids per population per time point.

Statistical analyses were carried out using JMP® Pro 12 (SAS Institute Inc., Cary, NC). Variation in the proportion of aphids infected with Arsenophonus, out of all successful extractions, was compared across sampling dates (week 0–16) and host plants (fava, black locust) using a generalized linear model. Each sampled aphid was treated as a separate data point within each population (jars 1–12). We used a model with a binomial distribution, a logit link function and firth adjusted maximum likelihood, with sampling date and host plant included as fixed effects and population as a random effect.

RESULTS AND DISCUSSION

Over the course of the experiment, frequency of Arsenophonus infection in cowpea aphid populations diverged depending on host plant. Over 16 weeks, prevalence of Arsenophonus‐infected aphids increased from 50% to 89 ± 7% in populations feeding on black locust, but dropped to 23 ± 15% in populations feeding on fava (Figure 1). Individual populations varied widely in infection trajectory, with most, but not all, black locust populations approaching fixation of Arsenophonus infection by the end of the experiment and most fava populations exhibiting declines in frequency of Arsenophonus‐infected aphids (Table S1). In two fava populations, Arsenophonus‐infected individuals appeared to be lost from the population entirely, whereas one population neared fixation for Arsenophonus infection, presumably through the vagaries of chance and the tight bottlenecks the populations underwent every two weeks. Nevertheless, proportional infection rates of black locust populations were significantly higher than fava populations from 8 weeks onward and over 3 × higher on black locust than fava by the last week of the experiment (Host Plant: χ² = 44.9, df = 1, p < .001, Week: χ² = 4.1, df = 1, p = .042, Plant × Week: χ² = 18.2, df = 1, p < .001; Figure 1).

Our findings are not unexpected, given that previous research has shown Arsenophonus provides a net fitness benefit for A. craccivora on black locust and a net fitness cost on fava (Wagner et al., 2015). However, it is useful to demonstrate that these fitness differentials clearly translate into shifts in population demographics across host plants over a short time span. The 16‐week experimental duration corresponds to approximately 11 aphid generations (Rani & Remamony, 1998), and the bottlenecks imposed by our experimental protocol were not unrealistic for ephemeral aphid populations that are regularly decimated by natural enemies before being re‐initiated by a few surviving colonizers (J. White, personel observations). Field populations of A. craccivora collected from locust versus fava routinely presented infection frequencies comparable to those generated by our laboratory experiment (Brady et al., 2014; Brady & White, 2013). The present results suggest that selection arising from differential fitness could be a sufficient explanation for observed field distributions, without invoking additional assortative mechanisms such as differential host plant preferences or movement patterns for infected versus uninfected aphids.

Mechanistically, the change in infection frequency is almost certainly attributable to differential survival and/or reproduction of infected versus uninfected aphids on the two plant species (Vorburger & Gouskov, 2011; Wagner et al., 2015). It is highly unlikely that the ‘spread’ of Arsenophonus in aphid populations on locust was due to horizontal transmission among aphids. Although some facultative symbionts of insects have been shown to be transmitted among hosts via shared plants (Caspi‐Fluger et al., 2012), the strain of Arsenophonus infecting the cowpea aphid has a substantially eroded genome that apparently limits bacterial motility and functionality (Genbank #CP038155‐6, J. White unpublished data). It is possible that reduced symbiont titre and reduced vertical transmission efficiency on fava may have contributed to lower overall rates of Arsenophonus infection in the fava populations (Desneux et al., 2018; Hansen, 2017; Serbus et al., 2015). However, such a vertical transmission effect is likely to be minor when compared to the large differential in population growth rates observed between Arsenophonus‐infected and uninfected aphids on locust versus fava (Wagner et al., 2015). Regardless of mechanism, the net observed effect was a dramatic increase in Arsenophonus infection frequency on black locust and decrease on fava, all within a time span equivalent to a growing season.

Our results are particularly interesting in light of the presumed newness of the association between the cowpea aphid and black locust. The cowpea aphid is thought to be European in origin (Blackman & Eastop, 2007), whereas black locust is a North American tree (Isely & Peabody, 1984). Thus, cowpea aphid likely did not come into contact with black locust until sometime within the last several hundred years. Both aphid and plant have become established on all continents except Antarctica, and both are considered invasive pests throughout portions of their range (cabi.org/isc/datasheet/6192, cabi.org/isc/datasheet/47698). Cowpea aphid does not undergo sexual reproduction throughout most of its range (Blackman & Eastop, 2007), and there is little genetic variation among either the Arsenophonus‐infected aphid clones, nor the Arsenophonus strains infecting them (Brady et al., 2014; White unpublished data), We therefore hypothesize that Arsenophonus infection predated contact between the cowpea aphid and black locust, and the infection acts as a preadaptation that favours the infected aphids whenever they encounter this host plant. In several other aphid taxa, formerly facultative symbionts have evolved to become co‐obligate symbionts along with the aphid's usual obligate symbiont Buchnera, compensating for loss of nutritional provisioning due to erosion of the Buchnera genome (Lamelas et al., 2011; Manzano‐Marín & Latorre, 2016; Manzano‐Marín, Simon, & Latorre, 2016; Meseguer et al., 2017). Here, it is possible that Arsenophonus acts instead to shift the nutritional repertoire of the cowpea aphid, allowing use of a host plant that would otherwise be inadequate. This symbiosis may represent a very early step in the evolution of host plant specialization as facilitated by microbial symbionts.

ACKNOWLEDGMENTS

We thank A. Dehnel, B. Griffis, M. Reams, M. Rogers and A. Styer for invaluable and interminable laboratory support. This work was supported by the National Institute of Food and Agriculture, U.S. Department of Agriculture, under Awards no. 2014‐67013‐21576 and Hatch no. 0224651.

PEER REVIEW

The peer review history for this article is available at https://publons.com/publon/10.1111/jeb.13697.

DATA AVAILABILITY STATEMENT

The dryad link to the article JEB 13697 https://doi:10.5061/dryad.2jm63xsmn.

REFERENCES