Many insects harbor various types of maternally inherited microbial symbionts (Baumann, 2005). A classic model for heritable symbiosis is the association of aphid and its primary (P-) symbiont Buchnera aphidicola, many aphid species first colonized an aphid ancestor 150 million years ago and which persists in almost all of the 5,000-aphid species (Baumann et al., 1995; Clark et al., 2000). B. aphidicola plays a prominent role in insect nutritional ecology by providing essential nutrients that are not obtained in sufficient amounts from a restricted diet of plant phloem (Baumann et al., 1995). Previous studies have shown that in addition to B. aphidicola, a wide range of heritable secondary (S-) symbionts are found in aphids (Buchner, 1965; Oliver et al., 2006; Vorburger and Gouskov, 2011). In contrast to B. aphidicola, which is strictly housed in bacteriocytes, S-symbionts are scattered in sheath cells and aphid hemolymph. They are not thought to be essential to host survival (Douglas, 1989). However, recent studies have revealed the important role of S-symbionts. These symbionts may confer conditional adaptive advantages to their host, such as protecting the insect host against pathogens and natural enemies (Piel, 2002; Oliver et al., 2003, 2005; Guay et al., 2009), ameliorating the detrimental effects of heat (Douglas, 1989, 1998; Ohtaka and Ishikawa, 1991; Montllor et al., 2002), enhancing insecticide resistance, aiding in adaptation to the plant (Leonardo and Muiru, 2003; Tsuchida et al., 2004; Guidolin and Cônsoli, 2017), and mediating insect host metabolism and biosynthesis (Akman and Douglas, 2009; Benoit et al., 2017). With the continuous development of molecular biology and omics technologies, greater knowledge of the evolutionary relationships between symbionts and host insects has been revealed (Degnan et al., 2009, 2010; Manzanomarín and Latorre, 2014; Cassone et al., 2015).

There are many S-symbionts that have been identified in aphids, including Serratia symbiotica, Hamiltonella defensa, Regiella insecticola, Rickettsia, Wolbachia, Spiroplasma, Arsenophonus (Tsuchida et al., 2002; Oliver et al., 2010), and SMLS (Sitobion miscanthi L type symbiont) (Li et al., 2016). H. defensa is a well-studied S-symbiont of insects and a gammaproteobacterium informally known as the pea aphid Bemisia-like symbiont or T-type (Darby et al., 2001), which is widely, but not universally, distributed in natural populations of aphid, especially in the grain aphid, Sitobion miscanthi Takahashi.

Sitobion miscanthi, a dominant grain aphid species in China, is a major wheat pest that frequently harbors S-symbionts. H. defensa has been comprehensively studied in Acyrthosiphon pisum (Nyabuga et al., 2010; Łukasik et al., 2013; Mclean and Godfray, 2015), but most of these studies focused on the role of H. defensa in helping the insect during pathogen or parasite invasion via the presence of a lysogenic lambdoid bacteriophage designated APSE (Moran et al., 2005a; Degnan and Moran, 2008). Hamiltonella-induced tolerance to high temperature has also been reported (Russell and Moran, 2006). Sporadic reports have noted the positive effect of H. defensa on aphid fecundity (Łukasik et al., 2013), but there is still relatively little information available on the ecological adaptation of H. defensa to the aphid host, especially in the wheat aphid S. miscanthi.

In general, artificial infection and antibiotic elimination studies offer a powerful opportunity to study the influence of symbionts on the host. Thus, in the case of presented study system, H. defensa was transferred by microinjection and selectively reduced by antibiotic treatment to provide a host line for the experimental assays. Different geographical populations of aphids naturally harbor various symbionts and provide a source of symbiont diversity for aphid-symbiont interaction research (Tsuchida et al., 2002; Zhao et al., 2016). Aphids with a single strain of S-symbiont are scarce in nature and hence the collection and identification of secondary symbionts in different geographical isofemale strains of S. miscanthi were the foundation for the target symbiont screening.

This study was conducted to assess the diversity of S-symbionts in 17 different geographical populations of S. miscanthi in China, from which the Hamiltonella-infected and Hamiltonella-free strains were obtained. A new Hamiltonella-re-infected strain and a Hamiltonella-significantly decreased strain with the identical genetic backgrounds was generated to evaluate the impact of Hamiltonella on host ecological fitness. Furthermore, the relative abundance of the B. aphidicola in each development stage was measured in Hamiltonella-free and Hamiltonella-re-infected strains. Our study clearly identifies the diversity of S-symbionts in S. miscanthi and reveals the effect of H. defensa on host aphid ecology adaptation and the correlation with B. aphidicola.

Research on the role of S-symbionts in the aphid S. miscanthi, one of the most important pests of gramineous crops in China, is still incomplete. In order to obtain target S-symbionts and understand their functions, stable geographical populations with different symbiotic infections are needed. We screened the diversity of S-symbiont infections from established laboratory populations of 17 geographical clones. A remarkably variable composition of S-symbiont infection was evident in S. miscanthi. Interestingly, the infection of S-symbionts among the different wheat-producing provinces differed somewhat, but samples from different places in the same province showed a high degree of consistency. It is noteworthy that H. defensa has only been detected in Yunnan province with an average elevation of 2,000m, and S. symbiotic has only been detected in the Tibet autonomous region with an average elevation of 4,000m. Interestingly, both areas belong to the plateau region of southwest China that features complex geological and topographical conditions. Our geographic sampling data suggest elevation and topography may contribute to the symbiotic diversity. As the mechanism of symbiont distribution in S. miscanthi are unknown, however, biotic and/or abiotic factors could produce the regional differentiation, as proposed by Oliver et al. (2014) and Tsuchida et al. (2002). Indeed, other factors that are not mediated by natural selection such as genetic drift, founder effects (Russell et al., 2013) and host plant (Guidolin and Cônsoli, 2017) can also affect the diversity of S-symbionts between geographical populations of S. miscanthi in different locations in China. Although we did not investigate the infection rate of symbionts in natural populations, our results, combined with information reported in the literature (Sepúlveda et al., 2016; Zytynska and Weisser, 2016), show that multiple S-symbionts coexist within local populations of S. miscanthi in different geographical strains.

Hamiltonella defensa is a well-known S-symbiont that confers protection to the aphid against parasitoid wasps (Oliver et al., 2005; Degnan and Moran, 2008; Łukasik et al., 2013; Vorburger and Rouchet, 2016), with a high infection rate of 31.3% (5/16) in US strains (Sandström et al., 2001) and 39.5% (15/38) in UK strains of A. pisum (Darby et al., 2001). However, it was not detected (0/119) in Japanese strains of A. pisum (Tsuchida et al., 2002; Rothacher et al., 2016). In our study, H. defensa was also not widely distributed in S. miscanthi from the screened area. It was only present in four clones collected in Yunnan Province. Even though we did not obtain a single Hamiltonella-infected S. miscanthi strain, two interesting strains (YX and DZ) were collected from Yuxi and Dezhou wheat fields in China. Both strains had the same S-symbiont infection profiles, except for the presence of H. defensa (YX: Hamiltonella-infected and DZ: Hamiltonella-free). By artificial infection, Hamiltonella-free strain (DZ) and Hamiltonella-re-infected strain (DZ-H) with the same genetic background were generated. After antibiotic treatment, a Hamiltonella-decreased strain (DZ-HT) was also obtained.

Fitness measurements revealed that the Hamiltonella-re-infected treatment increased the fitness of S. miscanthi, as evidenced by greater numbers of offspring, increased weight, and higher wingless aphid rate. Meanwhile, decreasing the abundance of H. defensa with antibiotics resulted in lowered adult aphid weight and percentage of wingless aphids. Similar results have been reported for the whitefly (Su et al., 2013), while Oliver et al. (2008) reported that H. defensa shortened the mean generation time of A. pisum. However, in S. miscanthi, research on the influence of H. defensa in improving aphid fecundity is sporadic (Łukasik et al., 2013) and other S-symbiont infections were not identified. In our research, we used four fitness indices (number of offspring, weight of adult aphids, percentage of wingless aphids, and longevity) to evaluate the effect of H. defensa on the ecological fitness of S. miscanthi strains. Surprisingly, the difference in longevity found between DZ and DZ-H strains were not significant, but was only slightly decreased. A close relationship with symbionts can be costly for the host (Polin et al., 2014), with these direct costs caused by a trade-off between allocating resources to symbiosis and other adaptive functions. A previous study found that aphids exhibited many behavioral defenses against enemies and these behaviors had some associated costs, leading to a reduction in aphid reproduction (Dion et al., 2011). We hypothesize that the benefit of H. defensa infection in terms of number of offspring, weight, and wingless aphid rate represents a trade-off against the cost of longevity, with an overall benefit of H. defensa infection.

Buchnera aphidicola, a primary symbiont in almost all aphids, appears to benefit their hosts primarily through the provision of nutrients that are limiting for growth and reproduction (Clark et al., 2000; Oliver et al., 2010). In this study, the relative abundance of B. aphidicola in the DZ-H strain was significantly higher than the DZ strain in all but the first developmental stage. This phenomenon implies a linkage of B. aphidicola with H. defensa, and the difference abundance of B. aphidicola in aphid offspring might not be caused by microinjection. The increased in B. aphidicola could result from increased nutritional requirements of the whole system. B. aphidicola may be “feeding” H. defensa since the latter does not seem to be able to manufacture many essential amino acids (Degnan et al., 2009). Therefore, we speculate that infection with H. defensa may indirectly improve the fitness of S. miscanthi by stimulating the abundance of B. aphidicola. In addition, based on the results of the current study and combined with the viewpoint that H. defensa inhabits secondary bacteriocytes intercalated between primary bacteriocytes that harbor B. aphidicola (Moran et al., 2005b), it can be hypothesized that B. aphidicola and H. defensa may perform complementary functions in the aphid host and facilitate biological interactions between them. More interestingly, several reports using high-throughput sequencing methods have suggested that S. symbiotica and Wolbachia, which are facultative symbionts in many aphid species, have evolved to form a deep and co-obligate association with B. aphidicola in different aphids (Manzanomarín and Latorre, 2014; De et al., 2015). Although H. defensa is not necessary for S. miscanthi, it affects the host aphid ecological adaptation to some extent. Further studies employing metagenomics and proteomics techniques will be needed to reveal the underlying mechanism of how H. defensa improves the fitness of S. miscanthi, as well as to examine the interactions between H. defensa and B. aphidicola and the interactions among the symbionts themselves.

In summary, we established a feasible method to obtain aphid strains with and without a target symbiont under the same other S-symbiont infection profiles and genetic background by screening different geographical populations and microinjection. In addition, infection with H. defensa increased the fitness of S. miscanthi and enhanced the relative abundance of B. aphidicola. Our results establish a basis for functional studies by revealing the effect of H. defensa on host aphid ecology adaptation and the correlation with B. aphidicola. A better understanding of the insect and endosymbionts interaction will supply information on the potential value of manipulating microbiota to control insect pests. The diversity and ecological function of symbionts may provide important guidance to field arrangements of wheat varieties in different wheat production areas.

LQ, FJ, CJ, and WM conceived and designed the experiments. LQ and SJ performed the experiments. LQ analyzed the data. LQ and CJ wrote the paper. All of the authors read and approved the final version of the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.