Effect of the Secondary Symbiont Hamiltonella defensa on Fitness and Relative Abundance of Buchnera aphidicola of Wheat Aphid, Sitobion miscanthi

Bacterial symbionts associated with insects are often involved in host development and ecological fitness. In aphids, the role of these symbionts is variable and not fully understood across different host species. Here, we investigated the symbiont diversity of the grain aphid, Sitobion miscanthi (Takahashi), from 17 different geographical areas. Of these, two strains with the same symbiont profile, except for the presence of Hamiltonella defensa, were selected using PCR. The Hamiltonella-infected strain, YX, was collected from a Yuxi wheat field in Yunnan Province, China. The Hamiltonella-free strain, DZ, was collected from a Dezhou wheat field in Shandong Province, China. Using artificial infection with H. defensa and antibiotic treatment, a Hamiltonella-re-infected strain (DZ-H) and Hamiltonella-significantly decreased strain (DZ-HT) were established and compared to the Hamiltonella-free DZ strain in terms of ecological fitness. Infection with the DZ-H strain increased the fitness of S. miscanthi, which led to increases in adult weight, percent of wingless individuals, and number of offspring. Meanwhile, decreased abundance of H. defensa (DZ-HT strain) resulted in a lower adult weight and wingless aphid rate compared to the DZ-H strain. However, the indices of longevity in both the DZ-H and DZ-HT strains decreased slightly, but were not significantly different, compared to the DZ strain. Furthermore, quantitative PCR showed that the relative abundance of the primary symbiont Buchnera aphidicola in the DZ-H strain was significantly higher than in the DZ strain in all but the first developmental stage. These results indicate that H. defensa may indirectly improve the fitness of S. miscanthi by stimulating the proliferation of B. aphidicola.


INTRODUCTION
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., 2003Oliver et al., , 2005Guay et al., 2009), ameliorating the detrimental effects of heat (Douglas, 1989(Douglas, , 1998Ohtaka 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., 2009Manzanomarín and Latorre, 2014;Cassone et al., 2015).
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 Ssymbionts in 17 different geographical populations of S. miscanthi in China, from which the Hamiltonella-infected and Hamiltonella-free strains were obtained. A new Hamiltonellare-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.

Aphid Rearing
The strains of S. miscanthi used in this study are shown in Figure 1 and Table 1. These samples were collected from 17 locations covering a wide range of the main wheat-producing areas in mainland China. All of these isofemale strains were established from different collections, feeding separately on aphid-susceptible wheat seedlings (Triticum aestivum L) in the culture room at 20 ± 1 • C with a 75% relative humidity and a light: dark photoperiod of 16:8 h. After 10 generations, the aphids were used for the following experiments.

Aphid DNA Extraction, Sequencing, and Secondary Symbiont Identification
Total aphid DNA was extracted from a single adult aphid with the DNeasy Blood & Tissue Kit (Qiagen, Valencia, CA) following the manufacturer's protocol. DNA was extracted from a total of 20 replicates for each geographical strain. DNA quality was evaluated using a NanoDrop-1000 spectrophotometer (Nano-Drop Technologies, Wilmington, DE). These samples were then screened for the seven known S-symbionts of pea aphids, S. symbiotica, H. defensa, R. insecticola, Rickettsia, Wolbachia, Spiroplasma, and Arsenophonus (Tsuchida et al., 2002;Oliver et al., 2010) with PCR using the specific symbiont primers listed in Table 2. Cycling conditions were 94 • C for 4 min, followed by 35 cycles at 94 • C for 30 s, 60 • C for 45 s, 72 • C for 1 min, and 4 • C for the final elongation. The reaction products were analyzed with FIGURE 1 | Sampling sites for 17 geographical populations collection to investigate the diversity of secondary symbionts in different geographical S. miscanthi strains in China Numbers on the map correspond to locality numbers in Table 1. a model 3500 ABI PRISM DNA sequencer (Perkin-Elmer, New York, NY). The nucleotide sequence of the H. defensa 16S rRNA partial gene from S. miscanthi described in this paper has been deposited in GenBank under accession number MG721025.

Artificial Infection of H. defensa by Hemolymph Injection
The injection of H. defensa followed a previously described method (Koga et al., 2003). Briefly, 0.5 µl of hemolymph was extracted from the YX strain and diluted with 0.5 µl of 0.01 M PBS. A 1 µl dilution was injected into the body of third-instar nymphs of the DZ strains. We detected the infection of H. defensa in the newborn nymphs from the offspring of the treated aphids 2 weeks after injection. Presence of H. defensa was confirmed again after three generations and before the start of the fitness experiment.

H. defensa Elimination by Antibiotic Treatment
To verify the fitness effect of H. defensa on S. miscanthi, a Hamiltonella-significantly decreased strain was established via antibiotic treatment. With a litter modified by the previous treatment, a mixture of the antibiotics ampicillin and gentamycin (Dykstra et al., 2014), each at 100 µg/ml, was added to a 20% sucrose solution as the artificial diet, with the control treatment containing antibiotic-free sucrose solution, for two days. The aphids that fed on the antibiotic artificial solution are referred to as Hamiltonella-significantly decreased aphids. Following antibiotic treatment, all of the adult aphids were placed on wheat seedlings to produce new nymphs. After rearing for 10 generations, the offspring of aphids were screened for abundance of B. aphidicola and S-symbionts by quantitative PCR. The feeding apparatus was prepared according to Chen et al. (2000) with 250 µl of solution diet sandwiched between two layers of parafilm membrane and stretched to a glass tube of 21-mm diameter under sterile conditions.

Fitness Measurement
The nymphs of DZ (collected from Dezhou wheat filed), DZ-H (a new strain created by DZ strain with H. defensa artificial infection) and DZ-HT (a new strain created by DZ-H strain with antibiotic treatment) were collected over 24 h from adult aphids. Thirty nymphs from strains were selected at random and individually placed in petri dishes containing wheat seedlings whose roots were inserted into water in 1.5 ml tubes and kept at 20 • C under a long-day (16 h) light cycle. These nymphs were allowed to develop to adults in the petri dishes, and the fitness indices wingless aphid rate, total number of offspring and longevity were checked daily until all nymphs completed their whole lifecycle. We measured the weight of 30 newly emerged adults collected from each strain and performed 6 replications. To eliminate any adverse effects of the injection or antibiotics on aphids, the DZ-H and DZ-HT strains were reared for 10 generations before performing the fitness experiment.  Quantitative PCR (QPCR) was carried out in 20-µl volumes containing 10 µl of SYBR Green PCR Mix (Takara Bio, Shiga, Japan), 0.4 µl of 50x ROX Dye II, 1 µl of each primer, and 1 µl of DNA. Cycling conditions were 95 • C for 15 min, then 40 cycles of 15 s at 95 • C, 30 s at the annealing temperature at 60 • C, and 30 s at 72 • C. The relative abundance of aphid B. aphidicola was normalized to the aphid housekeeping gene ß-actin and calculated using the comparative C t method 2 − ct method (Vandesompele et al., 2002).

Statistical Analyses
The treatment effects were analyzed by nonparametric tests. Kruskal-Wallis was used for multiple comparisons and Mann-Whitney was used for two sample comparisons at α = 0.05. All of the data analyses were performed using the IBM SPSS statistics Version 21 (ver. 21, SPSS Inc., Chicago, IL, USA) software.

Diversity of Secondary Symbiont Infection in Different Geographical Isofemale Strains of S. miscanthi
The diversity of S-symbiont infections of S. miscanthi was investigated from 17 isofemale strains by specific PCR detection. S. symbiotica, H. defensa, R. insecticola, Rickettsia, Arsenophonus, and Spiroplasma exhibited partial infections in different geographical strains. However, Wolbachia was not detected from any of the Chinese strains of S. miscanthi (Table 1). When an S-symbiont was detected in a population, infection rate was close to 100%. In these isofemale strains, no single S-symbiont infected strain was detected.

H. defensa Artificial Reinfection and Antibiotic Elimination
Because no H. defensa single infected strains were found in different isofemale strains and given the negative impact of using multiple antibiotics to remove different symbionts, we performed the following experiments using the two strains with the same symbionts background, except H. defensa. Twenty third-instar nymphs of the Hamiltonella-free DZ strain of S. miscanthi were injected with hemolymph obtained from the YX strain of S. miscanthi, and the Hamiltonella-re-infected strain DZ-H with the identical background of DZ strain was stably established. To evaluate whether H. defensa had been transferred into the aphids, DNA samples obtained from sets of 10 firstinstar nymphs were subjected to PCR detection based on the H. defensa 16S rRNA gene with two replications (data not shown). The sequence of H. defensa in our study, 1380-bp in length, exhibited 99.8% similarity to the sequence of H. defensa from A. pisum. After antibiotic treatment, DZ-HT strains were reared for 10 generations and, then, we estimated the relative abundance of symbionts in DZ-HT strains to ensure that the antibiotic did not affect the other symbionts, except H. defensa. Compared to the DZ-H strain, the abundance of H. defensa diminished by approximately 4.15-fold (P < 0.05; Figure 2A), while B. aphidicola, R. insecticola, and Spiroplasma showed almost no change (Figures 2B-D). These results showed that a moderate concentration of mixed antibiotics could specifically decrease targeted symbionts without affecting the other symbionts.

Role of H. defensa on Host Aphid Fitness
Aphid demographic parameters, including adult aphid weight, percentage of wingless aphids, total number of offspring, and longevity were compared among the DZ-H (Hamiltonellare-infected), DZ (Hamiltonella-free), and DZ-HT strains (Hamiltonella-decreased). The DZ-H strain had a significant positive effect on partial fitness indices compared to the DZ and DZ-HT strains (Figure 3). The total number of offspring per five adults of the DZ-H strain was 112.3, which was significantly higher than the DZ strain (89.3) (t = 3.326, P < 0.05), but it was not significantly different from the DZ-HT strain (Figure 3A). For the DZ-H strain, the adult aphid weight was 1.93 g, which was significantly higher than the DZ (1.68 g) (t = 5.677, P < 0.05) and DZ-HT strains (1.808 g) (t = 2.814, P < 0.05) (Figure 3B). The wingless aphid rate in the DZ-H strain was 67%, which was significantly higher than in the DZ (42%) (t = 3.727, P < 0.05) and DZ-HT strains (46%) (t = 3.316, P < 0.05) ( Figure 3C). However, the longevity of the DZ-H strain was 22.9 days, which was not significantly different from the DZ (26.4 days) (t = 1.705, P = 0.09) and DZ-HT strains (22.4 d) (t = 0.278, P = 0.79) (Figure 3D).

Relative Abundance Difference in B. aphidicola Between DZ-H and DZ Strains
The relative abundance of B. aphidicola was measured by quantitative PCR. The abundance of B. aphidicola in the DZ-H strains was significantly higher than in the DZ strain during the entire development stage, except for the first instar nymphs. The comparison of the abundance ratios of B. aphidicola between DZ-H and DZ in different developmental stages of S. miscanthi are listed as follows: 1.139-fold (t = 0.519, P = 0.631), 3.415fold (t =10.645, P < 0.05), 4.712-fold (t = 31.191, P < 0.05), 6.381-fold (t = 50.261, P < 0.05) for the first-fourth instar nymph stage, respectively (Figure 4A), and 3.281-fold (t = 19.345, P < 0.05) and 9.671-fold (t = 21.363, P < 0.05) in winged and wingless adult aphids, respectively ( Figure 4B). These results further confirmed that the relative abundance of B. aphidicola was affected by H. defensa infection.

DISCUSSION
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 FIGURE 2 | Relative symbiont abundance change with antibiotic treatment designed to specifically eliminate H. defensa QPCR analysis of DNA extracted from antibiotic-treated aphids and controls. (A) The abundance of H. defensa significantly declined in DZ-HT strains treated with antibiotics compared to the control DZ-H strains (P < 0.05). Antibiotic treatment had no effect on the abundance of B. aphidicola, R. insecticola, and Spiroplasma (B-D). The asterisk indicates significant differences based on the Mann-Whitney U-test for two sample comparison at P < 0.05 and ns indicates no significant difference. been detected in Yunnan province with an average elevation of 2,000 m, and S. symbiotic has only been detected in the Tibet autonomous region with an average elevation of 4,000 m. 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-reinfected 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.

AUTHOR CONTRIBUTIONS
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.