Detection and geographic distribution of seven facultative endosymbionts in two Rhopalosiphum aphid species

Abstract Study of the mutualistic associations between facultative symbionts and aphids are developed only in a few models. That survey on the situation and distribution of the symbionts in a certain area is helpful to obtain clues for the acquisition and spread of them as well as their roles played in host evolution. To understand the infection patterns of seven facultative symbionts (Serratia symbiotica, Hamiltonella defensa, Regiella insecticola, Rickettsia, Spiroplasma, Wolbachia, and Arsenophonus) in Rhopalosiphum padi (Linnaeus) and Rhopalosiphum maidis (Fitch), we collected 882 R. maidis samples (37 geographical populations) from China and 585 R. padi samples (32 geographical populations) from China and Europe. Results showed that both species were widely infected with various symbionts and totally 50.8% of R. maidis and 50.1% of R. padi were multi‐infected with targeted symbionts. However, very few Rhopalosiphum aphids were infected with S. symbiotica. The infection frequencies of some symbionts were related to the latitude of collecting sites, suggesting the importance of environmental factors in shaping the geographic distribution of facultative symbionts. Also, R. maidis and R. padi were infected with different H. defensa strains based on phylogenetic analysis which may be determined by host ×symbiont genotype interactions. According to our results, the ubiquitous symbionts may play important roles in the evolution of their host aphid and their impacts on adaptation of R. padi and R. maidis were discussed as well.

Most researchers agree with the idea that there exist costs for hosts to harbor the facultative symbionts (Oliver et al., 2008;Scarborough, Ferrari, & Godfray, 2005) and fitness reduction in aphids containing the facultative symbionts have been found in some cases (Laughton, Fan, & Gerardo, 2013;Vorburger & Gouskov, 2011) such as the infection of H. defensa could reduce aphid longevity (Vorburger & Gouskov, 2011). However, multiple infections of facultative symbionts are common in nature (Ferrari, West, Via, & Godfray, 2012;Oliver et al., 2014;Russell et al., 2013). The interactions between different symbionts coaffecting the host are complex. Some symbionts exhibit additive effects to the host: coinfection of S. symbiotica and H. defensa in A. pisum resulted in higher resistance to parasitism of Aphidius ervi (Oliver, Moran, & Hunter, 2006). However, inhibiting effects were found in another case: A. pisum coinfected with Rickettsiella viridis and H. defensa were more exposed to predation (Polin, Le Gallic, Simon, Tsuchida, & Outreman, 2015).
Several studies have assessed endosymbiont infections in R. padi to date. For instance, H. defensa-infected nymphs of R. padi collected from UK showed fivefold higher resistance to the parasitoid wasp Aphidius colemani (Viereck) than uninfected nymphs (Leybourne et al., 2018). De la Peña, Vandomme, and Frago (2014) found that R. padi collected from northwest of Belgium was only associated with S. symbiotica, whereas research showed five R. padi individuals collected from wheat harbored SMLS (Sitobion miscanthi L. type symbiont) but no Rickettsia (Li, Xiao, Xu, Murphy, & Huang, 2011) and an absence of targeted facultative symbionts was found in R. padi collected in Chile (Zepeda-Paulo, Ortiz-Martínez, Silva, & Lavandero, 2018). However, few research described the infection situation of symbionts in a particular region for R. maidis except one which reported that no facultative symbionts were detected from 25 R. maidis collected in Morocco (Fakhour et al., 2018). In this study, we conducted an extensive survey of seven facultative symbionts in hosts R. maidis and R. padi collected from the maize (Zea mays L.) in China and four European countries to assess geographic infection patterns of these facultative symbionts.

| Sample collection
We collected a total of 882 R. maidis from 37 geographical populations and 585 R. padi from 32 geographical populations. All aphids were collected from maize and the distance between each two samples was at least 10 m. All these collection sites (except four European populations) were selected to cover the comprehensive maize cultivating areas in China as much as possible and the collection work was done via random generation of co-ordinates within each site. More than 20 aphids per population were collected for most populations, although some populations may have fewer samples. All samples were identified by COI (mitochondrial cytochrome oxidase I) gene (Primers: LCO1490: 5′-GGTCAACAAA TCATAAAGATATTGG-3′; HCO2198: 5′-TAAACTTCAGGGTGACC AAAAAATCA-3′) (Folmer, Black, Hoeh, Lutz, & Vrijenhoek, 1994) and the information of aphid samples used in this study was listed in Tables A1 and A2 and the collecting locations were labeled on the maps (Figures 1 and 2). All collected aphids were preserved in absolute ethanol and stored at −20°C before molecular analysis.

| DNA preparation
Total DNA was extracted from single aphid using TEN buffer (10 mM Tris-HCl pH = 8, 2 mM EDTA pH = 8, 0.4 M NaCl), 20% SDS and 5 M NaCl solution according to the salting-out method (Sunnucks & Hales, 1996). 20-30 μl TE buffer was used to dissolve the DNA precipitate and the DNA quality was assessed with a Nanodrop 2000/2000C instrument. Then the DNA samples were kept at −20°C for further use.

| Symbionts detection
All 1,467 samples of the two aphid species were screened for the presence of seven facultative symbionts. Diagnostic PCR analysis was conducted using the specific primers listed in Table A3 to detect respective endosymbionts. PCR reactions of 20 μl volume for each sample were carried out under the following conditions: an initial denaturation at 94°C for 4 min, followed by 35 cycles of 94°C for 30 s, 55°C for 30 s, and 72°C for 30 s, and a final extension at 72°C for 5 min. DNA from aphids in laboratory of functional and evolutionary entomology (University of Liège) known to harbor a specific symbiont was used as a positive control and solution without DNA template was used as a negative control. The PCR products were detected by 1.5% agarose gel electrophoresis.  Table A1 F I G U R E 2 Collecting locations of Rhopalosiphum maidis in China. Numbers on the map correspond to locality numbers in Table A2 of other species were the source for multiple sequence alignment by DNAMAN and MEGA. The phylogenetic analyses were conducted using the Maximum likelihood methods with MEGA 4 software. Clade support was assessed with 1,000 bootstrap replicates (Stamatakis, Hoover, & Rougemont, 2008).

| Seven facultative symbionts were detected in R. padi and R. maidis
Both R. maidis (n = 882) and R. padi (n = 585) were frequently infected with various symbionts (Table 1). The infection frequencies for the seven targeted symbionts varied from 0.2% to 60.9% (Table 1) and only 20.2% of R. maidis and 17.1% of R. padi were not infected with any of the seven symbionts screened for (Table 2). Rickettsia ranked the highest frequency in the two aphid species (51.6% in R. maidis; 60.9% in R. padi) followed by R. insecticiola (34.1% in R. maidis; 40.7% in R. padi) and Spiroplasma (35.8% in R. maidis; 26.3% in R. padi), whereas both aphids had the lowest infection rate of S. symbiotica that only nine samples of R. maidis and one sample of R. padi were infected.
The trends of symbiont diversity per aphid were similar in both species (Table 2). Aphids infected with only one symbiont ranked the highest proportion of 29.0% in R. maidis and 32.8% in R. padi, respectively. Totally, around half of the tested aphids were infected with multiple symbionts (50.8% of R. maidis and 50.1% of R. padi).
The double infected samples occupied 25.1% in R. maidis and 30.4% in R. padi. In addition, two samples of R. maidis harbored as many as six facultative symbionts simultaneously and no R. padi was infected with six symbionts in a single aphid.

| Comparison of symbiont infection between R. maidis and R. padi from 19 common locations
The infection frequencies of each symbiont within 456 samples of R. maidis and 370 samples of R. padi from 19 common locations ( Figure 4) were compared using the method of Fisher's exact test (Table 3). Wolbachia between the two aphid species from 19 common locations.
The aphids infected with only one symbiont occupied the highest proportion from the 19 sites for both species ( Figure 5). However, the proportion of R. padi infected with single symbiont (37.6%) was significantly higher than that of R. maidis (30.0%) (p = 0.026). Significant higher proportions of R. maidis harbored three (20.0%) (p = 0.001) and four (6.4%) (p = 0.033) symbionts per aphid than that of R. padi  Table A1 and numbers on figure (b) correspond to locality numbers in Table A2 (11.1% and 3.0%). No significant difference (p > 0.05) was observed between R. maidis and R. padi of the aphid free of detected symbionts or infected with two, five and six kinds of the symbionts per aphid.

| Symbiont infection difference between China and Europe of R. padi
The infection frequencies of each symbiont in R. padi between 516 samples from China and 69 samples from four European countries were compared using the method of Fisher's exact test (Table 4).     (Russell et al., 2013). The infection frequencies of detected symbionts in this study ranged from 0.2% to 60.9%, these differences may result from the benefit-cost balance associated with harboring symbionts (Simon et al., 2007;Vorburger, Ganesanandamoorthy, & Kwiatkowski, 2013). Furthermore, non-selective factors such as transmission rates, migration, and drift may also affect the frequency and distribution of the symbionts (Oliver et al., 2014).

| Phylogenetic relationships
Interestingly, both R. maidis and R. padi were rarely infected with S. symbiotica (Table 1), whereas this bacterium was frequently detected in A. pisum (Sepúlveda, Zepeda-Paulo, Ramírez, Lavandero, & Figueroa, 2017;Tsuchida, Koga, Shibao, Matsumoto, & Fukatsu, 2002) and Aphis craccivora , which supports the result that symbiont combinations are mainly host specific (Fakhour et al., 2018).  Notes. These are the results of the statistical analysis which was carried out. a Means there is significant difference of the symbiont frequencies between two aphid groups. The group with higher average frequency is listed in the front.
Both R. maidis and R. padi were frequently infected with Rickettsia and R. insecticola, whereas previous study demonstrated that A. pisum both from pea and alfalfa were rarely infected with R. insecticola (Sepúlveda et al., 2017) and both symbionts showed a low frequency in A. craccivora from several host plants . In addition, European samples exhibited significantly higher frequencies of H. defensa than Chinese ones although Henry et al. (2015) found R. padi collected from UK harbored none symbionts of R. insecticola, H. defensa as well as S. symbiotica. Furthermore, R. padi collected from Western Europe were free-infected with four targeted facultative symbionts (Desneux et al., 2018) whereas in other cases, European R. padi lines were found infections with S. symiotica (De la Peña et al., 2014) and H. defensa (Leybourne et al., 2018). Also, research showed that Spiroplasma in A. pisum was rarely coinfected with other symbionts (Rock et al., 2017) whereas in our study, this bacterium was relative prevalent in both R. maidis and R. padi commonly coexisted with other symbionts. Moreover, infection frequencies of symbionts can also differ among host plants species. For instance, H. defensa was exclusively detected in A. craccivora collected from alfalfa  and there existed great diversity for the symbionts like R. insecticola in A. pisum collected from different host plants (Russell et al., 2013).
It is widely accepted that infection frequency and retention of an endosymbiont in insect are determined mainly by three aspects: first, the fidelity of maternal transmission (Luan et al., 2016(Luan et al., , 2018; F I G U R E 7 Maximum likelihood phylogenetic analysis inferred from Hamiltonella defensa 16S rDNA gene sequences. A bootstrap analysis was carried out and the robustness of each cluster was verified with 1,000 replicates. Values at the cluster branches indicate the results of the bootstrap analysis. Sequences are represented by the names of their host species. The GenBank numbers of the reference sequences are represented in Table A4 Uroleucon ambrosiae   (Fukatsu, Nikoh, Kawai, & Koga, 2000;Fukatsu, Tsuchida, Nikoh, & Koga, 2001 (Tsuchida et al., 2002). For example, the frequency of S. symbiotica in A. pisum increased in twothirds with increasing seasonal temperature in California (Montllor et al., 2002). Moreover, frequencies of symbionts with protective functions may also shift according to the changing of biotic factors such as parasitoid pressures (Smith et al., 2015).

| Geographic distribution of facultative symbionts
Wolbachia has been detected in R. maidis and R. padi with low frequencies of 2.8% and 3.3%, respectively. Also, this bacterium was distributed in northern China (above Henan province) and absent in R. padi collected from Europe, however, it has been found in other aphids from southern Europe (Greece, Portugal, Spain) (Gómez-Valero et al., 2004), Iran and Israel (Augustinos et al., 2011), China (Wang, Su, Wen, Jiang, & Qiao, 2014), USA (Russell et al., 2013),  (Fakhour et al., 2018) and China exhibits diverse ambient conditions from south to north of the latitude ranging from 4°N to 53°N which may affect the symbiont frequency but need further study to verify.

| CON CLUS ION
To conclude, both R. maidis and R. padi presented wide symbiotic relationship with the detected symbionts especially R. insecticola and Rickettsia, whereas these two Rhopalosiphum species were rarely infected with S. symbiotica. We hypothesize that the low infection frequency of S. symbiotica may be related to the environmental temperature of the collecting regions since S. symbiotica has been demonstrated to confer heat tolerance in aphid (Chen, Montllor, & Purcell, 2000;Montllor et al., 2002;Russell & Moran, 2006) which could be tested in the future. Multiple infections were common in these two aphid species, however, single or dou- Sciences, for comments and suggestions on this manuscript.

CO N FLI C T O F I NTE R E S T S
The authors declare that they have no competing interests.

E TH I C S S TATEM ENT
None required.

DATA ACCE SS I B I LIT Y
All data are available in the results section of this paper apart from the three fragments of the 16S rDNA sequence of H. defensa which were deposited at www.ncbi.nlm.nih.gov/genbank/ with accession numbers of KY550361, KY550362, and KY550363.