Coexistence of Three Dominant Bacterial Symbionts in a Social Aphid and Implications for Ecological Adaptation

Simple Summary Most insects are associated with a variety of symbionts that play a crucial role in insect life history. Symbiosis of aphids and their symbionts is a good model system to study insect–symbiont interactions. Pseudoregma bambucicola is a typical social aphid that lives parthenogenetically throughout the year on bamboos in subtropical areas, and it is the only aphid that exclusively feeds on the hard stalks of bamboo. In this study, we surveyed the symbiotic bacterial community associated with P. bambucicola. Our results showed that the diversity of P. bambucicola microbiome was low, but three symbionts, namely the primary endosymbiont Buchnera and two secondary symbionts (Pectobacterium and Wolbachia), were stable coexisting with a high infection rate. Combined with the biology of P. bambucicola, we speculate that Pectobacterium may help P. bambucicola feed on the stalks of bamboo, and Wolbachia may regulate the loss of sexual reproduction or has a nutritional role in P. bambucicola. These findings will advance our knowledge of the microbiomes of social aphids and set the foundation for further studies on the functional roles of P. bambucicola symbionts. Abstract Aphids are associated with an array of symbionts that have diverse ecological and evolutionary effects on their hosts. To date, symbiont communities of most aphid species are still poorly characterized, especially for the social aphids. In this study, high-throughput 16S rDNA amplicon sequencing was used to assess the bacterial communities of the social aphid Pseudoregma bambucicola, and the differences in bacterial diversity with respect to ant attendance and time series were also assessed. We found that the diversity of symbionts in P. bambucicola was low and three dominant symbionts (Buchnera, Pectobacterium and Wolbachia) were stably coexisting. Pectobacterium may help P. bambucicola feed on the hard bamboo stems, and genetic distance analysis suggests that the Pectobacterium in P. bambucicola may be a new symbiont species. Wolbachia may be associated with the transition of reproduction mode or has a nutritional role in P. bambucicola. Statistical tests on the diversity of bacterial communities in P. bambucicola suggest that aphid populations attended by ants usually have a significantly higher evenness than populations without ant attendance but there was no significant difference among aphid populations from different seasons.


Introduction
Associations between bacteria and insects are widespread in nature. Bacterial symbionts associated with insects may have diverse ecological and evolutionary effects on their hosts [1][2][3][4]. For examples, Wolbachia are widespread in insects as master manipulators of host biology [1], and Buchnera and Wigglesworthia provide essential nutrients to their host aphids and tsetses, respectively [5,6]. The microbiota of the insect gut often plays an important role in regulating the host's metabolism [7].
Phloem sap-feeding aphids (Hemiptera: Aphididae) feed on a wide variety of host plant species, they represent serious pests and act as vectors of phytopathogenic viruses probably the most specialized aphid species feeding on the hard stem of bamboo, while the other aphid species using bamboos as host plants usually feed on leaves or other softer parts [34]. We may therefore hypothesize that in P. bambucicola there should be a relation between specific symbiotic bacteria and utilization of the hard stem of bamboo. Although P. bambucicola can be a good choice to study ecological complexity and symbiotic association, most previous studies of this species have been conducted on field ecology [36] and behavior [32,35], but no study has specifically aimed to analyze the diversity of symbionts within this species.
The rapid development of high-throughput sequencing has made it easier to detect the entire bacterial communities of insects. In contrast, although the traditional PCR-based approach has been used to detect specific symbiotic bacteria within hosts, it may omit many other non-target cohabiting bacteria. In this study, the 16S rRNA gene amplicon sequencing was used to assess the entire symbiotic bacterial community of the social aphid P. bambucicola. We also examined whether the symbiotic bacterial composition can be stable along a time series (in different seasons) and whether it can be influenced by ant attendance or not.

Aphid Collection
Aphid samples from twenty-eight P. bambucicola populations on Bambusa bamboos in subtropical areas of southern China were collected from August 2015 to January 2018. Detailed sampling information is listed in Table 1. In order to examine temporal dynamics of the symbiotic bacterial community of P. bambucicola, the samples from Fujian province were assigned into four groups based on seasons of a year ('Time groups' in Table 1). These samples were also divided into two groups based on whether the collected aphid clones were attended by ants or not ('Ant groups' in Table 1). Each sample comprised of multiple individuals collected from one same aphid colony from a single bamboo stem. The number of aphids contained in each sample ranged from a few dozen to a few hundred, depending on the size of the clone. The specimens were kept in 95% ethanol and store at −20 • C after collection.

DNA Extraction
An apterous adult aphid from each sample was washed three times in ultrapure water. Total genomic DNA was extracted from whole individuals with the DNeasy Blood & Tissue kit (QIAGEN) according to the manufacturer's instructions. DNA extraction was carried out in an ultra-clean workbench to avoid contamination of environmental DNA. The bacterial universal primers 8F (5 -AGAGTTTGATCCTGGCTCAG-3 ) and 1492R (5 -GGTTACCTTGTTACGACTT-3 ) [37] were used to verify the success and quality of the DNA extractions. To assure the accuracy of results, sterile deionized water was used as a negative control. PCR amplifications were performed in 25 µL reactions containing 1 µL of DNA, 2.5µL 10× LA PCR buffer II (Mg 2+ plus), 0.5 µL dNTP mixture (2.5 mM each), 0.5 µL of each primer (10 µM), 0.5 µL of TaKaRa LA Taq (5 U/µL) (TaKaRa Bio Inc., Otsu, Japan), and 19.5 µL water. Amplification was performed in ProFlex TM Base (Applied Biosystems, Inc., Waltham, United States), using the following cycling conditions: 94 • C for 4 min, followed by 30 cycles of 30 s at 94 • C, 40 s at 65 • C, 90 s at 72 • C, and a final extension of 10 min at 72 • C. The PCR products were detected on 1% agarose gels, and the positive samples with a bright band of about 1500 bp were kept at −20 • C until 16S library preparation. The negative controls had no bands.

OTU Clustering and Taxonomic Assignment
According to the overlap relation between paired-end reads, the PE reads were merged into single, longer raw tags using FLASH v1.2.11 [39]. Raw tags were further quality trimmed to obtain clean tags using Trimmomatic v0.33 [40] ensuring > 20 quality scores on a sliding window of 50 bp. The chimera sequences were identified and removed using the UCHIME v8.1 [41]. Sequences with ≥ 97% similarity were assigned to the same operational taxonomic units (OTUs) using USEARCH v10.0 [42]. OTUs with a number of sequences < 0.005% of the total number of sequences were discard [43]. Representative sequences from each OTU were screened for further annotation. For each representative sequence, the Silva [44] was used with the RDP classifier v2.2 [45] to annotate taxonomic information. To further confirm the taxonomic assignment, the representative OTU sequences were compared to the sequences in GenBank using BLAST.
The results of taxonomic assignment of two dominant bacteria, Pectobacterium and Wolbachia, were further confirmed by reconstructing phylogenetic trees. The representative sequence of the most abundant OTU of Pectobacterium (OTU3) was extracted. Representative 16S rDNA sequences of Pectobacterium, Dickeya, and Erwinia were downloaded from GenBank. Sequences were aligned using MAFFT v 7.427 [46] with removal of sequences not in the same regions using MEGA 7.0 [47], and Escherichia coli (GenBank accession number, NR_024570.1) was chosen as the outgroup. A total of 10 Pectobacterium sequences, 11 Dickeya sequences, and 4 Erwinia sequences of 434 bp length were used to construct a maximum likelihood (ML) phylogenetic tree. In addition, the representative OTU sequence of Wolbachia (OTU5) was also extracted, and representative 16S sequences of Wolbachia were downloaded from GenBank. All sequences of Wolbachia were treated with the same steps as Pectobacterium, and Rickettsia endosymbiont (GenBank accession number, LN829697.2) was chosen as the outgroup. Finally, 29 Wolbachia sequences of 409 bp length were used to construct a ML phylogenetic tree. The phylogenetic trees of Pectobacterium and Wolbachia were reconducted by IQ-TREE v1.6.8 using the HKY+F+R2 model with 2000 SH-aLRT bootstrap replicates [48]. The pairwise genetic distances among those two sequence datasets were also calculated in MEGA 7.0 with the Kimura 2-parameter (K2P) model [49].

Microbiome Diversity
Species richness (Chao1 and ACE), diversity (Simpson and Shannon), and the coverage of library for each aphid sample were calculated in Mothur v.1.30 [50]. The Chao1 and ACE indices indicate species richness (number of microbial species). The Shannon and Simpson indices measure species diversity by taking into consideration species abundance and species evenness in the sample community. In the case of same species richness, the larger Shannon index and smaller Simpson index indicate the larger evenness of each species and the greater diversity of community [51]. These different indices for the 'Ant groups' and 'Time groups' were then compared, respectively, using one-way ANOVA with the IBM Statistical Package for the Social Sciences (SPSS) version 24.0 (Chicago, IL, USA).

Library Basic Statistics
Using Illumina's stringent quality control (QC), we obtained a total of 2,150,533 Clean tags. The mean number of Clean tags generated for each sample was 76,805 (SD = 19,897.854). After sequence filtering steps, we obtained 2,093,909 reads (Table S1). The assembly of paired sequences resulted in consensus sequences with an average length of 431 bp.

OTU Clustering and Taxonomic Assignment
High-quality reads were clustered using > 97% sequence similarity into 23 microbial OTUs (Table 2); the number of OTU in each sample is presented in Table S2. Table S3 shows the details of OTUs distribution in all samples, the raw sequences have been submitted to the NCBI Sequence Read Archive (accession number: PRJNA634633).
Of the total reads for 28 P. bambucicola samples, 99.94% OTUs were classified as Proteobacteria and most bacteria were from the Enterobacteriaceae (98.45% of the total reads). The obligate endosymbiont Buchnera was the most abundant taxon and was detected in all aphid samples (89.07% of the total reads) and was represented by two OTUs (OTU1, 56). Other taxa with relatively high abundance included Pectobacterium (6.56%, OTU3, 9, 104, 322, 427), Wolbachia (0.97%, OTU5), Serratia (1.72%, OTU2), and Arsenophonus (1.09%, OTU4). It is worth noting that Pectobacterium was detected in 26 of 28 aphid samples, and its relative abundance fluctuated between 0.73% and 14.41% in different samples. A similar situation occurred for Wolbachia which was detected in 23 of 28 aphid samples, with relative abundance fluctuating between 0.14% and 6.33%. Whilst another well-known insect symbiont, Serratia, was detected in 13 samples, its relative abundance was ≤0.01% (reads number < 5) in 11 of the 13 samples. Arsenophonus sequences were represented in six aphid samples, but four of which had a relative abundance ≤ 0.01% (reads number < 12). Table 3 shows the detailed information of the top five most abundant bacteria. The relative abundance of the top five genera in 28 samples are illustrated in Figure 1a, while the relative abundance of three dominant symbionts in all samples are shown in Figure 1b.

OTU ID Taxon Annotation
Bacteroides Table 3. Reads and relative abundance of symbiotic bacteria in each sample, emphasizing the major five bacteria.  In the Pectobacterium phylogenetic tree (Figure 2a), OTU3 and other Pectobacterium species clustered together, while Dickeya and Erwinia formed separate clades, respectively. The genetic distances between OTU3 and all other Pectobacterium sequences ranged from In the Pectobacterium phylogenetic tree (Figure 2a), OTU3 and other Pectobacterium species clustered together, while Dickeya and Erwinia formed separate clades, respectively. The genetic distances between OTU3 and all other Pectobacterium sequences ranged from 2.1% to 3.9%, which indicated that the Pectobacterium symbiont in P. bambucicola may represent a new bacterial species of this genus. As the Wolbachia phylogenetic tree (Figure 2b) shows, with a low support OTU5 and several other Wolbachia strains from aphids clustered together, and Wolbachia strains of other arthropods and nematodes separated from this clade. The genetic distances between OTU5 and all other Wolbachia sequences ranged from 0 to 3.8%, while the V3-V4 region of 16s rDNA sequence of Wolbachia from P. bambucicola was identical with three Wolbachia sequences from other aphids, indicating that the Wolbachia strain in P. bambucicola is a common one among aphid species.

Microbiome Diversity for the Ant and Time Groups
The coverage of library for each aphid sample was good, being more than 0.99 for all samples. We calculated the richness (Chao1 and ACE) and diversity (Simpson and Shannon) indices, which are shown in the Table S2. The mean Shannon diversity index was low, suggesting a low bacterial diversity in P. bambucicola and that each sample was dominated by only a few bacteria. The ant-attended aphid group had a significantly higher Shannon index (F 0.05 (1,24) = 12.233, p = 0.002) and lower Simpson index (F 0.05 (1,24) = 11.525, p = 0.002) than the aphid group not attended by ants, however, the Chao1 (F 0.05 (1,24) = 0.079, p = 0.781) and ACE (F 0.05 (1,24) = 0.002, p = 0.967) indices had no significant difference (Table 4). For the four time series groups, except the ACE index, there were no significant differences for the Chao1 (  (Table 5). shows, with a low support OTU5 and several other Wolbachia strains from aphids clustered together, and Wolbachia strains of other arthropods and nematodes separated from this clade. The genetic distances between OTU5 and all other Wolbachia sequences ranged from 0 to 3.8%, while the V3-V4 region of 16s rDNA sequence of Wolbachia from P. bambucicola was identical with three Wolbachia sequences from other aphids, indicating that the Wolbachia strain in P. bambucicola is a common one among aphid species.

Discussion
Our study examined the symbiotic bacterial community harbored by the social aphid P. bambucicola and compared the microbiomes in different aphid clones of four time groups and two ant groups. Overall, 16S rDNA sequencing showed a stable coexistence of three dominant bacterial symbionts, namely the primary aphid endosymbiont Buchnera, the Pectobacterium, and the Wolbachia. In addition to these three symbionts, Arsenophonus, Serratia, and Orbaceae sp., previously reported to be associated with insects, were also found with relative high abundance in our study but were only limited to a few samples (Table S3). It is notable that Arsenophonus was only infected with two samples in which Pectobacterium was absent ( Table 3). The genus Arsenophonus, a group of symbionts that kill male eggs of the wasp Nasonia vitripennis [52], has also been reported in other insects [53,54], but its function in many insects is far from clear.
The clones of P. bambucicola were stably infected by three symbiotic bacteria genera (Buchnera, Pectobacterium, and Wolbachia). Buchnera, the obligate endosymbiont of aphids, was the most abundant symbiotic bacteria of P. bambucicola and was detected in all samples. Fukatsu et al. (1994), using histochemical-based approaching analyses, revealed that P. bambucicola contain the typical intracellular symbiont Buchnera, rather than being replaced by a yeast-like fungus [55]. The stable coexistence of two secondary endosymbiotic bacteria is not common in aphids. For example, in spite of the most studied pea aphid containing eight to nine secondary endosymbionts, double infections with two secondary symbionts were rarely detected [56,57]. Gómez-Valero et al. (2004) found a coexistence of Wolbachia and a secondary symbiont (R type, Serratia) in the aphid Cinara cedri [58]. Interestingly, the coexistence of Pectobacterium and Wolbachia with high infection rate (78.57%) was detected in P. bambucicola. Meanwhile, compared to other aphid species with a variety of host plants, such as the pea aphid [13], the diversity of the bacterial community in P. bambucicola was low, likely because the habit (feeding on bamboo stems) of P. bambucicola is relatively simple.
High infection rates (26/28) of Pectobacterium have been detected in P. bambucicola. Such high prevalence suggests a potential functional role of Pectobacterium in this social aphid. The genus Pectobacterium are important causative agents of the soft rot disease with broad host plant ranges [59,60]. These bacteria produce a large amount of plant cell wall-degrading enzymes (PCWDEs) such as pectinases, cellulases, hemicellulases, and proteinases which makes plant tissue decay efficiently [61]. Among the over 5000 known aphid species, P. bambucicola is the only one that exclusively feeds on the hard stalks of bamboo, which consist of much harder cell walls than many other plants. To feed on the phloem sap of bamboo, aphid stylet must conquer the hindrance of the hard cell wall [62]. We speculate that Pectobacterium symbionts should help the host aphid to secrete PCWDEs, especially pectinases, to lyse and pierce the cell walls of bamboo. This may be one of the main reasons why P. bambucicola can adapt to the hard stalks of bamboo. In addition, our ongoing subsequent experiments of bacterial screening in different tissues of P. bambucicola found that Pectobacterium was mainly distributed in the ovary (unpublished data), indicating that Pectobacterium is likely a symbiotic bacterium and can vertically transmit in this aphid. It will be valuable to investigate how this symbiosis contributes to feeding adaptation of the host aphid at genomic levels for both Pectobacterium and P. bambucicola. Notably, genetic distances indicated that the Pectobacterium in P. bambucicola may be a new symbiont species and phylogenetic analysis showed that Pectobacterium as a symbiont in P. bambucicola may reflect an evolutionary transition from plant pathogen to insect symbiotic bacteria. Future localization observation such as FISH (fluorescence in situ hybridization) and comparative genomic study will also help to understand their symbiotic relationship.
Wolbachia is a widely distributed intracellular symbiont in arthropods and nematodes and is known as a reproductive manipulator [1]. It can alter the reproductive pattern of hosts in diverse ways, such as feminization, parthenogenesis, male killing, and spermegg incompatibility [1]. Wolbachia-induced female parthenogenesis has been documented in species such as mites, hymenopterans, and thrips [63][64][65]. In terms of aphids, a few previous PCR-based detections suggested Wolbachia does not have a high infection rate in aphids [66,67]. However, our current study indicates a high infection rate (23/28) of Wolbachia in the social aphid P. bambucicola. Although Wolbachia-induced reproductive manipulation has been demonstrated in many arthropods, this function has not been detected in aphid species. P. bambucicola typically live in subtropical areas and reproduce parthenogenetically, but rare records of sexual generation were reported in high altitude areas of Taiwan Island [34]. Therefore, it may be possible that this aphid has been experiencing the loss of the sexual generation in its distribution areas, and the highly infected Wolbachia may potentially have a role in regulating the loss of sexual reproduction in P. bambucicola populations. On the other hand, Wolbachia was demonstrated to act as a nutritional mutualistic bacterium in other insects, such as the aphid Pentalonia nigronervosa [68] and the bedbug Cimex lectularius [3]. Whether Wolbachia has a nutritional role in P. bambucicola needs to be investigated by further studies based on genome analysis, microscopic observation, and functional verification experiments.
Aphid groups visited by ants had a significantly higher diversity of bacterial communities than those not attended by ants. However, Chao1 and ACE indices, having no significant difference, suggested that ant attendance had no influence on the richness, while the symbiotic microorganisms of aphids attended by ants had higher community evenness. Ant-attended P. bambucicola populations having higher diversity of bacterial community may indicate that ant attendance could be a selection agent for shaping different symbiotic communities, which may influence the composition of aphid honeydew and thus adjust its attraction to ant partners. Besides, it has been demonstrated in some studies that ants can protect aphid clones from parasitoids and pathogens [69,70]. Henry et al. (2015) found that Hamiltonella defensa and R. insecticola, two protective aphid symbionts, were more likely to occur in aphid species not attended by ants [23]. Mandrioli et al. (2016) found that if aphid populations recurrently tended by ants were maintained in insectaries without ants, aphids increased the amount of H. defensa, making the composition of their microbiome context-dependent [28]. The associations with ants or protective symbionts may be two alternative ways that aphids can protect themselves from pathogens and nature enemies. In the field, P. bambucicola often encounter many kinds of natural enemies, and ants and soldier aphids alternately protect the clones from natural enemies [35]. This may be the potential reason why the number of microbial species in the P. bambucicola is not affected by ant attendance. In terms of the time series groups, there was no significant difference in the composition of bacterial communities, which confirms that the microbiota is dominated by three symbionts (Buchnera, Pectobacterium, Wolbachia) and time is not an important structuring factor for bacterial communities associated with P. bambucicola.

Conclusions
In this study, we reported the symbiotic bacterial community of the social aphid P. bambucicola. The diversity of the bacterial community in P. bambucicola was low and the microbiota of P. bambucicola was dominated by three symbiotic bacteria. In addition to the obligate endosymbiont Buchnera, high infection rates of Pectobacterium (92.9%) and Wolbachia (82.1%) in P. bambucicola may suggest a potential functional role in P. bambucicola. Combined with the biology of P. bambucicola, the symbiotic Pectobacterium may help P. bambucicola adapt its unique niche of bamboo stems, and the well-known reproductive manipulator Wolbachia may regulate the loss of sexual reproduction or have a nutritional role in P. bambucicola. The genetic distance analysis indicated that the Pectobacterium of P. bambucicola may be a new symbiont species which needs to be further validated with more data in future studies. Moreover, our study indicates that time is not an important structuring factor for bacterial communities associated with P. bambucicola, but the evenness of microbiota in P. bambucicola attended by ants is higher than that without ant attendance.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/insects12050416/s1, Table S1: The Illumina HiSeq results of bacterial 16S rRNA genes, Table S2: Detail information of microbiome diversity indices of 28 Pseudoregma bambucicola samples, Table S3: The details of OTUs distribution in all samples. Data Availability Statement: The raw sequencing data are deposited in the NCBI Sequence Read Archive (SRA) database with accession number PRJNA634633.

Conflicts of Interest:
The authors declare no conflict of interest.