Gut microbial communities associated with phenotypically divergent populations of the striped stem borer Chilo suppressalis (Walker, 1863)

Chilo suppressalis (Walker, 1863) is a serious stem borer of rice and water-oat plants, and has phenotypically diverged into rice and water-oat populations. Insect gut microbiota plays an important role in the host life and understanding the dynamics of this complicated ecosystem may improve its biological control. The effect of diet and gut compartments on the gut microflora of divergent populations of C. suppressalis is not fully clear. Herein, we characterized the gut microbiota of C. suppressalis populations fed on two hosts (i.e., water-oats fruit pulps and rice seedlings), by sequencing the V3–V4 hypervariable region of the 16S rRNA gene using the Illumina MiSeq platform. Gut bacterial communities showed variation in relative abundance among C. suppressalis populations fed on water-oats fruit pulps or rice seedlings. Proteobacteria and Firmicutes became the predominant phyla, and Enterobacteriaceae, Enterococcaceae and Halomonadaceae were the predominant family in all C. suppressalis populations. The highest bacteria diversity was found in the midgut of the rice population fed on water-oat fruit pulps. Bacterial communities in the midgut were more diverse than those in the hindgut. The bacterial genera distribution showed great differences due to diet types and gut compartments among populations. Our results demonstrated that the host plants tested had a considerable impact on gut bacterial composition of C. suppressalis populations. Additionly, the unique gut morphology and physiological conditions (viz., oxygen content, enzymes) also contributed to variation in microbiomes. In conclusion, our study provided an important insight into investigation of insect-bacteria symbioses, and biocontrol of this species and other related lepidopterans.

. Schematic diagram of the fully factorial experimental design used in the present study. Chilo suppressalis from water-oat field and rice field were reared on water-oats fruit pulps or rice seedlings for three continuous generations to examine the effects of host plant and population origin on the gut microbiota. There were two cross-rearing populations (i.e., non-original populations) comprised of water-oat population individuals reared on rice seedlings (j) and rice population individuals reared on water-oat fruit pulps (r). The corresponding original populations were water-oat population individuals reared on water-oat fruit pulps (J) and rice population individuals reared on rice seedlings (R). Processing of sequencing data. Raw fastq files were demultiplexed, quality-filtered using QIIME (version 1.17) with the following criteria: (i) The 250 bp reads were truncated at any site receiving an average quality score < 20 over a 10 bp sliding window, discarding the truncated reads that were shorter than 50 bp; (ii) exact barcode matching, 2 nucleotide mismatch in primer matching, reads containing ambiguous characters were removed; (iii) only sequences that overlap longer than 10 bp were assembled according to their overlap sequence. Reads which could not be assembled were discarded. Operational Units (OTUs) were clustered using UPARSE (version 7.1 http:// drive5. com/ uparse/) and chimeric sequences were identified and removed using UCHIME. The phylogenetic affiliation of each 16S rRNA gene sequence was analyzed by RDP Classifier (http:// rdp. cme. msu. edu/) against the silva (SSU115) 16S rRNA database.

Results
General structure of gut. The gut of C. suppressalis was a continuous tube running from the mouth to the anus. It was structurally divided into foregut, midgut and hindgut. The foregut (Fg) was a slender, elongate tube, expanding posteriorly and constricting at its ends. The midgut (Mg) was a well-developed saclike tube beginning from the end of the foregut and extending to the long, narrow hindgut (Hg). The freshly dissected foregut was translucent, the midgut was opaque white, and the hindgut was yellowish-brown ( Fig. 2).
Analysis of bacterial 16S rDNA gene sequences. Illumina sequencing obtained 861,370 sequences clustering into 3234 operational taxonomic units (OTUs) ( Table 2). Chao1 estimator and Shannon Index were calculated for the determination of the richness and homogeneity of the community. The relative bacterial abundance of 18 phyla differed significantly across the eight samples (Kruskal-Wallis test, p < 0.0001). The midgut and hindgut of the rice population feeding on water-oat fruit pulps (rMG, rHG), possessed the highest bacteria diversity. The bacteria in the midgut were more diverse than those in the hindgut ( Table 2).
A total of 44 and 66 OTUs were observed in the guts of original and cross-rearing populations, separately (Fig. 3C,D). These OTUs were pooled into 26 families for the midgut of original populations. The relative abundances of five families were Enterobacteriaceae, Halomonadaceae, Bacillaceae, Enterococcaceae and Streptococcaceae (S1C Table; S2 Table and S3 Table). However, in the hindgut of cross-rearing populations, the OTUs were   Table; S2 Table and S3 Table).
The bacterial genera from original populations showed distinct distribution according to diet types and gut compartments ( Fig. 5; S5 Table). Halomonas (69.9%) and Klebsiella (70.1%) were dominant in the midgut and hindgut of the rice population feeding on rice seedlings (RMG and RHG); but Bacillus (26.9%) and Klebsiella (35.14%) were prevailed in the midgut of the water-oat population feeding on water-oat fruit pulps (JMG), Citrobacter (40.8%) was enriched in the hindgut of the water-oat population feeding on water-oat fruit pulps (JHG). Enterococcus was dominant in the midguts of the two cross-rearing populations (jMG (64.8%) and rMG (45.9%)), and Citrobacter was prevailed in the hindguts of the two cross-rearing populations (jHG (43.7%) and rHG (37.1%)). However, the bacteria in cross-rearing populations showed different genus distributions based on

Diet-and compartment-related variations in the gut microbial composition. In all populations,
there were significant differences in the relative abundances at the family level (p < 0.0001, Kruskal-Wallis test). 95 bacterial taxa were identified at the genus level. Influence of compartment sampling proved significant with a well-defined cluster formed by the midguts of all original and cross-rearing populations (i.e., JMG, jMG, rMG and RMG). By contrast, bacteria from the hindguts of all populations (i.e., RHG, JHG, jHG, rHG) were more heterogeneous for constituting four different clusters (Fig. 6). All the midguts and hindguts exhibited a significant difference in bacteria abundance of three families: Enterobacteriaceae, Enterococcaceae and Bacillaceae. Enterobacteriaceae was dominant in the hindgut (66.4%), but was decreased to 24.6% in the midgut. In comparison, Enterococcaceae was less abundant in the hindgut (11.2%), whereas increased to 31.4% in the midgut; Bacillaceae (5.0%) resided in the hindgut was increased to 11.4% in the midgut ( Fig. 6; Table S3). The differences at the family level were (Fig. 6): (1) a higher abundance of Enterobacteriaceae in the hindguts (55.8%) than in the midguts of the rice population feeding on water-oat fruit pulps (8.6%) and the water-oat population feeding on water-oat fruit pulps (35.9%); (2) a higher presence of Enterococcaceae in the midgut of the water-oat population feeding on rice seedlings (64.8%) than in the midgut and hindgut of the water-oat population feeding on water-oat fruit pulps (12.4%%, (10.0%), midgut of the rice population feeding on wateroat fruit pulps (45.9%) and the hindgut of the water-oat population feeding on rice seedlings (18.9%); (3) and a higher presence of Halomonadaceae in the midguts of two original populations (RMG: 69.9%; JMG: 4.9%) and the rice population feeding on water-oat fruit pulps (6.0%), hindguts of the water-oat population feeding on water-oat fruit pulps and rice seedlings (3.4%, 0.3%). However, the Bacillaceae was higher in the midgut JMG midguts of water-oat population feeding on water-oat fruit pulps, RMG midguts of rice population feeding on rice seedlings, jMG midguts of water-oat population feeding on rice seedlings, rMG midguts of rice population feeding on wateroat fruit pulps, JHG hindguts of water-oat population feeding on water-oat fruit pulps, RHG hindguts of rice population feeding on rice seedlings, jHG hindguts of water-oat population feeding on rice seedlings, rHG hindguts of rice population feeding on water-oat fruit pulps. www.nature.com/scientificreports/ and hindgut of the water-oat population feeding on water-oat fruit pulps (26.9%, 11.3%) and midgut of the rice population feeding on water-oat fruit pulps (18.6%), than that in the midgut of the water-oat population feeding on rice seedlings (0.6%) and hindgut of the rice population feeding on water-oat fruit pulps (1.2%). A non-metric multidimensional scaling (NMDS) analysis was performed to analyze the influence of diet and compartment on the microbiota (Fig. 7A-D). The analysis revealed a clear separation of samples in accordance to the gut regions and a closer association among samples of the same gut region. At the midgut, the clusters were well defined and the highest variability was found in the RMG (i.e., midgut of rice population feeding on rice seedlings) cluster. The RMG and JMG (i.e., midgut of the water-oat population feeding on water-oat fruit pulps) clusters exhibited the most different taxa composition, followed by the rMG (i.e., midgut of the rice population feeding on water-oat fruit pulps) and jMG (i.e., midgut of the water-oat population feeding on rice seedlings) clusters, showing an intermediate composition (Fig. 7A). At the hindgut, there were clearly separated clusters: the RHG (i.e., hindgut of rice population feeding on rice seedlings) clusters exhibited a higher inter-sample variation; the JHG (i.e., hindgut of the water-oat population feeding on water-oat fruit pulps) cluster showed an intermediate composition respecting to the RHG, jHG (i.e., hindgut of the water-oat population feeding on rice seedlings) and rHG (i.e., hindgut of rice population feeding on water-oat fruit pulps) clusters (Fig. 7B). The midguts and hindguts clusters from all original populations (JMG, JHG, RMG, RHG) were well-defined, and the JMG, RHG clusters had similar homogeneity level (Fig. 7C). RMG was the most heterogeneous, followed by JMG and RHG. Clusters of cross-rearing populations were better defined than those of original populations. rMG was the most heterogeneous in taxa composition, followed by the jHG. suppressalis. The Y-axis represents the proportion of each taxon. JMG1-JMG3 midguts of water-oat population feeding on water-oat fruit pulps, RMG1-RMG2 midguts of rice population feeding on rice seedlings, jMG1-jMG3 midguts of water-oat population feeding on rice seedlings, rMG1-rMG3 midguts of rice population feeding on water-oat fruit pulps, JHG1-JHG3 hindguts of water-oat population feeding on water-oat fruit pulps, RHG1-RHG3 hindguts of rice population feeding on rice seedlings, jHG1-jHG3 hindguts of water-oat population feeding on rice seedlings, rHG1-rHG3 hindguts of rice population feeding on water-oat fruit pulps; Original populations: C. suppressalis collected from water-oat field and reared on water-oat fruit pulps; or C. suppressalis collected from rice field and reared on rice seedlings. Cross-rearing populations: C. suppressalis collected from water-oat field but reared on rice seedlings; or C. suppressalis collected from rice field but reared on water-oat fruit pulps. Abbreviations for each sample are explained in Tables 1 and 2

Discussion
To date, there were few documents on how gut microbial communities differ across divergent insect populations based on diet and gut compartments. Gut bacterial diversity overall was notably greater in the rice population feeding on water-oat fruit pulps compared to the water-oat population or rice population feeding on rice seedlings. Bacterial communities resided in the midgut were more diverse and variable than those in the hindgut. Only bacteria of Citrobacter, Enterococcus, Halomonas, and Klebsiella were shared by original populations of C. suppressalis, and they were core microbiota based on their relative distribution. The core bacteria was able to colonize in different gut regions 49 , and might have evolved in closely related to hosts and were potential symbiont or beneficial bacteria [50][51][52] . Since rice seedlings and water-oat were very different in nutritional ingredient and secondary compounds, it was probable that the bacteria inhabited in C. suppressalis gut were beneficial to their hosts. The gut bacterial composition and richness exhibited significant differences in the midgut and hindgut of different populations of C. suppressalis. Halomonas and Klebsiella dominated in the midgut and hindgut of rice seedlings-fed rice population, and Klebsiella and Citrobacter were prevailed in the midgut and hindgut of wateroat-fed water-oat population. Enterococcus were enriched in the midgut of cross-rearing populations, whereas Citrobacter was found exclusively in the hindgut of cross-rearing populations. According to Shao et al. And Zhang et al., the Enterococcus was associated with insecticide and pathogen resistances 53,54 , and the presence of this genus in C. suppressalis may enhance the immune of this pest during it host shift. suppressalis. The composition of each sample was based on the taxonomic assignment of the 16S rDNA sequences. The Y-axis represented the proportion of each taxon. JMG1-JMG3 midguts of the water-oat population feeding on water-oat fruit pulps, RMG1-RMG2 midguts of rice population feeding on rice seedlings, jMG1-jMG3 midguts of the water-oat population feeding on rice seedlings, rMG1-rMG3 midguts of rice population feeding on water-oat fruit pulps, JHG1-JHG3 hindguts of the water-oat population feeding on wateroat fruit pulps, RHG1-RHG3 hindguts of rice population feeding on rice seedlings, jHG1-jHG3 hindguts of the water-oat population feeding on rice seedlings, rHG1-rHG3 hindguts of rice population feeding on water-oat fruit pulps; Original populations: C. suppressalis collected from water-oat field and reared on water-oat fruit pulps; or C. suppressalis collected from rice field and reared on rice seedlings. Cross-rearing populations: C. suppressalis collected from water-oat field but reared on rice seedlings; or C. suppressalis collected from rice field but reared on water-oat fruit pulps. Abbreviations for each sample are explained in Tables 1 and 2 www.nature.com/scientificreports/ Our findings also showed that a remarkable different bacteria composition in the RMG and JMG and a intermediate bacteria composition in the rMG and jMG. The inter-individual variability was previously documented in honey bees Apis mellifera 55 , Anopheles 56 , and cockroaches Blattella germanica 38,57 , Shelfordella lateralis 58 and Periplaneta americana 59 . The divergence in taxon composition may reflect divergent functional roles in specific resource use. Gut harbored bacteria community of the water-oat population and rice population feed on their original hosts is closely adapted to the C. suppressalis's diets. With the change of diets, (i.e., water-oat population feeding on rice seedlings, and the rice population feeding on water-oat fruit pulps), the compositional change could be partly responsible for undergoing a recombination of the bacteria, accordingly. Curtis and Sloan suggested that the variation could be attributed to acquire microorganisms from a greatly diverse environmental reservoir microflora, randomly 60 . However, as the populations of C. suppressalis were reared many generations in Figure 6. Abundance and composition of gut microbiota of all populations. Heatmap represent the proportions of OTUs at the family level. The X-coordinate represents the sample of each population, and the Y-coordinate represents the taxon. The color code indicates relative abundance, ranging from blue (low abundance) to yellow to red (high abundance). JMG1-JMG3 midguts of the water-oat population feeding on water-oat fruit pulps, RMG1-RMG2 midguts of rice population feeding on rice seedlings, jMG1-jMG3 midguts of the water-oat population feeding on rice seedlings, rMG1-rMG3 midguts of rice population feeding on water-oat fruit pulps, JHG1-JHG3 hindguts of the water-oat population feeding on water-oat fruit pulps, RHG1-RHG3 hindguts of rice population feeding on rice seedlings, jHG1-jHG3 hindguts of the water-oat population feeding on rice seedlings, rHG1-rHG3 hindguts of rice population feeding on water-oat fruit pulps; Original populations: C. suppressalis collected from water-oat field and reared on water-oat fruit pulps; or C. suppressalis collected from rice field and reared on rice seedlings. Cross-rearing populations: C. suppressalis collected from water-oat field but reared on rice seedlings; or C. suppressalis collected from rice field but reared on water-oat fruit pulps.  61 Cluster analysis showed the jHG and rHG samples formed the most well-defined clusters, suggesting stable microbial profiles. The inter-individual differences suggested that SSB gut microbiome profiles may serve as useful biomarkers for bio-control in population-based studies. The oligophagous diet of stem borers provided suitable ecological niches for harboring bacteria in compared with monophagous lepidopterans 55 . As the phyla Proteobacteria were reported to be involved in carbohydrate degradation, such as starches and hemicellulose 62 , and can be involved in pectin-degrading 63 and nitrogen 64 . Firmicutes was suggested to take part in energy absorption from the diet and may influence the development 65 . The present results illuminated the abundance of two dominant phyla (i.e., Proteobacteria and Firmicutes) and the difference of three families (Enterobacteriaceae, Enterococcaceae and Halomonadaceae) in C. suppressalis populations. As the representative of the oligophagous, C. suppressalis feeding either on water-oat fruit pulps www.nature.com/scientificreports/ or rice seedlings. Both host plants shared the same family Gramineae, but their biochemical components and secondary substances were very different 9 . Our findings suggested that the rapid fluctuation of bacterial flora in larval gut was probably influenced by the biochemical components and secondary substances coming from the host plants; and the diet was an important factor in modulating the bacteria community, as was documented for other insect species 39,[66][67][68][69][70][71][72][73] . The gut bacterial genera were also varied, due to the difference of diets in C. suppressalis: in original populations, Halomonas was dominant in the RMG, Klebsiella was prevailed in the RHG and JMG, and Citrobacter was enriched in the JHG; in cross-rearing populations, Enterococcus was abundant in the midgut, and Citrobacter was predominant in the hindgut. Since diet and host taxonomy modulated bacteria community 71,74 , the successful expansion of bacteria over time probably in turn suppressed the bacteria growth from other phyla in the same habitat 66 . Therefore, we inferred that the different bacteria dominance might be related to successful reproduction of some bacteria genus and suppression of others. Whether the bacteria of Citrobacter, Enterococcus, Halomonas, and Klebsiella detected in the gut of original populations of C. suppressalis was truly associated with the host defense merits further investigation.
One interesting and unexpected result concerned the two compartments chosen for analysis, as we found that variability in microbial composition was higher in the midgut than in the hindgut, independently of diet. The obvious community difference indicated that only some specific groups of microorganisms were able to survive and colonize in the hindgut. However, Kacaniova et al. reported that the hindgut contained a higher number of anaerobic microorganisms than the midgut of honeybee 75 . Although the midgut and hindgut were alkaline, the unique morphology, favorable physiological conditions (viz., oxygen content, lack of unfavorable enzymes), and the availability of partially digested food could become a benign site for maintaining a special bacteria and quick proliferation in the hindgut of C. suppressalis. Indeed, this may be a controversial issue, and the different richness and colonization efficiency of the host symbiont indicated that further investigation should be done to understand their drivers.

Conclusion
We investigated the gut microbial communities of two phenotypically divergent populations of C. suppressalis. The comparison of the midgut and hindgut microbia of C. suppressalia fed on the same diet provided insights into the compartment changes in the gut microbiota of SSB. Analysis of microbial community supplied an initial step toward improving our understanding of the mechanisms underlying C. suppressalis adaptation to host plants at the microbiological level. The results showed that the highest bacteria diversity was found for the midgut of the rice population feeding on water-oat fruit pulps. The most dominant phyla were Proteobacteria and Firmicutes; and the enriched families were Enterobacteriaceae, followed by Enterococcaceae and Halomonadaceae. The microbial communities were highly diverse at the genera level due to diet types or gut compartments among populations. The bacterial community composition was driven mainly by diet types, and affected by other factors including gut compartments. These findings provided an important insight into investigation of insect-bacteria symbioses, and biocontrol of this species and other lepidopterans.