The bacteria in B. minax were quantified by 454 pyrosequencing of 16S rDNA gene amplicons. We obtained 100749 high quality pyrosequencing reads with an average read length of 400 bp of the 16S rDNA spanning the variable regions V3 and V4 from our samples. These reads were distributed among the samples as follows: male gut (21.90%), male testis (30.52%), female gut (23.40%) and female ovary (24.19%), and yielded 322 OTUs, 319 OTUs, 415 OTUs and 389 OTUs, respectively (Table 1) at the 97% identity threshold.

The number of bacteria species detected in a sample was strongly affected by the numbers of sequences analyzed [18]. We calculated the rarefaction curves at a 97% similarity level to verify whether the amount of sequencing reflected the diversity of the original microbial communities. The rarefaction analysis revealed that estimates of OTUs sharply increased before approaching a plateau. Our results showed that all bacterial libraries from our samples represented the bacterial communities well, as the rarefaction curves tended towards saturation (Figure 1A). The percentage coverage of FI, FR, MI and MR was 98.3%, 98.1%, 98.0% and 98.3%, respectively (Table 1). Rank-abundance curves indicate that a majority of reads belonged to rare organisms, whereas all samples contained relatively low proportions of highly abundant bacteria (Figure 1B).

We compared the compositions of OTUs and the relative abundance of sequences obtained from each OTU for every sample to calculate the similarity among the samples. The cluster analysis indicates that the bacterial communities in replicate samples were highly similar (Figure S1). Samples from different sexes were grouped into one large cluster, with no clear distinction between sexes. However, samples from different tissues were grouped into two large clusters, and those from the gut were more similar to each other than those from the reproductive organs.

Altogether, 17 bacteria phyla were detected in our samples. The relative abundances of different bacterial groups in each bacterial library were shown in Figure 2A. Proteobacteria was dominant in all of the libraries, composing 91.45%, 92.86%, 75.3% and 82.2% of the bacterial communities in the male gut, female gut, testis, and ovary, respectively, followed by Firmicutes, composing 4.42%, 5.80% and 8.48% in the male gut, female gut and male testis. Besides, Actinobacteria was also a major phylum in male testis and (7.97%) female ovary (8.64%). Reads belonging to Chlorobi, Chloroflexi, Fusobacteria, Deinococcus-Thermus, Nitrospira, Planctomycetes, Spirochaetes, Tenericutes and TM7 were found to be minor groups, as they appeared in various libraries with only a few reads (<1%). Another 5 groups including Acidobacteria, Bacteroidetes, Cyanobacteria/Chloroplast, Gemmatimonadetes and unclassified bacteria were recovered from our samples.

The relative abundance and distribution of the bacteria from each sample at genus level are shown in Figure 2B and Figure 3. Klebsiella constituted a dominant population in all samples, with 41.41% in female gut, 30.52% in male gut, 23.78% in female ovary and 35.28% in male testis. Other major bacteria corresponded to Citrobacter with 25.80% in female gut, 29.41% in male gut and 15.09% in male testis. But Serratia (19.53%) was strongly represented in female ovary.

The OUT sequences have been submitted to NCBI GenBank (KJ780722–KJ780731 and KM058914–KM061009). The blast results of the 10 most abundant OTUs were shown in table 2. The three species dominating most samples corresponded to Klebsiella, Citrobacter species: Klebsiella oxytoca, Raoultella terrigena and Citrobacter werkmanii. The relative abundance of these taxa varied with sexes and tissues. K. oxytoca dominated the male testis (34.01%) and showed a relatively high abundance in female gut (16.42%) and male gut (17.77%); R. terrigena accounted for 22.5% in male gut and 16.88% in female gut; C. werkmanii was strongly represented in both male gut (17.50%) and female gut (16.63%) (Table 3).

We compared the four samples to find out the influences of sex and tissue on bacterial communities in B. minax. Our results showed that the numbers of species in the female (667 in the gut and 1004 in the ovary) were similar to that in the male (654 in the gut and 960 in the testis) (Figure 4), and the abundance of major species was almost the same between them (Table 3 and Figure 3). When compared the bacterial communities in different tissues from the same sex we found that the reproductive organs had higher richness and diversity index (Table 1) and more bacteria species than the gut (Figure 4). At phylum level, Proteobacteria was dominant among all samples but more abundant in the gut than in the reproductive organs. Besides, most phyla had significantly different abundance among the four samples (Table 4). Furthermore, there were differences at genus level. Some genera such as Delftia and Herbaspirillum were only represented in reproductive organs. Most genera had different abundance among different samples. Some genera were rich in one sample but very low in another (Figure 2B and Figure 3). The PCA analysis showed a distinct clustering among the individual samples and the reproductive organ samples became more dispersed than the gut samples (Figure 5). The PCA yielded two main axes that accounted for 73.67% of the total variation in bacterial community structure (Figure 5). This result indicates the influence of the different variables on the various bacterial phylotypes. The reproductive organ samples were correlated with the presence of most phyla. However, correlations were detected between gut samples and the major phylum, Proteobacteria (Figure 5).

These results suggest that the factor of tissue has more significant influence on the bacterial community structure in B.minax than sex.

B. minax has not been notified under any act or laws and rules of the Government of China for its collection and observation since 2009. No permits were required for collecting the adults since they were obtained from our own land and reared in the laboratory.

The B. minax adults were obtained from our experimental field in Zigui country, Hubei province and were reared at 27±1°C and 70–80% relative humidity with a natural photoperiod in the insectary of the Institute of Urban and Horticultural Pests, Huazhong Agricultural University, Wuhan, China. The adults were fed 5% sugar water and sterile water.

B. minax adult samples were surface-sterilized with 75% ethanol for 2–5 min and rinsed three times in sterile water before dissection. The flies were dissected in a plate containing 10 ml sterile phosphate-buffered solution (PBS, pH 7.4) underneath a stereomicroscope. The dissected gut and reproductive organs were transferred separately to a tube containing 1 ml of extract solution (soil fastDNA extraction kit (BioTeke Corporation)) and were homogenized. Each sample comprised three biological replicates and every replicate contained 15 guts or reproductive organs. The homogenate was used for DNA extraction.

DNA extraction was performed using the soil fastDNA extraction kit (BioTeke Corporation) following the manufacturer's protocol. Briefly, 5 µl of solution A was added to every tube, and the tubes were vortexed for 1–2 min; then the samples were mixed and incubated in water at 37°C for 10 min. One hundred microliters of solution B was added to very tube, and the tubes were vortexed for 1–2 min; then the samples were mixed and incubated in water at 65°C for 10 min. The samples were spun at 10000 rpm for 10 min, and the supernatants were transferred to new 1.5 ml centrifuge tubes. Then, 1/3 volume of protein precipitating solution was added, and the samples were mixed and placed on ice for 8 min. The samples were then spun at 13000 rpm for 10 min. The supernatant was transferred to a pretreated purification column and spun at a low speed (≤2000 g). The filtrate was collected and a 0.6 volume of isopropanol was added. The samples were mixed and spun at 13000 rpm for 10 min. The supernatant was poured off carefully. The pellet was air-dried for 2 min at room temperature and the DNA sample was resuspended with 30 µl elution buffer.

We amplified the V3-V4 region of the bacterial 16S rDNA gene to assess the microbial diversity of the flies. The following primers were used for the PCR reaction: 5′-(CCATCTCAT CCCTGCGTGTCTCCGACGACT) XXXXXXXXTACGGRAGGCAGCAG-3′ and reverse 5′- (CCTATCCCCTGTGTGCCTTGGCAGTCTCAG) XXXXXXXX AGGGTATCT- AATCCT-3′. These primers were designed to contain an 8 nt barcode sequence for multiple samples. The adaptor sequences are in parentheses, and Xs represent the barcode sequences. The gene-specific primers 16S343F and 16S798R are underlined. The PCR was performed in triplicate for each sample in a total reaction volume of 20 µl. Each 20 µL reaction mixture contained 0.75 U Takara Ex Taq DNA polymerase, 2 µL 10 X Ex Taq buffer (Mg2+), 1.6 µL dNTP, and 0.4 µL each of the forward and reverse primers, 1 µL of extracted DNA template, 1 µL BSA and up to 20 µL with ddH₂O. The PCR conditions were the following: initial denaturing at 95°C for 5 min; 30 cycles of 94°C for 30 s, 56°C for 30 s and 72°C for 30 s; and a final extension at 72°C for 10 min. The PCR products were checked with a 2% agarose gel electrophoresis, purified using the AxyPrep DNA gel extraction kit and quantified using the fluorescence quantitative(Qubit 2.0 Fluorometer) with the Qubit dsDNA HS assay kit (Invitrogen, USA). A mixture of the PCR products was prepared by mixing an equal amount of the DNA of the purified 16S amplicons from each sample and then the mixture was cleaned using the Agencourt AMPure XP (Beckman, USA). The mixture was quantified again using the fluorescence quantitative (Qubit 2.0 Fluorometer) with the Qubit dsDNA HS assay kit (Invitrogen, USA) and pyrosequenced on the ROCHE 454 GS FLX platform at Shanghai Hanyubiotech Co., Ltd.

Raw sequencing reads were quality trimmed according to the flowing requirements: Pyrosequencing reads with ambiguous nucleotides, shorter than 90-nucleotides or without a complete barcode and primer at one end were removed. Qualified reads from each sample were aligned with MUSCLE [42]. A distance matrix was calculated from the alignment with PHYLIP version [43]. This matrix served as an input to DOTUR [44] for clustering the sequences into operational taxonomic units (OTUs) to generate rarefaction curves and to make calculations with the richness and diversity indexes at different dissimilarity levels. The tag pyrosequence reads were classified into different taxonomic groups using RDP classifier [45], [46]. The numbers of representative reads for each OTU determined at 3% dissimilarity at each taxonomic level was counted and the proportions of each group in a sample were calculated. A tree was built using the neighbor-joining method of Clustal_X [47]. The tree was displayed and edited using the Molecular Evolutionary Genetics Analysis (MEGA 5) [48]. A minimum of 75% sequence similarity was applied for assigning the tag sequences into the cluster. The relative abundance and occurrence of tag sequences assigned into clusters were visualized as a heatmap using the JColorGrid [49]. The pyrosequencing data has been submitted to the GenBank database as a file under accession number SRR1531158. The accession numbers of OTU sequences were KJ780722–KJ780731 and KM058914–KM061009.