Hongyu Zhang, State Key Laboratory of Agricultural Microbiology, Institute of Urban and Horticultural Pests, College of Plant Science and Technology, Huazhong Agricultural University, Wuhan 430070, China. E-mail: firstname.lastname@example.org
Aims: To (i) identify the bacterial communities in the gut of oriental fruit fly (Bactrocera dorsalis) adult and (ii) determine whether the different surroundings and diets influence the bacteria composition.
Methods and Results: Polymerase chain reaction-denaturing gradient gel electrophoresis (DGGE) fingerprinting was used to investigate bacterial diversity in the oriental fruit fly adult gut. The 16S rDNA cloned libraries from the intestinal tract of laboratory-reared (LR), laboratory sterile sugar-reared (LSSR) and field-collected (FC) populations of oriental fruit fly were compared. Phylogenetic analysis of 16S rDNA revealed that Gammaproteobacteria were dominant in the all samples (73·0–98·3%). Actinobacteria and Firmicutes were judged to be major components of a given library as they constituted 10% or more of the total clones of such library. The Flavobacteria, Deltaproteobacteria, Bacteroidetes and Alphaproteobacteria were observed in small proportions in various libraries. Further phylogenetic analyses indicated common bacterial phylotypes for all three libraries, e.g. those related to Klebsiella, Citrobacter, Enterobacter, Pectobacterium and Serratia. libshuff analysis showed that the bacterial communities of B. dorsalis from the three populations were significantly different from each other (P <0·0085).
Conclusions: (i) The intestinal tract of B. dorsalis adult contains a diverse bacterial community, some of which are stable. (ii) Different environmental conditions and food supply could influence the diversity of the harboured bacterial communities and increase community variations.
Significance and Impact of the Study: Comparison of the microbial compositions and common bacterial species found in this paper may be very important for the biocontrol of B. dorsalis.
Insects harbour diverse micro-organisms, many of which occupy their intestinal tract and have developed different interactions during the long period of coevolution (Dillon and Dillon 2004). Interactions between hosts and their microbes can range from mutualistic, such as the interaction between termites and their gut microbes (Breznak and Brune 1994; Schmitt-Wagner et al. 2003), to parasitic, such as the interaction of the bacterium Paenibacillus larvae (American foulbrood) in honeybees (Schmid-Hempel 1998). Some of these interactions have been relatively well studied for their economic importance or their remarkable biology. However, the exact nature of invasive carpophagous insect pests–microbe interactions, specifically those between microbes and the major fruit flies, remains poorly understood.
The oriental fruit fly (Bactrocera dorsalis Hendel) is a destructive polyphagous and invasive insect pest of tropical and subtropical fruits and vegetables (Clarke et al. 2005). In previous studies using the cultivable dependence techniques, members of the Enterobacteriaceae were found to be the dominant microbial population in the gut of Bactrocera fruit flies (Tsiropoulos 1983; Fitt and O’Brien 1985; Jang and Nishijima 1990; Drew and Lloyd 1991). However, there is still a dearth of knowledge on the bacterial communities harboured by the oriental fruit fly, and it is unclear whether the different surroundings and diets influence the bacteria composition.
In this paper, the composition and diversity of the microbial community in the gut of B. dorsalis obtained from three different populations were characterized by 16S rDNA-DGGE. The variations in bacterial communities among three libraries were compared by libshuff statistical method.
Materials and methods
Collection and rearing of insects
The LR and LSSR B. dorsalis adults were obtained from established laboratory colony (>13 generations) 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 LR adults were fed upon mixture of sucrose and yeast powder at ratio 3 : 1, while LSSR adults were reared on filtered sucrose solution (20%). The FC adults were collected thrice from the citrus orchard of Cadre Sanatorium in Wuhan, China (30°56′N and 114°31′E). Each time, the insects were brought to the laboratory within two hours and dissected.
Bactrocera dorsalis dissection
Before dissection, adult insects from each of the populations sampled were surface-sterilized with 70% ethanol for 2–5 min and rinsed three times in sterile distilled water. Approximately 30 adults from each population were dissected aseptically in a plate containing 10 ml sterile distilled water using two pairs of sterilized tweezers underneath a stereomicroscope. The dissected guts were transferred to a tube containing 100 μl of sterile phosphate-buffered solution (PBS) (10 mmol l−1, pH 7·4) and were homogenized. The homogenate was used for DNA extraction.
Total DNA extraction
To extract total DNA, dissected guts of B. dorsalis from each population were separately suspended in phosphate buffer, harvested by centrifugation, washed with the same buffer and then resuspended in 557 μl TE buffer [10 mmol l−1 Tris–HCl (pH 8·0), 50 mmol l−1 EDTA] with the presence of 10 μl lysozyme (5 mg ml−1). The reaction was incubated at 37°C for 20 min. After addition of 30 μl of 10% SDS and 3 μl of 20 mg ml−1 proteinase K, samples were incubated at 37°C for 40 min. We added 100 μl of 5 mol l−1 NaCl and 80 μl of CTAB/NaCl and incubated at 65°C for 10 min. Then, samples were extracted with phenol/chloroform/isoamyl alcohol [24 : 24 : 1 (v/v/v)] and centrifuged for 4 min at 13 400 g. Nucleic acids were precipitated with isopropyl alcohol, rinsed with 70% frozen ethanol and suspended in 30 ml of TE.
PCR amplification and cloning of bacterial 16S rRNA gene
PCR amplification was performed using Bacteria-universal primers 968-GClamp (5′-CGCCCGCCGCGCGCGGCGGGCGGGGCGGGGGCACGGGGGGAACGCGAAGAACCTTAC-3′) and L1401 (5′-CGGTGTGTACAAGACCC-3′) (Nubel et al. 1996). Amplification was carried out in 30-μl reactions with a thermal cycler. Each reaction tube contained 3 μl 10× PCR buffer (with Mg2+), 0·5 μl 10 mmol l−1 dNTP, 0·5 μl of each forward and reverse primer, 0·2 U of DNA polymerase and about 50 ng of sample DNA template. PCR was run under the following conditions: 94°C for 5 min followed by 30 cycles at 94°C for 30 s, 56°C for 30 s and 72°C for 1 min and a final extension period of 7 min at 72°C. PCR products were checked on an agarose gel (1·2% agarose, 1× TBE) stained with ethidium bromide. Negative controls (no DNA added) were routinely included as a contamination check. No products were obtained from these controls. Band was excised from the gel, and the DNA was purified with Omega DNA purification kit (Omega Inc., Norcross, GA, USA).
Purified PCR products were cloned into pMD18-T Vector using a rapid ligation kit according to the manufacturer’s instructions (TaKaRa Inc., Dalian, China). Ligation products were transformed into Escherichia coli DH5α (Promega, Madison, WI, USA). Ampicillin-resistant transformants were subcultured on Luria–Bertani (LB) agar plates containing ampicillin (100 μg ml−1; Sigma, St Louis, MO, USA), X-gal (20 mg ml−1) and IPTG (40 mmol l−1; Promega) by colour-based recombinant selection. Two hundred and fifty transformants were picked randomly from gut-cloned plates. Then, extracted DNA plasmids were examined by PCR with the vector-specific primers M13 and the same primers as in the PCR amplification (968-GC Clamp and L1401).
DGGE screening of 16S rDNA clones and sequencing
DGGE was carried out using a DCode™ Universal Mutation Detection System (Bio-Rad Lab., Hercules, CA, USA). DGGE marker was prepared from a selection of bacterial 16S rRNA gene products to enable gel-to-gel comparison. PCR products were loaded onto 8% (w/v) polyacrylamide gels with a denaturing gradient ranging from 30 to 70% (100% corresponding to 7 mol l−1 of urea and 40% (w/v) deionized formamide in 0·5× TAE buffer) and were run at a constant temperature of 60°C, the voltage at 200 V for 10 min and then 85 V for 16 h in 0·5× TAE buffer (pH 8·5). After electrophoresis, the gels were removed and stained with silver nitrate as described by Zhang and Jackson (2008). The DGGE bands were compared visually and clones showing identical DGGE profiles were grouped into the same phylotypes or operational taxonomic units (OTUs) (Simpson et al. 2002). One representative clone from each group was selected for sequencing. Sequencing was performed on both strands using vector-specific M13 forward and reverse primers by AuGCT Biotechnology Co., Ltd, Beijing, China.
Sequences were compared to the GenBank database by using the basic local alignments tool (blast) (Altschul et al. 1997) to identify their closely related 16S rDNA sequences. The nucleotide sequences from this study and GenBank databases were aligned pairwise with the clustal x program (Thompson et al. 1997), and the sequence was checked for chimeric properties using the RDP Chimera Check program (Maidak et al. 2001). Phylogenetic analysis was performed with phylip v. 3.6 software package (distributed by J. Felsenstein, University of Washington, Seattle). Pairwise evolutionary distances and similarity matrices were calculated using the dnadist program according to the Jukes–Cantor model (Jukes and Cantor 1969). The phylogenetic tree was constructed using the neighbour-joining (NJ) algorithms in Mega 4 software (Saitou and Nei 1987), and the confidence of the tree topology was tested by 1000 bootstrap replicates (Felsenstein 1985).
Rarefaction analysis and estimators of various diversity indices from sequence data were chosen to measure the microbial diversity of gut samples.
Shannon diversity index (H) was calculated with the equation: H = −∑ (Pi) (InPi), where Pi is the number of clones in each phylotype group divided by the total number of clones in each library (Hill et al. 2003).
Simpson’s index (D) was calculated according to the equation: D =1 −∑ (Pi)2
Coverage (C) was calculated using the equation as described by Good: C =1 − (n1/N), where n1 is the number of clones that occurred only once and N is the total number of clones (Good 1953).
Chao-1 estimator was determined using the equation: S1* = Sobs + [A (A − 1)/2 (B +1)], where Sobs is the number of 16S rDNA clones observed, A is the number of clones observed just once, and B is the number of clones observed twice (Chao 1984). The standard deviation (SD) was estimated using the equation SD = B [(A/4B)4 + (A/B)3 + (A/2B)2].
The phylotype compositions of libraries were compared using the Sorensen index, Cs = 2j/(a + b), where j was the number of common phylotypes in A and B, a was the number of phylotypes in sample A, and b was the number of phylotypes in sample B (Magurran 1996). The significance of difference in the composition between any of two cloned libraries (e.g. X and Y) was determined by using libshuff program (http://libshuff.mib.uga.edu/) according to the Singleton method (Singleton et al. 2001). Cramér-von Mises statistic was used to calculate differences (ΔC) between ‘homologous’CX (D) and ‘heterologous’ coverage curves CXY (D), and Monte Carlo resampling approach was applied to infer statistic significance. Because the libshuff does not correct experiment-wise error for multiple comparisons of libraries, the Bonferroni correction was used to calculate critical P-value. The two libraries were considered to be significantly different from each other if the lower of the two P-values generated by libshuff is below or equal to the critical P-value (the critical P-value for three cloned libraries is 0·0085).
DGGE screened a total of 546 clones, of which 178 clones were from LR, 181 from LSSR and 187 from FC. All these clones were classified into 150 unique phylotypes, with 55, 26 and 69 detected from LR, LSSR and FC populations, respectively (Table 1).
Table 1. Diversity indices for 16S rDNA libraries obtained from intestinal tract bacteria of Bactrocera dorsalis
Rarefaction analyses based on the different phylotypes and the analysed clone numbers revealed that the sampling size of the three individual cloned libraries was sufficient to cover the majority of the microbial diversity in the gut of B. dorsalis, as documented by the modest slope at the end of the rarefaction curves (Fig. 1). The percentage coverage of LR, LSSR and FC libraries was 81·8, 88·5 and 81·2%, respectively (Table 1), which indicated that each of the individual cloned libraries covered most of the micro-organisms in fruit fly gut. The rarefaction curves also suggested that the sequence population was the least diverse in the LSSR and the most diverse in the FC. This hypothesis was further supported by calculating Shannon (4·505–5·863) and Simpson’s indices (0·9583–0·9840) (Table 1). The Chao-1 estimator showed that the species richness of FC (74 ± 8) was the highest, followed by that of LR (58 ± 9), while LSSR (28 ± 30) had the lowest species richness (Table 1).
Taxonomic groups and phylogenetic distribution
All sequenced clones were grouped into either of the following phyla: Gammaproteobacteria, Actinobacteria, Alphaproteobacteria, Deltaproteobacteria, Bacteroidetes, Flavobacteria and Firmicutes. The relative abundances of different microbial phylogenetic groups in each cloned library are shown in Fig. 2. The Gammaproteobacteria dominated in all the libraries (73·0–98·3%). Besides this phylum, the following phyla were judged to be major components of each library as they constituted 10% or more of the total clones of a given library: Actinobacteria (10·1%) were the major component of LR and Firmicutes (12·3%) was the major component of FC. Flavobacteria, Deltaproteobacteria, Bacteroidetes and Alphaproteobacteria appeared in various libraries with a few clones (<8 per phylum) (Fig. 2). Among the 150 sequences of bacterial 16S rDNA, three distinct clusters were identified (labelled as unclassified groups in Figs 3–5).
Phylogenetic analysis placed 447 clones within Gammaproteobacteria (Figs 3 and 4). These clones (accounting for 81·9% of total clones) were from the LR (23·8%), LSSR (32·6%) and FC (25·5%) libraries and accounted for 73·0, 98·3 and 74·3% of the clones in LR, LSSR and FC libraries, respectively. The phylum Gammaproteobacteria contained members of Enterobacteriales, Pasteurellales, Pseudomonadales and Aeromonadales. Furthermore, Enterobacteriales clones were widely diverse in all three libraries and distributed among 12 distinct genera (Figs 3 and 4) in which Klebsiella ssp., Enterobacter ssp. and Cirtrobacter ssp. were the most often detected bacteria (Fig. 3). Two clones (LR178 and FC42) were distantly related to other sequences and represented a deep branch (Fig. 3). Five clones (LR37, 95, 134, 147 and FC85) formed a monophyletic clade and were distantly related to other known sequences (Fig. 4).
The closely related sequences (>97% sequence similarity) that had been retrieved from three libraries were grouped as one common species (Mohr and Tebbe 2006). Seven such groups (dotted text boxes) were found in the class Gammaproteobacteria (Enterobacteriales) (Figs 3 and 4). One of these groups (93 of 546, or 17·0%) was affiliated to the uncultured Klebsiella sp. (GQ416328.1) and Klebsiella pneumoniae (HM063413.1). Two others (67 of 546, or 12·3%) were identical or very similar to Citrobacter freundii (GQ983.53.1), Citrobacter sp. (EU031775.1) and uncultured Citrobacter sp. (GQ417858.1). A fourth symbiont group (32 of 546, or 5·9%) detected was closely related to members of the Enterobacteriales tree clustered with Enterobacter sp. (FJ588708.1) and Enterobacter hormaechei (HM771693.1). The last two groups (21 of 546, or 3·8%; 26 of 546, or 4·8%) were affiliated to the Pectobacterium carotovorum subsp. (FJ527484.1) and Serratia sp. (FJ862037.1), respectively. In addition, although LSSR150, LSSR155, LR241 and FC99 were not grouped into one clade, their sequences had very high homology (>97% sequence similarity) and were identical to the Enterobacter sp. (FJ890898.1) and Enterobacter cloacae (GQ406570.1).
Firmicutes, Actinobacteria and Deltaproteobacteria
Of 546 clones, 6·4% were affiliated with Firmicutes (23, 9 and 3 clones from FC, LR and LSSR, respectively) and fell into four families, Enterococcaceae, Streptococcaceae, Bacillaceae and Paenibacillaceae (Fig. 5). Twenty-five clones were affiliated with the phylum Actinobacteria and accounted for 4·6% of the total clones. Among them, 18 and seven clones were from LR and FC populations, respectively, but this phylum was absent in the LSSR library (Fig. 5). The phylum Actinobacteria contained members of Microbacterium, Agromyces and Leucobacter (Fig. 5). Sequences affiliated with Deltaproteobacteria accounted for 1·5% of the total clones and fell within a single family, Desulfovibrionaceae (Fig. 5), including five clones from FC and three from LR B. dorsalis populations.
Approximately, 0·7% of the total clones (four of 546 clones) were affiliated with the Flavobacteria group (Figs 2 and 5). They were detected within the LR and FC libraries. Two clones belonged to the phylum Alphaproteobacteria and were affiliated with the isolated strain Agrobacterium tumefaciens (Fig. 5). Clone FC41 fell into the phylum Bacteroidetes and clustered with a representative in Dysgonomonas (Fig. 5). Lastly, four clones (LR56, 109; and FC32, 100) were distantly related to other known sequences and represented a deep branch (Fig. 5).
Comparison of the microbial compositions between the cloned libraries
We compared the bacterial compositions in the cloned libraries by calculating the similarity indices and using the libshuff analysis. Results showed that the similarities of the phylotype populations in these three libraries ranged from 0·1818 to 0·4722 (Table 2). The LSSR library shared only fewer phylotypes with the libraries of LR and FC in common (the similarity index between the LSSR and the LR libraries was 0·2381, and that between LSSR and FC libraries was 0·1818). The highest similarity index (0·4722) was found between the libraries of LR and FC populations.
Table 2. Comparisons of the 16S rDNA cloned libraries of gut bacteria from the three Bactrocera dorsalis adult populations tested
Paired comparisons between the three cloned libraries indicated that the libraries are significantly different from each other (P <0·0085).
*Sorenson similarity index was determined as follow: Cs = 2j/(a + b), where a and b are the numbers of phylotypes in sample A and B, respectively, and j is the number of phylotypes found in both samples A and B.
By the paired comparisons method, both the homologous and heterologous coverage curves were similar in position and shape. All comparisons showed that the libraries were significantly different from each other (P <0·0085). Furthermore, comparisons of the calculated value of (CX − CXY)2 to the 95% value of (CX − CXY)2 from random shuffles showed that differences between the libraries were the greatest at genetic distances between 0·02 and 0·27 (Fig. 6), which indicated that differences in the libraries occurred in both shallow and moderately deep phylogenetic levels. Significant differences at P < 0·05 between homologous and heterologous coverage curve were indicated by values of ΔC (Table 2), which showed that the LR library differed most significantly from the LSSR library. The FC library was found to be significantly different from the other two libraries (Table 2).
In this study, 16S rDNA-DGGE and sequence analysis were used to compare the diversity of microbes associated with LR, LSSR and FC populations of the oriental fruit flies. Our results showed that the intestinal microbial community was dominated by Gammaproteobacteria, followed by members of the Firmicutes, Actinobacteria, Deltaproteobacteria, Flavobacteria, Alphaproteobacteria or Bacteroidetes. Other studies have also revealed the abundance of Gammaproteobacteria in the digestive tract of various invertebrates, such as Anopheles stephensi, a malaria vector (Rani et al. 2009); the earthworm, Lumbricus rubellus (Knapp et al. 2009); and the desert locust, Schistocerca gregaria (Dillon et al. 2010). In contrast, bacterial communities associated with certain soil insect pests such as scarabid beetles and termites are inclined to the Clostridia phylum and the so-called Termite group 1 (TG-1) phylum, respectively (Hongoh et al. 2005; Zhang and Jackson 2008).
Within the Gammaproteobacteria, members of the family Enterobacteriaceae have frequently been reported being dominant in the gut of several other tephritids, such as Dacus (Drew and Lloyd 1991), Bactrocera (Capuzzo et al. 2005), Anastrepha (Kuzina et al. 2001) and Ceratitis (Behar et al. 2008a). In this study, three oriental fruit fly populations shared common representatives of the Enterobacteriaceae, including Klebsiella, Citrobacter, Enterobacter, Pectobacterium and Serratia. Many of the enterobacterial species were diazotrophic (nitrogen fixing), notably the Klebsiella ssp. (Klebsiella oxytoca and Kl. pneumoniae), Cit. freundii, Pantoea ssp., Enterobacter ssp. and Pectobacterium ssp. (Zinder and Dworkin 2000; Bergey et al. 2001; Lauzon 2003). The role of Ent. cloacae may include the production of some metabolic by-products (e.g. vitamins) that is beneficial to the olive fruit fly (Hagen 1966) or produce a strong antifungal compound that is inhibitory to many fungi (Howell et al. 1988). In addition, Behar et al. (2008b) reported that the enterobacterial community within the medfly’s gut may have an indirect contribution to host fitness by preventing the establishment or proliferation of pathogenic bacteria. Serratia was a pathogen of other insects (Grimont and Grimont 1978), while our data showed that Serratia was the most common in three samples and did not have harmful effect on the oriental fruit fly’s health. Reports from Lloyd et al. (1986) have also revealed a similar phenomenon. These data indicated that Serratia may be present as symbionts in some fruit flies. More interestingly, Drew and Lloyd (1987) reported that some of enterobacterial species possibly played a role in stimulating the compounds that attract the foraging flies to the host fruit tree, which may be of agricultural significance.
Firmicutes was the major component in the gut of the FC B. dorsalis population. Among 35 clones of Firmicutes, eight clones showed high similarity with published sequences from the Lactococcus. This bacterium (Lactococcus) was also found in the Mexican fruit flies (Kuzina et al. 2001) and could stimulate the weak positive effects on immunity of Drosophila melanogaster (Burger et al. 2007). Thus, these clones may live commensally in the host gut and prevent the proliferation of pathogenic bacteria. In addition, Actinobacteria were the major component of LR population. Some species of Actinobacteria could generate biologically active secondary metabolites that degrade harmful substances (Hentschel et al. 2002). Therefore, we could draw such hypothesis that these clones in the gut of oriental fruit fly may play important roles in assisting their host to degrade toxic compounds.
Flavobacteria, Deltaproteobacteria, Bacteroidetes and Alphaproteobacteria appeared in various libraries with a few clones. Flavobacteria and Deltaproteobacteria were found in the libraries of FC and LR but were absent from LSSR library. Bacteroidetes and Alphaproteobacteria were exclusively observed in the sequences of the FC population, suggesting that they might have been recruited from the external environment. Their typical representatives were uncultured Dysgonomonas sp. (GQ891774.1) and Ag. tumefaciens (DQ466576.1), respectively. Many studies have shown that Dysgonomonas ssp. was anaerobic gram-negative cocci that could produce acid by fermenting monosaccharide and disaccharide (Hofstad et al. 2000) and Ag. tumefaciens existed widely in the soil or the surface of plant roots.
According to the analysis of diversity indices, phylotype composition and phylogenetic distribution of the 16S rDNA clones, we found that the bacterial communities of three libraries were significantly different from each other. The LSSR library was the least diverse and had fewer phylotypes in common with the other libraries, whereas the libraries of LR and FC were relatively more diverse. Reasons for the variations could be related to the oriental fruit flies living environment or food sources. Under laboratory conditions, the oriental fruit flies were reared in sanitary and controlled conditions, whereas reverse was the case under field conditions. LR and LSSR oriental fruit fly populations were fed upon mixture of sucrose and yeast powder and 20% sucrose sterile solution, respectively, while FC adult oriental fruit flies fed on uncontrolled natural diet. Thus, FC oriental fruit flies had more chances of having diverse gut flora as was observed in this study. Similar results were found in the mosquito adult midgut, where the bacterial community was variable between samples and changed with external factors, such as feeding and surroundings (Rani et al. 2009). Considerable impact of diet on the composition of the gut microbiota was also described among earthworms (Knapp et al. 2009), gypsy moth (Broderick et al. 2004) and wasps (Reeson et al. 2003).
In ecological aspects, we speculated that this fly-associated enterobacterial community was vertically transmitted from the female parent to its offspring, because enterobacteria were also the dominant community in the larva and pupae of oriental fruit flies (data was not shown). Similar results were found in the Mediterranean fruit flies midgut, where enterobacterial community may contribute to the fly’s nitrogen and carbon metabolism, affecting its development and ultimately fitness (Behar et al. 2008a). Some other potential function between the life cycle of the fly and its microbiota should be further studied.
In summary, our study unravelled the composition and diversity of the bacterial community in the gut of the oriental fruit fly by PCR-DGGE and compared the diversity of the harboured bacterial communities from the libraries of the LR, LSSR and FC populations. Among microbial communities, the most representative species were Klebsiella, Citrobacter, Enterobacter, Pectobacterium and Serratia and were found in three libraries. Further studies will be focused on identifying the function of every representative species and establishing whether these species could play important roles in the future as biocontrol agents.
The authors are thankful to Dr Akinkurolere R.O. of Shanghai Institute of Plant Physiology and Ecology, Shanghai Institute for biological Sciences, Chinese Academy of Science, China, for the language modification. This work is supported by the earmarked fund for Modern Agro-industry Technology Research System (MATS) and Special Fund for Agro-scientific Research in the Public Interest (no. 200903047).