Findings

The identification of bacteria in mosquito guts has relied on both culture-dependent and culture-independent techniques [1–3]. Molecular techniques for bacterial identification have received particular attention because they are more rapid than traditional culture methods and in addition can detect bacteria that cannot be cultured. Culture independent methods have mainly been based on the amplification of the 16S rRNA genes by PCR, followed by the identification of the amplified genes through nucleotide sequence comparisons [4].

The checkerboard DNA-DNA hybridization technique [5–8] is a semi-quantitative technique that has been extensively employed in odontology to detect and quantify a variety of bacterial species in dental samples and allows the simultaneous analysis of a large number of DNA samples against a range of DNA probes from different bacterial species on a single support membrane [8]. Here we have tested if this technique is suitable to detect and estimate the number of bacteria in total DNA samples extracted from both whole Aedes aegypti and from dissected A. aegypti midguts. In addition, we have also tested if we could detect and estimate the numbers of A. aegypti midgut-associated bacteria species in the midgut of other insect species.

In our experiments we employed a modified version [9] of the original DNA-DNA hybridization technique [8] (Additional file 1). As probes we used whole genomic DNA extracted from four bacterial species. Serratia sp. (FJ372764), Asaia sp. (FJ372770) and Klebsiella sp. (FJ372760) were isolated from laboratory-bred A. aegypti[1, 2]. Chryseobacterium sp. (EU169680.1) was isolated from wild-caught A. aegypti.

The results obtained after the hybridization of the phosphatase alkaline-labeled bacterial probes with defined amounts of total genomic DNA extracted from each bacterial species are shown in Figure 1. As can be observed, signals of increasing intensity are observed after the hybridization of defined amounts of the Asaia sp. and Serratia sp. probes with DNA amounts equivalent to 1 × 105, 5 × 105 and 1 × 106 cells of these two species (Figure 1A). Similar results are observed after the hybridization of the Klebsiella sp. and Chryseobacterium sp. probes with DNA amounts equivalent to 1 × 105, 5 × 105, 1 × 106 and 5 × 106 cells of these two species (Figure 1B). Together, these results reveal that the intensity of the signals is proportional to the amounts of DNA immobilized on the membranes, and further show that the genomic probes are specific and only detect the corresponding genomic DNAs. The sensitivity of our protocol, which enabled the detection of DNA amounts ranging from 105 to 106 cells, is similar to that described both in the original checkerboard DNA-DNA hybridization protocol [8] and in the modified versions [5, 9].

Figure 1
figure 1

Specificity of the genomic DNA probes. Vertical lanes contained genomic DNA amounts equivalent to different numbers of cells of each tested species (105, 5 × 105, 106, 5 × 106, as indicated). The samples labeled "Mix" are standards that contain DNA amounts equivalent to 105, 5 × 105, 106, 5 × 106, cells of each tested species. (A) The horizontal rows contain either Asaia sp. or Serratia sp. genomic DNA probes previously diluted in hybridization buffer, as indicated, and (B) the horizontal rows contain either Chryseobacterium sp. or Klebsiella sp. genomic DNA probes previously diluted in hybridization buffer, as indicated. A signal at the intersection of the horizontal and vertical lanes indicates the presence of a species.

Full size image

To test if this technique could detect the presence of bacterial species in A. aegypti samples, a membrane containing total genomic DNA extracted from single whole fourth instar larvae, old pupae, sucrose-fed adults and total genomic DNA extracted from dissected fourth instar larval and sucrose-fed adult midguts was hybridized to the four bacterial probes (Figure 2A). Different amounts of bacterial cells were present in the different A. aegypti samples (Table 1). Asaia sp., Klebsiella sp. and Serratia sp. were present in amounts of > 105 and < 6 × 105 cells in whole larvae, old pupae and adults and in larval midguts, with the exception of one whole larva (L2) and one pupa (P2), in which amounts < 105 cells of both Asaia sp. and Klebsiella sp. cells were detected (Table 1). The detection of Asaia sp. and Serratia sp. in larvae, pupae and adults of A. aegypti is consistent with previous studies that showed strong interactions between these bacteria and mosquito species [1, 10, 11]. In addition, even though a statistical analysis could not be performed due to the small number of samples investigated, our experiments suggest the presence of generally higher amounts of cells in A. aegypti larvae (whole larva L1; midguts L1 and L2) as compared to whole pupae (Figure 2A and Table 1). These results corroborate with other studies showing a reduction in bacterial numbers after the transition from the last feeding larval stage to the pupal stage [12]. On the other hand, lower amounts (<105 cells) of Chryseobacterium sp. cells were detected in whole larvae, pupae, adults and in larval midguts. These results might explain why this bacterial genus has not been identified in A. aegypti[1, 2], despite its high prevalence in Anopheles gambiae[13]. Finally, in adult midguts the investigated species were either not detected or detected at counts < 105 (Table 1). The higher number of bacterial cells detected in whole adults as compared to that observed in adult midguts could be attributed to bacterial colonization of other A. aegypti tissues as has been demonstrated for both A. aegypti and Anopheles stephensi[10, 14].

Figure 2
figure 2

A. aegypti midgut associated bacteria are detected both in whole animals and in dissected midguts. A. Analysis of DNA extracted from A. aegypti. Vertical lanes contain genomic DNA extracted from single whole fourth instar larvae (L1, L2), single whole old pupae (P1, P2), single whole sucrose-fed adult females (A1, A2), single midguts dissected from fourth instar larvae (L1, L2) and single midguts dissected from sucrose-fed adult females (A1, A2, A3). B. Analysis of DNA extracted from single midguts of L. longipalpis (Ll), D. melanogaster (Dm), B. hygida (Bh) and A. mellifera (Am). The horizontal rows contained the indicated DNA probes diluted in hybridization buffer, as indicated in the right hand side. The samples labeled "Mix" contain amounts of DNA equivalent to 105, 5 × 105 and 106 cells of each tested species.

Full size image
Table 1 Estimated numbers of bacterial cells in whole animals and dissected midguts
Full size table

Our results show that the checkerboard DNA-DNA hybridization technique can be employed to detect the presence of bacterial species known to be associated with A. aegypti in A. aegypti samples. This technique reveals differences in the counts of bacteria present in distinct life stages and is sensitive enough to detect differences in the amount of bacterial cells amongst individual samples [for example, Figure 2A, whole larvae (L1 and L2) hybridized to the Asaia sp. probe]. Overall, our results demonstrate that the checkerboard DNA-DNA hybridization is a suitable technique for routine investigation of mosquito samples.

The presence of these four bacterial species was also investigated in midguts dissected from another insect vector, Lutzomyia longipalpis, and from three other insect species Drosophila melanogaster, Bradysia hygida and Apis mellifera (Figure 2B, Table 1). Klebsiella sp. and Serratia sp. were both detected in all four insect species tested. Asaia sp. cells were detected in D. melanogaster, A. mellifera and B. hygida. Chryseobacterium sp. was the only bacterial species not detected in this group of insects. Klebsiella sp. and Serratia sp. have been previously reported in D. melanogaster, A. mellifera and L. longipalpis[15–18]. In addition, our results revealed the presence of A. aegypti midgut-associated bacteria species in the midgut of B. hygida, an insect species in which the indigenous microbiota has not previously been characterized.

The use of the checkerboard DNA-DNA hybridization technique to detect and estimate bacteria from insects is appealing since it can contribute to the characterization of insect microbiota without the need of employing culture dependent methods that are both laborious and time consuming. Sample preparation is simple, which enables the rapid and simultaneous investigation of numerous samples collected from distinct populations. In addition, this method has the sensitivity to detect bacteria in single individuals at different developmental stages (larval, pupal), as well as in a single organ such as the midgut, and therefore, can be employed to determine if there are differences amongst individuals in a single population. Finally, the use of this technique can contribute to the characterization of the microbial ecology associated with mosquitoes, elucidate intrinsic and extrinsic factors that influence bacterial composition and identify the bacteria that are implicated in vectorial capacity differences between mosquito populations.