Computational protocol: Bacterial Communities of Diverse Drosophila Species: Ecological Context of a Host–Microbe Model System

Similar protocols

Protocol publication

[…] Drosophila samples were collected with the help of many colleagues around the world (see Acknowledgments, and ). Flies were either washed in sterile water (to remove cuticular bacterial cells) or were dissected to obtain just their crops and digestive tracts. After DNA extraction, rDNA PCR amplification was done with bacterial specific primers. The 16S rDNA amplicons were cloned, transformed, and Sanger sequenced from both ends. For 50% of the clones, the two reads did not overlap and therefore a concatenated read was made by inserting gap characters in the space between the two reads.After all preliminary filtering, our dataset consisted of 3243 nearly full-length high quality sequences representing 39 host samples (which we refer to as libraries) ( and ). This dataset excluded 421 clones that were only sequenced from one end, 65 sequences with fewer than 300 non-gap characters, 76 sequences that were identified as chimeric, 9 that appeared to be chimeric based on conflicting taxonomy assignments of the 3′ and 5′ reads, 3 chloroplast sequences, and 351 sequences of likely endosymbionts such as Wolbachia and Spiroplasma (which will be addressed in a separate section). Because small sample sizes can lead to inaccurate diversity measures , two libraries containing a total of 28 sequences were removed completely. The 39 remaining libraries vary in size from 26 to 223 sequences, with an average of 83.2 ± 37.4. Most libraries (29 of 39) contain between 63 and 97 sequences. 20 libraries containing 1850 total sequences are from wild-caught hosts, while the remaining libraries and sequences came from laboratory samples and experiments. Full tables containing each library's identifier, size, the host species from which it was collected, location and date of collection, and other information are given in and .Clustering with mothur using the average neighbor algorithm with 0.03 cutoff (corresponding to 97% sequence similarity) creates 139 operational taxonomic units (OTUs), the largest of which contains 638 sequences. 66 OTUs are singletons (i.e., there is only a single sequence in the OTU) and 110 OTUs contain 10 or fewer sequences. The average OTU contains 23.3 sequences (standard deviation  = 78).Phylogenetic analysis was performed using FastTree . Included in this analysis (and many other comparisons throughout this study) were many previously identified Drosophila-associated bacteria , , , . [...] To estimate the role of host diet in shaping bacterial microbiome composition, we compared taxonomically diverse Drosophila species collected from different types of food sources. Our survey contains 11 samples of fruit-feeding flies and 6 samples of flower-feeders. UniFrac analysis , shows that flies subsisting on these two diets have significantly different bacterial microbiomes (p<0.01). One major difference is the absence of Acetobacteraceae in flower-feeding flies (). This may be due to the fact that Acetobacteraceae can thrive under the low pH and high ethanol conditions present in fermenting fruit. The same argument can be made for Lactobacillus, an acidophilic genus associated with high resource habitats, which is present at higher abundance in fruit-feeding flies (). In contrast, the genera Serratia and Pantoea (Enterobacteriaceae) are found in much higher proportions in flower-feeders (). Many of the largest OTUs are found only, or mainly, in association with one diet type (). Of the 14 largest OTUs, which contain 75% of all sequences, 10 derive >95% of their sequences from a single diet type (). In general, the difference between fruit- and flower-feeders is consistent and can be attributed to multiple host samples within each category. An exception to this pattern is Shigella, whose apparent abundance in fruit-feeding flies is due almost entirely to a single library (D. melanogaster from rotting grapes, Sample MAH) ().Similarities among the bacterial communities of wild populations and laboratory strains were summarized with PCA using UniFrac (). We find that the majority of fruit feeding flies occupy a distinct region within PC space, while the two mushroom feeders are mostly separated from the other samples. In congruence with the taxonomic similarity between the cactus feeding population and the fruit feeders, we find that the D. mojavensis sample clusters near the fruit associated flies.Some differences are also apparent within diet types. In particular, D. elegans was collected simultaneously from Alpinia and Brugmansia flowers (Samples ELA and ELD). These collections were made less than 10 meters apart and almost certainly represent a single fly population. Therefore, any differences in their bacterial communities are most likely due to the different food sources. We find that D. elegans collected on Alpinia has a much higher amount of Leuconostocaceae and Streptococcaceae (phylum Firmicutes), while those collected on Brugmansia are dominated by Enterobacteriaceae (phylum Proteobacteria) (). Alpinia-collected flies also show much higher bacterial microbiome diversity than Brugmansia-collected flies (Chao1  = 23 vs 7.5) (). Although it is possible that individual flies travel between host plants, these switches are clearly insufficient to overcome the effect of diet.Both mushroom-feeding populations were associated with a high amount of Lactobacillales, specifically D. falleni had 30% Vagococcus and Microdrosophila sp. had 57% Enterococcus (). D. falleni is also notable because its bacterial microbiome contains 16% each of both Bacillales and Burkholderiales, two orders that are otherwise rare in Drosophila bacterial microbiomes (). The mushroom-feeding species are also marked by relatively high community richness and diversity, especially compared to fruit-feeding Drosophila ( and ). The single cactus-associated population is very similar to many fruit feeders both in composition (84% Enterobacteriaceae Group Orbus) () and diversity (). […]

Pipeline specifications

Software tools mothur, FastTree, UniFrac
Applications Phylogenetics, 16S rRNA-seq analysis
Organisms Drosophila melanogaster, Bacteria