Consumption of dietary sugar by gut bacteria determines Drosophila lipid content

Abstract

Gut microorganisms are essential for the nutritional health of many animals, but the underlying mechanisms are poorly understood. This study investigated how lipid accumulation by adult Drosophila melanogaster is reduced in flies associated with the bacterium Acetobacter tropicalis which displays oral–faecal cycling between the gut and food. We demonstrate that the lower lipid content of A. tropicalis-colonized flies relative to bacteria-free flies is linked with a parallel bacterial-mediated reduction in food glucose content; and can be accounted for quantitatively by the amount of glucose acquired by the flies, as determined from the feeding rate and assimilation efficiency of bacteria-free and A. tropicalis-colonized flies. We recommend that nutritional studies on Drosophila include empirical quantification of food nutrient content, to account for likely microbial-mediated effects on diet composition. More broadly, this study demonstrates that selective consumption of dietary constituents by microorganisms can alter the nutritional balance of food and, thereby, influence the nutritional status of the animal host.

1. Background

Animal nutritional health is strongly influenced by the metabolic traits of associated microorganisms located in the gut lumen or other tissues of their animal host. These microorganisms can be a source of nutrients (e.g. B vitamins, essential amino acids) and use dietary constituents, including complex macromolecules intractable to the animal digestive enzymes; and they may also modify the animal regulatory networks controlling nutrient allocation and feeding [1,2]. However, the processes underlying microbial impacts on nutrient acquisition and allocation by animals are poorly understood.

An experimentally tractable system to investigate animal–microbial nutritional interactions is provided by Drosophila melanogaster and bacterial populations that cycle between the gut and external environment [3] including taxa which reduce fly lipid and glucose levels [4,5].

These nutritional effects have been attributed to bacterial modulation of insulin signalling in the fly and consequent shift in nutrient allocation patterns [4,6]. Recently, bacterial genes functioning in sugar catabolism have also been implicated in microbial-mediated reduction of fly lipid content [7], but the underlying processes were not investigated. In this study, we demonstrate that gut-associated bacteria consume dietary sugar, with major implications for the nutrient balance of the diet and host nutrition.

2. Material and methods

The Wolbachia-free Drosophila melanogaster strain Canton S was maintained at 25°C with 12 L : 12 D cycle on food comprising 10% yeast, 10% glucose and 1.2% agar, with 0.04% phosphoric acid and 0.42% propionic acid preservatives. For experiments (figure 1a), Drosophila eggs were surface-sterilized in 0.6% hypochlorite [5] and reared to adulthood on sterile food in 50 ml Falcon tubes. One day after eclosion (day-0 of the experiment), 12 replicate groups of 60 male flies were transferred aseptically to ca 100 ml containers (Kendall, #8889207026) containing 10 ml sterile food or sterile food inoculated 24 h previously with 5 × 10⁶ cells of either Acetobacter tropicalis DmCS_006 or Lactobacillus brevis DmCS_003 [5]. The test diets contained 10 or 2.5% glucose with 10% yeast. The microbiological status of flies on day-6 of the experiment was verified by plating fly homogenates (three flies per sample) on 1.2% MRS agar and incubating for 48 h at 30°C under aerobic condition for Acetobacter or micro-aerobic condition with CO₂ for Lactobacillus [5]. Any experiment in which bacteria were detected in the axenic flies was excluded.

Figure 1.

Bacteria-dependent Drosophila nutrition. (a) Experimental design. (b) Triglyceride (TAG) and glucose contents of flies (Ax: axenic; Lb: with Lactobacillus brevis; At: with Acetobacter tropicalis). (c) Glucose content of food at day-6. (d) Relationship between glucose content of food (x-axis) and TAG content of flies (y-axis) on day-6: mean ± s.e. (12 replicates) for each of three experiments (1–3) are shown. (e) Glucose assimilation by Ax-flies and At-flies (mean ± s.e.). Treatments with different letters denote significant differences.

For Drosophila nutritional assays, replicate groups of five male flies were homogenized in 125 µl ice-cold TET buffer (10 mM Tris pH8, 1 mM EDTA, 0.1% Triton X-100) with approximately 100 µl lysis matrix D (MP Biomedicals) and shaken for 45 s in a FastPrep-24 homogenizer (MP Biomedicals). After centrifugation of the homogenate for 3 min at 20 000g, the protein content of 10 µl supernatant was determined by the Bio-Rad DC kit according to manufacturer's protocol, and 50 µl supernatant was incubated at 72°C for 30 min to inactivate endogenous enzymes, flash frozen, stored at −80°C and subsequently used for quantification of triglyceride (TAG) and glucose, following the manufacturers' instructions (Sigma F6428 and GAGO-20, respectively). TAG and glucose contents were normalized to protein. (The protein content per unit weight does not vary with treatment (JH Huang 2014, unpublished data).) In parallel, eight food samples (each ca 10 mg) were taken from each vial in a sterile pipette tip and weighed to the nearest microgram using a Mettler-Toledo microbalance. Four samples were lyophilized at −80°C for dry weight determination, and the remainder were processed (as above) for nutrient content.

Glucose assimilation by Drosophila was determined using the capillary feeder system [8] with modification. At day-6, six replicate groups of five male flies per treatment were transferred to sterile 1.5 ml microcentrifuge tubes plugged with a sterile needle (BD, #305129) and provided with liquid food (10% yeast extract (Sigma, #Y1625) and 10 or 2.5% glucose) in a 5 µl microcapillary tube (WPI, #1B100-4) inserted into the plug. Preliminary experiments using flies colonized with green fluorescent protein-labelled A. tropicalis confirmed that bacteria are not regurgitated into the capillary and, therefore, that the food ingested is not modified by bacterial activity. The volume of food consumed by the flies was quantified from the difference between the change in height of liquid food in the microcapillary for tubes containing and lacking flies, from which the amount of glucose ingested (I) was calculated. The flies and microcapillary were then removed from the tube, and replaced by 50 µl filter-sterilized water. After vortexing the tubes for 30 s to disrupt the fly faeces, the glucose content of 15 µl sample per tube was determined as above, and corrected to the total volume to give the total glucose egested from flies (E) in each tube (fly-free tubes had undetectable glucose content). Supplementary experiments (electronic supplementary material, text) revealed that 79% of egesta is recovered by this method. The glucose assimilation efficiency (AE) was calculated as: (I – (E/0.79))/I.

Statistical analyses were performed using R, v. 3.0.2: t-test to compare the nutritional indices of the flies, two-way ANOVA for food composition and glucose assimilation experiments with Tukey's post hoc test to test for individual differences.

3. Results

Newly emerged male flies reared for 5 days on 10% glucose diet under axenic conditions (Ax) or in mono-association with Acetobactor tropicalis (At) or Lactobacillus brevis (Lb) displayed a net increase in TAG content, but the magnitude of the increase was greater for the Ax-flies and Lb-flies than the At-flies (figure 1b, electronic supplementary material, data S1a). The Ax-flies, but not the Lb-flies or At-flies, also increased significantly in glucose content (figure 1b). This nutritional response matches data obtained previously for Drosophila administered bacteria as hatching larvae [5], showing that the effect of the bacteria on adult Drosophila nutrition is not dictated by pre-adult processes.

The protein content and water content of the food did not differ between vials containing and lacking bacteria or flies (electronic supplementary material, figure S1), but the glucose content of vials with A. tropicalis was significantly reduced by 75%, irrespective of the presence of flies (figure 1c; electronic supplementary material, data S1b).

We hypothesized that the reduced glucose content of A. tropicalis-colonized food accounted for the depressed TAG content of At-flies relative to Ax-flies. To test this hypothesis, we reared flies for 5 days on sterile or A. tropicalis-colonized food containing 2.5 or 10% glucose. The TAG content of both Ax-flies and At-flies varied with dietary glucose content. Consistent with the data in figure 1b, the 10% glucose-food in vials administered A. tropicalis had an empirically determined glucose content of 2.5% (t₄ = 0.255, p = 0.812) (figure 1d, electronic supplementary material, data S1c). The TAG content of At-flies on this food did not differ significantly from Ax-flies on 2.5% glucose diet (t₄ = 0.0566, p = 0.96). These data indicate that A. tropicalis-mediated reduction in dietary glucose contributes to the reduced TAG content of At-flies, with the prediction that At-flies ingest and assimilate less glucose than Ax-flies.

To test whether glucose uptake by the flies can account for the difference in lipid levels between At-flies and Ax-flies, we raised flies for 5 days on 10% glucose diet, and then administered either 2.5 or 10% glucose solution via capillary tubes for 24 h. Two-way ANOVA yielded a significant interaction term between fly treatment and glucose concentration, with higher glucose assimilation by flies feeding from 10% than 2.5% glucose solution, and by At-flies than Ax-flies on the 10% glucose solution only (figure 1e; electronic supplementary material table S1 and data S1d). Noting that At-flies reared on 10% glucose diet reduced the food glucose content to 2.5% but Ax-flies did not modify the food glucose content (figure 1d), we compared glucose assimilation by Ax-flies feeding from 10% glucose solution and At-flies feeding from 2.5% glucose solution to test the prediction that At-flies ingest and assimilate less glucose than Ax-flies. Although the At-flies feeding from 2.5% glucose solution displayed significantly higher feeding rate and assimilation efficiency, they assimilated two-times less glucose than Ax-flies on 10% glucose. If At-flies and Ax-flies allocate the same proportion of assimilated glucose to lipid, the Ax-flies would have twice the lipid content of At-flies. This value is closely similar to the observed 3.4-fold difference in lipid accumulation (from day-1 to day-6) between Ax-flies and At-flies (figure 1b), indicating that bacterial modification of dietary glucose content can account quantitatively for the microbiota-dependent effect on TAG content of Drosophila.

4. Discussion

This study demonstrates that consumption of dietary sugar by A. tropicalis provides a quantitatively sufficient explanation for bacterial-mediated reduction in the TAG content of adult Drosophila on diets containing 10% sugar. This effect is likely advantageous to Drosophila feeding on many natural fruits (sugar content 2.5–16% (http://ndb.nal.usda.gov/ndb/foods)) because Drosophila reared on sugar-rich diets display reduced feeding rates, resulting in inadequate acquisition of other limiting nutrients, as well as hyperlipidaemia, hyperglycaemia and insulin resistance [9,10]. Our results are also relevant to the use of Drosophila in nutritional research, especially as a biomedical model for metabolic diseases, including obesity [9]. These experiments should include empirical determination of the nutrient content of food, to account for likely microbial-mediated changes to diet composition.

Although our data indicate that microbial modification of the relative composition of nutrients is mediated primarily by bacteria external to the gut of Drosophila (figure 1d), this interaction may also occur within the gut of many animals. For example, the ileum of humans and other mammals contains a substantial microbiota capable of using sugars and other nutrients generated by digestion [11]. There is increasing recognition that the relative abundance of dietary nutrients is an important determinant of animal nutritional health, partly because an animal can display nutritional deficiencies associated with processing of nutritional classes in dietary excess [12]. These considerations raise the possibility that microbial consumption of certain nutrients, such as sugars, may be advantageous to the host on nutritionally unbalanced diets.

Several lines of evidence suggest that bacterial consumption of dietary sugar, identified here, is one of multiple processes determining microbiota-dependent nutrition of Drosophila. The importance of complementary post-ingestive processes is indicated by the significantly reduced glucose assimilation efficiency in axenic flies (figure 1e), possibly determined by an overall difference in gut morphology or assimilatory function [4,13]. Furthermore, microbiota-dependent TAG content varies significantly among Drosophila genotypes, and many of the associated genetic variants function in cell signalling, likely including microbiota-mediated regulation of nutrient allocation patterns [14]. A key topic for future research is to understand how the effects of gut microorganisms on nutrient acquisition and nutrient allocation are integrated. As this study illustrates, Drosophila is a very tractable system for this research.

Data accessibility

The datasets supporting this article have been uploaded as part of the electronic supplementary material.

Authors' contributions

J.H.H. and A.E.D. conceived the study. J.H.H. conducted the experiments and wrote the manuscript. A.E.D. commented on manuscript drafts. Both authors approved the final manuscript.

Competing interests

The authors declare no competing financial interests.

Funding

This work was supported by NIH grant no. R01GM095372.