Diet influences host–microbiota associations in Drosophila

Gut microbes were previously suggested to influence mate preference in Drosophila melanogaster (1). Mate selectivity depended on the microbiota associated with flies after prior generations were maintained on different diets [cornmeal–molasses–yeast (CMY) versus starch] (1). Subsequent studies attempted to repeat these findings with contrasting success (2, 3). We suggest that a nonstandardized, transient microbiota—and how it is influenced by diet and dietary additives—might account for the conflicting results.

A point of contention between the studies (1–3) is that a fungicide used in fly media, methylparaben (mp)—also known as Tegosept or Nipagin M—might affect bacterial growth and thus microbiota composition (4, 5). Although Leftwich et al. (4) argued that moderate mp (up to 0.3%) should not alter microbiota, a previous study suggested that high mp levels (∼0.5%) can impact microbiota diversity (6). How mp concentration affects individual, fly-associated microbes on fly media has not been systematically addressed.

We found that using mp over 0.1% in fly medium severely restricts growth of some yeast and Acetobacter species, while the Lactobacilli tested were less affected even at 0.3%, the concentration at which mp solubility is compromised (Fig. 1). Differences in mp concentration in the CMY (0.1% mp) and starch (0% mp) diets, or in the medium used to maintain flies before experiments (0.3% mp), may therefore indeed perturb microbiota composition (1, 3, 5).

Fig. 1.

Methylparaben (mp) adversely affects bacterial and yeast growth. (A) Growth of microorganisms on fly food supplemented with increasing concentrations of mp. Colony diameters were measured for different microorganisms isolated from Drosophila (yeasts, Issatchenkia orientalis and Saccharomyces cerevisiae; Gram-negative bacteria, Acetobacter indonesiensis SB003 and Acetobacter cerevisiae; Gram-positive bacteria, Lactobacillus plantarum SB001, L. plantarum WF, and L. plantarum WCFS1), reflecting their growth over 2 (yeasts) or 3 (bacteria) days at 30 °C. Microorganisms (∼100 CFUs) were plated on Petri dishes (100 × 15 mm) containing CMYE medium (8.6% cornmeal, 5% sucrose, 0.5% yeast extract, and 0.5% agar) supplemented with 0.75% (vol/vol) ethanol and mp [0%, 0.05%, 0.1%, 0.15%, or 0.3% (wt/vol)]. Data show mean growth (±SEM) of 30 CFUs (independent duplicates) normalized to their mean growth in CMYE + 0% mp. (B) Number of microorganisms on spent fly food (CMYE with 0.3% mp or vehicle control). Microbes were grown overnight in De Man, Rogosa, and Sharpe (MRS) broth (1.5 mL medium in a 15-mL Corning tube) at either room temperature without shaking (I. orientalis and S. cerevisiae) or at 30 °C with shaking at 220 rpm in a New Brunswick Innova 44 shaker (A. indonesiensis SB003 and L. plantarum SB001). Axenic adults (3–5 d old, 20 flies per vial) were transferred to food vials containing ∼5 × 10⁵ CFUs (yeasts) or ∼5 × 10⁷ CFUs (bacteria). Flies were transferred to fresh sterile food at days 7 and 10, and microbes were collected from food surfaces at day 13. Dilution plating on MRS agar was used to determine CFUs. Data show mean CFUs of three vials (±SEM).

Our observation that preservatives impact microbe growth brings up an important point for Drosophila studies. In the laboratory, the fly gut microbiota is generally transient, meaning that microbes are ingested from the food and only associated with flies for the duration of meal passage. The reduction of microbial growth by mp likely alters the composition or quantity of bacteria and yeast in the laboratory fly gut. As Leftwich et al. (3) pointed out, a flexible fly–microbe symbiosis is not expected to support the evolution of host behavioral shifts.

Since microbes can contribute macronutrients and micronutrients to flies (7, 8), microbiota composition or quantity might influence mate selectivity by affecting nutrition and therefore host size, which has previously been associated with mate choice (9). This is consistent with the hypothesis that microbe-influenced mating preference stems from host vigor (2), which we suggest is a potential by-product of diets and associated microbes. Indeed, diet-induced mating preference was rapidly detected after only one generation (1), suggesting the possibility that immediate nutritional influences of microbes are the underlying cause. Previous studies did not quantify the number of microbes associated with flies, making it difficult to interpret the reported compositional differences based on relative abundances alone.

Additionally, the mate choice tests discussed here may need to be reexamined using wild isolates of the fly microbiota without mp. Accumulating evidence suggests that flies in the wild indeed have a more consistent or stably associated microbiota (10), which could considerably change the microbial load. Future behavioral investigations should consider differences in food recipes, which may influence the nutritional value of both the diet and associated microbes.