Is the Regulation of Insulin Signaling Multi-Organismal?

See allHide authors and affiliations

Sci. Signal.  13 Dec 2011:
Vol. 4, Issue 203, pp. pe46
DOI: 10.1126/scisignal.2002669

This article has a correction. Please see:


The sustained health of an individual animal depends on the composition and activities of its resident microbiota. A major challenge is to identify the processes by which the microbiota and animal interact, recognizing that this research should lead ultimately to novel strategies to promote human health. Drosophila is emerging as a tractable model system to investigate these interactions. New evidence reveals that the gut microbiota promotes insulin signaling in Drosophila, leading to increased growth and development rates. Different gut bacteria and mechanisms were implicated: acetic acid produced by Acetobacter pomorum and the promotion of amino acid nutrition by Lactobacillus plantarum. These findings raise the possibility that multiple bacterial effectors may interact with signaling networks to shape animal health.

Have we made a serious error of omission in our current understanding of the signaling networks that define the form and function of animals? Are we systematically underestimating the complexity of signaling by assuming that all components are of animal origin? These questions arise directly from two publications on the fruit fly Drosophila melanogaster (1, 2) that independently implicate the gut microbiota in the integration of cell growth and metabolism through the insulin-TOR (target of rapamycin) signaling cascades.

Drosophila is not only a superb genetic model for developmental and cell biology, it is also a tractable model for multi-organismal interactions between animals and their resident microbiota. The gut lumen of Drosophila is colonized by ~20 bacterial taxa—an order of magnitude lower than the 500 to 1000 taxa in the gastrointestinal tract of humans and other mammals (3, 4)—and is dominated by two bacterial groups that are readily culturable: the Gram-negative Acetobacter and related “acetic acid” bacteria (Proteobacteria) and the Gram-positive Lactobacillus (Firmicutes) (5, 6). Germ-free insects can be obtained by surface-sterilizing the eggs and rearing them under aseptic conditions, and associations with single bacterial taxa are generated by feeding the flies on the cultured bacteria. Germ-free Drosophila larvae tend to grow and develop more slowly than individuals bearing gut bacteria (conventional Drosophila), although the magnitude of the effect varies with diet (1, 2).

These experimental tools, combined with a simple but technically demanding genetic screen, enabled Shin et al. (1) to demonstrate that the gut microbiota modulates insulin signaling. The screen comprised a comparison between the growth of Drosophila larvae infected with a wild-type strain of the gut bacterium A. pomorum and each of ~3000 transposon mutants. It revealed that the beneficial effect of A. pomorum on Drosophila requires one metabolic complex: the pyrroloquinoline quinone–dependent alcohol dehydrogenase (PQQ-ADH)–dependent oxidative respiratory chain, which couples energy production with the fermentation of ethanol to acetic acid (7). The gut of Drosophila bearing the wild-type A. pomorum contains 25 to 30 pg of acetic acid, but acetic acid is undetectable in Drosophila infected with mutant bacteria lacking functional PQQ-ADH (1). The Drosophila with the mutant bacteria also displayed a decreased abundance of DILPs (Drosophila insulin-like peptides) that are homologous to the mammalian insulin, as well as multiple traits indicative of reduced insulin signaling, such as elevated sugar concentrations, depressed proliferation of intestinal stem cells, and reduced wing size and cell number. Two sets of analyses defined the bacterial effect on Drosophila insulin signaling more precisely. First, the phenotype of Drosophila infected with the mutant bacteria was recapitulated when Drosophila with genetically reduced insulin signaling were infected with wild-type bacteria. Second, the phenotype of larvae bearing wild-type bacteria could be rescued in larvae bearing the mutant bacteria, either by supplementing the medium with acetic acid or by using Drosophila genetically manipulated to overexpress DILPs. These experiments provide persuasive evidence that acetic acid produced by A. pomorum promotes insulin signaling in Drosophila.

Storelli et al. (2) also obtained evidence for reduced insulin signaling in germ-free Drosophila, providing assurance of the repeatability of the core conclusion of Shin et al. (1) in a different Drosophila strain and laboratory context. Nevertheless, the results of the two studies differ in detail, and the several discrepancies are informative (Fig. 1A). Each study identified a single bacterium that could recapitulate the effect of the full microbiota on insulin signaling and larval growth and development rates, but the bacterium was different. Storelli et al. obtained unambiguous effects with L. plantarum, which was one of four bacterial species derived from Drosophila guts that Shin et al. reported as supporting larval development at a rate significantly lower than that obtained with Acetobacter pomorum. This difference is important because it means that taxonomically divergent bacteria can affect insulin signaling in Drosophila. Lactobacillus is not known to release acetic acid but produces copious amounts of lactic acid (8). But we cannot infer acid production as a commonality between the two studies, because a strain of L. plantarum with minimal lactic acid production supports high development rates in Drosophila, indicative of “normal” insulin signaling (2). Instead, the key products of L. plantarum may be protein-derived branched-chain amino acids, which activate TOR signaling in the fat body and have downstream effects on DILP production (9).

Fig. 1

The impact of resident bacteria on insulin signaling in Drosophila. (A) Insulin signaling is promoted, with consequent enhanced larval developmental rates, by acetic acid produced by A. pomorum (1) and by L. plantarum (with proposed role of branched-chain amino acids, mediated by TOR signaling) (2). Dashed lines represent putative interactions. (B) Hypothesized role of bacterial products (A to E) as cues of resource availability, used by Drosophila to regulate developmental rates through insulin signaling, either by promoting (1, 2) or reducing Drosophila insulin signaling relative to germ-free animals. (C) Hypothesized conflict between the bacteria and animal over the setpoint of insulin signaling and correlated larval development rates, such that the realized setpoint for conventional Drosophila (bearing microbiota) is intermediate between the bacterial and animal optimum. The realized setpoint for germ-free Drosophila is lower than the optimal setpoint for the animal, which is calibrated constitutively to account for bacterial manipulation.


The key implication of these two studies is that at least two, and potentially multiple, bacterial products interact with Drosophila insulin signaling and the linked signaling networks that define nutrient sensing in the animal. This raises questions of mechanism. Which gut factor interacts with the bacterial products? How is the information conveyed to the signaling networks that integrate insulin, TOR, and other key components in nutrient sensing across multiple organs of the animal? There is also the issue of interactions. Multiple bacterial products may interact competitively, additively, or synergistically with Drosophila signaling networks, such that the impacts of A. pomorum, L. plantarum, and other bacteria differ between Drosophila bearing single species [as studied in (1) and (2)] and the full community of 20 to 30 bacterial taxa.

Not only is the mechanism unresolved, so is the ultimate explanation, “why?” Bacterial-mediated enhancement of larval growth and development rates is likely beneficial to Drosophila under natural conditions. The female flies deposit eggs onto the transient resource of rotting fruit, and larvae that develop fast are at a competitive advantage and are more likely to pupate before the fruit resource is exhausted (10). Why is the “desired” setpoint of the signaling cascades that control development rate dependent on input of bacterial products? There are two possibilities. Perhaps these bacterial products are cues that inform Drosophila of the likely abundance and persistence of resources in the rotting fruit, with the prediction that some bacterial products (indicative of a persistent resource or stressful conditions) may favor long developmental rates by suppressing insulin signaling (Fig. 1B). Alternatively, the difference in insulin signaling between conventional and germ-free larvae may be the evolutionary consequence of conflict between Drosophila and its microbiota. The optimal developmental rate may be lower for the animals, which are selected for lifetime reproductive success, than for the gut microbiota, selected to maximize dispersal through rapid larval development to adults that fly to (and defecate onto) multiple fruits. In this conflict scenario, the observed setpoint of insulin signaling is a titration between the high and low preferred setpoints of the bacteria and animal, respectively (Fig. 1C). If the intrinsic set point of the animal is calibrated constitutively to account for bacterial manipulation [as is likely because the bacteria are always present in naturally occurring Drosophila (6)], then the signaling would be depressed in the germ-free flies, which lack the manipulative up-regulation by the bacteria (Fig. 1C).

Associations with bacteria are common to animals and their protistan ancestors (11), and the signaling networks in animals presumably evolved in the context of cues and manipulative effectors produced by their associated bacteria. This raises the possibility that the modulation of Drosophila insulin signaling by bacterial products may be explicable not only in the context of fruit fly ecology but also in terms of the long evolutionary history of animal-microbial interactions. In other words, the fundamentals of the interactions between Drosophila and its gut microbiota could be general across the animals: set in deep evolutionary time and retained because their elimination would be disruptive to the complex, coevolved, multi-organismal networks. There is much to be done to understand the interactions between bacterial products and animal signaling networks, but one core point is becoming clear: bacterial intervention in animal signaling networks is part of how the resident microbiota keeps flies—and possibly us—healthy.

References and Notes

Funding: Supported by NIH grant 1R01GM095372-01, NSF grant IOS-0919765, and the Sarkaria Institute for Insect Physiology and Toxicology.
View Abstract

Navigate This Article