Is the refrain that “my microbes make me do it” true? Scientific reviews and the popular press often report that microbes can influence many aspects of the behaviour of their healthy animal or human hosts, from cognition to social interactions to emotional state1. If this is true, perhaps future microbial-based therapies might be used to improve mental health. Yet the evidence for most of these claims of microbial effects is limited. Experiments are urgently needed that definitively test whether bacteria can have a causal role in behaviour, and that identify the underlying mechanisms. Writing in Nature, Schretter et al.2 provide a superb example of rigorous scientific analysis, in which they demonstrate that the walking activity of the fruit fly Drosophila melanogaster can be affected by a specific gut bacterium. The authors identify a bacterial enzyme that mediates this effect, and establish aspects of the mechanism by which the fruit fly responds to the bacterium.
Schretter et al. used standard techniques to analyse the bacterial residents of the gut. The authors compared walking activity in flies harbouring their natural gut microbes (the microbiota) and flies that had been treated to eliminate the gut bacteria, and they observed that the treated flies were hyperactive in comparison to the others. These hyperactive flies walked faster and for longer than the others, but their daily (circadian) rhythms of activity and sleep were not perturbed. To determine the microbes associated with this effect, Schretter et al. supplied the hyperactive flies with various bacteria, and found that the bacterium Lactobacillus brevis restored walking activity to the level observed in flies that retained their complete microbiota.
There is a common expectation that gut microbes influence animal behaviour by producing small metabolites, including neurotransmitter molecules, which interact directly with the nervous system in the gut or that enter the bloodstream and from there reach the brain3,4. However, the bacterial product identified by Schretter and colleagues as involved in walking behaviour does not fit this paradigm. The authors provide persuasive evidence that the presence of the sugar-modifying enzyme xylose isomerase, which is produced by L. brevis, reduces locomotor activity in D. melanogaster (Fig. 1). Further experiments revealed that supplying xylose isomerase to flies whose bacteria had been eliminated was necessary and sufficient to modulate fly locomotion.
How does xylose isomerase cause D. melanogaster to slow down? The enzyme mediates the interconversion of certain sugar molecules — the change of glucose into fructose, for example. Schretter and colleagues found that the flies treated to remove their gut bacteria have a higher level of the sugar trehalose than do those that retain their usual microbiota. Perhaps this means that xylose isomerase decreases the availability of a glucose substrate needed for the synthesis of trehalose. The authors administered trehalose to flies that lacked gut bacteria and had been provided with xylose isomerase, and report that the trehalose treatment caused the flies’ walking speed to increase.
The authors proceeded to investigate the neural basis of the hyperactivity phenomenon. They used genetic approaches to activate neurons that regulate locomotion in D. melanogaster and the results focused their attention on a type of neuron called an octopaminergic neuron, which produces the neurotransmitter molecule octopamine. Schretter et al. found that the walking activity of flies that lacked gut bacteria but had been given xylose isomerase was increased by activation of the genes encoding enzymes needed for the synthesis of octopamine. Such an effect on locomotion was not observed for other neurotransmitters they tested. Furthermore, the authors observed that the flies with their natural microbiota and those that had been treated to remove gut bacteria but had received xylose isomerase both walked faster if they received octopamine.
Octopamine is a well-characterized regulator of locomotion in flies. In vertebrates, the neurotransmitter molecule noradrenaline, which is related in structure to octopamine, fulfils a similar role in promoting physical activity5. Work remains to be done to fill in the gaps in explaining how xylose isomerase affects the level of trehalose in the fruit fly and the activity of octopamine-producing neurons in the brain. Nevertheless, one key conclusion emerges: the effect of bacterial products on fly locomotion is mediated by the modulation of known circuits that control behaviour and not through previously unknown regulatory mechanisms.
Why does L. brevis make xylose isomerase? It should not be assumed that this is a specific adaptation to life in a D. melanogaster host. This bacterium is not specialized to exist only in the fruit fly gut. It maintains substantial free-living populations6 and is neither universally present nor abundant in D. melanogaster populations in the natural environment7. Xylose isomerase probably functions to increase the diversity of the carbon sources that L. brevis can exploit, as is the case for the many other bacteria that produce this enzyme. It would be interesting to learn the outcome of experiments comparing the abundance in D. melanogaster of resident wild-type L. brevis and of L. brevis mutants lacking xylose isomerase, to determine whether this enzyme enhances the fitness of the bacterium and whether any fitness effects depend on fly locomotor activity.
The most important question to ask next is whether the effect of L. brevis and xylose isomerase on the locomotor activity of D. melanogaster is relevant to animal behaviour in general, including that of humans and other mammals. As with many other discoveries first made in D. melanogaster8, perfect correspondence with mammalian systems is unlikely. Schretter and colleagues’ study does, however, alert microbiologists and those studying animal behaviour to pay attention to the enzymes of gut bacteria and their possible effects on sugar metabolism and on the neuronal circuits regulating walking activity.
Nature 563, 331-332 (2018)