Control experiments expressing GFP rather than G-CaMP in PERin neurons showed no fluorescent changes upon movement, showing that responses are not motion artifacts ( Figure 5B). Taken together, these experiments argue that PERin is activated upon movement, likely by mechanosensory inputs from multiple legs. If movement of the legs activates

PERin to inhibit proboscis AZD6244 ic50 extension, then one prediction would be that removing leg inhibition would promote extension and that this would require PERin. Flies whose legs were either removed (stumps) or immobilized with wax (wax) showed increased spontaneous proboscis extension, demonstrating that leg inputs inhibit extension (Figures 6A and 6B). Extensions were further enhanced in E564-Gal4, UAS-Shits flies, suggesting that tonic activity in PERin or nonleg inputs may also inhibit extension. Importantly, activation of PERin neurons with dTRPA1 in flies with stumps or immobilized legs prevented the increased spontaneous proboscis extension, suggesting that PERin neurons act downstream of leg inputs to inhibit extension ( Figures 6C and 6D). These studies suggest

that PERin neurons function to inhibit extension B-Raf inhibition while the animal is participating in other behaviors, such as locomotion. As PERin promotes behavioral exclusivity by altering the threshold for feeding initiation in response to mechanosensory-driven behaviors, we hypothesized that commitment to one behavior might more generally prevent other behaviors. Because E564-Gal4; UAS-Kir2.1, tub-Gal80ts flies display constitutive proboscis extension, we wondered whether engagement in this behavior might alter the probability of other behaviors. To test this, we monitored the activity of E564-Gal4; UAS-Kir2.1, tub-Gal80ts flies in a closed arena. Control flies, as well as E564-Gal4; UAS-Kir2.1, tub-Gal80ts flies not expressing Kir2.1, showed robust walking activity, whereas flies Bosutinib (SKI-606) expressing Kir2.1 in E564 neurons had greatly reduced activity, with some flies not taking a single step in the 60 s assayed ( Figures 7A and 7B). All flies were able to move when

presented with a startle stimulus. To test whether the movement impairment was a consequence of silencing PERin, we generated mosaic animals in which Kir2.1 and mCD8-GFP were expressed in subsets of E564 neurons, screened for constitutive proboscis extension, and assayed the extenders and nonextenders for movement (Figures 7A and 7B). Flies with extended proboscises displayed impaired locomotion. To ensure that the locomotion defect was a result of inactivating PERin, we screened mosaic animals for locomotor defects and determined the frequency distribution of neural classes in flies with normal locomotion (>250 mm/min traveled) or impaired locomotion (<200 mm/min traveled). PERin was enriched in flies with locomotor defects and no other cell-type correlated with locomotor defects (Figure 7C).