A novel assay reveals hygrotactic behavior in Drosophila

Abstract

Humidity is one of the most important factors that determines the geographical distribution and survival of terrestrial animals. The ability to detect variation in humidity is conserved across many species. Here, we established a novel behavioral assay that revealed the thirsty Drosophila exhibits strong hygrotactic behavior, and it can locate water by detecting humidity gradient. In addition, exposure to high levels of moisture was sufficient to elicit proboscis extension reflex behavior in thirsty flies. Furthermore, we found that the third antennal segment was necessary for hygrotactic behavior in thirsty flies, while arista was required for the avoidance of moist air in hydrated flies. These results indicated that two types of hygroreceptor cells exist in Drosophila: one located in the third antennal segment that mediates hygrotactic behavior in thirst status, and the other located in arista which is responsible for the aversive behavior toward moist air in hydration status. Using a neural silencing screen, we demonstrated that synaptic output from the mushroom body α/β surface and posterior neurons was required for both hygrotactic behavior and moisture-aversive behavior.

Fig 1. Hygrotactic behavior in thirsty Drosophila.

(A) Schematic diagram of the apparatus used in the hygrotactic behavior assay. (B) Flies dehydrated for 8 hours migrated toward and aggregated near the water source rapidly. (C) PER behavior induced by moist air in dehydrated wild type flies (female; eight-hour dehydration). Arrowheads indicate extended proboscises.

Fig 2. Quantification of hygrotactic behavior in thirsty Drosophila.

(A) Diagram for measuring aggregation value and hygrotaxis index in hygrotactic behavior. Aggregation values represent the aggregation strength at different time points during the test. NO_t denotes the number of flies within the red circle at the moment during the test; NO₀ is the number of flies in the red circle at the beginning of the test; NO_sum denotes the total number of flies in the test. Hygrotaxis index is used to represent the strength of hygrotactic behavior. (B) Plotting of the aggregation value as a function of time in wild type flies dehydrated for 8 hours. N = 12. (C) Hygrotaxis index of wild type flies dehydrated for 8 hours. ***, p < 0.001 (Student’s t test). N = 12. (D) The curve of hygrotaxis index vs. dehydration time in wild type flies. N = 12. (E) Relationship between hygrotaxis index and the shortest distance from flies to water source. Wild type flies were dehydrated for 10 hours before tests. N = 12. Data are presented as mean ± SEM.

Fig 3. Third antennal segments are required for hygrotactic behavior.

(A) Scanning electron microscope photograph showing the third antennal segment and arista of Drosophila. (B) The hygrotactic behavior was tested in wild type flies and the wild type flies that had removed aristae or both third antennal segments and aristae. Ablation of the third antennal segments and aristae abolished hygrotactic behavior, while removing the aristae alone did not affect hygrotactic behavior. All tested flies were dehydrated for 8 hours. (C) Ablation of the third antennal segments and aristae greatly reduced the hygrotaxis index in wild type flies dehydrated for 8 hours. N = 12. (D) Ablation of the third antennal segments and aristae significantly reduced moisture-induced PER rate in wild type flies dehydrated for 8 hours, while removing the aristae alone did not affect the PER rate. N = 8 trials with 50 flies per trial per condition. (E) Schematic diagram of the T-maze apparatus used in the humidity choice assay. Two tubes in the apparatus were filled with moist air (~99% RH) and dry air (~3% RH), respectively. Flies were placed between the two tubes, and allowed to make a choice between the two types of air. (F) The effect of sensory organ ablation on humidity choice behavior in wild type flies. Before dehydration, the intact wild type flies avoid moist air, while the wild type flies with the ablation of aristae or both the third antennal segments and aristae showed no bias towards dry or moist air. After dehydration for 8 hours, the intact flies and flies with aristae removed exhibited a strong preference for moist air, while flies with both third antennal segments and aristae removed showed no humidity preference. N = 15. NS, not significant (p > 0.05); ***, p < 0.001 (ANOVA with Tukey post hoc test). Data are presented as mean ± SEM.

Fig 4. MB α/βsp neurons are required for hygrosensation.

(A) Immunofluorescent detection of mCD8::GFP driven by NP3061-Gal4 and NP5286-Gal4 in the adult fly brain. Both Gal4 lines display specific GFP expression in MB α/βsp neurons, which had been characterized in a previous work [27]. Scale bar, 50 μm. (B) Inhibiting synaptic output from MB α/βsp neurons (NP3061-Gal4 > TNT, NP5286-Gal4 > TNT) impairs the hygrotactic behavior in flies dehydrated for 8 hours. N = 12. (C) Inhibiting synaptic output from MB α/βsp neurons significantly reduced the rate of moisture-induced PER in dehydrated flies. N = 8 trials with 50 flies per trial per genotype. (D) Blocking the activity of MB α/βsp neurons did not affect the water-drinking behavior of dehydrated flies. N = 8 trials with 50 flies per trial per genotype. (E) Blocking the activity of MB α/βsp neurons did not reduce weight loss during 8 hours of dehydration or water intake within 10 minutes of water-feeding following dehydration (flies carrying NP5286-Gal4 > TNT drank more water than wild type and control flies). Data represent the weight variations of 50 female flies. N = 12. (F) In T-maze humidity choice assays, the flies expressing TNT driven by NP3061-Gal4 or NP5286-Gal4 exhibited impaired avoidance behavior toward moist air before dehydration, and also showed reduced preference for moist air after eight-hour dehydration. N = 15. NS, not significant (p > 0.05); ***, p < 0.001 (ANOVA with Tukey post hoc test). Data are presented as mean ± SEM.