To test the response to visual light we followed a procedure by Vang et al. (2012). Briefly, flies were starved in vials containing 1.5% agar for 18 h. Then they were transferred to an empty vial by tapping, and the vial was connected to a 25 cm long test tube in a dark room 30 min prior to the test to adapt to darkness. The horizontal test tube was then gently pounded down to position the flies at one end, away from the light source. A perpendicular light source (15 cm distance) was placed at the other end of the test tube (Figure 2A). To start the test, the cold light source (Schott KL 1500 LCD, 15V, 150W, 3300K, position 5) was switched on thereby establishing a light gradient of about 90 lux at the nearest point to about 3 lux at the furthest point. Every minute in the 10 min test duration, the number of flies in each third of the test tube was scored. Data were expressed as attraction response index (RI%) corresponding to the percentage of flies in the segment closest to the light source relative to the total number of flies placed in the test tube. The test was repeated three times and data were plotted as mean ± SEM.

Response to volatile repellents was assayed as described in Vang et al. (2012). Briefly, flies were starved in vials containing 1.5% agar for 18 h before transferring them to an empty test vial 30 min prior to the test. To start the test, the vial was connected to a test tube containing 100 mM or 10 mM benzaldehyde in 1.5% agar at one end and the flies were gently tapped toward the repellent (Figure 2C). The number of flies in each third of the test tube was scored every minute for 10 min. Data were expressed as repulsion response index (RI%), corresponding to the percentage of flies in the segment furthest from the attractant relative to the total number of flies placed in the test tube. The test was repeated three times and data were plotted as mean ± SEM.

Response to non-volatile chemicals was tested as described by Vang et al. (2012). Briefly, flies were starved in vials containing 1.5% agar for 18 h before transferring them by tapping to an empty test vial 30 min prior to the test. The test was started by connecting the vial with a 25 cm long test tube containing 100 mM sucrose in 1.5% agar at one end and gently tapping the flies to the end of the empty vial (Figure 2E). Every minute for 10 min the number of flies in each third of the test tube was scored. Data were expressed as attraction response index (RI%), corresponding to the percentage of flies in the segment closest to the attractant relative to the total number of flies placed in the test tube. The test was repeated three times and data were plotted as mean ± SEM.

To evaluate the flies’ ability to display reflex-like response to attractive substances in the food and motivation to feed, we assayed the proboscis extension reflex. Thirty starved (24 h) female virgin flies of each genotype were presented with glass capillaries containing a solution of sucrose to the proboscis by making contact, and the number of proboscis extension responses was determined as described in Shiraiwa and Carlson (2007) and Qi et al. (2015). Presentation of stimulus was repeated in 10 trials per fly (n = 30/genotype). The experiment was performed once, and data were plotted as mean ± SD.

The startle-induced negative geotaxis assay was performed as described in Ismail et al. (2015) (Figure 3A). Flies of the indicated ages were transferred to climbing vials (two empty vials connected face to face) without anesthesia and left to acclimatize for 1 h. Testing was performed by tapping down the flies and counting the number of flies that climb past a mark on the vials (8 cm from the bottom) within 60 sec using a custom-made apparatus. In each experiment, the test was repeated 10 times (trials) with 1 min breaks in between. The results were analyzed as% of climbing flies at each time point (n = 5 batches of 10–12 flies/genotype/experiment). The experiment was repeated three times and data were plotted as mean ± SEM.

To assess walking behavior the exploratory walking assay was performed as previously described by Ismail et al. (2015) (Figure 3C). Briefly, female virgin flies were placed individually into the center of a 14.5 cm petri dish with a 1 cm square grid. The number of grid-line crossings during 1 min was scored and graphically presented (n > 40/genotype/experiment). The experiment was repeated three times and data were plotted as mean ± SEM.

Data from photo- and chemotaxis experiments were quantified and interpreted as response toward the non-volatile and volatile chemicals as well as light. To calculate the statistical significance of the experimental clustering (interaction trial × age × genotype effect), 3-way ANOVA analysis based on distance matrices (distances within all replicates versus distances between all replicates) was performed. Data from proboscis extension reflex (PER) experiments were presented as response frequency, i.e., percentage of responding flies per stimulation trial (total of 10) (Figure 2H). The experiment was performed once (n = 30/genotype) for each age (4 and 14 days). To estimate the statistical significance of the response frequency clustering (stimulation trial × genotype effect), 2-way ANOVA analysis based on distance matrices (distances within all replicates versus distances between all replicates) for each age timepoint (4 and 14 days) was performed.

Results obtained from the negative geotaxis assay were illustrated as climbing capability, and statistical analysis of differences between the two fly lines at the respective age and each given sequence cycle was performed by 3-way ANOVA based on distance matrices (distances within all replicates versus distances between all replicates). Exploratory walking behavior was scored as number of grid-line crossings per minute, and all data are graphically presented as individual values (n > 40/genotype/experiment, three experiments). Statistical analysis of explorative walking differences between the two fly lines at the respective age was performed by unpaired Student’s t-test.

To examine this idea, we performed various experiments to assess the behavioral response of the flies to sensory cues. We first tested response to light using a simple phototaxis assay (Vang et al., 2012). Mated female flies (4 or 14 days old) were placed in a 34 cm long tube, gently tapped to the bottom of the vial and scored for their tendency to move toward a light source that was placed at the opposite end of the tube (Figure 2A). The number of flies present in each area of the test tube, i.e., at the origin, in the middle and the furthest part (Figure 2A), was counted every minute. Interestingly, the experiment revealed no difference between Chd1^–/– and Chd1^WT/WT (Figure 2B). Young w¹¹¹⁸ flies (wildtype background control) also responded to the light by immediately walking toward it, but then exhibited a tendency to move backward again accounting for the slightly lower RI values detected with this line (Supplementary Figure 1A). Thus, Chd1 deletion did not affect light reception and positive phototaxis behavior, although we cannot rule out that other aspects of vision are compromised.

Next, we tested olfactory abilities by observing the behavior of Chd1-deficient and Chd1-wildtype flies in a similar experimental set-up, except that the strongly repellent odor benzaldehyde was added to an agar disc that was placed at one end of the apparatus (Figure 2C). As expected, wildtype flies (w¹¹¹⁸, Canton S) and Chd1^WT/WT flies at both ages showed strong repulsion by benzaldehyde by walking away from the repellent after exposure (Figure 2D and Supplementary Figure 1B). By contrast, about 70% of 4- and 14-day-old Chd1-deficient flies remained close to the benzaldehyde source, where they had been placed at the start of the assay (Figure 2D). Because Chd1^–/– flies showed similar walking behavior in the phototaxis assay as the wildtype flies, it is not likely that a potential impairment of locomotory abilities acted as a confounding parameter in this assay. Thus, the results suggest that absence of CHD1 strongly impacts on the sense of smell.

To assess the gustatory abilities of the flies, sucrose was offered as an attractant in the assay set-up shown in Figure 2E. Sucrose must be tasted because it is not volatile. Starved Chd1^WT/WT flies of both ages departed from the non-sucrose end, and the majority accumulated in the compartment nearest to the sucrose similar to w¹¹¹⁸ flies (Figure 2F and Supplementary Figure 1C). By contrast, the behavior of Chd1^–/– flies was comparable to the smell test with only about 20.2 ± 0.3% of the 4-day-old and 27.9 ± 0.5% of the 14-day-old mutant flies moving close to the sucrose source (Figure 2F). To ascertain that the observed aberrant behavior of Chd1-deficient flies was caused by defective gustatory perception and not by reduced motivation for foraging, we subjected the flies to the same assay set-up yet omitting the taste stimulus. The 4 days old starved Chd1^WT/WT flies showed roughly equal distribution across all sections of the tube at each time point with a slight preference for the middle part, whereas older wildtype flies slightly preferred the origin and furthest sections. Chd1^–/– flies were similar to the older wildtype flies except for a slight preference of the origin section at both ages (Figure 2G) suggesting that reduced motivation to forage might indeed contribute to the poor gustatory response observed in the tube assay. To examine taste abilities of the flies by another method, we employed the PER assay (Shiraiwa and Carlson, 2007). When the taste neuron-innervated sensilla of the fly’s labellum make contact to an attractive substance, such as sucrose solution, the proboscis is extended to consume the food. Monitoring the frequency of PER across 10 experimental trials, we observed that in young flies, the PER frequency progressively decreased in both Chd1^WT/WT and Chd1^–/– flies most likely reflecting progressive satiety (Wang et al., 2004; Shiraiwa and Carlson, 2007; Slone et al., 2007; Chen and Amrein, 2014). At 14 days of age, Chd1-wildtype flies showed 100% response frequency throughout the experiment, whereas Chd1-mutant flies started with about 75% response which gradually decreased to about 50% response (Figure 2H). Given that young Chd1-deficient flies exhibited almost normal PER, it is likely that their gustatory abilities are principally intact, at least at this young age. At 14 days of age, most of the flies initially responded to the stimulus but very quickly turned to satiety-like behavior. Together, the results from the gustatory tube assay and the PER assay suggest that absence of CHD1 does not cause a (complete) loss of taste perception. Rather, it appears that the flies can still taste but might have diminished motivation to feed. This is consistent with our previous observations that Chd1-deficient flies show strongly reduced food intake correlating with dysregulation of many neuropeptide genes responsible for feeding and foraging behavior (Schoberleitner et al., 2021).

Although walking abilities of Chd1^–/– flies appeared similar to wildtype flies in the phototaxis assay described above, we performed two additional locomotory assays. The negative geotaxis assay measures the startle-induced reflexive tendency of flies to move against gravity (Figure 3A), while in the exploratory walking assay (Figure 3C), more complex behavior involving the regulation of walking parameters, such as speed and direction, as well as orientation is assessed. Both, reflexive as well as decision-based locomotion, however, are substantially affected by age-related decline and are therefore also used to test for functional senescence (Grotewiel et al., 2005; Ismail et al., 2015).

We found that both 4 and 14 days old Chd1^–/– flies performed significantly worse in the negative geotaxis assay compared to wildtype (Figure 3B and Supplementary Figure 1D). Chd1-deficient flies were slow climbing upward and showed numerous direction changes compared to Chd1-wildtype and w¹¹¹⁸ flies that were moving only upward.

For the exploratory walking assay, flies were placed individually into the middle of a graduated petri dish, and their movement was monitored by counting the number of grid-line crossings (Figure 3D). Typically, the physical borders of an open field (borders of the Petri dish) are of main interest for the fly, and it will normally strive to head there (Soibam et al., 2012). Accordingly, Chd1-wildtype flies after placement into the center of the dish either remained rooted to the spot observing the environment for a short time or walked by the shortest path toward the border of the Petri dish and remained there. A similar, albeit not as pronounced behavior was displayed by w¹¹¹⁸ flies (Supplementary Figure 1E). By contrast, Chd1-mutant flies immediately started to walk in circles (many grid-line crossings) but did not head for the border (Figure 3D). When 14 days old flies were tested, we observed that the tendency to go to the border of the dish by the shortest way decreased also for Chd1-wildtype flies, but the difference between the genotypes remained (Figure 3D). The reduced performance of the flies in these test paradigms in the absence of CHD1 suggest an earlier entry into a state of age-related decline compared to wildtype flies.

The RNA-seq data had shown that most dysregulated genes in Chd1^–/– heads were rescued by both neuron-specific expression of Chd1 or expression under the control of its native promoter (Schoberleitner et al., 2021). To examine, if neuronal expression of Chd1 in the Chd1-deficient background would affect the sensory and locomotory abilities of the fly, we performed negative geotaxis, exploratory walking, gustatory, and olfactory assays also with Chd1^elav flies.

The results show that negative geotaxis, exploratory walking abilities, and gustatory response were completely rescued by expression of Chd1 with the elav driver (Figures 4A–C) indicating that activity of CHD1 in neurons is responsible for the respective defects in Chd1^–/– flies. Intriguingly, however, neuronal expression of Chd1 was not sufficient to rescue the olfactory deficits of Chd1^–/– flies (Figure 4D). Hence, for proper olfactory response, CHD1 activity in other cell types in addition to neurons is required.