Hungry flies switch from cold to warm preference upon food detection.

(A) A schematic diagram of the experimental conditions. Flies were reared on fly food. Emerging flies were kept on fly food for one or two days (orange) and then starved with water-absorbing paper for one or two nights (gray). Starved flies were refed with either fly food (orange), 2 mL sucralose (blue), or glucose solution (green) absorbed paper and then examined for temperature preference (Tp) behavioral assays. Details are provided in the experimental procedures.

(B-D) Comparison of Tp of white1118 (w1118) control flies between fed (white bar), starved (STV; gray bar), and refed (orange, blue, or green bar) states. Starved flies were refed with fly food (orange bar) for 5, 10, 30, or 60 min (1 hr) (B), 2.8 mM sucralose solution (blue bar) for 10 min or 1 hr (C), or 2.8 mM (equivalent to 5%) glucose solution (green bar) for 5 min, 10 min, or 1 hr (D). Behavioral experiments were performed at the specific time points, Zeitgeber time (ZT) 4-7. ZT0 and ZT12 are light on and light off, respectively. Dots on each bar indicate individual Tp in assays. Numbers in italics indicate the number of trials. The Shapiro-Wilk test was performed to test for normality. One-way ANOVA was performed to compare Tp between each refeeding condition. Tukey’s post hoc test was used to compare differences between experimental and starved conditions (green stars). Data are presented as mean Tp with SEM. *p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001. NS indicates no significance.

(E) Comparison of Tp among the following conditions in w1118 (control) flies: fed (white bar), starved overnight (O/N) (water or STV; gray bar), refed with fly food (fly food; orange bar), sucralose (blue bars), or glucose (green bars). The same data are used in Fig. 1B-D. The Shapiro-Wilk test was performed to test for normality. One-way ANOVA was performed to compare Tp between each refeeding condition. Red (or green) stars indicate Tukey’s post hoc test comparing differences between experimental and fed (or starved) conditions, respectively. *p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001. NS indicates not significant.

(F) Feeding assay: The number of touches to water (gray), 2.8 mM (equivalent to 5%) glucose (green), or 2.8 mM sucralose in each solution (blue) was examined using w1118 flies starved for 1 day. A cumulative number of touches to water or sugar solution for 0-30 min was plotted. Two-way ANOVA was used for multiple comparisons. Blue and green stars show Fisher’s LSD post hoc test comparing sucralose (blue stars) or glucose (green stars) solution feeding to water drinking. All data shown are means with SEM. *p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001.

Gustatory neurons are essential for taste-evoked warm preference.

(A-E) Comparison of Tp of flies between fed (white bar), starved (STV; gray bar), and refed (blue bar) states. Starved flies were refed with sucralose (blue bars) for 10 min. (F) Schematic of optogenetic activation assay.

(G-J) Comparison of Tp of flies between fed (white bar), starved (STV; gray bar), and starved with all-trans-retinal (ATR; yellow bars), which is the chromophore required for CsChrimson activation. Gustatory neurons in starved flies were excited by red light pulses (flashing on and off at 10 Hz) for 10 min. Behavioral experiments were conducted at ZT4-7. Dots on each bar indicate individual Tp in assays. Numbers in italics indicate the number of trials. Shapiro-Wilk test was performed for normality test. One-way ANOVA was used for statistical analysis. Green stars indicate Tukey’s post hoc test compared between each experiment and the starved condition. Pooled data are shown in Fig. S3. All data shown are means with SEM. *p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001. NS indicates not significant.

Both warm- and cold-temperature-sensing neurons are involved with taste-evoked warm preference.

(A-E) Comparison of Tp of flies between fed (white bar), starved (STV; gray bar), and refed (blue bar) conditions. Starved flies were refed with sucralose (blue bars) for 10 min. The Shapiro-Wilk test was performed to test for normality. One-way ANOVA or Kruskal-Wallis test was used for statistical analysis. Green stars indicate Tukey’s post hoc test or Dunn’s test compared between each experiment and the starved condition. Pooled data are shown in Fig. S3. (F and G) Comparison of Tp between starved (STV; gray bar) and all-trans-retinal (ATR; yellow bars) starved flies. Warm neurons (F) or cold neurons (G) in starved flies expressed CsChrimson, which was excited by red light pulses for 10 min. The Shapiro-Wilk test was performed to test for normality. Student’s t-test or Kolmogorov-Smirnov test was used for statistical analysis. (G) tubGal80ts; R11F02-Gal4>uas-CsChrimson flies were reared at 18°C, and emerged adults were collected and stored at 29°C. See Materials and Methods for details. These behavioral experiments were conducted on ZT4-7. All data shown are means with SEM. *p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001. NS indicates no significance. The same data from UAS-Kir/+ in Fig. 2E are used in Fig. 3E.

Hunger signals are involved in taste-evoked warm preference.

Comparison of Tp of flies between fed (white bar), starved (STV; gray bar), and refed states (orange, blue, green, or dark green). Starved flies were refed with fly food (orange bar), sucralose (blue bar), or glucose (green bar) for 10 min. These behavioral experiments were conducted on ZT4-7. Dots on each bar indicate individual Tp in assays. Numbers in italics indicate the number of trials. Shapiro-Wilk test was performed for normality test. One-way ANOVA was used for statistical analysis. Green stars indicate Tukey’s post hoc test compared between each experiment and the starved condition. Pooled data are shown in Fig. S3. All data shown are means with SEM. *p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001. NS indicates not significant.

Clock genes are involved in taste-evoked warm preference.

(A-D) Comparison of Tp of flies between fed (white circles), starved (STV; gray triangles), and refed (blue, green, or dark green squares) states. Flies were starved for 24 hrs. Starved flies were refed with sucralose (blue squares, A1-D1) or glucose for 10 min (green squares, A2-D2) or 1 hr (dark green squares, A3). After the 10 min or 1 hr refeeding, the temperature preference behavior assays were performed immediately at ZT1-3, 4-6, 7-9, and 10-12 in LD. The Shapiro-Wilk test was performed to test for normality. One-way ANOVA or Kruskal-Wallis test was used for statistical analysis. Stars indicate Tukey’s post hoc test or Dunn’s test compared between each experiment and starved condition at the same time point. All data shown are means with SEM. *p<0.05. **p<0.01. ***p<0.001. ****p < 0.0001.

(E) Comparison of Tp during daytime in each feeding state. One-way ANOVA or Kruskal-Wallis test was used for statistical analysis between ZT1-3 and ZT7-9 or ZT10-12. R and AR indicate rhythmic and arrhythmic, respectively, during the daytime.

The starvation period is unlikely to affect the ability to recover (Related to Fig. 3)

Comparison of Tp of w1118 control flies between fed (white bar), starved (STV; gray bar), and refed (STV; gray bar) and refed (blue or green bar). Flies were starved for overnight (STV1ON; gray bar) or 1.5 ON (STV1.5ON; dark gray bar), i.e. starved for 18-21 hrs. starved for 18-21 hrs or 26-29 hrs. Starved flies were refed with sucralose for 10 min (blue bar) or glucose for 10 min (green bar). These behavioral experiments were performed on ZT4-7. Dots on each bar indicate individual Tp in the assays. Numbers in italics indicate the number of trials. Per Shapiro-Wilk test was performed to test for normality. One-way ANOVA or Kruskal-Wallis test was used for statistical analysis. Green stars indicate Tukey’s post hoc test or Dunn’s test was used to compare between each experiment and starved condition. and starved condition. All data shown are means with sem. *p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001. NS indicates no significance. The same data from STV1ON, STV1ON+refed sucralose for 10 min and STV1ON+refed sucralose for 10 min in Fig. 1C-E are used in Fig. S1.

Number of touches to water, sucralose and glucose using yw, per01 and tim01 flies. (Related to Figures 1, 5, and 6)

Feeding assay: The number of touches to water (gray), 2.8 mM (equivalent to 5%) glucose (green), or 2.8 mM sucralose solution (blue) was examined using yw, per01, or tim01 flies starved for 1 day. Cumulative number of touches for 0-30 min was plotted. Two-way ANOVA was used for multiple comparisons. Blue and green stars show Fisher’s LSD post hoc test comparing sucralose (blue stars) or glucose (green stars) solution feeding to water drinking. Black stars show the statistical difference between sucralose and glucose refeeding. All data shown are means with SEM. *p < 0.05. *F*p < 0.01. ***p < 0.001. ****p < 0.0001.

Taste- or temperature-sensing neurons are not important for nutrient- (carbohydrate-) induced warm preference (Related to Figures 2 and 3).

Comparison of Tp of flies between fed (white bar), starved (STV; gray bar), and refed (orange, blue, green, or dark green bar) conditions. Starved flies were refed with fly food for 10 min (fly food; orange bar), sucralose for 10 min (blue bar), or glucose for 10 min (green bar) or 1 hr (dark green bar). These behavioral experiments were conducted on ZT4-7. The dots on each bar indicate individual Tp in the assays. Numbers in italics indicate the number of experiments. The Shapiro-Wilk test was performed to test for normality. One-way ANOVA or Kruskal-Wallis test was used for statistical analysis. Green stars indicate Tukey’s post hoc test or Dunn’s test compared between each experiment and the starved condition. All data shown are means with sem. *p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001. NS indicates no significance. The same data from UAS-Kir/+ in Fig. 2E are used in Fig. S3I.

Model

A schematic diagram of Tp recovery by taste-evoked warm preference (blue arrows) and nutrient-induced warm preference (green arrows).