Regulation of feeding dynamics by the circadian clock, light and sex in an adult nocturnal insect

Abstract

Animals invest crucial resources in foraging to support development, sustenance, and reproduction. Foraging and feeding behaviors are rhythmically expressed by most insects. Rhythmic behaviors are modified by exogenous factors like temperature and photoperiod, and internal factors such as the physiological status of the individual. However, the interactions between these factors and the circadian clock to pattern feeding behavior remains elusive. As Drosophila, a standard insect model, spends nearly all its life on food, we rather chose to focus on the adults of a non-model insect, Agrotis ipsilon, a nocturnal cosmopolitan crop pest moth having structured feeding activity. Our study aimed to explore the impact of environmental cues on directly measured feeding behavior rhythms. We took advantage of a new experimental set-up, mimicking an artificial flower, allowing us to specifically monitor feeding behavior in a naturalistic setting, e.g., the need to enter a flower to get food. We show that the frequency of flower visits is under the control of the circadian clock in males and females. Feeding behavior occurs only during the scotophase, informed by internal clock status and external photic input, and females start to visit flowers earlier than males. Shorter duration visits predominate as the night progresses. Importantly, food availability reorganizes the microstructure of feeding behavior, revealing its plasticity. Interestingly, males show a constant number of daily visits during the 5 days of adult life whereas females decrease visitations after the third day of adult life. Taken together, our results provide evidence that the rhythmicity of feeding behavior is sexually dimorphic and controlled by photoperiodic conditions through circadian clock-dependent and independent pathways. In addition, the use of the new experimental set-up provides future opportunities to examine the regulatory mechanisms of feeding behavior paving the way to investigate complex relationships between feeding, mating, and sleep-wake rhythms in insects.

Keywords: Agrotis ipsilon; Lepidoptera; automated feeding cage; circadian behavior; developmental timing; feeding behavior.

FIGURE 1

Photographs of the new device for automatic measurement of feeding behavior parameters of adult Lepidoptera, the moth-box. (A) Adult male Agrotis ipsilon, scale bar = 1 cm. (B) Open experimental device mimicking an artificial flower in the above view, including a sucrose dispenser and an infrared motion sensor detecting moth entrance in the mimetic short flower corolla. (C) Closed experimental device in side view. (D) Artificial flower visited by a moth during the scotophase (photo taken under red light). The infrared sensor detects the entry of the moth at its thorax.

FIGURE 2

Male moth’ feeding rhythm depends on the availability of sugar solution in artificial flower. The experiment was carried out over 5 days with a photoperiod 16L:8D (n = 9 for water cohort; n = 10 for sodium cohort). (A,B) Diagrams for males with a sugar solution available during the 5 days of experimentation (A), and for males with a sugar solution available during the first 2 days of experimentation (B) (full bar, flower visit during scotophase; empty bar, flower visit during photophase; red horizontal bar, stopping the delivery of sugar solution). (C) Mean visit number per moth for each of the two cohorts depending on the days (solid line, male with a sugar solution available from D1 to D5; dashed line, male with a sugar solution available on D1 and D2). Values correspond to the mean ± SEM. A Nemenyi’s test was performed to compare the visit number within a same cohort (#, p < 0.05). A Wilcoxon’s test (***, p < 0.001) was performed to compare the visit number between the two cohorts; statistical data are presented in Supplementary Table S1.

FIGURE 3

Feeding and locomotor rhythms can be dissociated from each other as soon as food distribution stops. Diagrams for feeding activity assessed in the moth-box (n = 10 for each cohort) and global locomotor activity assessed in the LAM25 (n = 14 for each cohort) in males with a sugar solution available from D1 to D5 (A), and in males with a sugar solution available on D1 and D2 (B) (full bar, event during scotophase; empty bar, event during photophase; red horizontal bar, stopping the delivery of sugar solution). Values correspond to the mean ± SEM. Diagrams for feeding activity are the same as in Figures 2A, B.

FIGURE 4

Duration of flower visit is regulated by the circadian clock and sugar solution availability. The experiment was carried out over 5 days with a photoperiod 16L:8D. A first cohort was with a sugar solution available during the 5 days of experimentation; a second cohort was with a sugar solution available during the first 2 days of experimentation (same data as in Figure 2). (A) Mean visit number per moth for each of the two cohorts according to the time of day, the 5 days of experimentation are pooled (solid line, male with a sugar solution available from D1 to D5; dashed line, male with a sugar solution available on D1 and D2; solid arrow, phase for males with a sugar solution available from D1 to D5; dashed arrow, phase for males with a sugar solution available on D1 and D2). A statistical analysis of the circadian rhythm was made to compare the two cohorts, statistical data are presented in Supplementary Table S2. Values correspond to the mean ± SEM (n = 10 for each cohort). (B) Distribution of flower visit durations (data from D2 and D3 are cumulated) for males with a sugar solution available from D1 to D5. Letters indicate the statistical differences after a Poisson mixed-model. (C) Duration of flower visit from D3 to D5 according to the two cohorts (solid line, male with a sugar solution available from D1 to D5; dashed line, male with a sugar solution available on D1 and D2). Curves represent normal laws fitted to the data (s² is the variance). A Fisher-Snedecor’s test was performed to analyze the variance of data.

FIGURE 5

Feeding behavior according to the sex of moths. Data from males in Figure 2A are compared to results from a cohort of female treated the same way. Values correspond to the mean ± SEM (n = 10 for each cohort, details as in Figure 2A). Diagram plots results for male (A) and female (B) with a sugar solution available during the 5 days of experimentation (full bar, flower visit during scotophase; empty bar, flower visit during photophase). (C) Mean visit number per moth for each of the two cohorts depending on the days (solid line, male; dashed line, female). A Nemenyi’s test was performed to compare the visit number within a same cohort (#, p < 0.05; ##, p < 0.01). A Wilcoxon’s test (*, p < 0.05) was performed to compare the two cohorts, and statistical data are presented in Supplementary Table S3. (D) Mean visit number per moth for each of the two cohorts according to the time of day, the 5 days of experimentation are cumulated (solid line, male; dashed line, female; solid arrow, phase for male; dashed arrow, phase for female). A statistical analysis of the circadian rhythm was made to compare the two cohorts, and statistical data are presented in Supplementary Table S4. (E) Proportions of food intake durations (all data from D1 to D5 are cumulated) for males and females. Letters indicate the statistical differences after a Poisson mixed-model. (F) Duration of flower visit from D1 to D5 according to the two cohorts (solid line, male; dashed line, female). Curves represent normal laws fitted to the data (s² is the variance). A Fisher-Snedecor’s test was performed to analyze the variance of data.

FIGURE 6

Feeding rhythms and feeding dynamics are gated by light. A first cohort was composed of males subjected to 16L:8D noted LD; a second cohort was composed of males subjected to 24L:0D noted LL; a third cohort was composed of males subjected to 0L:24D noted DD. Values correspond to the mean ± SEM (n = 10 for each cohort). Diagram for males subjected to 16L:8D (A) (same diagram as in Figure 2A), 24L:0D (B) and 0L:24D (C) (full bar, flower visit during scotophase; empty bar, flower visit during photophase). (D) Mean visit number per moth for the cohorts subjected to conditions LD, LL, and DD depending on the days (red, LD condition; black, DD condition; yellow, LL condition). Wilcoxon’s test (*, p < 0.05; **, p < 0.01; ***, p < 0.001) was performed to compare the cohorts, and statistical data are presented in Supplementary Table S5. (E) Mean visit number per moth for the cohorts subjected to conditions LD, LL, and DD according to the time of day; the raw data in Figure A, B and C acquired during the 5 days of experimentation were cumulated (red line, LD condition; red arrow, phase for LD condition; black line, DD condition; black arrow, phase for DD condition; yellow line, LL condition). Statistical analysis of the LD circadian rhythm was made are presented in Supplementary Table S6.

FIGURE 7

Interaction of the circadian clock with the acute effects of light patterns feeding behavior. Diagrams for males with a sugar solution available ad libitum and variation in number of visits and rhythm of feeding behavior according to photic environment. A first cohort was composed of males subjected to 16L:8D noted LD; a second cohort was composed of males subjected to 24L:0D noted LL; a third cohort was composed of males subjected to 0L:24D noted DD; a fourth cohort was composed of males subjected to 16L:8D on D1 and D5, and 24L:0D from D2 to D4 noted LD-LL; and a fifth cohort was composed of males subjected to 16L:8D on D1 and D5, and 0L:24D from D2 to D4 noted LD-DD. Values correspond to the mean ± SEM (n = 10 for each cohort). (A) Diagram for males subjected to 16L:8D on D1 and D5, and 24L:0D from D2 to D4 (full bar, flower visit during scotophase; empty bar, flower visit during photophase). (B) Mean visit number per moth for the cohorts subjected to conditions LD, LL, and LD-LL depending on the days (red, LD condition; black, DD condition; brown, LD-LL condition). Wilcoxon’s test was performed to compare the cohorts, and statistical data are presented in Supplementary Table S7. (C) Diagram for males subjected to 16L:8D on D1 and D5, and 0L:24D from D2 to D4 (full bar, flower visit during scotophase; empty bar, flower visit during photophase). (D) Mean visit number per moth for the cohorts subjected to conditions LD, DD, and LD-DD depending on the days (red, LD condition; black, DD condition; blue, LD-DD condition). Wilcoxon’s test was performed to compare the cohorts, and statistical data are presented in Supplementary Table S8.

FIGURE 8

Visit duration of artificial flower according to ambient photic environment. The experiment was carried out on males with a sugar solution available ad libitum over 5 days with a variable photoperiod. A first cohort was composed of males subjected to 16L:8D noted LD; a second cohort was composed of males subjected to 24L:0D noted LL; a third cohort was composed of males subjected to 0L:24D noted LL; a fourth cohort was composed of males subjected to 16L:8D on D1 and D5, and 24L:0D from D2 to D4 noted LD-LL; and a fifth cohort was composed of males subjected to 16L:8D on D1 and D5, and 0L:24D from D2 to D4 noted LD-DD. Curves represent normal laws fitted to the data (s² is the variance). (A) Duration of flower visit from D1 to D5 for the cohorts subjected to conditions 16L:8D (LD, n = 10), 24L:0D (LL, n = 6), and 0L:24D (DD, n = 9) (red, LD condition; black, DD condition; yellow, LL condition). A Fisher-Snedecor’s test was performed to analyze the variance of data. (B) Duration of flower visit from D2 to D4 for the cohorts subjected to conditions LD (n = 10), LL (n = 5), and LD-LL (n = 6) (red, LD condition; black, DD condition; brown, LD-LL condition). For all the conditions, a Fisher-Snedecor’s test was performed to analyze the variance of data. (C) Duration of flower visit from D2 to D4 for the cohorts subjected to conditions LD (n = 10), DD (n = 9), and LD-DD (n = 8) (red, LD condition; black, DD condition; blue, LD-DD condition). For all the conditions, a Fisher-Snedecor’s test was performed to analyze the variance of data.