Diet-induced changes to sleep architecture persist despite disruption of circadian rhythm and total sleep.

(A) Modulation of dietary sucrose content to 2.5% (LS) or 30% (HS) in a low (2.5%) yeast base increased sleep partitioning during both the day (lights-on) and night (lights-off) period during a 12∶12 hour light∶dark circadian cycle. (B) Disruption of circadian light cues with constant darkness or constant light for 5 days following a period of 12∶12 hour light∶dark entrainment did not disrupt the effects of diet on sleep (no significant effect of diet∶lighting interaction by two-way ANOVA). (C) Dietary sugar modulated sleep partitioning even in the absence of the circadian clock components cycle (cyc⁰¹), timeless (tim⁰¹), period (per⁰¹), and the circadian neuropeptide pigment dispersing factor (Pdf⁰¹). (D) Mutants containing the shaker minisleep (Sh^mns) variant showed low total sleep relative to genotype-matched controls (inset). However, the Sh^mns flies retained the ability to respond to dietary sugar. A significant diet∶genotype interaction (two-way ANOVA) reflects an exacerbation of diet effects by the Shaker mutation (E) Sleep deprivation was induced by one night (12 hours) of a shaking stimulus after 4 days on LS or HS food in control (yw) flies. During the 12 hour period following the deprivation, total sleep was increased relative to the identical period prior to deprivation (inset). A significant effect of diet persisted across conditions and there was no significant diet∶deprivation interaction (two-way ANOVA). (F) Flies were stimulated with a 5 minute light pulse 1 hour after lights-off during a 12∶12 hour circadian cycle. This is a time when >80% of flies are typically within a sleep bout and only flies exhibiting sleep at the time of the light pulse were used for analysis. LS-fed flies showed reduced latency to activity from the onset of the light pulse (upper left panel) and increased activity following the light pulse (2 representative traces from single flies, right panel). The number of flies showing no activity within 3 hours following the light pulse but resuming activity the following day (non-responders) was significantly higher in the HS-fed condition (Fisher's exact test). Error bars represent mean +/− SEM for each group. *** p<0.001, ** p<0.01,* p<0.05 for all statistical calculations. Significance values for t-tests between dietary conditions are shown above each set of bars and significance values for two-way ANOVA are shown above the graph, where applicable. See Figure S3 for additional supporting evidence.