The microbiome reduces the effect of timed feeding on the cycling of lipids and biogenic amines.

(A) Schematic showing the experimental protocol for 1) analysis of gut metabolite cycling in microbiome-containing (AM) and sterile (AS) iso31 flies fed on normal chow diets under ad lib and 2) timed feeding (TF) conditions, respectively.

(B-E) Pie chart depicting the fraction of cycling primary metabolites and polar histogram plots of the peak phase for cycling primary metabolites in microbiome-containing and sterile flies under ad lib and timed feeding conditions (AM, TM, AS and TS), respectively, using a JTK_cycle cutoff of P < 0.05.

(F-I) Pie chart depicting the fraction of cycling lipids and polar histogram plots of the peak phase for cycling lipids in AM, TM, AS and TS flies, respectively, using a JTK_cycle cutoff of P < 0.05.

(J-M) Pie chart depicting the fraction of cycling biogenic amines and polar histogram plots of the peak phase for cycling biogenic amines in AM, TM, AS and TS flies, respectively, using a JTK_cycle cutoff of P < 0.05.

(N-P) The ratio of primary metabolites, lipids and biogenic amines that cycle in microbiome-containing versus sterile flies under ad lib feeding (AM/AS) and timed feeding (TM/TS) conditions, respectively.

The effect of the microbiome on metabolite cycling depends on a functional circadian clock.

(A) Schematic showing the experimental protocol for analysis of gut metabolite cycling in microbiome-containing (AM) and sterile (AS) per01 flies fed on normal chow diets.

(B) Pie chart depicting the fraction of cycling primary metabolites, and (C) polar histogram plots of the peak phase for cycling primary metabolites in AM and AS per01 flies fed on normal chow diets, using a JTK_cycle cutoff of P < 0.05.

(D) Pie chart depicting the fraction of cycling lipids and (E) polar histogram plots of the peak phase for cycling lipids in AM and AS per01 flies fed on normal chow diets, using a JTK_cycle cutoff of P < 0.05.

(F) Pie chart depicting the fraction of cycling biogenic amines and (G) polar histogram plots of the peak phase for cycling biogenic amines in AM and AS per01flies fed on normal chow diets, using a JTK_cycle cutoff of P < 0.05.

The microbiome increases the number of cycling of metabolites in flies fed on high protein diets.

(A) Schematic showing the experimental protocol for analysis of gut metabolite cycling in microbiome-containing (AM) and sterile (AS) Iso31 flies fed on high protein diets.

(B) Pie chart depicting the fraction of cycling primary metabolites, and (C) polar histogram plots of the peak phase for oscillating primary metabolites in AM and AS iso31 flies fed on high protein diets, using a JTK_cycle cutoff of P < 0.05.

(D) Pie chart depicting the fraction of cycling lipids, and (E) polar histogram plots of the peak phase for oscillating lipids in AM and AS iso31 flies fed on high protein diets, using a JTK_cycle cutoff of P < 0.05.

(F) Pie chart depicting the fraction of cycling biogenic amines, and (G) polar histogram plots of the peak phase for oscillating biogenic amines in AM and AS iso31 flies fed on high protein diets, using a JTK_cycle cutoff of P < 0.05.

The microbiome decreases the number of cycling metabolites in iso31 flies fed on high sugar diets.

(A) Schematic showing the experimental protocol for analysis of gut metabolite cycling in microbiome-containing (AM) and sterile (AS) Iso31 flies fed on high sugar diets.

(B) Pie chart depicting the fraction of cycling primary metabolites, and (C) polar histogram plots of the peak phase for cycling primary metabolites in AM and AS iso31 flies fed on high sugar diets, using a JTK_cycle cutoff of P < 0.05.

(D) Pie chart depicting the fraction of cycling lipids, and (E) polar histogram plots of the peak phase for cycling lipids in AM and AS iso31 flies fed on high sugar diets, using a JTK_cycle cutoff of P < 0.05.

(F) Pie chart depicting the fraction of cycling biogenic amines and (G) polar histogram plots of the peak phase for cycling biogenic amines in AM and AS iso31 flies fed on high sugar diets, using a JTK_cycle cutoff of P < 0.05.

Amino acid metabolism is modulated by the microbiome, timed feeding and the circadian clock.

(A) KEGG enrichment analysis of oscillating metabolites unique to AM flies, shared in AM and AS flies and unique to AS flies. Data only show the P < 0.05 results.

(B) KEGG enrichment analysis of oscillating metabolites unique to AM flies, shared in AM and TM flies and unique to TM flies. Data only show the P < 0.05 results.

(C) KEGG enrichment analysis of oscillating metabolites unique to AS flies, shared in AS and TS flies and unique to TS flies. Data only show the P < 0.05 results.

(D) KEGG enrichment analysis of oscillating metabolites unique to TM flies, shared in TM and TS flies and unique to TS flies. Data only show the P < 0.05 results.

(E) KEGG enrichment analysis of oscillating metabolites unique to per01 AM flies, shared in AM and AS of per01flies and unique in per01 AS flies. Data only show the P < 0.05 results.

(F) KEGG enrichment analysis of oscillating metabolites unique to high sugar fed AM flies, shared in AM and AS of iso31 flies fed on high sugar diets and unique in high sugar fed AS flies. Data only show the P < 0.05 results.

Model showing effects of the microbiome, timed feeding, the circadian clock and diet composition on metabolite rhythms in the gut.

The microbiome broadly increases metabolite rhythms on a normal chow diet and its effects are enhanced by high protein diet and the circadian clock. Timed feeding promotes cycling itself, but attenuates the effect of the microbiome on cycling, as does a high sugar diet. Effects of each of these manipulations on phase are also indicated.

Timed feeding has less effect on cycling metabolites in microbiome-containing flies.

(A-C) A: Principal component analysis (PCA) of primary metabolites, lipids and biogenic amine abundance, respectively, at each time-point in a 12:12 LD cycle for AM, TM, AS, and TS flies.

(D-F) The ratio of primary metabolites, lipids and biogenic amines that cycle in microbiome-containing versus sterile flies under ad lib feeding (TS/AS) and timed feeding (TM/AM) conditions, respectively.

The amplitude of metabolite cycling is sometimes, but not always increased by the presence of a microbiome or by timed feeding.

(A-C) The comparison of amplitudes for cycling primary metabolites, lipids and biogenic amines, respectively, that overlap in AM and AS flies. Data are mean ± SEM, ****p < 0.0001 shown by Student’s t test.

(D-F) The comparison of amplitudes for cycling primary metabolites, lipids and biogenic amines, respectively, that overlap in AM and TM flies. Data are mean ± SEM, **p < 0.01 and ****p < 0.0001 shown by Student’s t test.

(G-I) The comparison of amplitudes for cycling primary metabolites, lipids and biogenic amines, respectively, that overlap in AS and TS flies. Data are mean ± SEM, ***p < 0.001 shown by Student’s t test.

Lipids and biogenic amines show higher abundance in microbiome-containing flies.

A-D: Average expression levels of gut primary metabolites that cycle in either microbiome-containing or sterile flies or both conditions (lost, shared and gained). The number indicates the average normalized counts in each condition for primary metabolites that cycle with p < 0.05, JTK_cutoff. Data are mean ± SEM, comparisons were with Student’s t test. A: AS versus AM; B: TM versus AM; C: TS versus AS; D: TS versus TM.

E-H: Average expression levels of gut lipids that cycle in either microbiome-containing or sterile flies or both conditions (lost, shared and gained). The number indicates the average normalized counts in each condition for lipids that cycle with p < 0.05, JTK_cutoff. Data are mean ± SEM, comparisons were with Student’s t test. E: AS versus AM; F: TM versus AM; G: TS versus AS; H: TS versus TM.

I-L: Average expression levels of gut biogenic amines that cycle in either microbiome-containing or sterile flies or both conditions (lost, shared and gained). The number indicates the average normalized counts in each condition for biogenic amines that cycle with p < 0.05, JTK_cutoff. Data are mean ± SEM, comparisons were with Student’s t test. I: AS versus AM; J: TM versus AM; K: TS versus AS; L: TS versus TM.

Abundance analysis of all metabolites in microbiome containing and sterile flies.

(A-C) Average expression levels of gut metabolites in microbiome-containing and sterile flies. iso31 flies are fed on normal chow diets. A: AS versus AM for primary metabolites; B: AS versus AM for lipids; C: AS versus AM for biogenic amines.

(D-F) Average expression levels of gut metabolites between microbiome-containing and sterile flies. per01 flies are fed on normal chow diets. D: AS versus AM for primary metabolites; E: AS versus AM for lipids; F: AS versus AM for biogenic amines.

(G-I) Average expression levels of gut metabolites between microbiome-containing and sterile flies. iso31 flies are fed on high protein diets. G: AS versus AM for primary metabolites; H: AS versus AM for lipids; I: AS versus AM for biogenic amines.

(J-L) Average expression levels of gut metabolites between microbiome-containing and sterile flies iso31 flies are fed on high sugar diets. J: AS versus AM for primary metabolites; K: AS versus AM for lipids; L: AS versus AM for biogenic amines.

Principal component analysis (PCA) of total expressed metabolites in per01 flies fed on normal chow diets, iso31 flies fed on high protein and high sugar diets.

(A-C) Principal component analysis (PCA) of metabolite abundance at each time-point in a 12:12 LD cycle for per01 AM and AS flies fed on normal chow diets.

(D-F) Principal component analysis (PCA) of metabolite abundance at each time-point in a 12:12 LD cycle for iso31 AM and AS flies fed on high protein diets

(G-I) Principal component analysis (PCA) of metabolite abundance at each time-point in a 12:12 LD cycle iso31 AM and AS flies fed on high sugar diets.

Phases also change for cycling metabolites shared between per01-AM and per01-AS.

(A-C) Polar histogram plots of the peak phase for shared oscillating primary, lipids, and biogenic amines, respectively, in per01 microbiome-containing and per01 sterile flies, using a JTK_cycle value of P < 0.05.

Effect of the microbiome on expression levels of cycling metabolites in clock mutants and under different dietary conditions.

A-C: Average expression levels of gut metabolites that cycle in microbiome-containing versus sterile per01 flies fed on normal chow diets. The middle group in each panel shows metabolites that cycle under both conditions, while those on either side indicate metabolites that cycle in only one condition. The number indicates the average normalized counts in each condition for metabolites that cycle with p < 0.05, JTK_cycle. Data are mean ± SEM, comparisons were with Student’s t test. A: AS versus AM for primary metabolites; B: AS versus AM for lipids; C: AS versus AM for biogenic amines.

D-F: Average expression levels of gut metabolites that cycle in microbiome-containing versus sterile iso31 flies fed on high protein diets. The middle group in each panel shows metabolites that cycle under both conditions, while those on either side indicate metabolites that cycle in only one condition. The number indicates the average normalized counts in each condition for metabolites that cycle with p < 0.05, JTK_cycle. Data are mean ± SEM, comparisons were with student’s t test. D: AS versus AM for primary metabolites; E: AS versus AM for lipids; F: AS versus AM for biogenic amines.

G-I: Average expression levels of gut metabolites that cycle in microbiome-containing versus sterile flies fed on high sugar diets. The middle group in each panel shows metabolites that cycle under both conditions, while those on either side indicate metabolites that cycle in only one condition. The number indicates the average normalized counts in each condition for metabolites that cycle with p < 0.05, JTK_cycle. Data are mean ± SEM, comparisons were with Student’s t test. G: AS versus AM for primary metabolites; H: AS versus AM for lipids; I: AS versus AM for biogenic amines.

Effects of the circadian clock and dietary composition on the phase of metabolite cycling in microbiome-containing and sterile flies (these data are the same as Figure1-4).

(A-D) Polar histogram plots of the peak phase for oscillating primary, lipids, and biogenic amines, respectively, in microbiome-containing and sterile flies, using a JTK_cycle value of P < 0.05. A: iso31 flies fed on normal chow diets; B: per01 flies fed on normal chow diets; C: iso31 flies fed on high protein diets; D: iso31 flies fed on high sugar diets.

Loss of the circadian clock, high protein or high sugar diets decrease metabolite cycling in microbiome-containing flies.

(A,B) Pie charts depicting the fraction of metabolites cycling in AM and AS iso31 flies fed on normal chow diets; (C,D) Pie chart depicting the fraction of metabolites cycling in TM and TS iso31 flies fed on normal chow diets; (E,F) Pie charts depicting the fraction of metabolites cycling in AM and AS per01 flies fed on normal chow diets; (G,H) Pie chart depicting the fraction of metabolites cycling in AM and AS iso31 flies fed on high protein diets; (I,J) Pie chart depicting the fraction of metabolites cycling in AM and AS iso31 flies fed on high sugar diets, using a JTK_cycle cutoff of P < 0.05.