Klf15 orchestrates circadian nitrogen homeostasis

Abstract

Diurnal variation in nitrogen homeostasis is observed across phylogeny. But whether these are endogenous rhythms, and if so, molecular mechanisms that link nitrogen homeostasis to the circadian clock remain unknown. Here, we provide evidence that a clock-dependent peripheral oscillator, Krüppel-like factor 15 transcriptionally coordinates rhythmic expression of multiple enzymes involved in mammalian nitrogen homeostasis. In particular, Krüppel-like factor 15-deficient mice exhibit no discernable amino acid rhythm, and the rhythmicity of ammonia to urea detoxification is impaired. Of the external cues, feeding plays a dominant role in modulating Krüppel-like factor 15 rhythm and nitrogen homeostasis. Further, when all behavioral, environmental and dietary cues were controlled in humans, nitrogen homeostasis exhibited an endogenous circadian rhythmicity. Thus, in mammals, nitrogen homeostasis exhibits circadian rhythmicity, and is orchestrated by Krüppel-like factor 15.

Figure 1. Klf15 exhibits 24-hour periodicity and is driven by the core clock machinery

(a) Klf15 mRNA accumulation from WT mice livers (n=5 per time point). (b) Representative KLF15 and U2AF65 protein expression from WT and Klf15-null liver nuclei. (c) KLF15 protein densitometry from three replicates. (d) Klf15-luciferase is induced on a dose dependent fashion by CLOCK/BMAL1, and inset illustrates four “E Box” motifs in the Klf15 promoter (−5 kb). (e) Klf15 mRNA accumulation in WT and Bmal1 KO livers. (g) (h) Rhythmic binding of BMAL1 on the Klf15 promoter (n=3 per time point). Data presented as mean ± SEM.

Figure 2. Nitrogen homeostasis exhibits 24-hour periodicity, driven by Klf15, in mice

(a–d) Plasma total amino acid pool, alanine, BCAA and urea measured every four hours over a circadian period after placing mice in constant darkness for 38 hours (n=5 per time point). The data are double-plotted, and ANOVA was used to determine rhythmicity. (e) Cumulative food intake measured every 5 minutes in WT and Klf15-null mice (n=4 per group). (f) Total body weights of WT and Klf15-null mice (n=4 per group). (g, h) Plasma total AA pool, urea from WT and Klf15-null mice measured every four hours under L/D and double plotted to illustrate rhythmicity (n=5 per group per time point). Data presented as mean ± SEM.

Figure 3. Klf15 regulates rhythmic amino acid utilization

Schematic illustrates the inter-organ transport, utilization of amino acids “the glucose-alanine cycle.” (a, b) Skeletal muscle Alt and Bcat2 expression in WT and KLF15-null mice (n=4 per group per time point). (c, d) Plasma alanine and BCAA in WT and Klf15-null mice (n=5 per group per time point) (e) Plasma glucose in WT and Klf15-null mice (n=5 per group per time point). (f) ChIP for KLF15 on Alt promoter (n=3 per time point). (# p<0.05 at all time points between WT and Klf15-null). Data presented as mean ± SEM.

Figure 4. Klf15 regulates rhythmic nitrogeneous waste excretion

Schematic illustrates the excretion of nitrogenous waste products, i.e., “the urea cycle.” (a, b) Plasma glutamate, ammonia in WT and Klf15-null mice (n=5 per group per time point). (c) Otc expression in WT and Klf15-null livers (n=4 per group per time point). (d) Plasma ornithine in WT and Klf15-null mice (n=5 per group per time point). (e) OTC enzymatic activity measured from liver mitochondrial extracts from WT and KLF15-null mice (n=4 per group). (f, g) Urinary levels of urea and ammonia in WT and KLF15-null mice (n=5 per group). (h) ChIP for KLF15 on the Otc promoter (n=3 per time point). (# p<0.05 at all time points between WT and Klf15-null). (i–k)Results of neurobehavioral testing for (i) Y-maze, a test of working memory (n=3 per group) (j) Fear conditioning during contextual changes (hippocampal function) and altered cues (amygdalar function) (n=8 per group) and (k) Morris water maze test, a test of hippocampal function (n=8 per group). Data presented as mean ± SEM.

Figure 5. Feeding alters Klf15 expression and nitrogen homeostatsis

Following ad-libitum or feeding restricted to the light-phase (ZT3-ZT9) for one month: (a, b & c) liver expression of Clock, Per2 and Cry1, (d, e) liver and skeletal muscle expression of Klf15, (f, g & h) total plasma AA, ammonia and urea concentrations (I, j) skeletal muscle Alt and Bcat2 expression (k) liver expression of Otc (l, m, n & o) plasma ornithine, alanine, BCAA, insulin (n=5 per group per time point). Data presented as mean ± SEM.

Figure 6. WT and Klf15-null adaptation to high protein diet

Following one week of high protein diet (70% casein) or normal diet (18% protein) WT and Klf15-null (a) blood glucose (n=5 per group), (b) plasma amino acids, (c, d) liver expression of Klf15, Alt, Otc (d) plasma ammonia and (e) plasma urea. Data presented as mean ± SEM.

Figure 7. Human endogeneous circadian nitrogen homeostasis

Fasting total plasma amino acid pool, alanine, BCAA and urea exhibit endogenous circadian rhythmicity in humans during the forced desynchrony protocol. The cosine models (black lines) and 95% confidence intervals (gray areas) are based on mixed model analyses and use precise circadian phase data. To show that these models adequately fit the actual data, we also plot the proportional changes across 60 circadian degree windows per individual multiplied by the group average with SEM error bars (open circles with error bars). Data are double plotted to aid visualization of rhythmicity. Lower X-axis, circadian phase in degrees, with fitted core body temperature minimum assigned 0-degrees, and 360 degrees equal to the individual circadian period (group average 24.09h); top X-axis, corresponding average clocktime for these individuals; vertical dotted lines, core body temperature minimum; horizontal gray bars, corresponding average habitual sleep episode in the two weeks prior to admission.