The Microbiome as a Circadian Coordinator of Metabolism

Abstract

The microbiome is critically involved in the regulation of systemic metabolism. An important but poorly understood facet of this regulation is the diurnal activity of the microbiome. Herein, we summarize recent developments in our understanding of the diurnal properties of the microbiome and their integration into the circadian regulation of organismal metabolism. The microbiome may be involved in the detrimental consequences of circadian disruption for host metabolism and the development of metabolic disease. At the same time, the mechanisms by which microbiome diurnal activity is integrated into host physiology reveal several translational opportunities by which the time of day can be harnessed to optimize microbiome-based therapies. The study of circadian microbiome properties may thus provide a new avenue for treating disorders associated with circadian disruption from the gut.

All forms of life on earth, from microscopic to more complex organisms, have adapted to the light and dark cycles in 24-hour periods. They have developed circadian rhythms—a coordinated and entrainable mechanism that orchestrates sleep-wake cycles, feeding patterns, hormone secretions, metabolic homeostasis and body temperature (1, 2).

The most primordial form of a circadian clock can be found in unicellular prokaryotic organisms like cyanobacteria (3). Cyanobacteria are very abundant in the oceans and make up a large proportion of the total global photosynthetic activity (4). Initially, it was believed that simple prokaryotes like cyanobacteria with dividing times shorter than the day-night cycle could not evolve elaborate circadian timing mechanisms (5). However, the circadian clocks of prokaryotes have emerged as prototypical examples for robust timekeeping systems in unicellular organisms, and the cyanobacterium Synechococcus elongatus PCC7942 has developed into a model for circadian research (6). Virtually all S elongatus promoters are under circadian control (7).

Interestingly, a critical feature in the cyanobacterial clock is rhythmic posttranslational modification. The central clock protein in cyanobacteria is the hexameric ATPase KaiC, and its phosphorylation status follows a circadian rhythm that is tightly regulated by KaiA and KaiB (8-10). KaiC consists of 2 P-loop ATPase domains, CI and CII (11). At dawn, KaiA binds to the CII domain of unphosphorylated KaiC, where it promotes autophosphorylation of the residues Ser-431 and Thr-432 (12, 13). When KaiC reaches the double phosphorylated state, its autokinase activity switches to an autophosphatase activity, leading to dephosphorylation of Thr-432 (13). In the resulting state (T432/pS431), which is usually reached in the evening, KaiB binds to KaiC at the CI domain and KaiA is recruited to the KaiB-KaiC complex, where it is inactivated and sequestered from CII during the night. In the absence of KaiA at CII, KaiC is further dephosphorylated (13). Around dawn, phosphorylation of KaiC reaches a minimum and the affinity for KaiB is lost (14). KaiA and KaiB are then released, allowing KaiA to bind to the CII domain of unphosphorylated KaiC again and start a new cycle. The molecular oscillator consisting of KaiA, KaiB, and KaiC can be reconstituted in vitro by mixing the 3 proteins with adenosine 5′-triphosphate (8). Because this central clock system functions in vitro without transcription or translation, it is considered to be a posttranslational oscillator.

Although the occurrence of circadian rhythms in photosynthetic cyanobacteria has been clearly demonstrated, it remains to be investigated in detail if nonphotosynthetic bacteria also possess a circadian timing system. Interestingly, homologues of KaiC have been found in 3 bacterial taxa apart from cyanobacteria, namely in Proteobacteria, Thermotogae, and Chloroflexi (15, 16). Some of these Proteobacteria form close associations with organisms possessing a circadian timing mechanism, like the nitrogen-fixing rhizobacterium Sinorhizobium medicae. This soil bacterium forms root nodules on plants of the Medicago genus in the legume family, leading to a symbiosis between the plant and the bacteria: The plant benefits from nitrogen fixation while S medicae obtains important nutrients from the plant (17). Moreover, it has been shown that plants possess a diverse microbiome that is crucial for their performance and exhibits a circadian rhythm. In plants as well as in humans, the microbiome is important for nutrient uptake, modulation of the host immune system, and prevention of colonization by pathogens (18). The plant circadian clock shapes the community structure of the microbiome, most likely via day-night differences in root exudates and water accessibility (19). Understanding circadian host-microbiome interactions in plants not only has important agricultural and economic implications, but may also serve as a blueprint for understanding this interplay in mammalian systems.

Similar to cyanobacteria, more complex multicellular organisms possess a clock composed of feedback loop mechanisms. In contrast to phosphorylation events, however, the main players of the mammalian clock are transcription factors: brain and muscle ARNT-like 1 (BMAL1) and circadian locomotor output cycles kaput (CLOCK) form heterodimers, translocate into the nucleus, and drive the expression of multiple clock-controlled genes including their own repressors Period and Cryptochrome (20). Several other loops that control BMAL1/CLOCK expression have been described, and hence through modulation of activators and repressors, cells maintain cyclical gene expression in response to environmental cues (21, 22).

In mammals, there is a master regulator located in the suprachiasmatic nucleus (SCN) of the hypothalamus (Fig. 1). It is composed of thousands of neurons, each with an intrinsic clock oscillating in harmony with each other. The SCN responds to signals transmitted from photosensitive receptors in the retina and other stimuli from other brain regions (23). Moreover, the SCN neurons project their axons to multiple regions of the brain to orchestrate circadian physiology and behavior, including locomotor activity, feeding, and hormone secretion (24). The SCN also projects to the paraventricular nucleus, which has been demonstrated to have a role in the corticotropin-releasing hormone–adrenocorticotropic hormone axis and ultimately secretion of corticosterone by the adrenal glands (25).

Figure 1.

Integration of the microbiome into the circadian network of host metabolism.

Undoubtedly, the SCN is critical for coordinating multiple functions throughout the body; however, most cells also contain endogenous clocks that can exhibit autonomous oscillations even in the absence of SCN synchronizing activity (see Fig. 1). For example, in studies in which liver tissue has been explanted from mice, circadian expression of a group of genes can be maintained without input from the SCN (26). Some of these tissue-intrinsic clocks possess key functions in the circadian regulation of metabolism, ranging from glucose homeostasis to lipid metabolism (22, 27-30). These activities are regulated by intrinsic clocks of the liver, heart, kidneys, skeletal muscle, and pancreas, among others (31). One of the first examples that highlighted the SCN-independent oscillation of metabolic genes in the liver focused on glycogen synthase, which showed robust fluctuations even in SCN-lesioned animals (32).

Environmental cues that entrain an organism’s biological rhythms to the earth’s 24-hour cycles are known as zeitgeber. As mentioned earlier, the SCN recognizes light and dark cycles that correspond to an organism’s active or restful periods depending on whether it is diurnal or nocturnal. During the active phase, an additional dominant zeitgeber for peripheral clocks is feeding, which stimulates hormones, produces food metabolites, and leads to body temperature variations that ultimately drive the expression of genes not only in organs directly involved in digestion of food but in the entire organism (33, 34).

In the following, we will discuss a new player in the circadian interplay between tissues and their involvement in the regulation of metabolism: the gut microbiome, the trillions of microorganisms inhabiting the gastrointestinal tract.

Rhythmic Features of the Microbiome: Moving Pictures of an Ecosystem

Among the many lifestyle factors involved in shaping the composition and function of the microbiome, diet has emerged as the most dominant one (35-37). Experiments in mice have demonstrated dramatic effects of various diets on microbiome composition, including a high-fat, high-sugar “Western” diet, and reduced-carbohydrate, vegetarian, and cooked diets (38, 39). Despite the apparent malleable nature of the microbiome, it has also been observed in long-term studies that the microbiome is remarkably stable (40-42). The short-term responsiveness and adaptable nature of the microbiome along with evidence for its long-term stability suggests that short-term fluctuations in microbiome composition may revolve around specific stable states that persist for extended periods of time (43).

Diet, similar to many of the other environmental factors that influence the microbiome, is inherently rhythmic in nature because of the diurnal human lifestyle. Consistently, recent studies have identified daily oscillations in many of the most important features of the intestinal microbiome (44-47). Through time-resolved longitudinal sampling of the microbiome, it was demonstrated that groups of bacteria exhibited distinct patterns of rhythmicity, with each time of day having a characteristic microbiota composition (48). The diurnal properties of the microbiome go beyond its taxonomic composition, and also include its localization within the gastrointestinal tract, and in particular the proximity of bacteria to mucosal surfaces, as well as its metagenomic function and metabolite secretion (46, 49).

The host circadian clock and the microbiome are tightly connected as seen in mouse models deficient of key circadian rhythm mediators. Both in Bmal1- and Per1/2-deficient mice, sequential sampling of the microbiome showed significant loss of rhythmicity and reduced numbers of cycling microbial species (or their computational proxy: operational taxonomic units) (45, 47).

What is the connection between the host circadian clock and microbiome rhythms? Numerous factors may contribute to imposing rhythmic pressures on the microbial ecosystem, including immune system rhythmicity (50, 51), intestinal epithelial rhythms (52), and rhythmic metabolic activity (53). First evidence came from mice fed a high-fat diet, which showed a significant reduction in oscillatory behavior of their microbiome (44). High-fat feeding, in addition to massively altering intestinal nutrient contents and microbiome composition, also modifies host behavior in that it dampens the diurnal rhythmicity of food intake (54). Similarly, the core clock is involved in maintaining a rhythmic structure of feeding times (55). In Per 1/2-deficient mice, in which rhythmic feeding was experimentally restored by limiting food intake to 12 hours per day, microbiome oscillations were likewise restored on the level of microbial taxonomy, biogeography, and metabolite secretion (45, 49). Time-restricted feeding under high-fat diet conditions partially restored daily microbiome fluctuations (44). Furthermore, restricting food intake to opposite times of the day provokes a 12-hour phase shift in microbiome oscillations (45, 49). Thus, although certain feeding-independent microbiome oscillations have also been observed (46), times of food intake appear to be the predominant zeitgeber for rhythmic microbiome activities.

What is driving these food-mediated bacterial oscillations? Little is known about the rapid mechanisms by which intestinal microbial ecology responds to daily cycles of nutrient influx and starvation. Initial insight came from estimates of bacterial proliferation rates in the microbiome using metagenomic sequencing. In time-resolved sequencing studies, examples were found in which bacterial proliferation rates mimicked their diurnal oscillation patterns and preceded the changes in relative abundances (56). These findings suggest that rhythmic proliferation of certain species in the microbiome drives oscillations in community structure. Other possible factors include rhythmic passage along the gastrointestinal tract and rhythmic death of cells due to nutrient fluctuations.

The Tripartite Interactions Between Host, Clock, and Microbiome

The downstream impact of these microbiome rhythms on the host is considerable (57). A sizable fraction of genes in a given tissue, typically around 15%, is undergoing rhythmic transcription, many of which are driven by the molecular clock and feeding rhythms (58, 59). A number of studies have determined that this set of rhythmic genes is also influenced by the state of microbial colonization. This phenomenon has so far been studied in the context of the gastrointestinal tract (49, 52, 60), the liver (46, 49, 60-62), white adipose tissue (60), the kidney (63), and the brain (46), although the latter 2 organs were not assessed in a genome-wide manner. These transcriptional alterations may quite profoundly influence the diurnal activity of the tissue. For instance, in the case of hepatic drug metabolism, it has been found that the time of day-dependency in acetaminophen-induced liver damage is mediated by the microbiome (49, 64).

How are microbiome-derived signals integrated into the circadian regulation of gene expression in different organs? Perhaps the most intuitive form by which the microbiome can shape rhythmic gene expression is by direct recognition of microbial molecules. This is typically accomplished by pattern recognition receptors (PRRs), whose activation leads to changes in transcription. PRR signaling in intestinal epithelial cells influences the circadian function of these cells, including day-night patterns in epithelial hormone secretion (52). Rhythmic Toll-like receptor expression and microbial detection lead to oscillating JNK and IKKβ signaling, which together with the clock elements RORα and Rev-erbα drive rhythmic gene expression in intestinal epithelial cells (52).

Microbial recognition can also influence rhythmic epithelial gene expression via an immune cell relay mechanism. One such signaling relay involves the leucine-zipper protein NFIL3. PRR signaling via the signaling adapter MyD88 on myeloid cells triggers interleukin (IL)-22 release from innate lymphoid cells via IL-23. IL-22, in turn, activates STAT3 signaling in intestinal epithelial cells, which feeds into oscillating gene expression via the clock component Rev-erbα and NFIL3 (65, 66). NFIL3 controls the expression of genes involved in lipid uptake, and thus the microbiome indirectly controls rhythmic lipid metabolism in epithelial cells.

In addition to direct microbial recognition, the microbiome also influences transcript oscillation patterns via the release of metabolites. Given the circadian oscillation of a myriad of microbial genes and pathways in the metagenome, it is perhaps not surprising that specific microbiota-associated metabolites likewise show diurnal variation (46, 49). Among the metabolites reported to undergo diurnal variations in abundance are short-chain fatty acids. Treatment of hepatic organoids with butyrate induced shifts in rhythmicity and enhanced the amplitude of circadian clock gene expression, revealing a mechanism by which the microbiome may act on peripheral organs (46). Within the gut lumen similar patterns are observed with cecal acetate and butyrate exhibiting rhythmicity, which is lost with antibiotic treatment (63). Administration of short-chain fatty acids at different times of day advances peak-phase expression of circadian clock genes in the kidney and liver in a time of day–dependent manner. Given that the production of the ketone body β-hydroxybutyrate is under circadian control and regulates food anticipatory activity (67), it is possible that the microbiome is also involved in the regulation of food anticipation. Another example of rhythmic metabolites with clock-modulating capacity is the group of polyamines. Modulation of core clock function by polyamines might be an additional mechanism by which the microbiome exerts its impact on global rhythmic transcription in the liver (49, 68).

As mentioned earlier, organisms acquire essential nutrients and energy through daily feeding. At the same time, however, macronutrient processing generates metabolites that need to be detoxified and excreted. In mammals, the liver, kidneys, and intestines function as filters that modify and/or remove metabolites. These organs possess receptors that recognize these xenobiotics. Interestingly, the expression of many of these receptors fluctuates in a circadian fashion. For example, the peroxisome proliferator activated receptor-γ (PPAR-γ) is activated by fatty acids in the liver and the intestine, and stimulates the expression of cytochrome P450 family of enzymes that mediate their metabolism (69). Interestingly, PPAR-γ mediates a portion of the de novo oscillatory gene expression in response to dietary and microbiome perturbations (62, 70). Another important xenobiotic receptor is the aryl hydrocarbon receptor (AhR), which is activated by polycyclic hydrocarbons that originate from digestion of fruits and vegetables (71). Similar to PPAR-γ, AhR exhibits diurnal expression under the control of BMAL1. Given that many metabolites regulating the activity of these receptors are microbiome derived, it will be intriguing to understand whether their circadian regulation is driven by the intestinal microbiome.

Furthermore, accumulating evidence suggests that the microbiome may influence oscillating gene expression via epigenetic modifications. The histone marks H3K4me2, H3K4me3, H3K27ac, and H3K9ac undergo rhythmic changes in abundance in intestinal epithelial cells over the course of a day, which are markedly disturbed in germ-free or antibiotic-treated mice (49, 72). This effect is mediated by microbiome-driven expression of the histone deacetylase HDAC3. Like NFIL3, HDAC3 controls the expression of epithelial genes involved in lipid metabolism, and mice lacking epithelial HDAC3 show reduced body fat content, similar to germ-free mice and NFIL3-deficient mice (65, 72). Of note, some of the metabolite-based and epigenome-based mechanisms described here might be functionally linked. For instance, short-chain fatty acids have been shown to have HDAC-inhibitory effects (73), although the underlying mechanisms remain to be defined.

Finally, recent data suggest that the microbiome influences rhythmic gene expression through the regulation of hormones. A number of hormones that undergo diurnal fluctuations in abundance are influenced by intestinal microbial colonization, including growth hormone and sex steroids (74). Interestingly, diurnal oscillations both of the microbiome (47) and the host transcriptome (60) show differences between male and female mice. Germ-free mice depict a marked dampening of these sexual differences, likely mediated by microbial regulation of growth hormone levels. This is accompanied by less pronounced differences in the abundance of the sex hormones testosterone and estradiol between male and female germ-free mice (60). In addition to these circulating hormones, several hormone receptors are also under circadian control and influenced by the microbiome. Examples include not only the receptors PPAR-γ and AhR as discussed earlier, but also the pregnane X receptor and farnesoid X receptor, both of which are dysregulated in germ-free mice (74).

In summary, these findings highlight that the microbiome generates a characteristic diurnal metagenome and metabolome and elicits specific circadian patterns of gene expression in the host via microbial recognition pathways, metabolic and epigenetic modulation, as well as hormonal signaling (see Fig. 1).

The Translational Aspects of Daily Host-Microbiome Interactions: It Is About Time

Disruption of the circadian clock is a common feature of multiple metabolic diseases (30). For instance, shift work has been linked to an enhanced risk for type 2 diabetes in epidemiological studies (75, 76), but the underlying mechanisms are still poorly defined. Studies have shown a link between hours of sleep and risk of obesity (77) as well as shift work and atherosclerotic disease (78). Similar associations between circadian disruption, the time of food intake, and metabolic disease have been revealed in animal models (79-81). Clock-mutant mice exhibit hallmarks of metabolic syndrome (82), and similar phenotypes are observed in Clock-deficient mice, albeit to a lesser extent (83, 84). Whole-body Cry1/2 knockout animals and liver-specific Bmal1 and Rev-erbα/β knockout mice exhibit higher body weights, hepatic steatosis, and dyslipidemia on high-fat diets (85). Time-restricted feeding is able to “rescue” these mice from metabolic syndrome, demonstrating that the rhythmicity of feeding may be crucial for maintaining metabolic homeostasis, independent of the molecular clock. Some of the most compelling evidence of a causal link between the circadian clock and metabolic dysfunction comes from experiments mimicking jet lag. “Jet-lagged” mice exhibit altered behavioral patterns and enhanced susceptibility to weight gain and systemic metabolic derangements (86). These studies demonstrate a strong relationship between metabolic syndrome, energy intake, and deviation from normal circadian rhythmicity. Given the recent data on diurnal properties of the microbiome and their influence on host metabolism, might the microbiome influence this connection between circadian disruption and disease (87)?

Indeed, circadian disruption perturbs the oscillations and composition of the intestinal microbial community (45, 88, 89), and microbiome transplantations from jet-lagged mice into germ-free mice recapitulate certain hallmarks of metabolic syndrome. Conversely, antibiotic depletion of the microbiome in jet-lagged mice prevents the occurrence of body fat accumulation and hyperglycemia (45). As mentioned earlier, germ-free mice or antibiotics-treated mice are characterized by reduced body fat content, which is recapitulated by mice lacking epithelial NFIL3 or HDAC3 (65, 72), further reinforcing the importance of epithelial lipid handling in mediating the impact of the microbiome on metabolic problems associated with clock disruption.

Whether the link between circadian disruption and other manifestations of metabolic syndrome and comorbidities of obesity, such as atherosclerosis, are likewise influenced by the microbiome remains to be investigated in detail.

The Future of Host-Microbiome Rhythmicity: Time in Motion

The role of the microbiome in circadian biology is a young field of study, and major questions remain unanswered. Among the most mesmerizing aspects of diurnal microbiome activity is the ecosystem-wide coordination of abundance fluctuations. The precise mechanisms underlying this coordination, the key microbial species driving rhythmicity, and the microbial codependencies that form an oscillating network of activity remain to be elucidated. Involved in all these aspects is the question of whether members of the commensal microbiome possess cell-intrinsic time-keeping mechanisms, analogous to those observed in cyanobacteria (90, 91). Combined time-resolved metagenomic and metaproteomic approaches will be needed to answer this intriguing question.

Likewise, the study of how microbiome rhythmicity influences the circadian biology of the host is still in its infancy. A few examples have been found of how the microbiome can shape the circadian activity of intestinal and extraintestinal tissues, but the universality of this effect is unclear (92). Similarly, it remains to be studied how many layers of circadian activity of a tissue are affected by the microbiome. The evidence that is available to date indicates that this effect could go far beyond rhythmic transcription. Ultimately, mechanistic insights into the metabolic circuits that are regulated by the microbiome in a circadian fashion are needed to bring us closer to understanding how the microbiome shapes metabolic disorders associated with circadian disruption.

Another major influencer of circadian rhythms is age. Several aspects of circadian molecular biology lose robustness with increasing age, as do behavioral outputs of circadian activity (68, 93). Some elements of this decline in circadian properties may be linked to age-associated alterations in the microbiome. Accumulating evidence indicates that the composition and function of the microbiome undergoes changes as the host ages, and that these age-associated alterations of the microbiome influence the course of host aging (94). Although the connection between the microbiome and the age-associated decline of circadian function has not been investigated, it is conceivable that microbiome oscillations are similarly affected by organismal aging and that impaired microbiome properties are causally involved in the circadian dysfunction of aged hosts.

An area of research that will strongly benefit from integration of knowledge about microbiome diurnal activity is chronopharmacology. Drug prescriptions in most cases do not include specific recommendations about the time of day that is optimal to maximize efficiency and minimize toxicity of the drug. Harnessing the increasing knowledge about circadian host-drug interactions is warranted to provide improvements and limit the interpersonal variability in pharmacological responses (95). The microbiome is a critical element of drug metabolism because it performs numerous chemical transformations that either activate or deactivate xenobiotic compounds (96, 97). Thus, the microbiome might constitute an underused element in chronopharmacology. A more detailed understanding of the time of day–dependency of microbial biotransformation reactions may greatly enhance our ability to predict the optimal time of drug consumption.

Finally, for all the listed points to come to fruition, the study of circadian host-microbiome interactions must increasingly focus on humans. Initial evidence suggests that various body-associated elements of the microbiome undergo rhythmic oscillations (45, 98-100), but the impact of these rhythms on the circadian biology of humans remains obscure. Time-resolved longitudinal sampling in large human cohorts is required to determine the strength of the circadian signal in diverse and nonsynchronized human populations with a large range of lifestyles. In addition, functional studies linking microbiome properties to daily activity patterns and molecular readouts will be necessary to probe the impact of the microbiome on human circadian biology. The next decade of basic science research is likely to reveal how much of the microbiome's impact on host physiology is driven by its diurnal properties.

Glossary

Abbreviations

AhR

aryl hydrocarbon receptor

BMAL1

brain and muscle ARNT-like 1

CLOCK

circadian locomotor output cycles kaput

PPAR-γ

peroxisome proliferator activated receptor-γ

PRR

pattern recognition receptor

SCN

suprachiasmatic nucleus

STAT

signal transducer and activator of transcription

Additional Information

Disclosure Summary: The authors have nothing to disclose.

Data Availability. Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.