Circadian rhythms are a hallmark of living organisms, observable in all walks of life from primitive bacteria to highly complex humans. They are believed to have evolved to co-ordinate the timing of biological and behavioural processes to the changing environmental needs brought on by the progression of day and night through the 24-h cycle. Most of the modern study of circadian rhythms has centred on so-called TTFLs (transcription–translation feedback loops), wherein a core group of ‘clock’ genes, capable of negatively regulating themselves, produce oscillations with a period of approximately 24 h. Recently, however, the prevalence of the TTFL paradigm has been challenged by a series of findings wherein circadian rhythms, in the form of redox reactions, persist in the absence of transcriptional cycles. We have found that circadian cycles of oxidation and reduction are conserved across all domains of life, strongly suggesting that non-TTFL mechanisms work in parallel with the canonical genetic processes of timekeeping to generate the cyclical cellular and behavioural phenotypes that we commonly recognize as circadian rhythms.
Circadian rhythms are arguably one of life's most basic mechanisms, regulating the timekeeping of a large swath of biological processes at the molecular, metabolic and behavioural level. It is widely believed that the evolution of these rhythms offers key biological advantages to organisms on Earth, conferring anticipation and the flexibility to adapt to cyclical changes in the surrounding environment, such as light, temperature and the availability of food. To this day, the study of circadian rhythms remains one of the most robust models, wherein genetic, cellular and environmental perturbations can be linked directly to behaviour.
The importance of circadian rhythms is underlined by daily oscillations of behaviour that occur that make up the temporal fabric of human life; every day and night, we experience wakefulness, sleep and hunger. The necessity of circadian rhythms for the continued function of the organism are many and varied; studies have shown that 7–21% of genes in the liver and adipose tissues follow a daily oscillatory pattern . In addition, chronic disruption of circadian rhythms by shiftwork, sleep deprivation and frequent jetlag result in higher incidences of metabolic disorders such as insulin resistance, obesity and diabetes [2,3], mood disorders [4,5] and certain types of cancer [6,7].
Circadian rhythms, autonomous and self-generating, are ubiquitous throughout multicellular and unicellular organisms, and have been largely demonstrated by the oscillatory expression of genes and proteins. From decades of research into biological clocks, a prevailing mechanistic model has emerged in the literature: circadian oscillations are generated by cycles of transcription and translation known as TTFLs (transcription–translation feedback loops), where rhythmic protein expression negatively regulates the expression of genes to form a negative-feedback loop that repeats its cycle every day.
In the present article, we re-examine the predominant TTFL paradigm of circadian rhythms in the light of the recent discoveries of oscillations in the activity of Prxs (peroxiredoxins) and redox homoeostasis. By doing this, we hope to provide a cohesive view of how circadian rhythms shape life on our planet. That these rhythms are conserved in all domains of life, an evolutionary consequence of the rotation of Earth on its axis, implies in itself that they are key in maintaining the well-being of organisms.
The origin and features of circadian rhythms
The origins of the study of biological timekeeping stem from observation made by the French astronomer Jean-Jacques d’Ortous de Mairan in 1729, who noted patterns in the movement of the leaves of the Mimosa plant with a periodicity of 24 h, even when the plant was kept in constant darkness , thus describing the concept that a 24-h rhythm was endogenous, i.e. that it persisted without external cues. The biological clock responsible for driving the synchrony of the ‘behaviours’ of organisms to anticipate solar cycles was conceptualized with subsequent research, and its features were expanded to include other criteria. These include that the clock must be ‘entrainable’ by external stimuli known as ‘zeitgebers’ (‘time-givers’ in German), such as light or feeding cues [8,9], and maintain 24-h periodicity over a range of physiological temperatures (known as ‘temperature compensation’) . Biochemically, the latter is a curious property, since most cellular reactions would be expected to double their rate for every 10°C increase in temperature, which is not the case for circadian clocks. The mechanistic basis of this property remains obscure.
Although the precise mechanisms driving these observed endogenous rhythms long remained elusive, evidence that genetic mutations were capable of regulating circadian behaviours was first discovered in Drosophila melanogaster , eventually leading to the discovery of the clock gene Period (Per) . Further genetic manipulations elucidated further ‘clock genes’, thereby cementing the concept that the clockwork in eukaryotic organisms is driven by cycles of gene transcription and subsequent translation. The mechanism behind these self-generating oscillations was first proposed by Hardin et al.  to be a negative-feedback loop, and has since formed the basis for the predominant mechanistic explanation for circadian oscillations in both plants  and animals .
Constructing a clockwork: TTFLs and post-translational modifications in eukaryotes
The general features of TTFLs are thought to be grossly similar between fruitflies and mammals. The clockwork comprises both positive and negative components, wherein the positive loops provide an activating transcriptional drive and the negative loops serve to block these positive components. In mammals, the positive elements of this feedback loop are the proteins BMAL1 and CLOCK. BMAL1 and CLOCK form heterodimers and bind to E-box-response elements in the promoter regions of the Per, Cryptochrome (Cry) and Rev-erb families of genes to activate transcription. The REV-ERB proteins inhibit Bmal1 transcription, and the PER–CRY heterodimers enter the nucleus and repress the activity of BMAL1–CLOCK, thereby inhibiting their own transcription. Decline in the transcriptional drive eventually results in the derepression of Bmal1 transcription, which is also activated by members of the ROR (retinoic acid-related orphan receptor) family of proteins, binding to Rev-erb/ROR-response elements in the Bmal1 promoter region. These events drive yet another feedback loop that is responsible for the positive drive of the transcriptional clockwork [15,16]. The TTFL model (Figure 1) provides a basic explanation of how cells can generate an autonomous oscillation, with the alternating expression of proteins such as PER and BMAL1 highlighting the peaks and troughs (or opposite phases) of the circadian cycle.
A simplified schematic diagram of the mammalian TTFL
Although the clock genes discovered originally in fruitflies are broadly homologous among many species of the kingdom Animalia, discrepancies exist when we begin to consider other eukaryotes. In the fungus Neurospora crassa, the basic architecture of the TTFL is also employed as the generator of the core circadian oscillator; however, the positive and negative components of the clockwork, white collar (wc) and frequency (frq) respectively, share little homology with their Animalia counterparts. Circadian clocks have also been rigorously studied in the small flowering plant Arabidopsis thaliana, in which the only conserved clock factors in the TTFL are cryptochrome and protein kinase CK2α . From these differences across phylogenetic kingdoms, one might infer that the lack of shared clock components indicates that circadian timekeeping evolved independently in these different lineages.
In addition to the arrangement of feedback loops, eukaryotic organisms also share a wide variety of post-translational mechanisms that are capable of modulating cellular timekeeping. Phosphorylation state is the best understood mechanism, with players such as the casein kinases able to target the negative arms of the TTFL in many organisms, across different phyla [17–19], regulating the cellular localization and degradation of clock proteins. Ubiquitylation, SUMOylation, proteasomal degradation, histone acetylation and chromatin remodelling are also known to play roles in the regulation of clock genes and proteins . The common interpretation of the effects of post-translational regulation on circadian rhythmicity is that they exist mainly as fine-tuning mechanisms, modulating the flexibility of the TTFL by shifting its phase, or by altering the period of the timing circuitry.
Oscillations in prokaryotes
It was widely thought that circadian rhythms did not exist in prokaryotic organisms, since their cell cycle is typically much shorter than a day, and therefore it seemed unlikely that they would need a 24-h clock to orchestrate their physiology. However, in the 1980s and 1990s, rhythms of nitrogen fixation and photosynthesis were discovered in cyanobacteria [20,21]. Subsequent studies revealed that a cluster of three adjacent genes, kaiA, kaiB and kaiC are essential for the generation of circadian rhythms in one such cyanobacterium, Synechococcus elongatus . Reiterating a similar mechanism echoing that in eukaryotes, a simplified TTFL was proposed, with KaiA positively regulating the transcription of the kaiBC operon, and KaiC serving as the negative arm of the feedback loop .
This model was called into question, however, when experiments revealed that rhythms of KaiC phosphorylation persisted independently of transcriptional and translational feedback mechanisms . A major breakthrough occurred when Kondo and colleagues reconstituted the system in vitro with the three Kai proteins and the addition of ATP . These sole constituents produced a self-sustainable 24-h oscillation of KaiC autophosphorylation and dephosphorylation that persisted when examined in vitro, a ‘test tube clock’. This identified the phosphorylation state of KaiC as the core pacemaker of the cyanobacterial system , and demonstrated for the first time that, at least in a prokaryote, gene transcription was dispensable for robust circadian rhythmicity.
The Kai proteins are, however, not highly conserved, or expressed as the obligatory triumvirate to generate circadian oscillations, even among other prokaryotes, and therefore the Kai-based phosphorylation model of non-transcriptional oscillators is restricted to cyanobacteria , and cannot be representative of a shared bacterial clock mechanism. This has resulted in a rather fragmented picture of the evolution of circadian rhythms throughout the domains of life, since only a few genes and mechanisms are shared between taxa (Figure 2).
Homologies and differences in core clock components between prokaryotes and eukaryotes
Redox mechanisms and circadian rhythms
Since Synechococcus obtains energy via photosynthesis, it was postulated that the cyanobacterial clock obtains information about the light environment by sensing cellular redox state through the proteins CikA and LdpA, and the redox-active quinone group of compounds [26,27]. Indeed, application of quinones to a mixture of Kai proteins in vitro has been shown to induce phase shifts, as well as synchronizing mixtures in different phases . A model was subsequently proposed wherein oxidized quinones bind to KaiA , resulting in its aggregation. KaiA subsequently modulates KaiC by ceasing the stimulation of its autophosphorylation. Although some aspects of the entrainment by the oxidation state of quinone remain unknown, it is clear that redox-sensitive mechanisms in cyanobacteria have been demonstrated to be a key component of how the oscillator senses its external environment.
Furthermore, it has been known for at least three decades that the cellular redox state in plants changes over circadian time , although many chronobiologists have long assumed that metabolic rhythms are a functional readout of the circadian clock, or that redox oscillators simply provide feedback upon the central TTFL pacemaker . Recent discoveries have, however, uncovered redox-based circadian oscillators that are conserved across both eukaryotic and prokaryotic species.
Redox reactions underpin many metabolic processes in cells. The redox state of a cell is a careful balance between the generation of oxidants [e.g. ROS (reactive oxygen species)] and the amount of reducing agent available. Since oxidants are generally regarded as damaging and toxic to cellular components, all organisms have evolved mechanisms, both enzymatic and non-enzymatic, to buffer oxidants. Well-known enzymes regulating cellular oxidative state include SOD (superoxide dismutase), which catalyses the dismutation of the superoxide radical (O2•−) into oxygen and hydrogen peroxide (H2O2), and catalase, which then brings about the decomposition of H2O2 into water and oxygen. Other systems, such those involving glutathione (GSH), Trx (thioredoxin) and Prx, complement these systems in different cellular compartments. GSH, for example, prevents oxidative damage to cell membranes by reducing lipid peroxides and maintaining the reduced state of protein thiol groups that are necessary for DNA repair . When oxidized, GSH forms a dimer (GS-SG), which is potentially toxic to the cell [33,34], but is normally reduced back to GSH by GRs (glutathione reductases) via an NADPH-dependent reaction . GPxs (glutathione peroxidases) are also well-understood antioxidant enzymes with an affinity for lipid peroxides, and may be involved in intracellular signalling [32,35]. Importantly, daily rhythms in the expression of GSH, and the activity of SOD, GPx and GR have been observed in several tissues in both fruitflies  and mammals [37,38].
Redox regulation by the peroxiredoxins
Prxs are a family of antioxidant molecules that modulate intracellular levels of H2O2, and are conserved in eukaryotes, prokaryotes and archaea . The Prxs are thought to have evolved from a Trx-like precursor , and are classed as 1-Cys or 2-Cys depending on the how the catalytic cysteine residues are provided. In the 1-Cys form, intramolecular disulfides are formed, whereas in the 2-Cys family, an intermolecular bond is formed, resulting in dimer formation. All peroxidase activity performed by Prxs can be summarized in three steps: peroxidation, resolution and recycling. They reduce peroxide through the oxidation (peroxidation) of a conserved ‘peroxidatic’ cysteine residue (CP), forming a sulfenic acid (CP-SOH) within the enzyme's active site. In the ‘typical’ 2-Cys class, CP forms a disulfide bond with a second free thiol group (CR, for resolving cysteine) near the C-terminus of a partner molecule, releasing a water molecule in the process (‘resolution’). The last step of the catalytic cycle is a recycling action performed by a Trx-like molecule, wherein the disulfide bond is broken and CP and CR are returned to their free thiol state  (Figure 3). In some eukaryotes, CP can undergo overoxidation to the sulfinic (CP-SO2H) form, and hyperoxidation into a sulfonic (CP-SO3H) state. The former can be reduced by Srx (sulfiredoxin) in an ATP-dependent reaction as part of redox regulation and repair mechanisms , whereas the sulfonic acid form is thought to be irreversible, and therefore potentially a marker of cumulative oxidative stress .
Activity of typical 2-Cys Prxs
Oxidation of Prxs: a new kind of circadian clock?
Circadian rhythms in Prx oxidation were first observed in human RBCs (red blood cells) , which are anucleate and are therefore a unique cellular model lacking transcriptional and translational mechanisms. Over- and hyper-oxidized Prx exhibit robust endogenous oscillations with a period of approximately 24 h. The period of these rhythms was also demonstrated to be temperature-compensated, i.e. the same at 37°C and 32°C, and entrainable by a temperature cycle, in which the peaks and troughs of Prx overoxidation coincide with low (36.8°C) and high (37.4°C) temperatures respectively .
In addition to human RBCs, circadian oscillation of Prxs have also been observed in Ostreococcus tauri , the smallest known alga. This organism has been previously notable in the chronobiology literature for sharing some circadian homology with the plant Arabidopsis, exhibiting oscillatory expression of TOC1 and CCA1 genes . The over- and hyper-oxidized forms of Prx were also found to be rhythmic in O. tauri, and these oscillations persisted in the absence of gene transcription, which shuts down shortly after the cells are transferred into total darkness, since the cells are phototrophic. If circadian rhythms were entirely dependent on TTFLs in this model, subsequently transferring these cells into a lit environment would result in a resetting of the circadian phase. This, however, was not observed, and cultured O. tauri exhibited a phase-dependent response to re-illumination, underlining that a non-transcriptional mechanism had been keeping time in the dark, when there is no de novo transcription. These experiments together thus demonstrated for the first time that cycles of Prx oxidation reflect circadian timekeeping in the absence of transcriptional mechanisms in eukaryotic systems.
Furthermore, Prx rhythms were found to be linked to TTFL oscillations, since long-period mutants exhibited similar periods of gene expression and Prx activity . So ubiquitous are Prxs and their functional activities that their circadian rhythms are found within representative archaea, prokaryotes and also higher-order eukaryotes. Robust oscillations of Prx proteins, as well as the over- and hyper-oxidized forms of Prx have been observed in mice, Drosophila, the fungus N. crassa, cyanobacteria and even the archaeon Halobacterium salinarum . Key questions remain, however, as to how the redox oscillator interacts with more complex circadian network, including the TTFL mechanism found in many higher-order eukaryotes.
Redox state regulates the clock genes
Mechanistic links between the redox clock and the TTFL framework of circadian rhythms have begun to emerge from research in a variety of organisms from bacteria to fruitflies to mammals. Although a single model has yet to emerge, studies suggest a plethora of ways in which redox state may interact with the circadian clockwork.
Perhaps the most direct way in which redox status affects the TTFL can be seen in the structure and function of the Arabidopsis, Drosophila and mouse variants of the CRY protein. CRY in insects and plants has been described to be directly photosensitive and functions as a circadian-dependent blue-light photoreceptor [46–48]. Homology between CRY and photolyases [49,50], a phylogenetically ancient enzyme that repairs DNA in response to UV light, have led to speculation that redox reactions may be critical for CRY function. Indeed, the CRY protein contains conserved motifs that bind flavin moieties [46,50,51], which are known for their prominent role in electron transport in many metabolic reactions. Blocking electron transport from reduced flavin using the drug diphenylene iodonium has been shown to attenuate the degradation of dCRY (Drosophila CRY) , and mutagenesis of the flavin-binding site in its residues abolishes light responses when expressed in Drosophila S2 cells . Purified dCRY binds to oxidized FAD, which is converted to the semiquinone radical FAD•− under blue-light illumination, as the protein undergoes a change in its absorption . These results strongly suggest that flavin-related redox activity can regulate the function of dCRY. Two recent studies have solved the crystal structure of dCRY, although significant differences between the two datasets exist, including the structures that mediate the reduction of FAD [55,56]. Thus the exact mechanism of redox regulation remains controversial.
In contrast, the flavin-binding site of mCRY (mouse CRY) does not appear to be redox-dependent, since mutations in the corresponding tryptophan residues critical for this function in dCRY do not abolish the transcriptional inhibition function of the mCRY protein . However, insights from the structure of mCRY reveal that it possesses the ability to bind FAD, but it is more likely that the FAD-binding pocket is involved instead in interactions with other proteins, such as FBXL3 (F-box and leucine-rich repeat protein 3), an E3 ubiquitin ligase, which binds across the FAD pocket . This F-box protein, along with FBXL21, regulates mCRY degradation, regulating the period of the transcriptional clockwork [57–59]. Thus, although dCRY possesses a site for redox interaction that mCRY does not, FAD-dependent electron transport does not represent the only mechanism by which redox systems can interact with the TTFL.
Biochemical assays have shown that NAD coenzymes, involved in a variety of metabolic oxidation–reduction reactions, are capable of modulating the DNA binding of circadian heterodimers such as BMAL1–CLOCK and NPAS2–BMAL1. The archetypal reducing factors, NADH or NADPH, were shown to promote DNA binding, whereas the oxidized forms, NAD and NADP, were inhibitory . Thus, at least by in vitro biochemical assays, redox state may influence transcription factor binding, but these observations have yet to be substantiated in a physiological setting.
Genetic studies highlight that mutations in the sod-1 gene in the fungus Neurospora result in a more robust and stable circadian rhythm of conidation banding (asexual spore formation) compared with wild-type fungi. These mutants also exhibit increased ROS levels and expression of the clock gene frq and cellular ROS, in the form of H2O2, also appear to regulate the function of the WC protein . This is further underlined in cyanobacteria and Arabidopsis deficient in 2-Cys Prxs, which still exhibit circadian rhythms, although they are somewhat different from wild-type cycles in either phase or amplitude . These findings suggest that a Prx-based redox oscillator may not be necessary for the generation of circadian rhythms when a functional TTFL oscillator is still present. This is perhaps not surprising, given that, by definition, all circadian clock model organisms have a well-described TTFL oscillator. Thus a challenge is to find novel model systems in which no such TTFLs are thought to exist, but robust circadian rhythms are nonetheless observed. One such model system is the nematode Caenorhabditis elegans, in which whole organism transcriptional cycling and redox oscillations take place , and behavioural readouts would allow more rigorous tests of the interplay between redox and transcriptional networks. Given the current evidence available, it appears most likely that TTFLs and redox oscillations interact with each other within a cell, thus implying that redox state and other metabolic mechanisms are integral to an organism's circadian timekeeping [31,63,64].
Clock genes regulate redox state
Circadian rhythms of ROS generation and scavenging have been observed in a variety of species, including plants , fungi , nematodes , fruitflies [36,66] and rodents . Further evidence of the involvement of the circadian clockwork in the regulation of redox systems is supported by studies in mutants where deletion or disruption of the clock genes also results in the perturbation of these systems. For example, long-period and null mutants of the frq locus in N. crassa display period and phase differences in their pattern of Prx oxidation , and arrhythmic mutants for the clock genes wc-1 (Δwc-1) and wc-2 (Δwc-2) have their rhythmic expression of the catalase gene, cat-1, abolished . However, the two Drosophila lines, ClkJrk and per01, although behaviourally arrhythmic, show cycles of Prx oxidation, albeit with altered parameters compared with their wild-type counterparts. Thus perturbation of the TTFL oscillator can reciprocally interact with the oscillatory activity of the redox system , but this seems to be variable in different species, and depending on the readout used to assess redox activity.
In mammals, 24-h oscillations of oxidized NADPH and the reduced form of FAD have been observed in organotypic slices of the rodent SCN (suprachiasmatic nucleus), a group of cells in the mammalian hypothalamus that function as the master pacemaker of the body by receiving photic information from the retina and synchronizing the rest of the body's physiological oscillations . Rhythms of these coenzymes are believed to be dependent on the molecular clockwork, since Bmal1−/− mice exhibit stochastic, but not circadian, rhythms of FAD and NADPH . These studies also elucidated links between redox state and membrane excitability of SCN neurons; oxidizing and reducing agents can produce hyperpolarization and depolarization respectively. Redox-dependent modulation of K+ channel conductance is believed to underlie these oscillations , implying a complex interdependence between redox state, cellular energetics and the circadian clockwork in mammals. Bmal1−/− mice are also known to have reduced lifespans and exhibit signs of early aging that correspond with increased levels of ROS with age in tissues such as the kidney and spleen .
Further evidence in mammalian models includes the discovery of clock-responsive E-box regulatory sites and RORα-response elements in the promoter sequences for rodent GR and GPx. Deficiency in vitamin A, an essential nutrient and antioxidant, abolishes the daily rhythms of BMAL1 in the rat liver, in addition to abolishing the rhythms of GSH and GR expression , suggesting that, in a peripheral organs such as the liver, changes in oxidative state may also mediate changes in the circadian regulation of metabolism.
Circadian involvement in oxidative challenge and intracellular signalling
H2O2 is a potent oxidizer, regarded generally as a toxic by-product of aerobic metabolism. It plays a key role in the oxidative damage of the cell, and is also often used as an exogenous agent to study cellular responses to oxidative challenge. If redox state and the circadian system regulate each other, H2O2 is also likely to be involved in this interplay. For example, addition of exogenous H2O2 to experimental cultures of Microcystis aeruginosa (a cyanobacterium) have been shown to alter the diurnal expression of kaiA, kaiB and kaiC through phase shifts. Furthermore, susceptibility to oxidative stress as a result of disruption of the circadian system has been proposed to be a mechanism by which H2O2 acts as an effective chemical algaecide .
Studies in wild-type and per01 mutant Drosophila have revealed similar results, with a circadian component to the regulation of the fruitfly's susceptibility to oxidative challenge. Wild-type Canton-S fruitflies experience lower mortality rates with nighttime exposure to H2O2, in contrast with those subjected to environmental conditions of constant light, which disrupted the clockwork. Moreover, per01 mutant fruitflies with disrupted circadian clockwork were also more susceptible to oxidative challenge overall. This phenomenon was coincident with increased endogenous mitochondrial production of H2O2, and enhanced protein carbonylation of catalase . Similar findings have emerged in clock mutant models in Arabidopsis. Plants with a mutation in the clock gene CCA1 have decreased catalase activity, and are more sensitive to ROS-generating agents. Overexpressing CCA1 results in the suppression of H2O2 levels . These studies suggest that the circadian clockwork may play a role in ROS regulation and the defence against exogenous ROS. It is unknown in these cases whether clock genes function as part of the molecular clock, or whether they take on non-clock roles.
In addition to its role in oxidative damage, H2O2 has also been recognized as an intracellular messenger in oxidation-dependent steps in signal transduction . Zebrafish have traditionally been an important animal model in chronobiology, since the clock genes involved in the negative arm of the zebrafish TTFL share homology with those in mammals , and they possess the additional advantage that the clocks in their peripheral tissues are light-sensitive . Studies in cultured zebrafish Z3 cells, a cell line derived from embryonic fish, have shown that light induces the production of H2O2, which in turn activates the expression of zCry1a and zPer2. This suggests that H2O2 may act as a signal transducer, relaying information about light to affect clock gene transcription . Furthermore, catalase activity appears to inhibit the expression of the zCry1a and zPer2 genes by reducing the amount of H2O2 available in the cell, underscoring the likely importance of ROS balance in regulating circadian gene expression .
From the diverse array of studies conducted in the literature, we know that various redox mechanisms in cells and tissues interact with the circadian machinery (Figure 4), be it in the form of KaiC protein phosphorylation, or the regulation of a TTFL encompassing ‘clock genes’. The variety of redox systems and pathways available in different cells have yet to point to a single coherent mechanism by which ROS, or redox state, regulate circadian rhythms, and indeed this mechanism may differ between organisms and tissue types. Nevertheless, the common narrative that emerges from these studies is that redox, ROS and cellular metabolism are a previously overlooked regulatory arm of an organism's circadian timekeeping mechanism, and further investigation is needed into how these systems may interlink and interact.
Interaction between the core circadian clock and redox mechanisms
Conclusions and perspectives
Circadian rhythms affect every aspect of biological function, influencing processes from the cellular to the organismal level. These oscillations not only govern daily cycles in human and animal behaviour and physiology, but also regulate some of the most vital molecular systems in the cell, including the cell division cycle [76,77]. Redox reactions are a fundamental chemical property, ubiquitous within the cell's many and varied metabolic processes. The discovery that redox influences, and is influenced by, other circadian oscillators suggests that circadian rhythms may be more important to the function of the cell than previously thought and capable of regulating basic biochemical systems. Recent medical evidence only furthers the notion that circadian rhythms play a key role in maintaining function, since disrupting these rhythms results in a wide variety of medical disorders and diseases. The study of circadian rhythms is therefore an interesting one, since disrupting daily oscillations has effects not only on the cell, but also on the wider cellular networks that influence physiology and behaviour. Indeed, it is rare in biology to find a process whose ramifications propagate so clearly from the level of gene expression and cellular metabolism to observable organismal phenotypes. Circadian rhythms are important at levels both great and small, and future studies linking redox and other metabolic processes to the cellular clock will surely cast new light on how biology proceeds with the ebb and flow of time.
Colworth Medal Lecture Delivered at the MRC Laboratory of Molecular Biology, Cambridge, on 13 December 2012 Akhilesh Reddy
F-box and leucine-rich repeat protein
retinoic acid-related orphan receptor
red blood cell
reactive oxygen species
transcription–translation feedback loop
Work by A.B.R. and his laboratory is supported primarily by the Wellcome Trust and the European Research Council.