Editor's evaluation: Quantification of protein abundance and interaction defines a mechanism for operation of the circadian clock
Year: 2022
Type: peer-review
Abstract: Article Figures and data Abstract Editor's evaluation Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract The mammalian circadian clock exerts control of daily gene expression through cycles of DNA binding. Here, we develop a quantitative model of how a finite pool of BMAL1 protein can regulate thousands of target sites over daily time scales. We used quantitative imaging to track dynamic changes in endogenous labelled proteins across peripheral tissues and the SCN. We determine the contribution of multiple rhythmic processes coordinating BMAL1 DNA binding, including cycling molecular abundance, binding affinities, and repression. We find nuclear BMAL1 concentration determines corresponding CLOCK through heterodimerisation and define a DNA residence time of this complex. Repression of CLOCK:BMAL1 is achieved through rhythmic changes to BMAL1:CRY1 association and high-affinity interactions between PER2:CRY1 which mediates CLOCK:BMAL1 displacement from DNA. Finally, stochastic modelling reveals a dual role for PER:CRY complexes in which increasing concentrations of PER2:CRY1 promotes removal of BMAL1:CLOCK from genes consequently enhancing ability to move to new target sites. Editor's evaluation The transcriptional negative feedback loop of the mammalian circadian clock is mainly regulated by interactions among BMAL1, CLOCK, PER1/2 and CRY1/2 in the nucleus. While the binding of CRY with BMAL1:CLOCK is known to block the transcriptional activity of BMAL1:CLOCK and the binding of PER:CRY dissociates BMAL1:CLOCK from DNA have been known, our understanding is limited in qualitative level. Koch et al., quantified the dynamic interactions among the core clock molecules such as their diffusion coefficients, binding affinity, and abundances in the nucleus. This greatly improves our understanding of the mammalian circadian clock. Importantly, this dynamic information is incorporated via a mathematical model to understand BMAL1-CLOCK binding to the target site (e.g., circadian proteins operate within an optimal range to modulate E-box binding), providing a coherent view on the mechanism driving the oscillation. https://doi.org/10.7554/eLife.73976.sa0 Decision letter Reviews on Sciety eLife's review process Introduction The 24 hr light-dark cycles inherent to our planet have led to the evolution of molecular circuits capable of conveying time of day information, commonly known as circadian clocks. In mammals, cell-autonomous circadian clocks operate in virtually all cells across tissues and enables coordination of numerous biological processes, including metabolism, immunity, and cell cycle progression (Bhadra et al., 2017; Gibbs et al., 2014). Autonomous cellular clocks are characterised by transcription-translation feedback loops (TTFLs), leading to cycles in protein and mRNA tuned to the 24 hr rhythms of the day-night cycle. Central to the mammalian circadian clock is the heterodimeric transcription factor comprised of CLOCK (circadian locomotor output cycles protein kaput) and BMAL1 (brain and muscle ARNT-like 1) that searches the genome to bind consensus sequence E-box sites (CANNTG), inducing expression of several hundred clock-controlled output genes every day. Targets include key circadian negative feedback regulators, Period (Per1, Per2, Per3), Cryptochrome (Cry1 and Cry2) and a secondary loop regulated by Nr1d1 and Nr1d2 (Buhr and Takahashi, 2013; Gekakis et al., 1998; Huang et al., 2012; Liu et al., 2008). These proteins act to repress the activity of CLOCK:BMAL1 to form a delayed negative feedback loop driving daily oscillations. In a current model, it is proposed that PER and CRY proteins dimerise to form a repressive complex with CK1 (casein kinase 1) to promote the removal of CLOCK:BMAL1 from DNA and thereby repress transactivation of target genes, while CRY1 independently binds the PAS domain core of CLOCK:BMAL1 and the BMAL1 transactivation domain leaving DNA binding intact whilst repressing recruitment of additional transcriptional coactivators (Chiou et al., 2016; Xu et al., 2015). An additional feedback loop is conferred by the protein REV-ERBα, which operates as a transcriptional repressor of Bmal1, resulting in a cycle of BMAL1 protein abundance (Liu et al., 2008). Ultimately, a prerequisite for generation and output of cellular circadian rhythms is the ability of a finite pool of CLOCK:BMAL1 heterodimer protein to bind rhythmically to specific target sequences leading to the regulation of circadian gene expression in cells (Koike et al., 2012). Currently, we have very little insight into the quantitative biology of this process. Heterodimeric formation of transactivating and repressive complexes is a well-defined feature of the circadian molecular circuit, including the formation of CLOCK:BMAL1 and PER:CRY complexes (Chiou et al., 2016; Huang et al., 2012; Xu et al., 2015). Recently, PER:CRY proteins have been described as part of very large macromolecular complexes within the cell (Aryal et al., 2017). We have previously visualised several core circadian proteins, and from this measured the spatiotemporal profile and protein abundance for BMAL1 and PER2 (Smyllie et al., 2016; Yang et al., 2020). PER2 was found to cycle with a maximum amplitude of 12,000 copies per cell in fibroblasts and without circadian gating of nuclear localisation, contrary to observations in the Drosophila clock (Shafer et al., 2002; Smyllie et al., 2016). Only a relatively small amount of CRY1 is needed to localise PER2 to the nucleus, as shown in live SCN studies, with PER2 localisation remaining predominantly nuclear throughout the day (Smyllie et al., 2022). A recent study using a cancer cell line model has also shown that CRY1 protein remains nuclear at all circadian phases and at markedly higher abundance than its partner protein PER2 (Gabriel et al., 2021). In order to gain insight into the operation of core circadian clock proteins, we generated a genetically modified mouse in which CRY1 has been C-terminally fused with a fluorescent protein. We crossed this line to a previously described strain of mice expressing fluorescent-tagged BMAL1. We then used advanced imaging in both ectopically transformed cell lines and endogenously modified mice to characterise governing parameters in the regulation of CLOCK:BMAL1 DNA binding, including repression by PER2 and CRY1. We constructed mathematical models of the complex interactions and phased timings from multiple molecular species and experimentally inaccessible complexes, demonstrating how DNA binding in the peripheral circadian clock is regulated. Using a combination of mathematical modelling and experimental validation, our data reveal that high-affinity interactions between circadian protein complexes serve to offset the low abundances of circadian proteins. In this way, the abundance of key components of the molecular clockwork is positioned optimally to regulate E-box binding. This is partly facilitated through PER2:CRY1 mediated displacement of CLOCK:BMAL1, such that PER2 protein serves a dual role, acting as both a component of the negative feedback arm but also to redistribute CLOCK:BMAL1 to new target sites. Thus, the stochiometric balance of PER:CRY with CLOCK:BMAL1 is critical for the elucidation of the full cellular circadian repertoire. Results BMAL1 determines nuclear localisation and mobility of CLOCK To quantify the properties of BMAL1 and CLOCK proteins (Figure 1A), we first used NIH/3T3 fibroblasts expressing fluorescent fusion proteins via a ubiquitin ligase C promoter, delivered by lentiviral transduction either singly (LV1) or as two sequential transductions (LV2) (Figure 1B; Bagnall et al., 2015). Expression of the transgene was in >10 fold excess over the native unfused protein, as determined by single molecule Fluorescence In Situ Hybridisation (Figure 1—figure supplement 1B). Confocal microscopy of tagRFP::CLOCK or BMAL1::EGFP showed BMAL1 expression to be strongly localised to the nucleus, whereas CLOCK was predominantly cytoplasmic when expressed alone (Figure 1C). Co-expression of both proteins in the same cells caused localisation of tagRFP::CLOCK to move to the nucleus, in agreement with earlier studies which showed cytoplasmic CLOCK localisation in BMAL1-deficient cells, and that circadian regulation of nuclear localisation of CLOCK correlated with BMAL1 availability (Kondratov et al., 2003; Stratmann et al., 2012). We also transduced cells with a fluorescent fusion of a DNA-binding mutant of BMAL1, in which a leucine is substituted for glutamic acid in the basic helix-loop-helix region of the protein; referred to as L95E. The mutant BMAL1 also re-localised tagRFP::CLOCK protein to the nucleus from the cytoplasm in an manner equivalent to WT BMAL1 (Figure 1C; Huang et al., 2012). We next performed Fluorescence Recovery After Photobleaching (FRAP) experiments to test the impact of CLOCK on the recovery dynamics of a bleached nuclear region of BMAL1::EGFP, by comparing responses with or without co-expressed tagRFP::CLOCK (Figure 1D–F). BMAL1 recovery half-life was found to be insensitive to the diameter of the bleached region, indicating that binding contributes to the recovery profile rather than this being a solely diffusion-led process (Figure 1—figure supplement 2; Sprague and McNally, 2005). Reaction binding equations were fitted to determine the rate of dissociation, kOFF, for BMAL1::EGFP, the reciprocal of which equates to an average characteristic duration of binding or residence time. Residence time of BMAL1::EGFP was increased in the presence of tagRFP::CLOCK (p < 0.0001), consistent with a requirement for CLOCK to bind DNA (Figure 1F). The mean residence time for the fluorescent CLOCK:BMAL1 complex was 4.13 s (95% CI, 0.57), a value consistent with DNA residence times for similar transcription factors (Hettich and Gebhardt, 2018; Stratmann et al., 2012). Using the L95E DNA-binding mutant protein, we saw significantly reduced residence time of 2.83 s (95% CI, 0.54; p = 0.0002). Notably, the initial publication of the BMAL1 L95E mutant showed a twofold reduction in PER2::LUC expression and so suggests a strong relationship between DNA binding and transcriptional output (Huang et al., 2012). Figure 1 with 2 supplements see all Download asset Open asset Short-lived DNA binding of BMAL1 and CLOCK. (A) Schematic representation of parameters regulating CLOCK:BMAL1 dimers binding to target DNA sites. (B) NIH/3T3 cells are either singularly or sequentially transduced to express fluorescent fusions with CLOCK or BMAL1 (wildtype and mutant variants).(C) Confocal microscopy images of cells solo-expressing (LV1) either tagRFP::CLOCK or BMAL1::EGFP or co-expressing (LV2) them together (including BMAL1 L95E DNA-binding mutant). (D) Confocal microscopy images for photobleaching of LV2BMAL1::EGFP-RFP::CLOCK labelled cells, either with wild-type or BMAL1 L95E DNA binding mutant. Images show nuclei and highlight region of bleaching. (E) Representative fluorescence recovery curves of bleach region for B. following normalisation. (F) Residence time calculated as the inverse of kOFF (s-1), determined from fitting the recovery data with a single component binding model (n = 69, 58, and 51 cells). Bar represents median values. Source data for panel F available as Figure 1—source data 1. Figure 1—source data 1 BMAL1 residence times. https://cdn.elifesciences.org/articles/73976/elife-73976-fig1-data1-v2.csv Download elife-73976-fig1-data1-v2.csv To investigate this further we used Fluorescence (Cross) Correlation Spectroscopy (F(C)CS) (Yu et al., 2021), a technique used to determine live-cell concentration and diffusion properties of individually fluorescent-labeled BMAL1 and CLOCK proteins, as well as their interactions when co-expressed (Figure 2A). A normal diffusion model fitted the majority of data collected from cells expressing free EGFP or nuclear only NLS::EGFP proteins, as previously reported (Dross et al., 2009), whereas anomalous diffusion models – sub-diffusion caused by a range of factors such as DNA interactions and molecular crowding – accounted for a > 20% fraction, which in this instance may be explained by molecular crowding (Tsekouras et al., 2015). In comparison, for the fusion proteins, anomalous diffusion models accounted for >80% of all BMAL1 data sets (Figure 2—figure supplement 1). We used an anomalous diffusion model for all further analyses of circadian proteins to calculate diffusion coefficients and protein concentrations. Figure 2 with 2 supplements see all Download asset Open asset Live-cell interaction measurements demonstrate BMAL1 and CLOCK mobility is regulated by dimerisation and DNA binding. (A) Schematic of confocal volume used in FCCS with corresponding photon count traces. Interaction may be seen by correlation between both channels. Representative auto- and cross- correlation data showing raw data and fit lines for monomeric and complexed fluorescent proteins. (B) FCS data showing diffusion for BMAL1 and CLOCK in solo- and co-expressed conditions (n = 173, 152, 198, and 185 cells). (C) FCS results for BMAL1::EGFP diffusion for NIH/3T3 cells that co-express tagRFP::CLOCK. Data shown is for comparison of BMAL1 as either wild-type of L95E DNA-binding mutant. Bars show median and interquartile range. (D) Correlation of nuclear protein quantification showing relationship of BMAL1::EGFP with tagRFP::CLOCK for both wildtype and DNA binding mutant (n = 221 cells from three biological replicates). (E) Average cross-correlation curves for BMAL1::EGFP (WT) with tagRFP::CLOCK (n = 140) compared to a non-interacting control of NLS::EGFP co-expressed with tagRFP::CLOCK (n = 408). (F) Dissociation plot from FCCS data for BMAL1::WT and tagRFP::CLOCK. (G) Summary of calculated dissociation constants across all conditions, including BMAL1 dimerisation mutant, V435R (n = 156, 274, and 244). Mann-Whitney non-parametric test to determine significance (values are denoted as p > 0.05 ns, p < 0.05 *, p < 0.01 **, p < 0.001 *** and p < 0.0001 ****). Source data for panels B,C available as Figure 2—source data 1 and panel E as Figure 2—source data 2. Figure 2—source data 1 BMAL1 and CLOCK FCS diffusion rates. https://cdn.elifesciences.org/articles/73976/elife-73976-fig2-data1-v2.csv Download elife-73976-fig2-data1-v2.csv Figure 2—source data 2 BMAL1 and CLOCK paired FCS concentrations. https://cdn.elifesciences.org/articles/73976/elife-73976-fig2-data2-v2.csv Download elife-73976-fig2-data2-v2.csv Singly expressed fluorescent BMAL1 and CLOCK were found to diffuse rapidly with a median coefficient of 9.2 µm2s-1 (SD, 3.3) and 12.6 µm2s-1 (SD, 6.1), respectively. In contrast, co-expression significantly reduced the rate of diffusion to 1.9 µm2s-1 (SD, 1.3; p < 0.0001) and 4.7 µm2s-1 (SD, 3.2; p < 0.0001) for BMAL1 and CLOCK ,respectively (Figure 2B). The L95E mutant diffused more rapidly than WT BMAL1, consistent with fewer interactions with DNA in the nucleus (Figure 2C). When co-expressed, WT BMAL1 and CLOCK exhibited a 2:1 concentration ratio in the nucleus (Figure 2D, Figure 2—figure supplement 2E), presumably arising from a combination of differences in protein turnover, shuttling and direct interaction. F(C)CS was then used to observe this interaction and determine a live-cell dissociation constant (KD; reciprocal measure of affinity) (Krieger et al., 2015). A positive cross correlation curve was observed between BMAL1::EGFP and tagRFP::CLOCK that was not apparent in cells expressing NLS::EGFP with tagRFP::CLOCK (Figure 2E). To calculate KD, we fitted a one-site saturating binding curve to the relationship between heterodimer and monomer which yielded a value of 148 nm (SD, 9.8) for WT BMAL1::EGFP and tagRFP::CLOCK (Figure 2F). The KD was measured for cells with the reverse fluorescent protein labelling, namely EGFP::CLOCK and BMAL1::tagRFP, finding similar a value of 145 nm (SD, 4.8), although a stronger interaction was found in vitro by surface plasmon resonance (Figure 2—figure supplement 2F). Previous work found that the V435R mutation of BMAL1 in the PAS-B domain, leads to reduced dimerisation with CLOCK (Huang et al., 2012). We used the V435R mutation to confirm our F(C)CS measurements by co-expressing V435R-BMAL1 and WT-CLOCK. This elicited a ≈1.5-fold reduction in interaction affinity, resulting in a KD of 201 nm (SD, 14) (Figure 2G) and a reduction from 2:1 to a 4:1 ratio of BMAL1 and CLOCK in the nucleus (Figure 2—figure supplement 2E). In contrast, the BMAL1 L95E DNA binding mutant showed no difference in interaction affinity compared to WT BMAL1 protein. These data demonstrate that BMAL1 is a critical determinant of the localisation, mobility and concentration of CLOCK in the nucleus. From this, we can infer the abundance of nuclear CLOCK from measurements of BMAL1, and make use of available endogenously labelled Venus::BMAL1 mice to measure remaining DNA binding parameters. First, we confirmed our cell line measurements for binding rates and diffusion using the Venus::BMAL1 mice (Yang et al., 2020), finding that they remain within a similar range across a number of primary cell types, including macrophages and pulmonary fibroblasts (Figure 3—figure supplement 1). We also measured protein number of the endogenous BMAL1, observing that total copies per nucleus vary from 1000 to 10,000 between individual cells, likely due to desynchrony and differing nuclear volumes. Moreover, a large overlap in nuclear copy numbers was observed across all cell types despite substantial changes in the mean. These data are critical in our understanding of the ratio of BMAL1 to target sites, effectively determining the capacity to regulate the full repertoire of target genes within a specific cell type. Quantification of strong rhythmic interaction of BMAL1 with CRY1 The ability to measure BMAL1 amounts to infer copy number of CLOCK, allows us to measure other critical pairings with BMAL1. This includes the repressive action of CRY1 binding to CLOCK:BMAL1, resulting in reduced transactivation. To explore the interaction between BMAL1 and CRY1, we generated a genetically modified mouse in which CRY1 has been C-terminally fused with the red fluorescent protein mRuby3 using CRISPR-mediated genomic editing to insert the coding sequence, replacing the endogenous CRY1 stop codon (Figure 3—figure supplement 3A; Bajar et al., 2016; Bennett et al., 2021). First, to test any potential impact on circadian pace-making, we measured CRY1::mRuby3 fluorescence in whole-field organotypic SCN slices (Figure 3—figure supplement 3B) which exhibited regular cycles in red fluorescence with a period of 23.9 hr (SD, 0.6) (Figure 3—figure supplement 3C; Smyllie et al., 2016). Additionally, wheel running measurements of these mice confirmed normal behavioural rhythmicity (Figure 3—figure supplement 2). We next crossed these mice to the Venus::BMAL1 mouse line (Yang et al., 2020), previously inter-crossed with a PER2::LUCIFERASE background (Yoo et al., 2004) to provide an independent circadian phase-reference marker (referred to as BMAL1xCRY1 labelled mouse). Using isolated lung fibroblasts from BMAL1xCRY1 mice we assessed bioluminescence in response to dexamethasone (DEX) synchronisation, and observed 23.3 hr cycles (SD, 0.6) which were sustained for (Figure 3—figure supplement Figure 3—figure supplement From this, we are that the fluorescent fusion proteins not the normal operation of the circadian Using the same we then measured BMAL1xCRY1 fluorescence in single cells every hr from 24 to hr synchronisation, using F(C)CS (Figure fluorescent were localised to the nucleus. Venus::BMAL1 showed a consistent diffusion over a circadian with a mean diffusion coefficient of µm2s-1 (SD, whereas CRY1 mobility exhibited circadian with diffusion hr and diffusion rates hr (Figure this in mobility is consistent with a binding to a equivalent to the molecular of protein concentrations of BMAL1 and CRY1 an and of hr et al., analyses the concentration of BMAL1 is on average fold (SD, higher than CRY1, with a mean concentration of nm and nm to and molecules per consistent with the range we reported earlier for PER2 (Smyllie et al., 2016). The amplitude of CRY1 was found to be cycling from nm (SD, at to nm (SD, at to the amplitude observed for CRY1 in the SCN (Figure 3—figure supplement BMAL1 a cycling from nm (SD, at to a of nm (SD, hr (Figure Figure with supplements see all Download asset Open asset A rhythmic and strong interaction observed between BMAL1 and CRY1 repression. (A) Schematic of mice from which primary lung fibroblasts were isolated (B) Confocal images of two cells shown for Venus::BMAL1 and CRY1::mRuby3 over time. FCS determined for diffusion coefficient (C) and protein concentration (D) of Venus::BMAL1 and CRY1::mRuby3 (n = 173, and line the mean and show the Interaction between BMAL1 and CRY1 was also measured over time as by the of affinity as well as of dissociation constant the test used to determine significance (values are denoted as p > 0.05 ns, p < 0.05 *, p < 0.01 **, p < 0.001 *** and p < 0.0001 ****). Source data for panels B,C available as Figure data Figure data Figure data Figure data Figure data 1 CRY1 FCS diffusion rates. Download Figure data 2 BMAL1 FCS diffusion rates. Download Figure data BMAL1 FCS Download Figure data CRY1 FCS Download We then the interaction affinity between BMAL1 and CRY1 over time (Figure This interaction exhibited changes over a 24 hr cycle (p = < 0.0001), with the interaction at KD = nm (SD, and at KD = nm (SD, (Figure These were found to with the diffusion profile of CRY1 (Figure the mean interaction between BMAL1 and CRY1 is fold stronger than that between BMAL1 and CLOCK (Figure A similar relationship was found in vitro when interactions using et al., 2020). these interaction measurements not between either proteins are complexed with other diffusion data is consistent with BMAL1 being to CLOCK, and are with a model in which the low abundance of the CRY1 repressor is offset by a affinity for the CLOCK:BMAL1 The changes in the diffusion profile of CRY1 are consistent with its association with an additional binding such as thereby the affinity of CRY1 for CLOCK:BMAL1 et al., et al., 2014). To measure the interaction between CRY1 and PER2 we transduced NIH/3T3 cells with so that cells express or fusion proteins. In both the expressed protein was found to localise to the nucleus, although cytoplasmic fluorescence was When co-expressed, localisation was although large of were observed (Figure PER2 was found to have the diffusion coefficient within all our F(C)CS when in the PER2 mobility was not following co-expression with CRY1, whereas CRY1 exhibited reduced mobility in the presence of (Figure The diffusion coefficient for CRY1 co-expressed with PER2 was similar to measurements of the endogenous protein (Figure PER2 and CRY1 similar within these The anomalous diffusion model fit the majority of data including and normal diffusion models accounted for of correlation analyses a in CRY1 following interaction with PER2 < 0.0001), from a of DNA binding of the CLOCK:BMAL1 complex (Figure supplement 1). fit models for data a strong affinity between PER2 and CRY1 with a KD of nm (SD, (Figure and consistent with in vitro measurements et al., 2014). Figure with 1 supplement see all Download asset Open asset PER2 CRY1 mobility via a high-affinity (A) Confocal images of transduced NIH/3T3 cells that either solo- or express PER2 and CRY1. (B) FCS data showing diffusion for PER2 and CRY1 in solo- and co-expressed conditions (n = 274, and diffusion rate of (C) Dissociation plot from nuclear FCS measurements for and (n = determined by Mann-Whitney test (values are denoted as p > 0.05 ns, p < 0.05 *, p < 0.01 **, p < 0.001 *** and p < 0.0001 ****). Source data for panels available as Figure data 1. Figure data 1 PER2 and CRY1 FCS diffusion rates. Download data are an and component of mathematical modelling of cell et al., quantified the we then to use them in a mathematical model of CLOCK:BMAL1 DNA binding, with the of understanding how the multiple of molecular interactions and binding to regulate DNA binding and transcriptional of BMAL1. We multiple that were to fit BMAL1 binding the model that incorporated our measured We the using a of equations to a current understanding of the BMAL1 with CLOCK, which may bind and to DNA target sites. To model CRY1 may either CLOCK:BMAL1 via direct binding via dimerisation to form PER2:CRY1 complexes with to CLOCK:BMAL1 from DNA (Figure et al., 2016; Huang et al., 2012; et al., 2012; Xu et al., 2015). The presumably to rhythmic changes in the residence time of BMAL1 and provide a to fit and the Figure with 1 supplement see all Download asset Open asset PER2 via CRY1 to rhythmic displacement of CLOCK:BMAL1 from DNA. (A) Schematic representation of model used for the model of CLOCK:BMAL1 DNA binding. (B) lung fibroblasts from BMAL1 CRY1 PER2 mice were with PER2 concentration as measured via FCS by Smyllie et al., as well as mean BMAL1 binding time time was measured by confocal measurements performed on the Venus::BMAL1 line the inverse of kOFF (s-1), determined from fitting the recovery data with a single component model (n = and cells). (C) model was fit to binding data from and using a measured for PER2 nuclear concentration previously determined in Smyllie et al., output showing (D) nuclear concentrations for molecular complexes (E) and CLOCK:BMAL1 without and with CRY1 to target sites for has been from Figure from Smyllie et al., To dynamic changes in DNA binding rates of BMAL1, we isolated lung fibroblasts from BMAL1xCRY1 mice and determined the kOFF by following of BMAL1 protein recovery were every hr from 24 hr to hr showing kOFF to be rhythmically regulated (Figure The BMAL1 kOFF profile was in to of nuclear PER2 concentrations from Smyllie et al., (Figure To fit all parameters to the model the measured concentrations of PER2 (Smyllie et al., CRY1 and BMAL1 were used as using data described in Figures and rates were to measured dissociation constants from F(C)CS with the KD value between BMAL1 and E-box sites at as measured by Huang et al., Using a mean value from multiple
Source: eLife
Author: Naama Barkai
Institution: Weizmann Institute of Science
Language: English
Subfield: Endocrine and Autonomic Systems
Field: Neuroscience
Domain: Life Sciences
Open Access status: gold
Estimated APC paid: $2,500
OpenAlex ID: W4225844774