Machine learning helps identify CHRONO as a circadian clock component

Abstract

Over the last decades, researchers have characterized a set of "clock genes" that drive daily rhythms in physiology and behavior. This arduous work has yielded results with far-reaching consequences in metabolic, psychiatric, and neoplastic disorders. Recent attempts to expand our understanding of circadian regulation have moved beyond the mutagenesis screens that identified the first clock components, employing higher throughput genomic and proteomic techniques. In order to further accelerate clock gene discovery, we utilized a computer-assisted approach to identify and prioritize candidate clock components. We used a simple form of probabilistic machine learning to integrate biologically relevant, genome-scale data and ranked genes on their similarity to known clock components. We then used a secondary experimental screen to characterize the top candidates. We found that several physically interact with known clock components in a mammalian two-hybrid screen and modulate in vitro cellular rhythms in an immortalized mouse fibroblast line (NIH 3T3). One candidate, Gene Model 129, interacts with BMAL1 and functionally represses the key driver of molecular rhythms, the BMAL1/CLOCK transcriptional complex. Given these results, we have renamed the gene CHRONO (computationally highlighted repressor of the network oscillator). Bi-molecular fluorescence complementation and co-immunoprecipitation demonstrate that CHRONO represses by abrogating the binding of BMAL1 to its transcriptional co-activator CBP. Most importantly, CHRONO knockout mice display a prolonged free-running circadian period similar to, or more drastic than, six other clock components. We conclude that CHRONO is a functional clock component providing a new layer of control on circadian molecular dynamics.

Figure 1. Integration of core clock features.

(A) List of exemplar core clock genes used as example models of core clock components. (B–E) Metric functions describing core clock features were generated from published data. Distributions of these metrics among nonclock genes (left panel) and exemplar clock genes (center panel) were used to construct evidence factors (right panel). (B) Cycling was evaluated using time-course microarray data from liver, pituitary, and NIH 3T3 cells. (C) Circadian disturbance metric quantifies the influence of RNAi-mediated gene knockdown on circadian dynamics in the U2OS model system. (D) The interaction metric counts the number of interactions inferred between each gene and the exemplar set of core clock genes. (E) The tissue ubiquity scores were taken from an EST database. (F) List of 20 genes most likely to have a core circadian function as determined by evidence factor integration. Genes highlighted in blue were included in the exemplar training set. Genes highlighted in purple were not in the training set but have been identified as having a role in the circadian clock. Gm129 was selected for further characterization.

Figure 2. Chrono transcript demonstrates circadian oscillations in peripheral tissues.

qPCR was used to measure transcript abundance of Chrono, Per2, and Nr1d1 in (A) liver, (B) skeletal muscle, and (C) adipose tissue. Circadian variation is observed in each tissue with the amplitude of Chrono oscillations comparable to that of Per2 and Nr1d1. Data shown are the average of 3–4 biological replicates.

Figure 3. Physical and functional interactions of CHRONO.

(A) Results from a matrix of mammalian two-hybrid assays between known circadian clock components fused to Gal4 DNA binding domain (Gal4 DBD) or VP16 activation domain (VP16 AD). Black and gold indicate bait–prey interactions that resulted in less or greater than 5-fold activation of the 4XUAS reporter, respectively. Co-IP with tagged CHRONO confirms complex formation with (B) BMAL1 and (C) PER2. (D) C- and N-terminal regions of Venus, an enhanced florescent protein, were fused with identified constructs. A yellow bi-molecular fluorescence signal identifies interactions. (E) HEK 293T cells were transiently transfected with a Per1:luc reporter, wild-type, or mutant Bmal1/Clock, and increasing amounts of Cry1 or Chrono. BMAL1/CLOCK point mutants are resistant to CRY1-mediated repression but sensitive to CHRONO. (F) The ability of native CHRONO to repress BMAL1/CLOCK activity was determined by transient transfection with two distinct shRNA constructs directed against Chrono. The indicated plasmids were co-transfected with the Per1-luc reporter into HEK 293T cells. Average activities and standard deviations from reporter assays were determined from independent biological triplicates.

Figure 4. Influence of CHRONO on in vitro and in vivo rhythms.

(A–D) Raw bioluminescence data from NIH 3T3 fibroblasts expressing BMAL:dLUC reporter are plotted after transfection with four shRNA constructs targeted against Chrono. Control and Chrono knockdown tracings are depicted in blue and red, respectively. Two replicates are shown. The period (E) and amplitude (F) of the observed rhythms are plotted. Representative wheel-running activity records for (G) wild-type control and (H) Chrono^flx/flx knockout mice. Blue shading indicates light exposure during the initial 12∶12 h, L∶D cycle. Arrows indicate transition to constant darkness. Regression lines fit to activity onset and computed period are shown. (I) Periodogram estimates of observed periods from wild-type (n = 5), Chrono^flx/+ (n = 8), and Chrono^flx/flx mice (n = 6). Error bars indicate standard error of the mean.

Figure 5. CHRONO interacts with the C-terminus of BMAL1 but not BMAL2.

(A) Overexpression of either BMAL1 or BMAL2, along with CLOCK, activates Per1:Luciferase reporter activity. Both are repressed by overexpression of CRY1. CHRONO specifically represses BMAL1-induced reporter activity. (B) BMAL1 and BMAL2 have similar structures with conserved bHLH DNA binding domains and PAS A and B interaction domains. BMAL1 contains a unique C-terminal region. Chimeric proteins were constructed by swapping corresponding domains from each protein as shown. Two-hybrid screening in HEK 293T cells demonstrates that BMAL1 truncation mutants (C) and chimeric proteins (D) that contain the 487–586 region of BMAL1 bind CHRONO and induce UAS:Luc reporter expression. This region is adjacent to but distinct from the annotated CRY1 binding site. (E) All BMAL1–BMAL2 constructs induce Per1-luc reporter activity in HEK 293T cells. In all constructs, reporter signal is repressed by the addition of CRY1. Functional repression by CHRONO is limited to BMAL constructs containing the implicated binding domain. (F) In cells overexpressing MYC–CHRONO along with BMAL1, BMAL2, or a chimeric BMAL2–BMAL1 construct, co-IP confirms complex formation between CHRONO and proteins containing the implicated BMAL1 C-terminal region.

Figure 6. CHRONO interferes with BMAL1–CBP binding.

(A) BiFC was used to observe BMAL1–CBP interactions in the nuclei of HEK 293T cells. Co-expression of intact or S-tagged CHRONO reduced the complementation signal. Expression of the 212–385 CHRONO truncation mutant had no discernable effect. (B) IP confirms CHRONO-mediated interference in BMAL1/CBP complex formation. Endogenous protein was immunoprecipitated with anti-CBP antibody followed by immunoblotting as indicated. (C) ChIP qPCR analysis was used to evaluate the effect of CHRONO on the acetylation of histone H3–K9 near the Per1 promoter E-box region. Schematic diagram of the human Per1 promoter and primers used for ChIP assay are shown. Lysates obtained from control U2OS cells and those stably expressing CHRONO were collected 24 and 36 h after dexamethasone synchronization. ChIP DNA samples were quantified by quantitative real-time RT-PCR. Data are mean ± standard error of biological triplicates. (D) Various S-tagged, N-, and C-terminal CHRONO truncation mutants were generated. (E) Percent of cell nuclei demonstrating complementation after overexpression of various CHRONO constructs. (F) Per1:luciferase reporter signal in unsynchronized cells overexpressing BMAL1/CLOCK is enhanced by the transient overexpression of CBP. The effect of the overexpression of CHRONO constructs on reporter activity is shown.