CREBH Couples Circadian Clock with Hepatic Lipid ... - Diabetes

1 downloads 0 Views 2MB Size Report
Aug 9, 2016 - Paul. D. Walker. 4. , Gregory Kapatos. 1, 3. , and Kezhong Zhang ...... Zhang D, Tong X, Arthurs B, Guha A, Rui L, Kamath A, Inoki K, Yin L: Liver ...
Page 1 of 51

Diabetes

CREBH Couples Circadian Clock with Hepatic Lipid Metabolism

Ze Zheng1 #, Hyunbae Kim1 #, Yining Qiu1, Xuequn Chen5, Roberto Mendez1, Aditya Dandekar2, Xuebao Zhang1, Chunbin Zhang1, Andrew C. Liu8, Lei Yin7, Jiandie D. Lin6, Paul * D. Walker4, Gregory Kapatos1, 3, and Kezhong Zhang1, 2 Center for Molecular Medicine and Genetics, 2 Department of Immunology and Microbiology, 3 Department of Pharmacology, 4 Department of Anatomy & Cell Biology, 5Depatment of Physiology, Wayne State University School of Medicine, Detroit, MI 48201, USA; 6 Life Sciences Institute and Department of Cell & Developmental Biology, 7 Department of Molecular and Integrative Physiology, University of Michigan Medical School, Ann Arbor, MI 48109, USA; 8 Departments of Biological Sciences, University of Memphis, Memphis, TN 38152, USA 1

Running title: CREBH regulates circadian metabolism

* Corresponding Athour: Kezhong Zhang, Ph.D., Center for Molecular Medicine and Genetics, Wayne State University School of Medicine, 540 E. Canfield Avenue, Detroit, MI 48201, USA Tel: 313-577-2669; FAX: 313-577-5218; Email:[email protected]

# ZZ and HK contribute equally to this work.

The manuscript has word count of 4603, 8 figures, and Supplemental information.

Keywords: Circadian regulation; hepatic lipid metabolism; transcriptional regulation; nuclear receptors.

Diabetes Publish Ahead of Print, published online August 9, 2016

Diabetes

Page 2 of 51

Abstract

Circadian clock orchestrates diverse physiological processes that are critical for health and disease. Cyclic AMP responsive element-binding protein, hepatocyte-specific (CREBH) is a liverenriched, endoplasmic reticulum (ER)-tethered transcription factor known to regulate hepatic acutephase response and energy homeostasis under stress conditions. Here, we demonstrate that CREBH is regulated by the circadian clock and functions as a circadian regulator of hepatic lipid metabolism. Proteolytic activation of CREBH in the liver exhibits typical circadian rhythmicity that is controlled by the core clock oscillator BMAL1 and AKT/GSK3β signaling pathway. GSK3β-mediated phosphorylation of CREBH modulates the association between CREBH and the Coat Protein Complex II (COPII) transport vesicle, and thus controls the ER to Golgi transport and subsequent proteolytic cleavage of CREBH in a circadian manner. Functionally, CREBH regulates circadian expression of the key genes involved in triglycerides (TG) and fatty acid (FA) metabolism, and is required to maintain circadian amplitudes of blood TG and FA in mice. During the circadian cycle, CREBH rhythmically regulates, and interacts with, the hepatic nuclear receptors PPARα and LXRα as well as the circadian oscillation activator DBP and the repressor E4BP4 to modulate CREBH transcriptional activities. In conclusion, our studies revealed that CREBH functions as a circadianregulated liver transcriptional regulator that integrates energy metabolism with circadian rhythm.

1

Page 3 of 51

Diabetes

Introduction

Mammalian circadian rhythms are biological processes that exhibit endogenous oscillations over a 24-hour light-dark cycle and are entrainable by internal and external stimuli. Circadian rhythms are generated at the level of gene transcription by a network of clock-controlled genes that form an auto-regulatory feedback loop (1). The CLOCK/BMAL1 heterodimer drives circadian expression of many other transcription factors, thereby extending and enhancing other circadian regulatory functions. Peripheral organs, such as the liver, have local rhythms that are synchronized by master clock oscillators located in the suprachiasmatic nuclei (SCN) of the anterior hypothalamus (2). Previous work has demonstrated the intimate and reciprocal interaction between the circadian clock system and fundamental metabolic pathways (3; 4). In the liver, nuclear receptors or transcription factors are inducible by metabolites or hormones, while half of them exhibit rhythmic expression (5). Therefore, the liver nuclear receptors or transcriptional regulators may serve as direct links between metabolic pathways and circadian regulation. We recently reported that the ER-tethered, liver-enriched transcription factor CREBH regulates energy homeostasis under metabolic stress. The expression and activation of CREBH in the liver are regulated by a variety of inflammatory and metabolic signals, such as pro-inflammatory cytokines, saturated FA, insulin, fasting, and atherogenic high-fat diets (6; 7). Activated CREBH acts as a multifaceted activator of transcription that induces expression of the genes involved in hepatic acute-phase response, FA oxidation, lipolysis, lipogenesis, and gluconeogenesis (6-10). CREBH-null

mice

develop

profound

non-alcoholic

steatohepatitis

(NASH)

and

hypertriglyceridemia when fed an atherogenic high-fat diet (6). Importantly, recent studies

2

Diabetes

Page 4 of 51

confirmed that human patients with hypertriglyceridemia exhibit a high-rate of nonsense mutations or rare genetic variant accumulation in the human CREBH gene (9; 11; 12). In this study, we demonstrate that CREBH is an organ-specific circadian regulator of lipid metabolism whose activation is regulated by the circadian oscillation in the liver. CREBH plays an indispensable role in maintaining lipid homeostasis under the circadian control, and dysfunction of CREBH in mice leads to impaired rhythmic profiles of TG and FA. Our finding that CREBH functions as a liver circadian metabolic regulator has important implications in the understanding of the molecular basis of circadian metabolism and the development of metabolic disorders.

3

Page 5 of 51

Diabetes

Research Design and Methods

Animal model - All animal experiments were performed with the approval of the Institutional Animal Care and Use Committee (IACUC) of Wayne State University. Male wild-type and CREBH-null C57BL6 mice (6) of 4-month-old were housed in 12hr light/12hr dark (LD) cycles with free access to food and water for at least 2 weeks before switching to constant darkness (DD) for 24hr to allow endogenous clocks to free run. Mice were euthanized with isoflurane followed by rapid cervical dislocation. Liver samples from 3-5 mice per time point per genotype group were collected in constant darkness every 4hr for a 24hr period.

In vitro circadian synchronization of mouse primary hepatocytes - Primary hepatocytes isolated from C57BL/6J mice were infected with recombinant adenovirus expressing Bmal1 shRNA, the dominant-negative or constitutively-activated AKT, or the dominant-negative or constitutivelyactivated GSK3β for 24 hrs before they were subjected to serum-shock (50% horse serum) for 2 hrs for circadian synchronization (13). After serum shock synchronization, the shock medium was replaced with serum free medium. Cell lysates were collected at 8-hr intervals between 24-hr (circadian 0 hr) and 72-hr (circadian 48 hr) post-serum shock for Western blot analysis.

ChIP assays with mouse liver chromatin - Mouse liver chromatin was fragmented to an average size of 500 bp by sonication as previously described (14). Fragmented chromatin was pre-cleared by incubating with the rabbit IgG, followed by incubation with protein G agarose. CREBH-binding complexes were pulled down by using a rabbit anti-CREBH antibody (8). As controls, the precleared chromatin samples were pulled down using a rabbit anti-HA antibody. Immunoprecipitated chromatin fragments were reverse cross-linked and digested by proteinase K. Presence of CREBH 4

Diabetes

Page 6 of 51

in gene promoters under different circadian phases were quantified by qRT-PCR and expressed relative to the input genomic DNA. The primer sequences for ChIP-PCR assays are described in STable 1.

Site-directed mutagenesis - Site-directed mutations were introduced into the putative GSK3β phosphorylation sites, Serine (S)-256 and S-260, of human CREBH using a QuikChange II sitedirected mutagenesis kit (Stratagene). The S256A or S260A mutation (where S was changed to A) was achieved by site-directed mutagenesis PCR using a human full-length CREBH expression plasmid as the template and the primers: 5’- GAACAAGCAGGCGGCGCAAGAAA-3’ and 5’TTTCTTGCGCCGCCTGCTTGTTC-3’

for

S256A,

or

5’-

GGCGCAAGAAGCCAGGAAGAAGA-3’ and 5’- TCTTCTTCCTGGCTTCTTGCGCC-3’ for S260A. All constructs were confirmed by sequencing analysis.

Statistics - The results of experiments were analyzed by several statistical methods. Unpaired Mann Whitney U test was used for non-parametric comparisons. One-way ANOVA test was used for parametric comparisons. Two-way ANOVA was used to distinguish the effects of genotypes from the effects of circadian time on gene expression, levels of mouse blood lipids, and quantification of food intake. When multiple testing procedures were implemented, the Bonferroni correction was used.

5

Page 7 of 51

Diabetes

Results

CREBH is a clock-regulated diurnal regulator in the liver. To test whether circadian-dependent regulation for CREBH is present in the liver, we examined expression and proteolytic activation profiles of CREBH in the livers of wild-type (WT) mice under a 24-hr circadian cycle. Expression of CrebH mRNA in mouse liver showed a trend to increase during the late phase of daytime and decrease during the late phase of nighttime, but did not exhibit typical circadian rhythmicity (Figure 1A). We next examined levels of precursor and activated forms of CREBH proteins in mouse liver across the day-night cycle. Production of the activated CREBH protein involves translocation of CREBH precursor from the ER to Golgi where it is cleaved by S1P and S2P proteases(7), and therefore, levels of the activated CREBH can be evaluated by examining cleaved CREBH proteins. Western blot analyses with the membrane and nuclear protein fractions prepared from the pooled liver tissues of the mice under the circadian circle showed that levels of the membrane-bound CREBH precursor protein during the “daytime” phases were lower than those of the “nighttime” phases, which displayed a peak at CT20 and a trough at CT8 (Figure 1B). In an opposite manner, levels of the activated, nuclear CREBH protein reached a trough at CT20 and peaked at CT8, 12 hrs after the peak production of CREBH precursor protein (Figure 1C). Apparently, the decreases in the levels of CREBH precursor were coincident with the increase in the levels of the cleaved CREBH, which exhibited typical circadian rhythmicity. The rhythmicity in the production of CREBH precursor and cleaved proteins was confirmed by the Western blot analysis with total liver protein lysates from WT and CREBH-null animals under the circadian cycle (S-Fig 1A). Taken together,

6

Diabetes

Page 8 of 51

these results indicate that the proteolytic cleavage/activation process of CREBH in the liver are rhythmically regulated by the circadian clock during the day-night cycle.

Circadian rhythmic activation of CREBH is regulated by BMAL1-AKT-GSK3β signaling pathway. BMAL1, the core circadian oscillator, plays a central role in regulating expression or activities of circadian output regulators (1). To determine whether BMAL1 regulates circadian rhythmic activation of CREBH, we examined expression and proteolytic cleavage of CREBH protein in liver-specific Bmal1 conditional knockout (Bmal1 LKO) and control mouse livers collected every 6 hrs during a 24-hr circadian period (15). Western blot analysis showed that the levels of the cleaved/activated CREBH protein, but not the CREBH precursor, were reduced in the livers of Bmal1 LKO mice across the day-night periods, compared to those in the control fl/fl mouse livers (Figure 2A-B). Additionally, we examined activation of CREBH in livers of Bmal1 LKO and control mice under the normal feeding condition or after 16-hr fasting. The levels of the activated CREBH protein was significantly decreased in the livers of Bmal1 LKO mice under the feeding, but not the fasting condition (S-Fig 1B), confirming that BMAL1 plays a major role in regulating CREBH activation under the normal physiological condition. However, under metabolic stress (fasting), CREBH is likely activated through alternative regulatory mechanism(s) independent of BMAL1, an interesting question to be elucidated in the future. To gain further insights into the regulatory mechanism underlying circadian activation of CREBH, we explored the signal transduction pathway through which BMAL1 regulates CREBH activation in primary hepatocytes during circadian oscillation. Bmal1 was knocked down in mouse primary hepatocytes using adenoviral-based expression of Bmal1 shRNA (16), and then subjected to horse serum shock for circadian synchronization (13). Western blot analysis showed that the

7

Page 9 of 51

Diabetes

amplitudes of CREBH cleavage in hepatocytes across a 48-hr circadian period were significantly repressed by Bmal1 knockdown (Figure 2C). Indeed, the cleaved/activated form of CREBH was diminished in Bmal1-knockdown hepatocytes during the circadian cycle except the circadian time point 8 or 32 hr. AKT, a serine/threonine protein kinase, plays a key role in regulating energy homeostasis (17). Recent studies showed that circadian oscillation of AKT activities is impaired in Bmal1 KO mice (18; 19) and that BMAL1 regulates hepatic lipogenesis through AKT2 signaling pathway (16). Furthermore, GSK3β, a constitutively active serine/threonine kinase that is negatively regulated by AKT2-mediated phosphorylation (17), can regulate energy homeostasis and circadian function through phosphorylation-mediated suppression of its downstream substrates (20). Through bioinformatics analysis we identified that mammalian CREBH proteins possess conserved GSK3β phosphorylation sites within their b-ZIP domains (S-Fig 1C). These information prompted us to speculate that BMAL1 may regulate CREBH circadian activation through AKT-GSK3β signaling pathway. To test this hypothesis, we examined hepatic AKT and GSK3β phosphorylation states in Bmal1-knockdown primary hepatocytes under the circadian clock. Phosphorylation at serine 473 of AKT2, an indication of AKT activity in insulin signaling and circadian regulation (17; 19), was significantly reduced in Bmal1-knockdown hepatocytes (Figure 2C). Bmal1 knockdown largely repressed phosphorylation of GSK3β at serine 9, the inhibitory phosphorylation site that is regulated by AKT (17). Interestingly, the circadian time points where the cleaved CREBH protein was elevated were coincident with the time points when phosphorylation of AKT or GSK3β was upregulated (Figure 2C), implicating that circadian-regulated cleavage of CREBH may be regulated through a BMAL1-AKT-GSK3β regulatory axis. To validate this possibility, we examined proteolytic activation of CREBH in AKT- or GSK3β- dominant negative primary hepatocytes across the circadian cycle. The levels of cleaved CREBH in AKT-dominant negative hepatocytes

8

Diabetes

Page 10 of 51

during the 48-hr circadian cycle were reduced approximately 20-70%, compared to the control hepatocytes (Figure 2D). In contrast, over-expression of a dominant negative GSK3β (GSK3β-DN) led to increased proteolytic cleavage of CREBH in hepatocytes over the circadian cycle (Figure 2E), thus supporting a regulatory role of AKT-GSK3β signaling in BMAL1-controlled CREBH cleavage. To validate that BMAL1 regulates CREBH cleavage through AKT/GSK3β signaling under the circadian clock, we performed the reconstitution experiments by over-expressing a constitutively active AKT (AKT-CA) (16) or GSK3β-DN in Bmal1-knockdown primary hepatocytes. Expression of AKT-CA increased the levels of cleaved CREBH in the Bmal1-knockdown hepatocytes, compared to expression of GFP control (Figure 3A). Similarly, expression of GSK3β-DN can rescue the decreased CREBH cleavage activity in the Bmal1-knockdown hepatocytes under the circadian clock (Figure 3B), thus confirming that BMAL1 controls circadian-regulated CREBH cleavage/activation through AKT/GSK3β signaling in hepatocytes.

GSK3β-mediated phosphorylation modulates CREBH association with COPII vesicle proteins and subsequent CREBH cleavage under the circadian clock. CREBH cleavage process involves transport of the ER-bound CREBH from the ER to Golgi where it is cleaved by S1P and S2P proteases (7). To undergo the sequential proteolysis at the Golgi, ER-bound nascent proteins need to be transported to the Golgi via Coat Protein Complex II (COPII) transport vesicles (21; 22). Formation of COPII vesicles is initiated by the small GTPase Sar1, which sequentially recruits the heterodimeric protein complexes Sec23/24 and Sec13/31 to the ER membrane to facilitate vesicle formation (21). We reasoned that GSK3β-mediated phosphorylation of CREBH may serve as a control point for circadian-regulated transport of CREBH from the ER to Golgi and subsequent proteolytic cleavage of CREBH. To test this possibility, we first over-expressed a constitutively-

9

Page 11 of 51

Diabetes

active GSK3β kinase (GSK3β-CA), in which the regulatory phosphorylation site, serine 9, was mutated to an alanine and thus resistant to inactivation mediated by Akt phosphorylation (23), or GFP control in mouse primary hepatocytes. After circadian synchronization by serum shock, we examined the interaction between CREBH and Sec23/24, the core component proteins of COPII transport vesicles, in the hepatocytes. CREBH complexed with Sec23A/Sec24A in the primary hepatocytes expressing GFP control in a circadian-dependent manner (Figure 4A). However, the circadian-regulated association between CREBH and Sec23A/Sec24A in the hepatocytes expressing GSK3β-CA was significantly repressed. Moreover, the CREBH-Sec23/24 interactive activities were in compliance with the levels of cleaved CREBH protein in the hepatocytes expressing GFP or GSK3β-CA under the circadian clock (Figure 4B). These results implicated that CREBH is associated with the COPII vesicle proteins in a circadian-dependent manner and that GSK3β modulates circadian-regulated association of CREBH with the COPII vesicle. To further delineate the regulation of CREBH-COPII interaction by GSK3β, we generated two CREBH mutants in which the putative GSK3β phosphorylation sites within the bZIP domain, Serine (S) 256 and S260, were changed to Alanine (A) (S256A and S260A), respectively (Figure 4C, S-Fig 1B). The vectors expressing the wild-type CREBH (CREBH-WT), the CREBH mutant S256A, or the CREBH mutant S260A was transferred into the mouse hepatoma cell line Hepa16, followed by serum shock for circadian synchronization. Confirming the circadian-regulated association of CREBH with the COPII vesicle proteins, CREBH-WT was rhythmically associated with Sec23A and Sec24A in a circadian-dependent manner (Figure 4D). In comparison, the CREBH mutant S260A exhibited increased affinity to complex with Sec23A and Sec24A, indicating that GSK3β-mediated phosphorylation of CREBH at S260 repressed the association of CREBH with Sec23/24 (Figure 4D). Interestingly, the site mutation at S256 (S256A) only slightly increased the

10

Diabetes

Page 12 of 51

affinity of CREBH to complex with Sec23A or Sec24A, suggesting that S260, but not S256, is the primary GSK3β phosphorylation site within CREBH that affects the interaction between CREBH and Sec23/24. Furthermore, we examined the levels of cleavage of CREBH-WT, S256A, and S260A in hepatocytes under the circadian clock. Consistent the circadian-regulated associations between CREBH isoforms and Sec23/24, the CREBH mutant S260A displayed increased cleavage activities, compared to CREBH-WT, under the circadian clock (Figure 4E). Together, these results indicated that GSK3β-mediated phosphorylation of CREBH at S260 alters the affinity of CREBH to complex with the COPII vesicular proteins, and thus modulates the ER to Golgi transport and subsequent proteolytic cleavage of CREBH under the circadian clock.

CREBH regulates circadian rhythmic levels of TG and FA by activating the genes encoding functions in lipolysis, FA oxidation, and lipogenesis. To elucidate whether CREBH regulates energy homeostasis under the circadian cycle, we characterized rhythmic profiles of circulating lipids in CREBH-null mice. Compared to WT mice, CREBH-null mice exhibited significantly higher amplitudes of serum TG and FA, but not cholesterol, over a 48-hr circadian period (Figure 5A-B, S-Fig 1D). However, hepatic TG levels of CREBH-null mice were reduced, compared to that of WT mice, at CT20 (equal to nighttime 2am) when mice usually take most of their meals of the day (Figure 5C). Furthermore, serum levels of the metabolic hormone FGF21, a known target of CREBH under metabolic stress (8), were reduced in CREBH-null mice across the circadian period (Figure 5D). Additionally, we examined body weights, body composition, blood glucose, and food consumption of CREBH-null and WT control mice. CREBH-null mice displayed modest reduction in body fat mass while they consumed modestly increased amounts of food, compared to the WT

11

Page 13 of 51

Diabetes

mice (S-Fig 2). These metabolic phenotypes may be partially attributed to the defects in lipolysis and FA oxidation in CREBH-null mice as we previously demonstrated (6; 8). To understand the molecular basis underlying the lipid-associated phenotype of CREBHnull mice under the circadian clock, we examined whether CREBH, as a transcription factor, rhythmically regulates expression of the genes involved in hepatic lipid metabolism. Quantitative real-time PCR (qRT-PCR) analysis indicated that rhythmic expression of the following genes were significantly repressed in CREBH-null mice (Figure 6A, S-Fig 3): 1) the genes encoding the coactivators or enzymes of lipolysis, including apolipoprotein C2 (ApoC2), ApoA4, and FA desaturase 2 (Fads2); 2) the genes encoding the enzymes or regulators in FA oxidation, including carnitine palmitoyltransferase 1A (CPT1α), 3-hydroxybutyrate dehydrogenase 1 (BDH1), and FGF21; and 3) the genes encoding the enzymes in lipogenesis, including Acetyl-CoA Carboxylase 1 (ACC1) and FA Elongase 6 (Elvol6). Consistent with the mRNA expression profiles, levels of ApoC2, ApoA4, CPT1α, BDH1, FADS2, Elovl6, and ACC1 proteins were decreased in the livers of CREBH-null mice (Figure 6C, S-Fig 4). Additionally, rhythmic expression of other key metabolic genes involved in lipolysis, FA oxidation, and lipogenesis, including Dhcr24, Lcat, Elvol2, Acot4, Hmgcl and Dgat2, was repressed in CREBH-null livers (S-Fig 3). These results support that CREBH functions as a critical regulator of circadian TG and FA metabolism by regulating expression of the key enzymes or regulators involved in lipolysis, FA oxidation, and de novo lipogenesis. To determine whether CREBH directly regulates its target genes involved in lipid metabolism during the circadian cycle, we performed ChIP-qPCR analysis to determine CREBH enrichment in the promoter regions of metabolic genes whose rhythmic expression profiles were altered in CREBH-null mouse livers. ChIP-qPCR analyses with WT mouse livers collected at

12

Diabetes

Page 14 of 51

different circadian phases indicated that CREBH binds in a circadian phase-dependent manner to the ApoC2, Bdh1, Cpt1a, Fgf21, Fads2, or Acc1 gene promoter that possesses one or multiple CREbinding elements (Figure 6B)(S-Fig 5). Increased enrichment of CREBH in the ApoC2 gene promoter was detectable at CT16 and peaked at CT4, which is consistent with the rhythmic expression profile of the ApoC2 mRNA in the liver. Similarly, consistent with the mRNA expression profiles, the enrichments of CREBH in the promoters of the genes encoding functions in lipogenesis, including Fads2 and Acc1, reached peak levels at CT8 and CT16, respectively (Figure 6B). Taken together, these results suggest that CREBH activates expression of genes involved in the metabolic pathways of both energy utilization (lipolysis and FA oxidation) and storage (lipogenesis) depending upon the circadian periods.

CREBH rhythmically regulates, and interacts with, the nuclear receptors PPARα and LXRα across the circadian cycle. To elucidate whether CREBH regulates, and/or interacts with, other local circadian regulators, we first examined expression of the genes involved in hepatic lipid metabolism in the livers of CREBH-null and WT control mice. Among others, PPARα is a liverenriched, clock-regulated nuclear receptor that plays key roles in regulating lipid utilization pathways during the starvation phase (24). Liver X receptor α (LXRα) is a nuclear receptor that regulates de novo lipogenesis and lipid uptake (25). Rhythmic levels of the Pparα and Lxrα mRNAs were significantly reduced in CREBH-null livers, compared to those in the control mice (Figure 7AB). ChIP-qPCR analysis indicated that CREBH binds in a day-night dependent manner to the Pparα or Lxrα gene promoters in mouse livers (Figure 7A-B), suggesting that Pparα and Lxrα are circadian-dependent targets of CREBH. Furthermore, Western blot analysis indicated that rhythmic levels of PPARα and LXRα proteins were decreased in CREBH-null livers (Figure 7C, S-Fig 6A).

13

Page 15 of 51

Diabetes

Circadian transcriptional regulators may interact with local transcription factors or nuclear receptors to control the circadian output of gene expression (26). IP-Western blot analyses with the mouse livers collected across the day-night cycle indicated that CREBH interacts with PPARα and LXRα at the different circadian phases (Figure 7D, S-Fig 6B). The interaction between CREBH and PPARα was detected from CT24/0 to CT12 and peaked at CT4, the day-time phases when lipolysis and FA oxidation are activated in mice upon energy demands. CREBH also interacted with LXRα, a nuclear receptor involved in lipogenesis and lipid uptake, during the circadian cycle (Figure 7D, S-Fig 6B). The interaction between CREBH and LXRα peaked from CT20 to CT24/0, a night-time period when lipogenesis is more activated in mice. Apparently, the phases of the interactions between CREBH, PPARα, and LXRα correlate with the time of feeding in mice, and support the dual functions of CREBH in energy utilization (lipolysis and FA oxidation) and storage (lipogenesis) during the day-night cycle.

The circadian output activator DBP and the repressor E4BP4 interact with CREBH to modulate CREBH transcriptional activities. Next, we examined whether CREBH regulates and/or interacts with the core or output circadian oscillation regulators to modulate its transcriptional activity. CREBH-deficiency resulted in marginally altered rhythmic expression of the genes encoding the core circadian oscillators including Bmal1, Clock, Per2, and Rev-erbα in mouse livers (S-Fig 7A). DBP, a direct target of CLOCK/BMAL1, function as circadian output transcriptional activator to activate expression of genes involved in energy metabolism (3; 27). E4BP4 is a basic-leucine-zipper transcription factor that function as a repressor of circadian oscillations (28). CREBH-null mice exhibited decreased expression of Dbp mRNA at the circadian time CT8, while CREBH deficiency resulted in a phase-inversed expression pattern of E4bp4

14

Diabetes

Page 16 of 51

mRNA in the liver (S-Fig 7B). Western blot analysis showed that the rhythmic levels of the circadian oscillator BMAL1, but not CLOCK, in CREBH-null liver were modestly decreased, compared to those in the WT livers (Figure 8A, S-Fig 8A). Rhythmic expression of E4BP4 protein was repressed in CREBH-null livers, compared to that in WT mouse livers. Expression levels of DBP was not affected by CREBH deletion, but CREBH-null mice exhibited a phase-shift in DBP expression during the circadian period from CT16 to CT4 (Figure 8A). Further, we examined the interactions between CREBH, DBP and E4BP4 in mouse liver during the circadian cycle. The interaction between CREBH and DBP was detected at CT12, CT8, and CT12, while the interaction between CREBH and E4BP4 was detected from CT16 to CT8 (Figure 8B, S-Fig 8B). Interestingly, robust interactions between CREBH and E4BP4 were detected at the circadian period from CT20 to CT4 when the interaction between CREBH and DBP was diminished (Figure 8B). Overall, the phase and amplitudes of the CREBH-DBP interaction roughly opposes that of the CREBH-E4BP4 interaction, implying that DBP and E4BP4 may compete to interact with CREBH in a circadian phase-dependent manner. To verify the suppressive effects of E4BP4 on CREBH interaction with its partners, mouse primary hepatocytes were infected with the adenovirus expressing E4BP4 or GFP control, followed by serum shock for circadian synchronization. Upon E4BP4 over-expression, the circadian-regulated interactions between CREBH and PPARα or DBP were repressed in the hepatocytes (Figure 8C, S-Fig 8C), thus confirming that E4BP4, as a circadian output repressor, can suppress CREBH interactions with its partners under the circadian clock. To explore the functional significance of the interactions between CREBH and DBP or E4BP4, we performed reporter analysis with the Fgf21 gene promoter, a common target of CREBH and its interaction partner PPARα (8). While over-expression of the activated CREBH or DBP

15

Page 17 of 51

Diabetes

alone can significantly increase Fgf21 gene promoter activity, co-expression of CREBH with DBP or PPARα further augmented the reporter activity (Figure 8D). In contrast, co-expression of CREBH with E4BP4 significantly decreased Fgf21 promoter activity, compared to expression of CREBH alone or co-expression of CREBH with GFP. These results suggest that DBP and PPARα function as co-activators of CREBH in driving Fgf21 gene transcription, while E4BP4 acts as a repressor of CREBH-dependent Fgf21 gene expression. Moreover, we observed that co-expression of E4BP4 with the combination of CREBH and DBP or PPARα repressed expression of the Fgf21 gene reporter, compared to co-expression of CREBH with DBP or PPARα (Figure 8D), thus supporting the suppressive effect of E4BP4 on CREBH transcriptional activity through competition with the co-activator DBP or PPARα. Given that the rhythmic expression of E4BP4 is regulated by CREBH (Figure 8A), the repressive effect of E4BP4 on CREBH activity may represent a negative feedback regulation of CREBH activity under the circadian constrain.

16

Diabetes

Page 18 of 51

Discussion

In this study, we demonstrated that CREBH is a liver circadian oscillator that plays key roles in integrating energy metabolism with circadian rhythm (Figure 8E). Our study revealed that proteolytic cleavage of CREBH, but not CrebH mRNA transcription, exhibits typical circadian rhythmicity (Figure 1), suggesting that CREBH is an output circadian factor that is regulated by the clock at the post-translational level. This is consistent with the recent rhythmic proteome studies showing that approximately one-half of rhythmic proteins are under significant translational or posttranslational diurnal controls and have no corresponding rhythmic mRNAs (29; 30). Our finding that BMAL1, the core circadian oscillator, regulates cleavage/activation of CREBH through AKTGSK3β signaling implicated a major molecular network through which BMAL1 regulates circadian energy metabolism in the liver. Importantly, GSK3β-mediated phosphorylation of ER-bound CREBH, which is under the control of Bmal1-AKT regulatory axis, is a critical regulatory event for the ER-to-Golgi transport and subsequent proteolytic cleavage of CREBH (Figures 3-4). Upon GSK3β-mediated phosphorylation, CREBH protein exhibits decreased affinity to complex with Sec23/24, the core protein components of the ER-to-Golgi transport vesicle COPII, under the circadian clock. It is possible that phosphorylation of CREBH by GSK3β leads to altered CREBH conformation with resulting decreased affinity toward the COPII-coated transport complex. To assemble the COPII complex, Sec23 and Sec24 form heterodimer where Sec24 is mainly responsible for cargo recognition and Sec23 binds Sar1-GTP (21). CREBH may indirectly interact with Sec24 through a potential scaffold protein like the SREBP-escort protein SCAP (31), an interesting question to be elucidated in the future. Our studies demonstrated that CREBH, as a hepatic transcription factor, is required to maintain homeostasis of circulating TG and FA by regulating expression of the key enzymes or 17

Page 19 of 51

Diabetes

regulators in lipolysis, FA oxidation, and de novo lipogenesis in a circadian-dependent fashion. Because CREBH is required for expression of Apoc2, Apoa4, Fads2, Dhcr24, and Lcat (Figure 6, S-Fig 3), all of which play important roles in TG lipolysis, CREBH-null mice have the defect in the clearance of TG-rich lipoproteins from the circulation and therefore display hypertriglyceridemia. Moreover, reduced expression of the regulators or enzymes in FA metabolism, including FGF21, CPT1α, BDH1, Fads, Elovls and PPARα, could be partially responsible for elevated serum FA in CREBH-null mice. Interestingly, during the nighttime, CREBH regulates expression of the genes encoding the functions in hepatic de novo lipogenesis, including Acc1, Elovl2, Elovl6, Lxrα, Hmgc1, and Dgat (Figures 6-7, S-Fig 3). This is consistent with the observation that CREBH-null mice preserved reduced levels of hepatic lipids in the nighttime phase when mice take most of the meal of the day (Figure 5C). The circadian-dependent, CREBH-regulated gene expression profiles suggested that CREBH regulates multiple metabolic pathways involved in both energy utilization and storage across the circadian cycle. CREBH has reciprocal interactions with the circadian transcriptional regulators PPARα and LXRα as well as the circadian oscillation activator DBP and repressor E4BP4 (Figures 7-8). CREBH regulates and interacts with PPARα or LXRα to enhance CREBH transcriptional activity, which oscillates in-phase with expression of the CREBH-target genes involved in lipolysis, FA oxidation, and lipogenesis. On the other hand, CREBH interacts with DBP or E4BP4 to modulate CREBH transcriptional activity during the night-to-day transition period. Interestingly, the phase of CREBH-DBP interaction is complimentary to that of CREBH-E4BP4 interaction, suggesting that DBP and E4BP4 may compete to interact with CREBH and thereby modulate CREBH activities during different circadian phases. As a co-activator of CREBH, PPARα interacts with CREBH in the circadian phase that partially overlaps with the CREBH-LXRα interaction (Figure 7D). It is

18

Diabetes

Page 20 of 51

possible that the interactions between CREBH, PPARα, and LXRα may represent enhancing mechanisms that facilitate CREBH peak activity. In summary, the core circadian oscillation regulates CREBH activity through two layers (Figure 8E): 1) the core circadian oscillator BMAL1 regulates proteolytic activation of CREBH; and 2) the output circadian modulators E4BP4 or DBP interact with the activated CREBH protein to exert suppressive or synergizing effect on CREBH activity. Conversely, CREBH also regulates expression of the core or output circadian components, including CLOCK, BMAL1, DBP, and E4BP4. The reciprocal regulation between CREBH and the key circadian regulators may provide an avenue through which local and central circadian regulators are integrated to influence whole body physiology. In the future, it is important to test whether modulation of CREBH activity represents an effective approach to intervene metabolic disorders.

19

Page 21 of 51

Diabetes

Authors’ contributions Study concept and design: ZZ, KZ, HK; Data acquisition, analysis and interpretation: KZ, ZZ, HK, YQ, RM, AD, XZ, CZ, LY, JL, PW, GK; Writing or revising the manuscript: KZ, ZZ, GK; Obtained funding: KZ; Technical or material support: XC, AL, LY, JL, PW, GK; Study supervision: KZ.

Acknowledgement Portions of this work were supported by National Institutes of Health (NIH) grants DK090313 and ES017829 (to KZ) and American Heart Association Grants 0635423Z and 09GRNT2280479 (to KZ). We thank Dr. Tianqing Peng of University of Western Ontario, Canada, for providing the recombinant adenovirus expressing GSK3β, and Dr. Xiao-Wei Chen and Dr. David Ginsburg of University of Michigan for providing the anti-Sec24A antibody. We thank Dr. Todd Leff of Wayne State University, USA, and Dr. Kenji Fukudome of Saga University, Japan, for their critical comments on this work. K.Z is the guarantor of this work and, as such, takes full responsibility for the work as a whole, including the study design, access to data, and the decision to submit and publish the manuscript. Conflict of interest statement: the authors ZZ, HK, YQ, XC, RM, AD, XZ, CZ, AL, LY, JL, PW, GK, and K.Z. have no conflict of interest to declare.

Abbreviations ER, endoplasmic reticulum; CT, circadian time; LD, 12hr light:12hr dark; DD, constant darkness; ChIP, chromatin immunoprecipitation; FA, fatty acids; TG, triglyceride. A full list of gene or protein name abbreviations is in Supplemental Information.

20

Diabetes

Page 22 of 51

References 1. Lowrey PL, Takahashi JS: Mammalian circadian biology: elucidating genome-wide levels of temporal organization. Annu Rev Genomics Hum Genet 2004;5:407-441 2. Reppert SM, Weaver DR: Molecular analysis of mammalian circadian rhythms. Annu Rev Physiol 2001;63:647-676 3. Bass J, Takahashi JS: Circadian integration of metabolism and energetics. Science 2010;330:1349-1354 4. Hatori M, Vollmers C, Zarrinpar A, DiTacchio L, Bushong EA, Gill S, Leblanc M, Chaix A, Joens M, Fitzpatrick JA, Ellisman MH, Panda S: Time-restricted feeding without reducing caloric intake prevents metabolic diseases in mice fed a high-fat diet. Cell Metab 2012;15:848860 5. Yang X, Downes M, Yu RT, Bookout AL, He W, Straume M, Mangelsdorf DJ, Evans RM: Nuclear receptor expression links the circadian clock to metabolism. Cell 2006;126:801-810 6. Zhang C, Wang G, Zheng Z, Maddipati KR, Zhang X, Dyson G, Williams P, Duncan SA, Kaufman RJ, Zhang K: Endoplasmic reticulum-tethered transcription factor cAMP responsive element-binding protein, hepatocyte specific, regulates hepatic lipogenesis, fatty acid oxidation, and lipolysis upon metabolic stress in mice. Hepatology 2012;55:1070-1082 7. Zhang K, Shen X, Wu J, Sakaki K, Saunders T, Rutkowski DT, Back SH, Kaufman RJ: Endoplasmic reticulum stress activates cleavage of CREBH to induce a systemic inflammatory response. Cell 2006;124:587-599 8. Kim HB, Mendez R, Zheng Z, Chang L, Cai J, Zhang R, Zhang K: Liver-Enriched Transcription Factor CREBH Interacts with Peroxisome Proliferator-Activated Receptor alpha to Regulate Metabolic Hormone FGF21. Endocrinology 2014:en20131490 9. Lee JH, Giannikopoulos P, Duncan SA, Wang J, Johansen CT, Brown JD, Plutzky J, Hegele RA, Glimcher LH, Lee AH: The transcription factor cyclic AMP-responsive element-binding protein H regulates triglyceride metabolism. Nat Med 2011;17:812-815 10. Lee MW, Chanda D, Yang J, Oh H, Kim SS, Yoon YS, Hong S, Park KG, Lee IK, Choi CS, Hanson RW, Choi HS, Koo SH: Regulation of hepatic gluconeogenesis by an ER-bound transcription factor, CREBH. Cell Metab 2010;11:331-339 11. Johansen CT, Wang J, McIntyre AD, Martins RA, Ban MR, Lanktree MB, Huff MW, Peterfy M, Mehrabian M, Lusis AJ, Kathiresan S, Anand SS, Yusuf S, Lee AH, Glimcher LH, Cao H, Hegele RA: Excess of rare variants in non-genome-wide association study candidate genes in patients with hypertriglyceridemia. Circ Cardiovasc Genet 2012;5:66-72 12. Cefalu AB, Spina R, Noto D, Valenti V, Ingrassia V, Giammanco A, Panno MD, Ganci A, Barbagallo CM, Averna MR: Novel CREB3L3 Nonsense Mutation in a Family With Dominant Hypertriglyceridemia. Arterioscler Thromb Vasc Biol 2015; 13. Balsalobre A, Damiola F, Schibler U: A serum shock induces circadian gene expression in mammalian tissue culture cells. Cell 1998;93:929-937 14. Kapatos G, Vunnava P, Wu Y: Protein kinase A-dependent recruitment of RNA polymerase II, C/EBP beta and NF-Y to the rat GTP cyclohydrolase I proximal promoter occurs without alterations in histone acetylation. J Neurochem 2007;101:1119-1133

21

Page 23 of 51

Diabetes

15. Molusky MM, Ma D, Buelow K, Yin L, Lin JD: Peroxisomal localization and circadian regulation of ubiquitin-specific protease 2. PloS one 2012;7:e47970 16. Zhang D, Tong X, Arthurs B, Guha A, Rui L, Kamath A, Inoki K, Yin L: Liver clock protein BMAL1 promotes de novo lipogenesis through insulin-mTORC2-AKT signaling. J. Biol. Chem. 2014;289:25925-25935 17. Saltiel AR, Kahn CR: Insulin signalling and the regulation of glucose and lipid metabolism. Nature 2001;414:799-806 18. Rudic RD, McNamara P, Curtis AM, Boston RC, Panda S, Hogenesch JB, Fitzgerald GA: BMAL1 and CLOCK, two essential components of the circadian clock, are involved in glucose homeostasis. PLoS Biol 2004;2:e377 19. Shi SQ, Ansari TS, McGuinness OP, Wasserman DH, Johnson CH: Circadian disruption leads to insulin resistance and obesity. Curr Biol 2013;23:372-381 20. Sahar S, Zocchi L, Kinoshita C, Borrelli E, Sassone-Corsi P: Regulation of BMAL1 protein stability and circadian function by GSK3beta-mediated phosphorylation. PloS one 2010;5:e8561 21. Lee MC, Miller EA: Molecular mechanisms of COPII vesicle formation. Sem Cell Dev Biol 2007;18:424-434 22. Zanetti G, Pahuja KB, Studer S, Shim S, Schekman R: COPII and the regulation of protein sorting in mammals. Nature cell biology 2012;14:20-28 23. Eldar-Finkelman H, Argast GM, Foord O, Fischer EH, Krebs EG: Expression and characterization of glycogen synthase kinase-3 mutants and their effect on glycogen synthase activity in intact cells. Proc Natl Acad Sci USA 1996;93:10228-10233 24. Oishi K, Shirai H, Ishida N: CLOCK is involved in the circadian transactivation of peroxisomeproliferator-activated receptor alpha (PPARalpha) in mice. Biochem J 2005;386:575-581 25. Willy PJ, Umesono K, Ong ES, Evans RM, Heyman RA, Mangelsdorf DJ: LXR, a nuclear receptor that defines a distinct retinoid response pathway. Genes Dev 1995;9:1033-1045 26. Schmutz I, Ripperger JA, Baeriswyl-Aebischer S, Albrecht U: The mammalian clock component PERIOD2 coordinates circadian output by interaction with nuclear receptors. Genes Dev 2010;24:345-357 27. Yamaguchi S, Mitsui S, Yan L, Yagita K, Miyake S, Okamura H: Role of DBP in the circadian oscillatory mechanism. Mol Cell Biol 2000;20:4773-4781 28. Clayton JD, Kyriacou CP, Reppert SM: Keeping time with the human genome. Nature 2001;409:829-831 29. Reddy AB, Karp NA, Maywood ES, Sage EA, Deery M, O'Neill JS, Wong GK, Chesham J, Odell M, Lilley KS, Kyriacou CP, Hastings MH: Circadian orchestration of the hepatic proteome. Curr Biol 2006;16:1107-1115 30. Mauvoisin D, Wang J, Jouffe C, Martin E, Atger F, Waridel P, Quadroni M, Gachon F, Naef F: Circadian clock-dependent and -independent rhythmic proteomes implement distinct diurnal functions in mouse liver. Proc Natl Acad Sci USA 2014;111:167-172 31. Sun LP, Seemann J, Goldstein JL, Brown MS: Sterol-regulated transport of SREBPs from endoplasmic reticulum to Golgi: Insig renders sorting signal in Scap inaccessible to COPII proteins. Proc Natl Acad Sci USA 2007;104:6519-6526

22

Diabetes

Page 24 of 51

Figure legends

Figure 1. Activation of CREBH in the liver is regulated by circadian rhythm. (A) Circadian oscillations of CrebH mRNA expression levels in WT mouse liver tissues collected every 4 hrs over a 24-hr period in constant darkness, determined by qRT-PCR. Fold changes of mRNA levels are shown by comparing to the nadir mRNA levels at CT8. Each point denotes the mean ± SEM (n=3 mice/time point). (B-C) Western blot analysis of levels of membrane-bound CREBH precursor (B) and nuclear forms (C) in mouse livers over the circadian circle. Cellular membrane and nuclear protein fractions were prepared from pooled liver tissues of WT mice collected every 4 hrs over a 24-hr circadian cycle (n=3 mice/time point). Levels of the ER chaperone BiP or the nuclear protein Lamin B1 were determined as controls. The graphs beside the images show the quantifications of CREBH precursor and nuclear forms in the mouse livers under the circadian clock. CREBH protein signals in the pooled liver membrane and nuclear protein fractions, determined by Western blot densitometry, were normalized to that of the ER membrane protein control BiP or the nuclear protein control Lamin B1. Fold changes of protein levels are shown by comparing to that at CT12.

Figure 2. Circadian activation of CREBH is regulated by BMAL1-AKT-GSK3β signaling pathway. (A) Western blot analysis of CREBH precursor and activated form in the livers of Bmal1 LKO and fl/fl control mice during different circadian phases. The liver protein lysates were prepared from pooled liver tissues of Bmal1 LKO and fl/fl mice collected every 6 hrs over a 24-hr circadian period (n=3-4 mice/genotype/time point) (15). Levels of BMAL1 and β-actin were determined as controls. A, am; P, pm. (B) Rhythmic fold changes of CREBH precursor and activated proteins in

23

Page 25 of 51

Diabetes

the livers of Bmal1 LKO and fl/fl mice during the circadian cycle. The CREBH protein signals, determined by Western blot densitometry, were normalized to that of β-actin. Fold change of the normalized CREBH precursor or activated protein levels was determined by comparing to that at 10 am. (C-E) Western blot analysis of CREBH precursor and activated protein in Bmal1-knockdown (C), AKT-dominant negative (AKT-DN) (D), GSK3β-dominant negative (GSK3β-DN) (E), and control primary hepatocytes during the 48-hr circadian cycle. Mouse primary hepatocytes were infected with adenovirus expressing Bmal1 shRNA, AKT-DN, GSK3β-DN, control LacZ shRNA, or control GFP for 24 hrs before subjected to horse serum shock for circadian synchronization. Cell lysates were collected at 8-hr intervals for the 48-hr circadian cycle for Western blot analysis to determine levels of CREBH, BMAL1, phosphorylated AKT2 (S473), total AKT2, phosphorylated GSK3β (S9), total GSK3β, phosphorylated Glycogen Synthase (GS) (S641), total GS, or β-actin. The graphs below the images show the rhythmic fold changes of cleaved CREBH proteins in Bmal1-knockdown, AKT-DN, GSK3β-DN, and control primary hepatocytes over the circadian cycle. CREBH protein signals were normalized to that of β-actin. Fold change of the normalized CREBH protein levels at each circadian time point was determined by comparing to that of the starting circadian time.

Figure 3. Expression of a constitutively active AKT or a dominant negative GSK3β can rescue decreased CREBH cleavage activities in Bmal1-knockdown primary hepatocytes. Western blot analyses of CREBH precursor and cleaved protein in Bmal1-knockdown and WT control mouse primary hepatocytes expressing the constitutively active AKT (AKT-CA) (A) or the dominant negative GSK3β (GSK3β-DN) (B) under the circadian clock. Mouse primary hepatocytes were infected with adenovirus expressing Bmal1 shRNA, AKT-CA, GSK3β-DN, or GFP control before

24

Diabetes

Page 26 of 51

subjected to horse serum shock for circadian synchronization. Cell lysates were collected at 8-hr intervals for Western blot analysis to determine levels of CREBH, BMAL1, phosphorylated AKT, or β-actin. The graphs on the right side show the rhythmic fold changes of cleaved CREBH proteins in Bmal1-knockdown and WT control hepatocytes expressing AKT-CA, GSK3β-DN, or GFP control. CREBH protein signals were normalized to that of β-actin. Fold change of the normalized CREBH protein levels at each circadian time point was determined by comparing to that of the starting circadian time.

Figure 4. GSK3β-mediated phosphorylation modulates CREBH association with the COPII vesicle proteins Sec23/24 and subsequent CREBH cleavage under the circadian clock. (A) IPWestern blot analysis of interactions between CREBH, Sec23A, and Sec24A in mouse primary hepatocytes expressing the constitutively active GSK3β (GSK3β-CA) or GFP control. Mouse primary hepatocytes were infected with adenovirus expressing GSK3β-CA or GFP, followed by horse serum shock for circadian synchronization. Cell lysates collected post the serum shock were pulled down by the anti-CREBH antibody and then probed with the Sec23A or Sec24A antibody. As the control, the CREBH-pull down lysates were probed with the CREBH antibody. The graphs below the image show the rhythmic fold changes of CREBH-associated Sec23A or Sec24A in hepatocytes expressing GSK3β-CA or GFP control. The immunoprecipated Sec23A and Sec24A protein signals were normalized to that of CREBH. Fold change of the normalized Sec23A and Sec24A protein levels at each circadian time point was determined by comparing to that of the starting circadian time. (B) Western blot analysis of CREBH, GSK3β, and β-actin in the mouse primary hepatocytes expressing GSK3β-CA or GFP control as described in the panel A. The graphs below the image show the rhythmic fold changes of cleaved CREBH protein in the hepatocytes

25

Page 27 of 51

Diabetes

expressing GSK3β-CA or GFP control. (C) Illustration of wild-type CREBH protein (CREBH-WT) and two CREBH mutants, S256A and S260A, in which the GSK3β phosphorylation residue S256 or S260 was mutated to A. (D) IP-Western blot analysis of interactions between CREBH, Sec23A, and Sec24A in Hepa16 cells expressing CREBH-WT, S256A, or S260A under the circadian clock. Hepa16 cells were transfected with plasmid vector expressing CREBH-WT, S256A or S260A, followed by horse serum shock for circadian synchronization. Cell lysates collected post the serum shock were pulled down by the anti-CREBH antibody and then probed with the Sec23A, Sec24A, or CREBH antibody. The graphs below the image show the rhythmic fold changes of CREBHassociated Sec23A or Sec24A in Hepa16 cells expressing CREBH-WT, S256A or S260A. The protein signal normalization and fold change calculation methods were as described in the panel A. (E) Western blot analysis of CREBH, GSK3β, and β-actin in the Hepa16 cells expressing CREBHWT, S256A, or S260A as described in the panel D. The graphs below show the rhythmic fold changes of cleaved CREBH protein in the Hepa16 cells expressing CREBH-WT, S256A, or S26A.

Figure 5. CREBH is required to maintain rhythmic levels of lipids and FGF21. Levels of serum TG (A), serum FA (B), hepatic TG (C), and serum FGF21 (D) in CREBH-null and WT mice under the circadian clock. Blood samples were collected every 6 hrs for 48 hrs in constant darkness for measuring TG, FA and FGF21. Liver tissue samples of CREBH-null and WT control mice were collected every 4 hrs for 24 hrs in constant darkness for measuring hepatic TG. Data was presented as mean ± SEM (n=8 mice/time point). * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.

Figure 6. CREBH regulates expression of the genes involved in lipolysis, FA oxidation and lipogenesis. (A) Rhythmic expression of CREBH-target genes involved in lipolysis, FA oxidation

26

Diabetes

Page 28 of 51

and lipogenesis in CREBH-null and WT mouse livers across the circadian clock. Levels of mRNAs were determined by qRT-PCR. Fold changes of mRNA levels are shown by comparing to that of one of WT mice at the starting circadian time point. Each bar denotes mean ± SEM (n = 3-5 mice/time point). (B) CREBH enrichment in CREBH-target gene promoters in WT mouse livers under different circadian phases determined by ChIP-qPCR. CREBH-null liver nuclei were used as negative control for the endogenous CREBH ChIP assays. Quantification of CREBH enrichment in the gene promoters at different circadian phases was determined by comparing ChIP-qPCR signals from the samples pulled down by the anti-CREBH antibody to that pulled down by a rabbit anti-IgG antibody. Each bar donates mean ± SEM (n=3 mice/time point). (C) Circadian rhythmic levels of the proteins encoded by the CREBH-target genes, including ApoC2, ApoA4, BDH1, CPT1α, FADS2, Elovl6, and ACC1, in the livers of CREBH-null and WT control mice. The liver tissue samples from CREBH-null and WT control mice were collected every 4 hrs in a 24-hour circadian period. Pooled liver protein lysates from 3-5 mice per genotype group per time point were used for the Western blot analyses. Levels of GAPDH were determined as loading controls. The quantifications of the rhythmic fold changes of these proteins were in S-Fig 4.

Figure 7. CREBH rhythmically regulates, and interacts with, circadian transcription regulators involved in hepatic lipid metabolism. (A-B) Rhythmic expression levels and CREBH enrichment in the promoters of the Pparα and Lxrα genes in the CREBH-null and WT mouse livers under the circadian clock. Fold changes of mRNA levels were determined by qRT-PCR (upper panel). Rhythmic enrichment of endogenous CREBH in the target gene promoters in WT mouse livers under different circadian phases were determined by ChIP-qPCR (lower panel). Each bar donates mean ± SEM (n=3 mice/time point). (C) Western blot analysis of rhythmic levels of PPARα and

27

Page 29 of 51

Diabetes

LXRα proteins in CREBH-null and WT mouse livers collected every 4 hrs in a 24-hr circadian period. Pooled liver protein lysates from 3-5 mice per time point per genotype group were used. (D) IP-Western blot analysis of interactions between CREBH and PPARα or LXRα in the liver nuclear fractions of WT mice in different circadian phases of a 24-hr circadian period. Liver nuclear proteins pooled from 3 WT mice per time point was pulled down by the anti-CREBH antibody and then probed with the PPARα, LXRα, or CREBH antibody. Levels of Lamin B1 were determined as controls. The quantifications of protein rhythmic fold changes for the panels C-D were in S-Fig 6.

Figure 8. CREBH rhythmically interacts with the circadian activator DBP and repressor E4BP4 to modulate CREBH transcriptional activity. (A) Western blot analysis of rhythmic levels of CLOCK, BMAL1, DBP, and E4BP4 proteins in CREBH-null and WT mouse livers collected every 4 hrs in a 24-hr circadian period. Pooled liver protein lysates from 3-5 mice per time point per genotype were used. Levels of GAPDH were included as loading controls. (B) IP-Western blot analysis of interactions between CREBH and DBP or E4BP4 in the liver nuclear fractions of WT mice under different circadian phases of a 24-hr circadian period. Liver nuclear proteins pooled from 3 WT mice per time point was pulled down by the anti-CREBH antibody and then probed with the DBP or E4BP4 antibody. Levels of Lamin B1 were determined as controls. (C) IP-Western blot analysis of interactions between CREBH, PPARα, and LXRα in mouse primary hepatocytes expressing E4BP4 or GFP control. Mouse primary hepatocytes were infected with adenovirus expressing E4BP4 or GFP, followed by serum shock for circadian synchronization. Cell lysates collected post the serum shock were pulled down by the anti-CREBH antibody and then probed with the PPARα, LXRα, or CREBH antibody. The same protein samples were subjected to Western blot analyses of E4BP4 and β-actin levels (the image below). For the panels A-C, the

28

Diabetes

Page 30 of 51

quantifications of protein rhythmic fold changes were in S-Fig 8. (D) Luciferase reporter analyses of transcriptional activation of the human Fgf21 gene promoter by CREBH alone or in combination with DBP, PPARα, and/or E4BP4. Hepa1-6 cells were transduced with the Fgf21 reporter vector or vehicle. After 24 hrs, the transfected cells were infected with adenovirus expressing GFP (control), DBP, PPARα, and/or E4BP4. Renilla reporter plasmid was included in the co-transfection for normalization of luciferase reporter activities. The same amounts of adenovirus titers were used for individual infections. Each bar denotes the mean ± SEM. * p < 0.05. (E) A working model for CREBH function as a circadian metabolic regulator.

29

Page 31 of 51

Diabetes

Fig 1

A 1.6

CREBH mRNA (fold changes)

1.4 1.2 1.0 0.8 0.6 0.4

12

16

20

24/0

4

8

12

Circadian Time (CT)

CT 12

16

20 24/0 4

8

12 Membrane_CREBH (precursor) Membrane_BiP

Membrane CREBH (fold changes)

B 1.5

Pooled membrane proteins

1 0.5 0

CT 12

16

20

24/0

4

8

12

8

12

CT 12

16 20 24/0

4

8

12 Nuclear_CREBH (activated form) Nuclear_Lamin B1

Nuclear CREBH (fold changes)

C 1.5

Pooled nuclear proteins

1 0.5 0

CT 12

16

20

24/0

4

Diabetes

A

fl/fl CREBH Precursor

Page 32 of 51

Fig 2

D

Bmal LKO

AKT DN

GFP

10A 4P 10P 4A 10A 4P 10P 4A

Circadian (h) CREBH Precursor

0 8 16 24 32 40 48 0 8 16 24 32 40 48

Cleaved Cleaved Bmal1 Bmal1 Actin GSK3 (S9)

B

AKT Pooled liver proteins

fl/fl ko

1.5 1.2 0.9 0.6 0.3 0.0

10A

4P

10P

4A

1.8

Actin

fl/fl ko

1.5 1.2

Cleaved CREBH rhythmic profile (fold changes)

Cleaved CREBH (fold changes)

CREBH precursor (fold changes)

Pooled liver proteins 1.8

0.9 0.6 0.3 0.0

10A

4P

10P

4A

4

LacZ shRNA

AKT DN

2 1

0 CT (h) 0

C

GFP

3

8

16

24

32

40

48

Bmal1 shRNA

Circadian (h) 0 8 16 24 32 40 48 0 8 16 24 32 40 48 CREBH Precursor

E GFP GSK3 DN Circadian (h) 0 8 16 24 32 40 48 0 8 16 24 32 40 48 CREBH Precursor

Cleaved

Bmal1 Cleaved p-AKT (S473) Bmal1 AKT GSK3

p-GSK3 (S9)

p-GS GSK3 GS Actin

8

sh LacZ

sh Bmal1

6 4 2 0 CT (h) 0

8

16

24

32

40

48

Cleaved CREBH rhythmic profile (fold changes)

Cleaved CREBH rhythmic profile (fold changes)

Actin 4

GFP

GSK3 DN

3 2 1

0 CT (h) 0

8

16

24

32

40

48

Page 33 of 51

Diabetes

Fig 3

A WT

Bmal1 shRNA

GFP CT (h)

0

8 16 24 32 0

GFP 8

AKT-CA

16 24 32 0

8 16 24 32

Cleaved CREBH rhythmic profile (fold changes)

CREBH Precursor Cleaved

BMAL1

P-AKT

Bmal1 shRNA

WT GFP

40

GFP

AKT-CA

30 20 10 0 CT (h)

0

8

16

24

32

-actin

B CT (h)

CREBH Precursor Cleaved

BMAL1

GSK3

-actin

0

8 16 24 32 0

Bmal1 shRNA GFP 8 16 24 32 0

GSK3-DN 8 16 24 32 Bmal1 shRNA

Cleaved CREBH rhythmic profile (fold changes)

WT GFP

WT GFP

40

GSK3-DN

GFP

30 20 10 0 CT (h)

0

8

16

24

32

PageFig 34 of451

Diabetes

A

D GFP CT (h) 0

GSK3-CA

8 16 24 32 0

8

CREBH WT

16 24 32

CT (h) 0

CREBH IP Sec23A IB

CREBH IP Sec24A IB

CREBH IP Sec24A IB

CREBH IP CREBH IB

CREBH IP CREBH IB GFP

CT (h) 0

8

15

GSK3-CA

16

GFP

24

32

9 6 3 0

8

S260A

0 8 16 24 32 0

30

8 16 24 32

WT S256A S260A

25 20 15 10 5 0

CT (h) 0

8

16

24

32

GSK3-CA

12

CT (h) 0

8 16 24 32

CREBH-Sec23A interaction (fold change)

8 7 6 5 4 3 2 1 0

CREBH-Sec24A interaction (fold change)

CREBH-Sec24A interaction (fold change)

CREBH-Sec23A interaction (fold change)

CREBH IP Sec23A IB

S256A

16

24

32

3.0

WT S256A S260A

2.0

1.0

0.0

CT (h) 0

8

16

24

32

B GFP CT (h) 0

GSK3-CA

8 16 24 32

0

8 16 24 32

CREBH Precursor

E

CREBH WT

S256A

CT (h) 0 8 16 24 32 0

S260A

8 16 24 32

0

8 16 24 32

CREBH Precursor

Cleaved GSK3

Cleaved

Actin GSK3

Cleaved CREBH (fold changes)

4.0

GFP

GSK3-CA

3.0

Actin

2.0 1.0 0.0

8

16

C S256

CREBH WT

S260

32

50

ER-TM

CREBH 256A

ER-TM S260A

bZIP

S256A

S260A

30 20 10

CT (h) 0

S260

bZIP S256

WT

40

0

bZIP S256A

CREBH 260A

24

Cleaved CREBH (fold changes)

CT (h) 0

ER-TM

8

16

24

32

Page 35 of 51

Diabetes

Fig 5

A

B 250

Night Day Night***Day

**

200 150 100

****

**** ***

*** ***

*

50 0

CT

12

18 24/0 6

12

18 24/0 6

1800 1600

KO

Night** Day Night

1400

*

1200

* Day

*

1000

**

800 600 400 200 0

CT

12

C

12 18 24/0 6

12

18 24/0 6

12

D WT

WT

KO

Serum FGF21 (pg/mL)

Hepatic TG (g/mg liver)

Serum FFA (mg/dL)

Serum TG (mg/dL)

300

WT

KO

WT

5.0

Night

Day

4.0 3.0

*

2.0 1.0 0.0 CT 16

20

24/0

4

8

12

KO

1600 1400

Night

Day Night

1200

Day

*

1000

*

800 600 400 200 0 CT

12

18 24/0

6

12

18 24/0 6

12

Diabetes mRNA (fold changes)

A

ApoC2 *

2.0

* 0.9 0.6

0.4

0.3

*

16

20

24/0

4

8

12

Fgf21

2.5

*

2.0

** **

*

0.2 0.0

CT 12

16

20 24/0

4

8

12

CT 12

Fads2

**

***

*

16

1.8

*** *

0.6

*

1.6

20 24/0

4

8

12

Acc1

WT KO

*

1.4 1.2

*

1.0

0.4

1.0

*

0.6

0.8

1.5

*

0.8

0.2

0.5

B

16

20 24/0

4

8

12

CREBH in ApoC2 promoter ***

13 11

0.4

CT 12

7

5

7

4

5

3

*

3

KO

CT20

CT4

CT8

20 24/0

1

CREBH in Fgf21 promoter

6

**

5

***

4 3 2 1

KO

CT16

C

CT20 CT4 WT

CT8

WT CT 12 16

20 24/0 4

17 15 13 11 9 7 5 3 1

8

* * * KO

CT16

CT20

WT 7

4

CREBH in Bdh1 promoter

2

CT16

16

6

9

1

0.6

0.0

0.0

CT 12

Fold enrichment

0.8

0.4

1.0

*

**

1.0

*

WT KO

Cpt1

1.2

0.0

0.0

mRNA (fold changes)

*

1.2

0.8

CT 12

Fold enrichment

*

**

1.2

Bdh1 ** *

1.5

1.6

Page 36 of 51

CT4 WT

CT8

CREBH in Fads2 promoter

*

*

*

KO

CT16

CT20

CT4

CT8 WT

12

9 8 7 6 5 4 3 2 1

10 9 8 7 6 5 4 3 2 1

CT 12

16

20 24/0

4

8

12

CREBH in Cpt1 promoter *

* *

KO

CT16

CT20

CT4 WT

CT8

CREBH in Acc1 promoter

**

* KO

CT16

CT20

* CT4 WT

CT8

CREBH KO 8

12

12 16 20 24/0 4

8

12

ApoC2 ApoA4 BDH1 CPT1 FADS2 Elovl6 ACC1 GAPDH

Fig 6

Page 37 of 51

Diabetes

Fig 7

A

B *

1.4 1.2

**

1.0

*

0.8 0.6 0.4

CT 12

16

20 24/0

4

8

12

*

*

CT 12

16

20 24/0

4

1.0 0.8 0.6 0.4

8

12

CREBH in Lxr promoter 7

**

7 6

fold enrichment

fold enrichment

*

1.2

CREBH in Ppar promoter 8

*

5 4 3 2 1

**

1.6 1.4

WT KO

Lxr

1.8

mRNA (fold changes)

Ppar

1.6

mRNA (fold changes)

WT KO

KO

T16

CT20 CT4 WT

5 4

*

3

*

2 1

CT8

*

***

6

KO

T16

CT20 CT4 WT

CT8

D

C WT CT 12

PPAR

16 20 24/0 4

CREBH KO 8

12

12 16

20 24/0 4

8

12

CT

IP: CREBH IB: PPAR

LXR

IP: CREBH IB: LXR

Actin

IP: CREBH IB: CREBH WB: Lamin B1

12

16

20 24/0

4

8

12

Fig 8 Page 38 of 51

Diabetes

A

WT

B

CREBH KO

CT 12 16 20 24/0 4

8 12 12 16 20 24/0 4

CT 12 8 12

16

20 24/0

4

8

12

IP: CREBH IB: DBP

CLOCK

IP: CREBH IB: E4BP4

BMAL1

IP: CREBH IB: CREBH

DBP

WB: Lamin B1

E4BP4 GAPDH

0

8 16

24 32

WB: E4BP4

FP

+CREBH

*

CREBH precursor

+GFP Active CREBH

LXRD + CREBH

E4BP4

DBP PPARD + CREBH

Nucleus Lipogenesis

Lipolysis, FA oxidation

*

C G R FP EB C H R EB +D B H +P P PA C R R EB D C H R + EB D B H +P P PA R D

no

n-

tr

an

ER

Ve c

Day – Energy utilization

50 45 40 35 30 25 20 15 10 5 0

s

Night – Energy storage

FGF21 reporter (RFU)

WB: Actin

E

D B P E4 B P4

8 16 24 32

no

Hours 0

E4BP4

G

GFP

*

D B P E4 B P4

IP: CREBH IB: CREBH

*

*

G FP C R EB H

IP: CREBH IB: DBP

*

50 45 40 35 30 25 20 15 10 5 0

Ve c

8 16 24 32

s

0

an

8 16 24 32

D

tr

Hours 0 IP: CREBH IB: PPARD

E4BP4

FGF21 reporter (RFU)

GFP

n-

C

+E4BP4

Page 39 of 51

Diabetes

Supplemental Information “CREBH Couples Circadian Clock with Hepatic Lipid Metabolism” by Zheng et al.

A list of gene or protein abbreviations used in this manuscript: CREBH, cyclic AMP responsive element-binding protein, hepatocyte-specific. Also known as CREB3L3 (Cyclic AMP responsive element-binding protein 3-like 3) BMAL1, Aryl hydrocarbon receptor nuclear translocator-like DBP, albumin D-element-binding protein E4BP4, nuclear factor interleukin-3-regulated protein CLOCK, Class E Basic Helix-Loop-Helix Protein 8 PER2, Period Circadian Clock 2 REV-ERBα, nuclear Receptor Subfamily 1, Group D, Member 1 (also called NR1D1) PPARα, peroxisome proliferator-activated receptor α LXRα, liver X receptor-alpha FGF21, Fibroblast growth factor 21 ApoC2, apolipoprotein C2 Elovl5, elongation of very long chain fatty acids protein 5 Bdh1, 3-Hydroxybutyrate Dehydrogenase, Type 1 Cpt1α, carnitine palmitoyltransferase 1α Fads2, fatty acid desaturase 2 Acsl1, long-chain-fatty-acid-CoA ligase 1 Acc1, acetyl CoA carboxylase 1

Resources of antibodies and the other bio-reagents used in this study Synthetic oligonucleotides were purchased from Integrated DNA Technologies, Inc. (Coralville, IA). The affinity-purified rabbit polyclonal anti-CREBH antibody was developed in our laboratory (Endocrinology 2014;155:769-782). The commercially available antibodies were used to detect endogenous protein levels of DBP, E4BP4, FADS2, CPT1α, BDH1 (Santa Cruz Biotech), LXRα (Invitrogen), HNF4α (Invitrogen), CLOCK (Cell Signaling), BMAL1 (Novus Biologicals), PPARα (Millipore), ACC1 (Epitomics), Sec23A (Genetex), Sec24A (kindly provided by Dr. David Ginsburg) (1), phospho-AKT (Ser473) and total AKT (Cell Signaling), phospho-GSK3β (Ser9) and total GSK3β (Cell Signaling), β-actin (Sigma), phosphoGlycogen synthase (GS) and total GS (Cell Signaling), and GAPDH (Sigma), respectively, in mouse liver lysates or nuclear protein fractions by Western blot or IP-Western blot analysis. Kits for measuring TG, FA, and Cholesterol were from BioAssay System (Hayward, CA). Mouse FGF21 ELISA kit was from R & D System, Inc (Minneapolis, MN). The adenovirus expressing Bmal1 shRNA was previously described in details (2). The adenovirus expressing the dominant negative AKT (AKT-DN) and the adenovirus expressing the constitutively-activated AKT (AKT-CA) were kindly provided by Dr. Morris Birnbaum at 1   

Diabetes

Page 40 of 51

the University of Pennsylvania. The adenovirus expressing the dominant negative GSK3β (GSK3β-DN) and the adenovirus expressing the constitutively- activated GSK3β (GSK3β-CA) were kindly provided by Dr. Tianqing Peng at University of Western Ontario, Canada (3-5).

Supplemental figure legends S-Fig 1. (A) Circadian oscillations of CREBH protein levels in pooled WT and CREBH-null mouse liver tissues collected every 4 hrs over a 24-hr period in constant darkness, determined by Western blot analysis. Levels of GAPDH were determined as the loading controls. (B) Levels of CREBH precursor and activated and protein in the livers of Bmal1 LKO and flox/flox (fl/fl) control mice under the feeding condition or after 16-hr overnight fasting. Liver protein lysates were prepared from pooled livers of Bmal1 LKO or fl/fl mice (n=3-4 mice per genotype per treatment) for Western blot analysis to determine the levels of CREBH, BMAL1, and β-actin. (C) Potential phosphorylation sites within the mammalian CREBH proteins, analyzed by PhosphoSitePlus (http://www.phosphosite.org). The mammalian CREBH proteins possess conserved GSK3β phosphorylation site within their bZIP domain. The GSK3β phosphorylation sites in human, mouse, and rat CREBH proteins are S256/S260, S252/S256, and S252/S256, respectively. TM, transmembrane domain. S, Serine; T, Threonine; Y, Tyrosine; N, Asparagine. (D) Levels of circulating cholesterols in CREBH-null and WT control mice under the circadian clock. Blood samples were collected every 6 hrs for 48 hrs in constant darkness for measuring cholesterols. Data was presented as mean ± SEM (n=8 mice per time point) at each time point. S-Fig 2. Body weights, body compositions, and food consumption of CREBH-null and WT control mice. (A-D) Body compositions, including body fat mass (A), body lean mass (B), body fluid (C), and body water (D), in male CREBH-null and WT control mice of 3-months old (n=6). The body compositions were measured by EchoMRI (EchoMRI LLC, Houston, USA). (E) Body weights of male CREBH-null and WT control mice of 3-months old (n=6). (F) Levels of blood glucose of male CREBH-null and WT control mice of 3-months old. (G) Levels of food consumption of male CREBH-null and WT control mice of 3-months old over a 48-hour circadian phase. S-Fig 3. Quantifications of the rhythmic fold changes of the protein levels in Figure 6C. The ApoC2, ApoA4, BDH1, CPT1α, FADS2, Elovl6, and ACC1 protein signals, as determined by Western blot analyses in Figure 6C, were normalized to that of GAPDH. Fold change of the normalized protein levels at each circadian time point was determined by comparing to that of the starting circadian time. S-Fig 4. Expression profiles of the genes encoding key enzymes involved in lipolysis, FA oxidation, and lipogenesis, including 24-Dehydrocholesterol Reductase (Dhcr24), Lecithin-Cholesterol Acyltransferase (Lcat), ELOVL fatty acid elongase 2 (Elovl2), Elovl6, Acyl-CoA Thioesterase 4 (Acot4), ApoA4, 3hydroxymethyl-3-methylglutaryl-CoA lyase (Hmgcl), and diacylglycerol O-acyltransferase 2 (Dgat2), in the livers of CREBH-null and WT control mice under circadian clock. The liver samples from CREBHnull and WT control mice were collected every 4 hrs over a 24-hr period. These RNAs were subjected to 2   

Page 41 of 51

Diabetes

quantitative real-time RT-PCR analysis. Expression values were normalized to the Arbp mRNA levels. Fold changes of mRNA levels are shown by comparing to that of one of the WT control mice at the starting circadian time point. Asterisks indicate significant differences (* p < 0.05, ** p < 0.01) between WT and CREBH-null mice by post-hoc analyses followed by two-way ANOVA. Data represent mean ± SEM (n=3-5 mice per group per time point). S-Fig 5. CRE binding motifs in the promoter regions of mouse ApoC2, ACC1, Fgf21, Bdh1, Cpt1α, Fads2, Pparα, Lxrα, and E4bp4 genes. The binding motifs are highlighted (red underline). The complementary sequences (blue underline) are presented if the binding motifs locate in the negative strand. S-Fig 6. Quantifications of the rhythmic fold changes of the protein levels, determined by Western blot and IP-Western blot analyses in Figure 7. The PPARα and LXRα protein signals, as determined by Western blot analysis in Figure 7C, were normalized to that of Actin (A). The CREBH-associated PPARα and LXRα protein signals, as determined by IP-Western blot analysis in Figure 7D, were normalized to that of CREBH (B). Fold change of the normalized protein levels at each circadian time point was determined by comparing to that of the starting circadian time. S-Fig 7. Rhythmic expression levels and amplitudes of core clock genes in the livers of CREBH-null and WT control mice. Expression profiles of the core clock genes and direct output circadian regulator genes Bmal1, Clock, Rev-erbα, and Per2 (A) as well as the circadian oscillation modulator genes Dbp and E4BP4 (B) in the livers of CREBH-null and WT control mice under the circadian clock. The liver samples from the CREBH-null and WT control mice were collected every 4 hrs over a 24-hr period in constant darkness. Expression values of mRNAs were determined by qRT-PCR and normalized to the Arbp mRNA levels. Fold changes of mRNA levels are shown by comparing to that of one of the wild-type control mice at the starting circadian time (CT) point. Asterisks indicate significant differences (* p < 0.05) between WT and CREBH-null mice by post-hoc analyses followed by two-way ANOVA. Data represent mean ± SEM (n=3 mice per group per time point). S-Fig 8. Quantifications of the rhythmic fold changes of the protein levels, determined by Western blot and IP-Western blot analyses in Figure 8. The CLOCK, BMAL1, DBP, and E4BP4 protein signals, as determined by Western blot analysis in Figure 8A, were normalized to that of GAPDH (A). The CREBHassociated DBP and E4BP4 protein signals, as determined by IP-Western blot analysis in Figure 8B, were normalized to that of CREBH (B). The CREBH-associated PPARα and DBP protein signals, as determined by IP-Western blot analysis in Figure 8C, were normalized to that of CREBH (C). Fold change of the normalized protein levels at each circadian time point was determined by comparing to that of the starting circadian time. S-Table 1. Sequences of the primers used in this study.

3   

Diabetes

Page 42 of 51

References 1. Chen XW, Wang H, Bajaj K, Zhang P, Meng ZX, Ma D, Bai Y, Liu HH, Adams E, Baines A, Yu G, Sartor MA, Zhang B, Yi Z, Lin J, Young SG, Schekman R, Ginsburg D: SEC24A deficiency lowers plasma cholesterol through reduced PCSK9 secretion. eLife 2013;2:e00444 2. Zhang D, Tong X, Arthurs B, Guha A, Rui L, Kamath A, Inoki K, Yin L: Liver clock protein BMAL1 promotes de novo lipogenesis through insulin-mTORC2-AKT signaling. The Journal of biological chemistry 2014;289:25925-25935 3. Shen E, Fan J, Peng T: Glycogen synthase kinase-3beta suppresses tumor necrosis factor-alpha expression in cardiomyocytes during lipopolysaccharide stimulation. Journal of cellular biochemistry 2008;104:329-338 4. Eldar-Finkelman H, Argast GM, Foord O, Fischer EH, Krebs EG: Expression and characterization of glycogen synthase kinase-3 mutants and their effect on glycogen synthase activity in intact cells. Proc Natl Acad Sci USA 1996;93:10228-10233 5. Summers SA, Kao AW, Kohn AD, Backus GS, Roth RA, Pessin JE, Birnbaum MJ: The role of glycogen synthase kinase 3beta in insulin-stimulated glucose metabolism. The Journal of biological chemistry 1999;274:17934-17940

4   

Page 43 of 51

Diabetes

A

S-Fig 1

B Feed WT CT 12 16 20 24/0 4

fl/fl

CREBH KO 8

12

12 16 20 24/0 4

8 12

CREBH Precursor

CREBH precursor Activated CREBH cleaved GAPDH

BMAL1 Actin

C

T273 Y267 S260 S173 T229 S256 S276

Human CREBH bZIP

TM

N269 Y263 S256 S166 T225 S252 S272

Mouse CREBH bZIP

TM

N269 Y263 S256 S166 T225 S252 S272

Rat CREBH bZIP

TM

D Serum Cholesterol (mg/dL)

WT

KO

100

Night Day Night Day 80 60 40 20 0

CT 12 18 24/0

6

12 18 24/0 6

12

ko

Fast fl/fl

ko

Diabetes

2.0 1.0

WT KO

16.0 12.0 8.0 4.0 0.0

20.0

p=0.37

16.0 12.0 8.0 4.0 0.0

WT KO

Food consumption (g)

G 48-h food intake 12

p=0.07

10 8 6 4 2 0

WT KO

0.25

p=0.40

0.2

p=0.07

0.15 0.1 0.05 0

WT KO

E

WT KO

F 30.0

Body weight (g)

D

20.0

Body fluid (g)

p=0.18

C

25.0

p=0.19

20.0 15.0 10.0 5.0 0.0

WT KO

Blood glucose (mg/dL)

3.0

Body lean mass (g)

4.0

0.0

Body water (g)

S-Fig 2

B

Body fat mass (g)

A

Page 44 of 51

200

p=0.84

160 120 80 40 0

WT

KO

S-Fig 3

Diabetes

mRNA (fold changes)

Page 45 of 51

**

0.7

0.7

*

0.6

**

0.6

0.5

0.5

0.4

0.4

*

0.3

12

16

20

24/0

4

8

12

CT

12

16

20

24/0

Elovl2

3.0

4

8

12

Elovl6

WT KO

2.0

2.5 1.6

2.0

*

1.5

1.2 0.8

1.0

0.0

12

16

20

24/0

4

8

12

Acot4

CT

12

16

1.7

0.9

1.4

0.6

1.1

20

24/0

4

8

12

ApoA4

2.0

1.2

WT KO

0.8

0.3 0 CT

*

0.4

0.5

1.5

mRNA (fold changes)

WT KO

0.8

0.8

0.0 CT

0.5

12

16

20

24/0

4

8

12

Hmgc1 mRNA (fold changes)

Lcat

0.9

0.9

CT

mRNA (fold changes)

Dhcr24

1.0

12

16

20

*

1.4

1.4

1.2

1.3

24/0

4

8

12

Dgat2

1.6

1.6 1.5

CT

WT KO

*

*

1.0

1.2 1.1

0.8

1.0

0.6

0.9 0.8 CT

12

16

20

24/0

4

8

12

0.4 CT

12

16

20

24/0

4

8

12

Page 46 of451 S-Fig

Diabetes

A

B

ApoC2 WT

1.5

KO ApoA4 protein levels (fold changes)

ApoC2 protein levels (fold changes)

5.0

ApoA4

4.0 3.0 2.0 1.0

WT

KO

1.2 0.9 0.6 0.3 0

0.0

CT 12

16

20

24/0

4

8

CT 12

12

C

16

20

24/0

1.5

12

4

8

12

CPT1

WT

5

KO Cpt1 protein levels (fold changes)

Bdh1 protein levels (fold changes)

8

D Bdh1

1.0

0.5

CT 12

16

20

24/0

4

8

WT

12

E

KO

4 3 2 1 0 CT 12

0.0

16

20

24/0

F Fads2

Elovl6 5

WT

KO

Elovl6 protein levels (fold changes)

3.0 Fads2 protein levels (fold changes)

4

2.5 2.0 1.5 1.0 0.5

KO

3 2 1 0

0.0

CT 12

16

20

24/0

4

8

CT 12

12

G ACC1 1.5 ACC1 protein levels (fold changes)

WT 4

WT

KO

1.0

0.5

0.0 CT 12

16

20

24/0

4

8

12

16

20

24/0

4

8

12

Page 47 of 51

Diabetes

S-Fig 5 Apoc2 promoter TGGCCTCTGACTGTCACTGT (nt -117 to nt -113) Acc1 promoter CTAACGCTGACCTTCTTTAC (nt -315 to nt -310) CTTTCTCATGAACTTTATTT (nt -271 to nt -265) Fgf21promoter CCACTCCTGACGCGTGATAT (nt -63 to nt -67) Bdh1 promoter GTGAGGTGACCAATCCCCCT (nt -452 to nt-434) Cpt1a promoter TCATTCTCTGATGTTAGACAAGC (nt -568 to nt -562) TTCCTTACTGACCTCCTCCCCGCA (nt -245 to nt -240) Fads2 promoter AGGTCAGACACGTCGCCGACCG (nt -599 to nt -594) Ppara promoter ACAGGGGTGACGGGGGC (nt -323 to nt -319) Lxra promoter GGAACGCTGACTCTGGAGGCT(nt -184 to nt -180) GTGGGGGTGACTGAGAAGCAG (nt -151 to nt -147) E4bp4 promoter CCGCCGCCCGTCACGGCGGG G (nt -160 to nt -156). GCAGT

Diabetes

Page 48 of 51

S-Fig 6

A LXR

WT

LXR protein levels (fold changes)

PPAR protein levels (fold changes)

PPAR 4 KO

3 2 1 0 CT

12

16

20

24/0

4

8

1.5 WT

0.9 0.6 0.3 0.0 CT

12

KO

1.2

12

16

20

24/0

4

8

12

800

CREBH-PPAR interaction LXR protein interaction profile (fold change)

PPAR protein interaction profile (fold change)

B 700 600 500 400 300 200 100 0

CT 12

16

20

24/0

4

8

12

CREBH-LXR interaction

45 40 35 30 25 20 15 10 5 0

CT 12

16

20

24/0

4

8

12

Page 49 of 51

Diabetes

S-Fig 7

A Bmal1 mRNA (fold changes)

mRNA (fold changes)

12.0 10.0 8.0 6.0 4.0 2.0 0.0

12

16

20

24/0

4

8

Reverb

16.0 14.0 12.0 10.0 8.0 6.0 4.0 2.0 0.0 CT

12

16

20

24/0

4

8

3.0

*

2.5 2.0 1.5

*

1.0 0.5 0.0 CT

12

mRNA (fold changes)

mRNA (fold changes)

CT

WT KO

Clock 3.5

14.0

12

12

16

20

4

8

12

4

8

12

Per2

1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0

CT

24/0

12

16

20

24/0

Dbp

3.0

6.0

*

mRNA (fold changes)

mRNA (fold changes)

B 2.5 2.0 1.5 1.0 0.5 0.0 CT

E4bp4

**

5.0

*

*

4.0

WT KO

* *

3.0 2.0 1.0 0.0

12

16

20

24/0

4

8

12

CT

12

16

20

24/0

4

8

12

Page 50 of S-Fig 851

Diabetes

A

Bmal1 10

WT

Bmal1 protein levels (fold changes)

CLOCK protein levels (fold changes)

CLOCK 4 KO

3 2 1

0 CT

12

16

20

24/0

4

8

WT

6 4 2

0 CT

12

12

16

20

WT

KO

0.9 0.6 0.3 0 CT

12

16

20

24/0

24/0

4

8

12

4

8

12

E4BP4 E4BP4 protein levels (fold changes)

DBP protein levels (fold changes)

DBP 1.5 1.2

KO

8

4

8

12

1.5 WT

1.2

KO

0.9 0.6 0.3 0 CT

12

16

20

24/0

B DBP protein interaction profile (fold change)

E4BP4 protein interaction profile (fold change)

CREBH-DBP interaction

1.2 1.0 0.8 0.6 0.4 0.2 0.0

CT 12

16

20

24/0

4

8

CREBH-E4BP4 interaction

80 70 60 50 40 30 20 10 0

CT 12

12

16

20

24/0

4

8

12

C CREBH-DBP interaction DBP protein interaction (fold change)

PPAR protein interaction (fold change)

CREBH-PPAR interaction 200 GFP E4BP4 over-expression

160 120 80 40 0

CT

30

GFP E4BP4 over-expression

25 20 15 10 5 0

0

8

16

24

32

CT

0

8

16

24

32

Page 51 of 51

Diabetes

S-Table 1. Sequences of the primers used in this study   Group

Target region Rplp0 promoter Fgf21 promotor Acc1 promotor Apoc2 promoter

ChIP-qPCR

Fads2 promoter Bdh1 promoter Cpt1α promoter LXRα promoter E4BP4 promoter PPARα promoter Arbp Fgf21 Per2 RevERBα Bdh1 Acc1 E4bp4 Clock Bmal1

Gene expression qPCR

Pparα Lxrα ApoC2 ApoA4 Cpt1α Fads2 Elovl6 Elovl2 Dhcr24 Lcat

 

Primer name msRplp0NS2-F msRplp0NS2-R msFGF21-F msFGF21-R msAcc1-F msAcc1-R msApoc2-F msApoc2-R msFads2-F msFads2-R msBdh1-F msBdh1-R msCpt1a-F msCpt1a-R msLXRa-F msLXRa-R msE4BP4-F msE4BP4-R msPPARa-F msPPARa-R ms-Arbp-F ms-Arbp-R msFgf21-F msFgf21-R msPer2-F msPer2-R msRevERBa-F msRevERBa-R msBDH1-F msBDH1-R msAcc1-F msAcc1-R msE4bp4-F1 msE4bp4-R1 msClock-F1 msClock-R1 msBmal1-F1 msBmal1-R1 msPPARa1-F msPPARa1-R msLXRa-F msLxRa-R msApoC2-F msApoC2-R msApoA4-F msApoA4-R msCpt1a-F msCpt1a-R msFads2-F msFads2-R msElovl6-F msElovl6-R msElovl2-F msElovl2-R msDhcr24-F msDhcr24-R msLcat-F msLcat-R

5'- Sequence -3' CTTCTCCCTCCCTCACCCC CTTCTTGGCCCTCAGCAGTG CGC CCT GGC CAC GGT GGA AT CTC CGG TGC CCA GCA GGG AT ATTCATCAGCCCAGGGACTG CTTGTGAAGGCAGCAGCTGT CACACTGTTTAGGAAAGGAGGCA CTGCTGTACTCCACTCTTTCAC CCAGCAGGGCTTAACTCCAT AGGATCTTTCGAAGGCCAGC TGCTTGCCAGAGGGTCAAAT CGTGTTTGTCATCGAACGGG CAGAGAAGTTTACGGGCGGA TAAGTCCCGAGCTTGCCAAC CAAAGAGCCTCCAGGGTGAG CCCTGTCCCCTACCCTCTAC AATGGGCAAAAGGGTCCTGG CAGTCCGCGTCCTTCTCTG GCAGTCCCTTCACCTAACCC CTGGACGGCAGTGTCTGATT CCGATCTGCAGACACACACT ACCCTGAAGTGCTCGACATC GCTGCTGGAGGACGGTTACA CACAGGTCCCCAGGATGTTG TGTGCGATGATGATTCGTGA GGTGAAGGTACGTTTGGTTTGC CTACTGGCTCCCTCACCCAGGA GACACTCGGCTGCTGTCTTCCA AGATGCGGCTAGTGGCAAAG CAGTTCCTTGACCCCAGCAT CAGTAACCTGGTGAAGCTGGA GCCAGACATGCTGGATCTCAT GAGCAGAACCACGATAACCCA AGGACTTCAGCCTCTCATCC CAGGCACGTGAAAGAAAAGCA GCCGTCTTCTGTGTGACTGA GCAACTACAGTGGCCCTTTG TCCACAGGATTTGACTGGGG GGGAACTTAGAGGAGAGCCAAG CCATGTTGGATGGATGTGGC ACGCGACAGTTTTGGTAGAGG AACTCCGTTGCAGAATCAGGA CTCTGCTGGGCACGGTGC A GCCGCCGAGCTTTTGCTGTAC GCATCTAGCCCAGGAAACTG ATGTATGGGGTCAGCTGGAG AGAATCTCATTGGCCACCAG CAGGGTCTCACTCTCCTTGC GCTCTCAGATCACCGAGGAC AGTGCCGAAGTACGAGAGGA AGCACCCGAACTAGGTGACACGA TGAACCAACCACCCCCAGCGA CAACATGTTTGGACCACGAG GATGCCCCTGAGAGACAGAG GGCGAGACGCTACGCAAGCT TGGGCACAGCCAGATGGGGT GCTCTGTGGCCAGTGGCAGG AGGAGTGCGGTAGGCACCCA