Background
Body weight homeostasis is maintained through complex communications between the brain and peripheral organs [
1‐
8]. Declining body weight during fasting promotes increased food intake and decreased energy expenditure. By contrast, weight gain following several large meals is compensated by decreased food intake and increased energy expenditure. Many of the orexigenic and anorexigenic signals providing dynamic control of energy and body weight homeostasis are conveyed by G protein coupled receptors (GPCRs) in the brain and periphery [
5,
9,
10]. Here, we investigate a novel approach to understand how G protein signaling in liver regulates metabolic activity to maintain body weight and energy balance.
The activity cycle of heterotrimeric G proteins revolves around receptor-catalyzed guanine nucleotide exchange and GTP hydrolysis on the Gα subunit. In the inactive state, Gα
GDP forms a heterotrimeric complex with Gβγ. Hormone binding to GPCRs activates intracellular signaling by catalyzing guanine nucleotide exchange on the Gα subunit [
11]. Active Gα
GTP and Gβγ subunits dissociate to regulate effector proteins and the subsequent production of second messengers that provoke cellular responses to physiologic stimuli. GTP hydrolysis on Gα restores the inactive heterotrimeric complex of Gα
GDPβγ. Regulators of G protein signaling (RGS) proteins are GTPase activating proteins (GAPs) for Gi and Gq class α subunits [
12‐
15], and are distantly related to rgRGS proteins that accelerate GTP hydrolysis on G12 class α subunits [
16,
17]. RGS proteins regulate the specificity, intensity and duration of Gi and Gq signaling [
18]. RGS proteins of the R4 family, such as Rgs16, are feedback inhibitors that can terminate signaling by uncoupling hormone binding from effector protein activation [
19‐
21].
A useful characteristic of RGS gene expression is that it can be induced by GPCR agonists and second messengers [
22,
23]. A paradigm for feedback regulation of G protein signaling by RGS proteins was established by analysis of the yeast mating response [
24]. Mating pheromones are GPCR ligands that stimulate cell cycle arrest in haploid cells. The G protein alpha subunit (GPA-1) releases Gβγ to stimulate a MAP kinase cascade resulting in the transcriptional activation of the mating response pathway [
25‐
27]. Interestingly, transcription of the yeast RGS gene
Sst-2 is also induced by mating pheromone [
28]. SST-2 is the GAP for GPA-1 and it is the most important gene product for pheromone desensitization and re-entry into the cell cycle [
28,
29]. We applied this paradigm of GPCR-ligand induced RGS gene expression to identify and characterize G protein signaling in liver during fasting and refeeding because we had made several observations indicating that Gq and RGS proteins influenced and responded to changes in liver metabolism.
We found that knockout mice deficient in either Gαq or its close paralog, Gα11 [
30], exhibited abnormalities in liver regeneration following partial hepatectomy (Yu and Wilkie, unpublished observation). Given that RGS proteins are essential regulators of Ca
+2 signaling evoked by Gq/11-coupled agonists [
19,
21,
31], we reasoned that a RGS gene might be induced in response to activation of Gq/11 signaling during liver regeneration. We found Rgs16 was rapidly up regulated following partial hepatectomy, suggesting a role in the metabolic response to the sudden loss of 70% of the liver. Therefore, we screened the livers of fasted and refed mice for differential regulation of RGS genes. Interestingly, of the 20 cloned RGS genes, only Rgs16 mRNA was induced in liver during fasting. Our studies described herein demonstrate that Rgs16 is a diurnally regulated gene in periportal hepatocytes of the liver, its expression is regulated by feeding and dietary constituents, and Rgs16 can be used as a biomarker to investigate G-protein pathways in liver regulating energy homeostasis.
Discussion
Rgs16 is one of the few genes whose expression oscillates in a circadian pattern in both liver and SCN (Figs.
1,
2,
3,
4) [
35]. Rgs16 and other oscillatory genes in liver are synchronized by feeding time, whereas SCN neurons presumably respond to light stimulated neurotransmitter release at synaptic junctions with neurons from the retinal-hypothalamic tract. Many of the genes that oscillate both in liver and SCN are transcription factors integral to the circadian clock [
35,
36,
38,
39], whereas Rgs16 is a GAP for G alpha proteins of the Gi and Gq class [
31,
40,
41]. Rgs16 and closely related RGS proteins in the R4 family negatively regulate G protein signaling [
18], and can feedback inhibit Gi/Gq signaling by uncoupling hormone binding from effector protein activation in a variety of primary cell types [
19,
21,
31]. The expression pattern and enzymatic activity of Rgs16 identifies it as a candidate regulator of Gi/Gq signaling during the transitions from fasting to feeding in liver and between light and dark phases in SCN.
G protein-coupled ligands stimulate expression of RGS genes in eukaryotes from mammals to yeast [
18,
22‐
24]. For example, the yeast mating pheromones are GPCR-ligands that induce both cell cycle arrest and the transcription/translation of the yeast RGS protein
SST-2, which is required for rapid desensitization to mating pheromone and re-entry to the cell cycle [
28]. Because the early phase of liver regeneration is impaired in mutant mice deficient in Gq signaling (data not shown), we applied the yeast paradigm as our rationale for screening the expression of all RGS genes in liver of fed and fasted mice to identify Gi and/or Gq signaling pathways regulating liver metabolism.
We found that Rgs16 is the only RGS gene (of 20 total [
17]) that is diurnally expressed in liver of
ad libitum fed mice (Fig.
1). Mice are nocturnal, and when provided
ad libitum access to food and water, they eat nearly 80% of their daily food during the dark phase in a 12 hr light:dark (12hL:D) cycle. Rgs16 mRNA accumulates in liver in anticipation of a meal, either at the end of the light phase in mice maintained on a 12 hr L:D cycle (Figs.
1,
2), or prior to feeding at ZT5 on day 2 and subsequent days of restricted feeding (RF; Fig.
4). Furthermore, the amplitude of daily oscillations in Rgs16 mRNA and protein expression are modulated by energy deficiency, either increased in fasted mice whose body weight is significantly below set point (Fig.
4) or decreased in over weight mice maintained on a high fat diet (Fig.
6A). Importantly, the rate of Rgs16 gene transcription is induced by fasting and declines to basal levels shortly after feeding (Fig.
5). Coincident with these changes in transcription, Rgs16 mRNA and protein levels decline within 20 minutes after feeding begins, and drop to basal levels within 120 minutes (Fig.
6). Interestingly, if mice are not allowed to continue eating, but are restricted to the amount they can consume within 30 minutes of refeeding, Rgs16 mRNA expression returns within 120 min (data not shown). The tight localization of expression to Zone 1 periportal hepatocytes (Fig.
3) places Rgs16 at an ideal location to be regulated by orexigenic and satiety signals from the gut, the peripheral and/or central nervous system. The dynamic and localized expression of Rgs16 could be regulated by hormones and/or metabolites controlling liver metabolism.
The liver helps maintain whole body energy homeostasis in part by metabolism of glucose and fatty acids. Catabolic and anabolic metabolism in liver is regulated in response to daily repetitions of fasting and feeding, and the availability of energy stores in liver, adipose, and muscle. Glucagon and insulin are classical hormone regulators of the fasted and fed states 42, 43]. During fasting, glucagon simultaneously stimulates gluconeogenesis while inhibiting fatty acid synthesis in liver [
44]. Glucagon signaling in hepatocytes is transduced by Gαs, which stimulates adenylyl cyclase to produce the second messenger cAMP, thereby activating the cAMP responsive transcription factor CREB [
45,
46]. However, Rgs16 is not a Gs-GAP and does not directly regulate Gs activity [
18]; rather, Rgs16 may help integrate Gi- and/or Gq-signaling with glucagon or insulin signaling in liver.
Guided by the paradigm of the yeast mating pathway, we hypothesize that Rgs16 expression is induced in liver during fasting by activation of the Gi/Gq pathway(s) that Rgs16 protein negatively regulates. We propose that a preprandial agonist activates a hypothetical GPCR pathway that induces Rgs16 gene transcription and, possibly, promotes Rgs16 mRNA stability. Curiously, Rgs16 protein is not abundantly expressed in liver of fasted mice. We note that Rgs16 and other RGS proteins of the R4 subfamily typically are weakly expressed in many cell types and tissues despite abundant mRNA expression [
47]. One possible explanation for this pattern is Rgs16 mRNA accumulates during fasting and is poised for translation, dependent on other dietary factors.
Rgs16 is specifically expressed in periportal hepatocytes, the oxygen-rich zone of the liver where lipolysis and gluconeogenesis predominates, suggests Rgs16 may regulate Gi/Gq pathway(s) that stimulate fatty acid oxidation and/or glucose production in liver. Because refeeding rapidly terminates Rgs16 gene transcription and allows Rgs16 mRNA and protein degradation, Rgs16 might inhibit Gi/Gq signaling during the transition from the fasted to fed states. In this model, Rgs16 functions as a switch, turning off preprandial signaling in liver once feeding has commenced. Given that the hepatic expression of Rgs16 and circadian clock genes rapidly and coordinately adjust to changes in feeding schedules, Rgs16 might regulate the feeding cues that reset the circadian oscillator in liver.
Materials and methods
Materials
[32P]-dCTP was purchased from Amersham Bioscience (Piscataway, NJ). TRIZOL was purchased from Invitrogen (Carlsbad, California) and GeneScreen nylon membranes were obtained from New England Nuclear (Shelton, CT). All other lab supplies were purchased from Sigma (St Louis, MO) or Fisher (Hampton, NH).
Animal and colony conditions
Mice were maintained at 20°C under a standard 12-hour light:dark cycle (12h L:D, lights on at 5 am). C57BL/6 mice were purchased from Jackson Laboratories (Bar Harbor, Maine). Female C57BL/6 mice in feeding studies were pair-caged unless otherwise indicated; male mice were caged individually during restricted feeding studies. Mice were ad libitum fed standard mouse chow (Normal Chow) containing 6% total energy as fat (Teklad 7002, Harlan Teklad Laboratories, Indianapolis, IN), except where noted. The high-fat diet contains 41% calories as fat-40% carbohydrate-19% protein (Teklad #96001). Mice were six to eight weeks of age before being switched from a normal chow to high fat diet. Experimental research on animals followed internationally recognized guidelines and had UT Southwestern IACUC approval (APN#0602-05-01-2).
Real-time quantitative PCR
mRNA expression levels were determined using real time quantitative polymerase chain reaction. Cyclophilin was used as the normalizing gene. QPCR primers were designed using the Primer Express Software v 2.0 (Applied Biosystems, Foster City, CA) from published mRNA sequences. Rgs16 (NM_011267) QPCR primers: Forward – 5'cctggtacttgctactcgctttt3'; Reverse – 5'agcacgtcgtggagaggat3'; Cyclophilin (M60456) QPCR primers: Forward – 5'tggagagcaccaagacagaca3'; Reverse – 5'tgccggagtcgacaatgat3'. Total RNA was extracted from 50 mg liver as described below in Northern Blot Hybridization. Following DNase treatment, cDNA was synthesized from total RNA (2 μg) in 100 μL reactions using Superscript II reverse transcriptase kit (Invitrogen, Carlsbad, CA) and stored at -20°C until use. Polymerase chain reaction amplifications were performed in 96-well optical reaction plates (ABI) on the ABI Prism 7000 Sequence Detection system using the SYBR-Green (ABI) reaction conditions (1 cycle at 50°C for 2 min, 1 cycle at 95°C for 10 min, 40 cycles at 95°C for 15 sec and 60°C for 1 min), the baseline and threshold were set to experimentally determined values and the data were analyzed using the Comparative C
T method (
) as described [
32]. Relative Rgs16 mRNA fold induction was calculated compared to pooled RNA from fed mice (as in Fig.
2A, condition 2).
Northern blot hybridization
At time of collection, mice were sacrificed by cervical dislocation and individual livers were dissected, frozen in liquid nitrogen immediately, and stored at -80°C for future analysis. RNA extraction was performed using TRIZOL according to manufacture's protocol and liver RNA (20 μg per lane) was used for Northern analysis. Briefly, RNA samples were size separated by electrophoresis on a 1.0% agarose denaturing gel, transferred to a nylon membrane overnight, and probed with Rgs16 cDNA in 50% formamide. Radionucleotide hybridization probes were either random primed (Rgs16 cDNA; complete open reading frame) or end-labeled (18s rRNA oligonucleotide [GCCGTGCGTACTTAGACATGCATG]) as described [
48]. Northern blot hybridization filters were prehybridized at 42°C for six hours, hybridized overnight at 42°C, and then the membrane was washed twice in 2X SSC at 25°C, once in 2X SSC/2% SDS at 55°C, and once in 0.1X SSC at 25°C; filters were dried and exposed to Fuji RA film for at least 16 hours. Fold change (Δ) in Rgs16 mRNA levels are relative to basal expression at D12 (assayed by densitometry).
In situ hybridization
Mice were sacrificed by cervical dislocation and the livers were rapidly frozen on dry ice and stored at -80°C. Fresh frozen sections were cut in a cryostat and thaw-mounted onto SuperFrost Plus glass slides (Fisher Scientific, Pittsburgh, PA). Sections were pre-treated as described, including fixation, acetic anhydride and defatting steps [
49]. Slides with cover slip were incubated for 18 hrs at 60°C in a humidified chamber with buffer containing denatured salmon sperm DNA (0.033 mg/mL), yeast tRNA (0.15 mg/mL), dithiothreitol (40 μM), and a cRNA at 1 × 10
7 cpm/mL, other conditions as described [
49]. Riboprobes were generated using an in vitro transcription kit (Ambion; Austin, TX) by using T3 or T7 RNA polymerase in the presence of
32P-UTP, and purified using RNA quickspin columns. Sections were then treated with RNase A (20 mg/mL, 30 min at 45°C) and washed in descending concentrations of sodium citrate buffer to a stringency of 0.1X SSC at 60°C, and air-dried. Tissue sections were exposed to Biomax MR film (Kodak, Rochester, NY) followed by dipping in autoradiographic emulsion (NTB2, Kodak), exposed for an appropriate duration, developed, fixed, counterstained with cresyl violet acetate, a cover slip applied with DPX mounting media and visualized under bright and dark field microscopy (Olympus, Melville, NY). Adjacent sections from fresh-frozen livers were used as controls for the effects of perfusion on mRNA integrity and hybridization to sense probes.
Western blots
Livers were rapidly excised and immediately frozen in liquid nitrogen. Liver (50 mg) was resuspended in 1% SDS that contained a cocktail of protease inhibitors, including leupeptin (10 μg/mL), soybean trypsin inhibitor (10 μg/mL), MG132 (16 μg/mL), phenylmethylsulfonyl fluoride (16 μg/mL), tosyl lysine choromethyl ketone (16 μg/mL), and tosyl phenylalanine choromethyl ketone (16 μg/mL). The samples were sonicated (Virtis, Gardiner, NY) and immediately boiled for three minutes. A Lowry protein assay was employed to quantitate the protein and ensure equal loading. The samples were separated by SDS-PAGE and transferred to nitrocellulose paper. Proteins were detected using Rgs16 antiserum (a gift from Carol Beadling, Cornell University). Rgs16 antiserum does not cross react with recombinant mouse Rgs4 or Rgs8 protein. Enhanced Chemiluminescence (Amersham-Pharmacia) was utilized to detect the rabbit HA-tagged secondary antibody. The membrane was exposed to ML film (Kodak Biomax) for 2–15 minutes. Autoradiographs were scanned with a BioRad Fluor-S MultiImager to determine signal intensities.
Transcription run-on
Nuclei from mouse livers were isolated as described [
50], resuspend in glycerol storage buffer (50 mM Tris-HCl pH 8.3, 5 mM MgCl
2, 0.1 mM EDTA pH 8.0 and 40% glycerol) and flash-frozen in liquid nitrogen. Nuclei were stored at -80°C and used within 4 weeks. Mouse Rgs16, mouse Rgs8, rat GAPDH cDNA or empty plasmid vectors (5 μg) were slot-blotted onto nylon GeneScreen membrane (NEF 983, NEN life Science Products, Inc) according the manufacturer's protocol. The nuclear RNA elongation reaction, isolation of newly synthesized
32P-RNA, and hybridization to cDNA plasmids were performed as described 51]. The amount of hybridizing
32P-RNA was quantitated by densitometric scanning (using a Fujifilm FLA-5100 image reader) of phosphorimager screens exposed for 24 h. Data from different fasting and refeeding conditions were normalized to transcription of the GAPDH gene.
Statistics
Experiments were conducted in groups of two or three mice per condition and repeated at least twice. Quantitative data for each condition or time point are represented as mean values ± SEM. GraphPad Prism software (GraphPad, San Diego, CA) was used to perform all statistical analyses. The two-tailed, unpaired Student's t-test was used to determine statistically significant differences (P < 0.05) between mean values.
Acknowledgements
We thank John Shelton, Jim Richardson, Liping Zhao, and Brian Potts for in situ hybridization, Bob Hammer for providing the PEPCK in situ probe, Carol Beadling for Rgs16 antiserum, Susanne Mumby for advice on protein preparation from tissues for Westerns, Joyce Repa for generously sharing expertise and machine time for Real-Time quantitative PCR, David Sierra for technical assistance, and the many colleagues mentioned above or in Discovery Biology for discussions and comments on the manuscript. This work was supported by NIH (GM61395; PAR-98-057) and Welch Foundation (I-1382) grants to TMW and a NIH postdoctoral fellowship (HL072551) to DMK.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
JH contributed to the QPCR analysis and did Western blots of fasting and refeeding experiments, and final formatting of all figures. VP carried out nuclear run-on and restricted feeding experiments. DMK did restricted feeding experiments. KY did the Northern blots and co-ordinated in situ hybridization. SJG did SCN in situ hybridization. TMW conceived of the study, participated in its design, execution, and coordination. All authors contributed figures, helped draft portions of the manuscript, and read and approved the final manuscript.