Targeted deletion of one or two copies of the G protein β subunit Gβ5 gene has distinct effects on body weight and behavior in mice

Abstract

We investigated the physiological role of Gβ5, a unique G protein β subunit that dimerizes with regulators of G protein signaling (RGS) proteins of the R7 family instead of Gγ. Gβ5 is essential for stability of these complexes, so that its knockout (KO)causes degradation of the entire Gβ5-R7 family. We report that the Gβ5-KO mice remain leaner than the wild type (WT) throughout their lifetime and are resistant to a high-fat diet. They have a 5-fold increase in locomotor activity, increased thermogenesis, and lower serum insulin, all of which correlate with a higher level of secreted epinephrine. Heterozygous (HET) mice are 2-fold more active than WT mice. Surprisingly, with respect to body weight, the HET mice display a phenotype opposite to that of the KO mice: by the age of 6 mo, they are ≥15% heavier than the WT and have increased adiposity, insulin resistance, and liver steatosis. These changes occur in HET mice fed a normal diet and without apparent hyperphagia, mimicking basic characteristics of human metabolic syndrome. We conclude that even a partial reduction in Gβ5-R7 level can perturb normal animal metabolism and behavior. Our data on Gβ5 haploinsufficient mice may explain earlier observations of genetic linkage between R7 family mutations and obesity in humans.—Wang, Q., Levay, K., Chanturiya, T., Dvoriantchikova, G., Anderson, K. L., Bianco, S. D. C., Ueta, C. B., Molano, R. D., Pileggi, A., Gurevich, E. V., Gavrilova, O., and Slepak, V. Z. Targeted deletion of one or two copies of the G protein β subunit Gβ5 gene has distinct effects on body weight and behavior in mice.

G protein-coupled receptors (GPCRs) regulate numerous body functions, including feeding and physical activity behaviors and metabolism. Many hormones, such as insulin, do not bind to GPCRs, but their secretion is ultimately regulated by GPCRs (1, 2). In the canonical GPCR pathway, receptors activate heterotrimeric G proteins, which, in turn, regulate effector enzymes and ion channels. This cascade returns to its resting state when GTP bound by the G protein is hydrolyzed. This crucial signal termination step involves regulators of G protein signaling (RGS; ref. 3), a family of GTPase-activating proteins (GAPs) that accelerate GTP hydrolysis by as much as several hundred fold (see reviews; refs. 4, 5).

There are >30 members of the RGS homology superfamily, defined by the presence of a 120-aa RGS domain. In most of these proteins, this domain is shown to be responsible for the interaction with Gα subunits and the GAP activity (6). Many of the RGS proteins have additional structural motifs and a variety of other functions (see reviews; refs. 7–9). For example, the R7 family of RGS proteins (RGS6, RGS7, RGS9, and RGS11) contains three such domains: DEP (also found in disheveled, egl10, pleckstrin), DHEX (DEP helical extension), and GGL (Gγ-like), which binds to the G protein subunit Gβ5 (10–12). The Gβ5 and R7 family RGS subunits stabilize each other against proteolysis, as demonstrated by the coexpression of Gβ5 and R7 cDNAs (13) and gene knockouts (KOs) in mice (14) and Caenorhabditis elegans (15). Although complexes of Gβ5 with Gγ subunits and monomeric R7 protein have been expressed and studied in vitro, all currently available evidence indicates that in native tissues they exist only as Gβ5-R7 heterodimers (see reviews; refs. 16, 17). These heterodimers can be found in cytoplasm or associate with membranes via anchoring proteins R9AP and R7BP (see reviews; refs. 17, 18).

Gβ5-R7 complexes were shown to possess GAP activity toward the Gi but not Gq family of proteins, and accordingly, they negatively regulate Gi/o-mediated signal transduction in reconstituted systems (19–21). Studies in C. elegans show that Gβ5-R7 complexes can also attenuate Gq-mediated signaling, possibly via non-GAP mechanisms (22). In transfected mammalian cells, the Gβ5-RGS7 complex can selectively inhibit Gq signaling by means of direct interaction with muscarinic M3 receptor, rather than GAP activity (23). Like some other RGS proteins (24–26), the R7 family members can associate with GPCRs (27, 28).

Members of the Gβ5-R7 family are considered to be neuronal proteins because their expression at the protein level was reliably detected only in the brain, spinal cord and retina. Studies of RGS9-KO mice implicated the R7 family in regulation of several CNS functions, including sensory signaling, addiction, and drug-induced hyperactivity (29, 30). Recently, the Gβ5-RGS6 protein complex was detected in cardiac tissues, and the KO of the RGS6 gene highlighted the role of this protein in regulation of heart contractility (31, 32). In this work, we investigated mice with the targeted deletion of Gβ5 gene, a model lacking the entire family of R7 proteins, and discovered novel phenotypes in both the homo- and heterozygote KOs.

MATERIALS AND METHODS

Animals

All of the procedures with the mice used in this study were performed according to the Guidelines for the Care and Use of Laboratory Animals of the National Institutes of Health and protocols approved by the University of Miami Committee on Use and Care of Animals. Gβ5-KO mice were provided by Dr. Ching-Kang Chen (Virginia Commonwealth University, Richmond, VA, USA); the method used for generation of these KO mice has been previously described by Chen et al. (14). Mice were subsequently backcrossed for ≥5 generations from the original hybrid 129SvJ/C57BL/6J background onto the C57BL/6J strain. They were housed under controlled temperature conditions and a 12-h light-dark cycle with free access to a normal chow (13.5% calories from fat; Lab Diet; PMI Nutrition International, St. Louis, MO, USA) and water. If required by the experiment, mice were switched to a high-fat diet (HFD; 60% kcal from fat; D12492; Research Diets, New Brunswick, NJ, USA), as described in the corresponding figure legends. Genotyping was performed by PCR using a standard procedure described by Chen et al. (14). The wild-type (WT), heterozygous (HET), and KO animals were generated by breeding the heterozygotes. Cohorts consisting of age-matched males were used for all experiments.

Body weight, food intake, and body composition

After weaning at 4 wk of age, mice were group housed with littermates of the same sex and given a weighed amount of regular chow. Animals and the amount of food remaining in their cages were weighed weekly. Alternatively, food intake was measured in 31-wk-old mice housed individually in regular home cages with Carefresh paper bedding (International Absorbents, Inc., Ferndale, WA, USA). Mice were allowed to adapt to individual caging for a week prior to data recoding. The amount of regular chow diet (4.08 kcal/g) in the feeding rack was measured weekly for 6 wk. Food spillage measurement was performed on wk 6 by sifting the bedding, separating feces, and collecting and weighing shredded food from the bottom of the cages. At the age of 38 wk, the same cohort of mice was switched to an HFD (D12451, 45 kcal%, 4.73 kcal/g; Research Diets), and food intake was measured for an additional 4 wk. In a separate experiment, food consumption and feces production were determined within 24 h using individual mice kept in Techniplast metabolic cages (LabPlanet, Northbrook, IL, USA). Body composition was measured using double-energy X-ray assay (DEXA).

Indirect calorimetry

Oxygen consumption and CO₂ production were measured at 24°C using an 8-chamber Oxymax system (Columbus Instruments, Columbus, OH, USA) as reported previously (33). Mice were allowed to adapt to testing chambers for 24 h, and data were recorded for the following 24 h.

Serum analysis

Blood was obtained from the tail vein in non-food-deprived mice at 9 AM. Blood or serum glucose levels were measured using an Elite glucometer (Bayer, Elkhart, IN, USA). Serum insulin, leptin, and adiponectin were assayed using radioimmunoassay (Linco Research, St. Charles, MO, USA). Serum triglycerides (Thermo DMA, Louisville, CO, USA) and free fatty acid (FFA; Roche Applied Science, Indianapolis, IN, USA) were measured using colorimetric assays. Triiodothyronin (T3), thyroxine (T4), and TSH were measured using the thyroid hormone panel kit (Millipore, Billerica, MA, USA) in blood collected by cardiac puncture from isoflurane-anesthetized mice.

Glucose and insulin tolerance tests

Following overnight food deprivation (16 h), glucose tolerance test was performed by intraperitoneal injection of glucose (2 mg/g body weight). Insulin tolerance was performed in non-food-deprived state. Insulin (Humulin R, 0.75 U/kg; Lilly, Indianapolis, IN, USA) was injected intraperitoneally. Blood samples were taken from the tail vein before and at the indicated time points postinjection, and glucose was measured using a portable Elite glucose meter (Bayer).

Body temperature

The core body temperature was measured with a probe RET-3-ISO (Physitemp Instruments, Clifton, NJ, USA) connected to a high-precision TH-5 thermalert monitoring thermometer (Physitemp Instruments). The probe was inserted in the rectum 1 cm from the anus, and the measurement was recorded within ∼5 s, after the signal stabilized.

UCP1 gene expression assay

Total RNA was extracted from brown adipose tissue using TRIzol (Life Technologies, Carlsbad, CA, USA), according to the manufacturer's instructions. The extracted RNA was quantified with a NanoDrop spectrophotometer (Thermo Fisher Scientific, Wilmington, DE, USA), and 1.0 μg of total RNA was reverse transcribed by using high-capacity cDNA reverse transcription kit (Applied Biosystems, Foster City, CA, USA). UCP1 mRNA levels were measured on the Bio-Rad iCycler iQ real-time PCR detection system using the iQ SYBR Green Supermix (Bio-Rad, Richmond, CA, USA) under the following conditions: 15 min at 94°C (hot start), 45 s at 94°C, 45 s at 55°C, and 45 s at 72°C for 40 cycles. Standard curves consisted of 4 or 5 points of serially diluted mixed experimental and control group cDNA. Cyclophilin A was used as a housekeeping internal control gene. The coefficient of correlation (r²) was >0.98 for all curves, and the amplification efficiency ranged between 80 and 110%. Results are expressed as the ratios of test mRNA/cyclophilin A mRNA.

Analysis of pancreatic islets and total insulin content

Islets were isolated by the means of a combined enzymatic (collagenase type V; Sigma, St. Louis, MO, USA) and mechanical dissociation followed by purification on Euro-Ficoll gradients (Mediatech, Herndon, VA, USA), as described previously (34, 35). After purification, islets were counted and scored for size using microscopic analysis. For total insulin content measurement, animals were euthanized after overnight food deprivation, and pancreata were dissected and frozen at −80°C. Pancreata were homogenized in 100% ethanol and extracted by 2 cycles of incubation with 80% phosphoric acid. The extract was kept at −80°C until the insulin content was measured by enzyme-linked immunosorbent assay (ELISA; Mercodia, Winston Salem, NC, USA), as described previously (36).

Histology

Animals were euthanized by CO₂ inhalation followed by cervical dislocation, and tissues were collected. The epididymal, inguinal, and intrascapular brown adipose depots were dissected and weighed. After fixing with paraformaldehyde and embedding in optimal cutting temperature (OCT) compound, the tissue was sectioned and stained with hematoxylin and eosin (H&E) using a standard protocol. Livers were snap-frozen on dry ice, embedded in OCT compound, prefrozen in anhydrous ethyl alcohol, and stained with H&E and Oil Red.

Locomotor behavior

The locomotor activity was measured in transparent rodent cages (42×24×20 cm) using an activity meter (Opto-Varimex-Mini Model B; Columbus Instruments). Both the total counts and the ambulatory counts were recorded to a computer. The Opto-Varimex-Mini Model B separates counts (beam interruptions) of total activity from counts that correspond to ambulatory (horizontal) movements, which is accomplished by blocking additional counts from a particular beam from being scored until a different beam is broken. Counts were registered every 5 min for a total of 60 min, and results are presented as mean ± se cumulative horizontal counts.

Catecholamines

Mice were housed in metabolic cages (LabPlanet) with food and water provided ad libitum. Urine samples were collected for 24 h, supplemented with HCl to a final concentration of 45 mM, and stored at −80°C. Catecholamine concentration was measured at the Mouse Metabolic Phenotyping Center (Vanderbilt University) using HPLC.

Antibodies and Western blot analysis

Antibodies against RGS7, Gβ5, Gβ1, and Gαo were raised in rabbit against synthetic peptides and were described earlier (13). Antibodies against actin, tyrosine hydroxylase, Tau1, and β3 tubulin and secondary antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA) or Jackson Immunologicals (West Grove, PA, USA). Western blot analyses were performed using a standard protocol with enhanced chemiluminescence (ECL) detection, as described earlier (13, 23).

Statistics

Data were analyzed by ANOVA using Prizm (GraphPad, San Diego, CA, USA) or Excel (Microsoft, Redmond, WA, USA) software packages, with values of P < 0.05 considered statistically significant.

RESULTS

Body weight of the Gβ5-KO and HET mice

In the course of this study, we investigated 9 cohorts, each consisting of 5–10 age-matched males of 3 genotypes: Gβ5-KO (Gβ5^−/−), HET (Gβ5^−/+), and WT (Gβ5^+/+). Western blot confirmed the absence of Gβ5 and RGS7 in the brain of the KO mice (Fig. 1A). In the HET mice, the expression of both Gβ5 and RGS7 was reduced by ∼50% (47.8±2.4, n=4, and 46.9±4.4, n=3, respectively). The levels of Gαo, Gαq, and Gβ1 were unaffected either in HET or KO mice.

Figure 1.

Body weight of Gβ5-KO and HET mice. A) Brain lysates from age-matched WT, KO, HET mice were analyzed by Western blot using antibodies against Gβ5, RGS7, Gβ1, Gαq, Gαo, and actin, as described in Materials and Methods. B) Body weight of WT, HET, and KO mice was monitored as described in Materials and Methods. Values are from 6 independent cohorts at the age of 6 mo. Mice per genotype: WT, n = 35; HET, n = 61; KO, n = 38. C) After weaning at the age of 4 wk, a representative cohort of WT, HET, and KO mice was fed a regular diet, and body weight was determined weekly. WT, n = 7; HET, n = 9; KO, n = 6. D) Mice were fed a regular diet as in C, but then were switched to the HFD (60% calories from fat) at the age of 12 wk, as indicated by the arrow. WT, n = 10; HET, n = 10; KO, n = 6. E) Ratio of organ weight to body weight was determined using 3 cohorts of 6- to 9-mo-old males. WT, n = 17; HET, n = 28; KO, n = 13. All the animals were fed regular diet. Br, brain; H, heart; Lv, liver; K, kidney; T, testis; Sp, spleen; Pan, pancreas; QM, quad muscle; Ing, inguinal fat pad; Epi, epididymal fat pad; B, brown fat pad. Values are means ± sd in B and E; means ± se in C and D. *P < 0.05; **P < 0.001.

In agreement with the original report (14), the KO mice showed a dramatically reduced body size at birth. By 4 wk of age, their body weight was only ∼60% of the WT. To determine whether the KO animals could reach normal weight at an older age, we monitored them for several months (Fig. 1B, C and Supplemental Fig. S1A). We found that the KO mice gained up to 80% of the normal weight within 2 mo, but remained significantly smaller than the WT through their lifetime.

Monitoring the HET animals led us to a very surprising finding: instead of the expected partial reduction or no change in body weight, they became heavier than the WT. Figure 1B shows that at 6 mo of age, the HETs were >16% heavier than the WT and >35% heavier than the KO mice. At birth and until ∼3 mo of age, the HET mice were indistinguishable from the WT, but the difference became significant (P<0.001, n=24 for WT; n=36 for HET) at about 5 mo and continued to increase with age (Fig. 1C and Supplemental Fig. S1).

Visual inspection suggested that the HET mice had a higher adiposity. Indeed, gravimetry of the dissected fat pads (Fig. 1E) confirmed the increase in their relative mass, while DEXA showed an ∼60% increase in percentage fat in the HET mice (Supplemental Fig. S1B). The percentage of fat in the KO mice was less than in WT, and accordingly, the weight of fat pads was markedly reduced. Histological analysis of fat tissue showed that adipocytes were larger in the HET mice and smaller in the KO compared to the WT (Fig. 2A). The livers in the HET mice were enlarged (Fig. 1E) and steatotic (Fig. 2B). The significant increase in relative mass of the liver in the KO mice that is evident in Fig. 1E results from their smaller total body mass, while the increase in the HET mice occurs despite the larger body mass. The sizes of brain, heart, and kidney were the same in the 3 genotypes; the apparent changes in the percentage mass of these organs are due to the changes of total body weight.

Figure 2.

Histological analysis of fat pads and liver in Gβ5-KO and HET mice. One group of mice was challenged with 10 wk of HFD at the age of 12 wk; control group was fed regular diet for 22 wk total. A) Mice were sacrificed, and their fat pads were stained by H&E. B) Livers were stained by Oil Red, as described in Materials and Methods. Representative images (×10 lens) from 3 animals of each genotype. Scale bars = 50 μm.

Effects of HFD treatment

We compared the effects of an HFD on the WT, HET, and KO mice (Fig. 1D). Mice were subjected to HFD at 12 wk of age, when the body weight of HET and WT mice was still similar and the KO mice were ∼20% lighter than the WT. We made two observations. First, the KO mice were totally resistant to HFD and stopped gaining weight on this diet. Second, while both WT and HET mice gained weight on HFD, the rate of this increase was ∼2-fold higher for the HET mice than the WT. Consistent with its influence on the body weight, HFD increased the size of adipocytes and content of fat in the livers of WT and HET animals, but it had no such effect in the KO mice, which had no signs of liver steatosis on any diet (Fig. 2B).

Glucose and other metabolites

Because changes in adiposity are often associated with shifts in glucose and lipid metabolism, we surveyed serum levels of glucose, insulin, leptin, adiponectin, triglycerides (TGs), and FFAs. Tests were performed on mice that were maintained on regular diet at 6 mo, the age when the HET mice become obese. As shown in Table 1, there was a slight but statistically significant increase in leptin level in the HET mice compared to WT or KO. The concentrations of FFA and TG were slightly increased in the KO mice, which also had a significant increase in glucose and a decrease in insulin. Most notably, the HET mice had a 2.7-fold increase in insulin levels, suggesting insulin resistance.

Table 1.

Serum hormone and metabolite levels in Gβ5-KO and HET mice

Genotype	Body weight (g)	FFA (mM)	TG (mg/dl)	Adiponection (μg/ml)	Serum glucose (mg/dl)	Insulin (ng/ml)	Leptin (ng/ml)	T3 (ng/ml)
WT	33.03 ± 1.11	0.45 ± 0.04	100.13 ± 8.95	4.85 ± 0.44	149.54 ± 19.6	0.59 ± 0.05	8.75 ± 2.44	2.21 ± 1.11
HET	36.27 ± 1.34*	0.37 ± 0.07	92.68 ± 9.81	5.52 ± 1	159.94 ± 18	1.52 ± 0.32**	10.84 ± 2*	2.60 ± 0.93
KO	26.20 ± 1.47**	0.58 ± 0.08	135.09 ± 2.87	6.24 ± 1	195.44 ± 20*	0.45 ± 0.10	7.60 ± 2	5.02 ± 1.06*

Values are expressed as means ± se. Serum was collected from a cohort of age-matched males (WT, n=10; HET, n=10; KO, n=7) at the age of 6 mo. Mice were fed the regular diet for their entire lifetime and kept in individual cages. The indicated metabolites were measured using standard techniques, as described in Materials and Methods and Supplemental Data.

^*P < 0.05;

^**P < 0.001.

To test this directly, we performed insulin and glucose tolerance tests (Fig. 3). In response to acute insulin bolus, the WT and KO mice cleared glucose with a comparable rate, indicating similar insulin sensitivity; however, the HET mice showed a significantly blunted response to insulin (Fig. 3A). The glucose tolerance test showed that HET mice also had a slightly slower glucose clearance rate, compared to the WT controls (Fig. 3B). Unexpectedly for a lean and insulin-sensitive animal, the KO mice had significantly impaired glucose clearance. This impaired glucose tolerance was likely caused by the reduced insulin levels in the KO mice (Fig. 3C and Table 1).

Figure 3.

Glucose and insulin levels in Gβ5-KO and HET mice. A) Insulin tolerance test was performed as described in Materials and Methods. Mice (WT, n=10; HET, n=10; KO, n=5) were injected with insulin, and tail vein blood glucose concentration was measured at the indicated times postinjection. B) Same cohort shown in A was subjected to glucose tolerance test (GTT), where tail vein blood glucose concentration was measured after intraperitoneal glucose injection. C) Concentration of insulin during the GTT test in B was determined in tail vein plasma fractions collected at the indicated time. D) Top panels: Western blot analysis of human and rat islets and mouse insulinoma cell line Min6 was probed with antibodies against Gβ5, β3 tubulin, Tau1, and actin. Mouse brain homogenate (Br) was used as a positive control. Bottom panels: mouse islets isolated from WT and KO male mice (n=3/genotype, sacrificed at 2.5 mo of age; normal diet). Lysate obtained from ∼50 islets was loaded per well; actin antibody was used for loading control. E) Mouse islets were isolated, and their size distribution was measured and scored by microscopy, as described in Materials and Methods. Data from two experiments (WT, n=10; HET, n=10; KO, n=3) on mice that were fed regular chow until sacrifice at age of 3 mo. F) Insulin was extracted from dissected whole pancreata and measured by ELISA as described in Supplemental Data. Data show mean ± se values determined for 3-mo-old mice (n=3/). Values are means ± se. *P < 0.05; **P < 0.001.

To investigate whether the glucose intolerance in the KO mice could represent a primary defect in the pancreatic function, we tested whether Gβ5 was expressed in the pancreas of WT mice. Consistent with original studies showing the absence of Gβ5 in splanchnic organs (37), we could not detect Gβ5 in pancreatic homogenate (not shown). However, in isolated islets from humans, rats, and mice, anti-Gβ5 antibodies revealed a prominent band corresponding to Gβ5 (Fig. 3D), evidently due to enrichment compared to the whole tissue. Neuronal markers Tau1 and β3 tubulin were not detected in our islet preparations, and Gβ5 was also present in the Min6 murine insulinoma cell line, which rules out contamination with neurons. The Gβ5-immunoreactive protein band was not detected in islets obtained from the KO mice, supporting specificity of staining. Although the expression level of Gβ5 in the islets was ≥5-fold lower than in the brain, the presence of appreciable amount of Gβ5 protein in pancreatic islets indicates that it can be involved in regulation of their function. We also examined the morphology of pancreata from the WT, HET, and KO mice using light and immunofluorescence microscopy, but did not detect a notable difference between the 3 groups (not shown). The number of isolated islets, their size distribution, and the amount of total insulin extracted from pancreata were also similar among the 3 genotypes (Fig. 3E, F). These results indicate that Gβ5 KO does not cause a gross change in the structure of the pancreas or insulin production, but it is likely to impair signaling within the islets and/or islet regulation.

Food intake, energy balance, and activity

To investigate the reason for the changes in body weight in HET and KO mice, we measured food intake using 3 types of experiments. In one, animals were housed individually in home cages, fed a normal chow diet (Fig. 4A) or an HFD (Supplemental Fig. S2A), and food consumption was measured for several weeks (Fig. 4B). In another, we studied individual mice in metabolic cages for 48 h (Supplemental Fig. S2B). Finally, we studied regular chow and HFD consumption by mice housed in groups of 3–5/cage (Supplemental Fig. S2C). All of these assays showed that the KO mice consumed ∼20% more food per gram of body weight than their WT or HET counterparts on both a normal chow diet and an HFD. The amount of feces was consistent with food intake. Interestingly, the KO mice spilled about 4 times more food than WT or HET mice (Fig. 4B and Supplemental Fig. S2B), suggesting increased activity levels in KO mice.

Figure 4.

Energy balance, activity, and catecholamine levels in Gβ5-KO and HET mice. A) Food intake was measured starting from 31 wk of age in individually caged mice fed normal chow (4.08 kcal/g) and housed in regular home cages with Carefresh bedding. Mice were allowed to adapt to individual caging for a week prior to data recoding. Amount of food in the feeding rack was measured weekly for 5 wk and normalized to body weight. WT, n = 10; HET, n = 10; KO, n = 5. B) Food spillage measurement was performed at wk 6 by sifting the bedding, separating feces, and weighing shredded food from the bottom of the cages. C) Total Vo₂ was measured at 24°C using the Oxymax system, as described in Materials and Methods. WT, n = 9; HET, n = 10; KO, n = 5. D) Locomotor activity was monitored in 3-mo-old mice in an open-field test as described in Materials and Methods; y axis represents cumulative beam interruptions (n=10/ genotype). E) Same cohort of animals was reanalyzed at the age of 9 mo. F) Analysis of urinary catecholamines. Mice were kept in metabolic cages for 48 h with food and water provided at libitum. Urine samples were collected and analyzed by HPLC for norepinephrine (NE), epinephrine (Epi), and dopamine (DA), as described in Materials and Methods. WT, n = 12; HET, n = 16; KO, n = 11). G) Adrenal glands were dissected from the WT, HET, and KO mice to obtain total (Ad) and adrenal medulla (M) homogenates, which were analyzed by Western blot using anti-Gβ5 antibody. H) Rat adrenal glands were dissected to obtain medulla (M) and cortex (C), and analyzed by immunoblot using anti-Gβ5 and anti-tyrosine hydroxylase (TH) antibodies. Values are means ± se. *P < 0.05; **P < 0.001.

Since food intake in the KO mice was not reduced, the reduction in body weight might be the result of increased energy expenditure. Indeed, total and resting oxygen consumption were significantly higher in the KO mice (Fig. 4C and Supplemental Fig. S2D), whereas no difference between the HET and WT mice was detected. Analysis of core body temperature (Fig. 5A) showed that it was similar in all 3 genotypes when measured at ambient room temperature. However, the difference was revealed when the mice were challenged by cold. After 2 h of exposure to 5°C, core temperature in WT and HET mice was reduced by 1.6–1.7°C. This drop was <1°C in the KO mice, indicating an increased thermogenic response. In agreement with increased thermogenesis, the KO mice had smaller brown adipocytes (Fig. 5B), increased UCP1 mRNA (Fig. 5C) and elevated serum triiodothyronin (Table 1).

Figure 5.

Thermogenesis and brown fat in the Gβ5-KO mice. A) Core body temperature in the WT, HET, and KO mice was measured as described in Materials and Methods. After 1 h at room temperature (1 animal/cage), mice were transferred to a cold room (5°C), and intrarectal measurements were taken at indicated times. Values are for 1 cohort at the age of 3 mo (WT, n=5; HET, n=10; KO, n=5). B) Brown fat was dissected from mice fed a regular diet at the age of 3 mo (WT, n=5; HET, n=10; and KO, n=5), fixed in paraformaldehyde, embedded in OCT, sectioned, and stained with H&E. Scale bars = 50 μm. C) UCP1 mRNA levels in brown adipose tissue dissected from WT, HET, and KO mice. Mice were fed regular diet and group housed at ambient room temperature until sacrifice at age of 3 mo. Total RNA extraction and RT-qPCR analysis were performed as described in Materials and Methods. Data show ratio of UCP1 mRNA to cyclophilin A determined for each animal (WT, n=5; HET, n=4; and KO, n=5). Values are means ± se. *P < 0.05; **P < 0.001.

To test whether the reduction of body weight in the KO mice could be also associated with increased physical activity, we analyzed their locomotor behavior in the open-field test and found a striking difference between the 3 genotypes (Fig. 4D, E). At the age of 12 wk, the KO mice moved about 5 times more than the WT. The HET mice were also more active than the WT, but only by 2-fold. Interestingly, both the WT and HET mice slowed down after ∼20 min of recording, but the KO mice kept moving at the same rate for the entire hour of the recording (Fig. 4D). When we reanalyzed the same cohort of mice 6 mo later, the activity of the WT mice was reduced by ∼50%, which is an expected effect of aging; the HET mice had a similar change. In sharp contrast, the KO mice maintained their remarkable hyperactivity (Fig. 4E).

Catecholamine levels

One potential mechanism explaining the increased metabolism and thermogenic response, hyperactivity, and reduced insulin level of Gβ5-KO mice could be an up-regulation of sympathetic tone. To test this hypothesis, we measured the levels of urine catecholamines (Fig. 4F). The results showed a 2-fold higher level of epinephrine in the KO mice. The presence of Gβ5 in the adrenal medulla (Fig. 4G, H) suggests that the increase in epinephrine could be due to a primary effect on signaling and secretion in the chromaffin cells. However, increased urine catecholamines may also result from alterations in other parts of the nervous system.

DISCUSSION

G protein-mediated signal transduction is a universal molecular mechanism involved in the regulation of numerous physiological functions. In this study, we investigated the effects of the KO of the gene encoding G protein subunit Gβ5 in mice and discovered novel phenotypes pertaining to control of body weight and behavior.

Perhaps the most surprising finding in our study is that the deletion of a single copy of the Gβ5 gene causes increase in body weight, while the loss of both alleles has the opposite effect. How could this be explained? We propose that the full KO causes many disruptions in the neuroendocrine system, resulting in multiple negative effects, including growth retardation. Indeed, Gβ5 KO was originally thought to be lethal, since most newborns died before weaning, and the survivors were runty (14). These mice are also characterized by poor breeding (according to our data, average 3.3 pups/litter and ≤2 pregnancies/female). The KOs also consistently displayed some neurological abnormalities, such as excessive grooming and impaired placing reflex (failure to stretch out their front limbs when lifted by the tail). Therefore, we think that the low body weight of the KO mice is not a specific effect, but is rather one of many manifestations of the overall poor health of these mice. In contrast, the 50% reduction in the Gβ5 level in the HET mice does not drastically impair their condition. For example, reproductive functions of the HET mice are normal, and we did not notice any obvious neurological defects in the HET mice. Therefore, we conclude that haploid insufficiency of the Gβ5 gene in mice causes a specific phenotype: late-onset obesity. Such characteristics of this mouse model as insulin resistance and liver steatosis developed on normal diet strongly resemble human metabolic syndrome.

Hundreds of genes have been associated with body weight abnormalities. However, monogenic mutations causing obesity are relatively rare, and there are only ∼10 genes that cause obesity on the loss of one normal allele. These genes include leptin, its receptor, POMC, and other genes involved in the production and signaling of melanocortin hormones (see reviews; refs. 38, 39). Thus, our study identifies Gβ5 as one of the rather small number of genes that can cause obesity in a heterozygous state.

The second new phenotype discovered in our study is the increase in locomotor activity (Fig. 4). With respect to this phenotype, the HET mice displayed the expected “intermediate” characteristics: they move more than WT and less than KO mice. It is worth noting that the HET mice gain the extra weight despite the increased locomotor activity at the young age; the animals slow down as they become heavier. This supports our conclusion that the late-onset obesity in Gβ5 heterozygotes represents a specific phenotype.

Earlier studies implicated another member of the R7 family, RGS9, in locomotor behavior: it was shown that RGS9 KO potentiated responses to cocaine (29). However, RGS9 KO had no effect on noninduced locomotor activity. Thus, it appears that the effect of elimination of the entire R7 family in the Gβ5 KO is distinct from the selective ablation of RGS9.

Control of behavior and metabolism involves direct communication between neurons and input from circulating hormones, both of which involve multiple GPCRs. The absence of the R7 family RGSs could lead to an increased activity of multiple G proteins and/or compensatory down-regulation of signaling. Therefore, understanding molecular mechanisms leading to the Gβ5-KO and HET phenotypes will require a systematic analysis of global changes of the neuroendocrine system, as well as signaling in specific types of neurons and/or endocrine cells that express Gβ5-R7.

Our study identified elevated epinephrine as a potential cause of some of the KO phenotypes (Fig. 4F). The lean body and resistance of these mice to an HFD can be explained by adrenergic receptor-mediated acceleration of metabolism, increased thermogenesis, and intensified lipolysis. The higher FFA and TG levels in the KO mice (Table 1) indicate an increase in fat utilization and are consistent with this hypothesis. An increase in sympathetic tone could also explain the striking hyperactivity in the open-field test and food spillage (Fig. 4), as well as the increased glucose mobilization and reduced insulin response in the glucose tolerance test assay (Fig. 3). The lean phenotype and HFD resistance of the Gβ5-KO mice resemble the phenotype of KO of M3 muscarinic receptor (40). Because the M3 receptor was shown to directly interact with Gβ5-RGS7 (23), it is tempting to speculate that the mechanisms underlying the phenotypes in these mouse models are related.

Despite the obvious obesity in the HET mice, we did not detect significant hyperphagia or reduction in oxygen consumption. In contrast to models of morbid obesity, such as leptin-deficient mice, the obesity of Gβ5-HET animals is quite modest and could be caused by only a few percentage changes in energy balance. The incremental changes that occur in HET mice can only be detected in long-term assays that allow their accumulation. In fact, increased adiposity, insulin resistance, and liver steatosis become obvious only as these mice age. Thus, Gβ5-HET mice bear phenotypic characteristics of highly prevalent type of obesity in humans, which is associated with type 2 diabetes and metabolic syndrome.

Indeed, human genetics studies established a link between obesity in a French Canadian cohort and a quantitative trait locus containing the RGS7 gene (41). More recently, researchers linked the RGS6 gene and obesity in a Mexican-American group (42). Although the exact nature of the human RGS7 and RGS6 mutations has not been determined, the mechanism causing the human phenotypes is likely to be similar to that in our Gβ5-HET mice. Since both Gβ5 and R7 subunits are necessary for stability of their complex, the deletion of a single copy of the Gβ5 gene should have the same effect as the partial loss of R7 protein expression or function. On the basis of these considerations, we predict that heterozygous mice lacking single alleles of some R7 family members, e.g., RGS7, will display a similar late-onset obesity phenotype. It is also likely that association between human obesity and Gβ5 mutations will be found in the future.

In summary, we discovered novel mouse phenotypes associated with Gβ5 depletion. The complete KO has a striking increase in activity, is lean, and extremely resistant to HFD. At least as interesting is the late-onset obesity and insulin resistance caused by the Gβ5 haploid insufficiency. Thus, our findings highlight new roles of the Gβ5-R7 family and underscore the importance of careful analysis of heterozygote animals in gene-knockout studies.