Grafts of immortalized chromaffin cells bio-engineered to improve met-enkephalin release also reduce formalin-evoked c-fos expression in rat spinal cord.

Transplantation of adrenal medullary tissue for terminal cancer pain has been tested clinically, but this approach is not practical for routine use because of the shortage of organ donors and lack of tissue homogeneity. As a first alternative step, we have generated immortalized chromaffin cells over-expressing opioid peptides, namely met-enkephalin. Rat chromaffin cells have been genetically modified with vectors containing expression cassettes with either synthetic met-enkephalin or pro-enkephalin gene coding regions, fused with the nerve growth factor signal peptide for secretion. After stable transfection and differentiation in vitro, met-enkephalin and pro-enkephalin cells had higher met-enkephalin immunoreactivity and secreted met-enkephalin levels, compared to control cells containing the expression vector only. In the formalin hindpaw-injection model, 15 days after subarachnoid transplant of cells, grafts of met-enkephalin and pro-enkephalin cells significantly reduced the number of formalin-evoked c-fos immunoreactive spinal neurons in the spinal cord, compared to grafts of vector-alone chromaffin cells. The use of such expandable cell lines, for chronic spinal delivery of opiates, could offer an attractive and safe alternative strategy based on ex vivo gene therapy for the control of opioid-sensitive chronic pain.

Lack of palmitoylation redirects p59Hck from the plasma membrane to p61Hck-positive lysosomes.

Hck, a protein-tyrosine kinase of phagocytes, is the unique member of the Src family expressed under two alternatively translated isoforms differing in their N-terminal site of acylation: p61(Hck) has an additional 21-amino acid sequence comprising a single myristoylation motif, whereas p59(Hck) N terminus has myristoylation and palmitoylation sites. To identify the molecular determinants involved in the targeting of each isoform, they were fused to GFP and expressed in HeLa and CHO cells. p61(Hck) was associated with lysosomal vesicles, whereas p59(Hck) was found at the plasma membrane and to a low extent associated with lysosomes. Their unique N-terminal domains were sufficient to target GFP to the corresponding intracellular compartments. Mutation of the palmitoylation site of p59(Hck) redirected this isoform to lysosomes, indicating that the palmitoylation state governs the association of p59(Hck) with the plasma membrane or with lysosomes. In addition, both isoforms and the nonpalmitoylated p59(Hck) mutant were found on the Golgi apparatus, suggesting a role of this organelle in the subcellular sorting of Hck isoforms. Regarding their subcellular localizations, we propose that bi-acylated p59(Hck) might transduce plasma membrane receptor signals, whereas p61(Hck) and the nonpalmitoylated p59(Hck) might control the biogenesis of phagolysosomes, two functions yet proposed for Hck in phagocytes.

Morphine-like activity of natural human IgG autoantibodies is because of binding to the first and third extracellular loops of the mu-opioid receptor.

We have previously demonstrated that randomly selected healthy individuals express anti-human mu-opioid receptor antibodies which behave as agonist in vitro. In this study, we show that the activity of these antibodies was not affected by the deletion of the amino-terminal region of the receptor. Using agarose-bound peptide columns, we affinity-purified IgG specifically directed toward each extracellular loop. Whatever its specificity, each anti-human mu-opioid receptor (hMOR) extracellular loop peptide IgG preparation was unable, when examined individually, to reduce adenylate cyclase activity. Activation of the hMOR was, however, achieved by the simultaneous binding of IgG to the first and third extracellular loops of the receptor. Our results suggest that the simultaneous binding of IgG antibodies to these two loops mimics morphine-induced receptor activation by triggering a coordinated shift of the third and sixth transmembrane helices.

Nonselective coupling of the human mu-opioid receptor to multiple inhibitory G-protein isoforms.

The human mu-opioid receptor was expressed in Saccharomyces cerevisiae. Binding of [3H]diprenorphine to yeast spheroplasts was specific and saturable (Kd = 1 nm, Bmax = 0.2-1 pmol x mg-1 of membrane proteins). Inhibition of [3H]diprenorphine binding by antagonists and agonists with varying opioid selectivities (mu, delta and kappa) occurred with the same order of potency as in mammalian tissues. Affinities of antagonists were the same with yeast spheroplasts as in reference tissues whereas those of agonists, except etorphine and buprenorphine, were 10-fold to 100-fold lower. Addition of heterotrimeric Gi,o-proteins purified from bovine brain shifted the mu-opioid receptor into a high-affinity state for agonists. Using individually purified Galpha-subunits re-associated with betagamma-dimers, we showed that alphao1, alphao2, alphai1, alphai2 and alphai3 reconstituted high-affinity agonist binding with equal efficiency. This suggests that the structural determinants of the mu-opioid receptor responsible for G-protein coupling are not able to confer a high degree of specificity towards any member of the Gi,o family. The selective effects of opioid observed in specialized tissues upon opioid stimulation may be a result of regulation of G-protein activity by cell-specific factors which should conveniently be analysed using the reconstitution assay described here.

Isolation and characterization of natural human IgG with a morphine-like activity.

Although naturally occurring antibodies have been associated with numerous biological activities, their functional relevance is still a matter of debate. The presence of natural autoantibodies towards immune-related molecules such as cytokines and antibodies suggests a physiological immunomodulatory role. The neuroendocrine opioid system participates in the immune homeostasis. We report here the presence of antibodies with an agonist-like activity towards the human mu-type opioid receptor within a normal human IgG pool. Starting from an IgG pool, autoantibodies were affinity purified using Chinese hamster ovary cells expressing the human mu-opioid receptor. Their specificity was assessed by cytofluorometry and pharmacological analyses. The potency of these antibodies to recognize the mu-opioid receptor was similar to mu-opioid selective agonists. Furthermore, the functional opioid-like activity of the anti-opioid receptor IgG was demonstrated by their ability to inhibit adenylate cyclase activity by a Gi/o-protein-mediated mechanism as indicated by abrogation of the effect by either opioid antagonist or pertussis toxin. Five IgG pools, each from four unrelated healthy blood donors, and single IgG preparations from six other donors were prepared. Antibodies directed against the mu-opioid receptor were found in all IgG samples.

Implication of the first and third extracellular loops of the mu-opioid receptor in the formation of the ligand binding site: a study using chimeric mu-opioid/angiotensin receptors.

Recent studies on chimeric mu/delta-, mu/kappa- and delta/kappa-opioid receptors have suggested that extracellular loops of the receptors were involved in the discriminatory binding of selective ligands by controlling their entry into the transmembrane binding site. Since homochimeric opioid receptors are mostly informative in terms of selectivity, the role of extracellular loops was examined here by studying heterochimeric mu receptors where the totality or parts of extracellular loops were replaced by the corresponding regions of the receptor for angiotensin II. Chimeric mu receptors with extracellular loop EL1 or EL3 originating from the angiotensin receptor had 100-fold decreased affinities for opioids; the length of the first extracellular loop, which is one residue longer in angiotensin than mu receptors, was shown to be responsible for this situation. Substitution of the mu receptor second extracellular loop by that of the angiotensin receptor diminished by approximately 10-fold the affinities for opioids. Since all chimeras had altered affinities for selective and nonselective ligands, we propose that extracellular domains of the mu receptor, particularly the first and third loops, constrain the relative positioning of the connected transmembrane domains where selective as well as nonselective contact points form the opioid binding site.

Agonist-induced signaling and trafficking of the mu-opioid receptor: role of serine and threonine residues in the third cytoplasmic loop and C-terminal domain.

The human mu-opioid receptor and a mutant form, muS/ T[i3+Cter]A, in which all Ser and Thr residues from the third cytoplasmic loop and C-terminal domain were changed to Ala, were studied after expression in CHO-K1 cells. Although the mutant receptors had similar affinities for agonists and EC50 values for inhibition of adenylyl cyclase as compared to wild-type receptors, the Emax were almost 2-fold decreased, suggesting a role of the mutated residues in G-protein coupling. After chronic morphine or etorphine, the EC50 values of the agonists were about 5-fold increased at both receptors but the Emax values were not altered; upon agonist withdrawal forskolin-stimulated cAMP levels were increased to almost 200% of control levels. Sequestration and rapid down-regulation of the mu-opioid receptor were induced by DAGO and etorphine but not morphine. In contrast, the muS/T[i3+Cter]A receptor was not sequestered and was up-regulated (150-380%) after treatment with agonists. The results indicate that the Ser and Thr residues in the third cytoplasmic loop and C-terminus of the mu-opioid receptor are not involved in the limited desensitization or in the adenylyl cyclase superactivation promoted by agonists but that their integrity and/or their phosphorylation is required in the intricate and coordinately regulated pathways involved in receptor signaling and trafficking.

Identification in the mu-opioid receptor of cysteine residues responsible for inactivation of ligand binding by thiol alkylating and reducing agents.

Inactivation by thiol reducing and alkylating agents of ligand binding to the human mu-opioid receptor was examined. Dithiothreitol reduced the number of [3H]diprenorphine binding sites. Replacement by seryl residues of either C142 or C219 in extracellular loops 1 and 2 of the mu receptor resulted in a complete loss of opioid binding. A disulfide bound linking C142 to C219 may thus be essential to maintain a functional conformation of the receptor. We also demonstrated that inactivation of ligand binding upon alkylation by N-ethylmaleimide occurred at two sites. Alteration of the more sensitive (IC50 = 20 microM) did not modify antagonists binding but decreased agonist affinity almost 10-fold. Modification of the less reactive site (IC50 = 2 mM) decreased the number of both agonist and antagonist binding sites. The alkylation site of higher sensitivity to N-ethylmaleimide was shown by mutagenesis experiments to be constituted of both C81 and C332 in transmembrane domains 1 and 7 of the mu-opioid receptor.

Expression and pharmacological characterization of the human mu-opioid receptor in the methylotrophic yeast Pichia pastoris.

The human mu-opioid receptor cDNA from which the 32 amino-terminal codons were substituted by the Saccharomyces cerevisiae alpha-mating factor signal sequence has been expressed in the methylotrophic yeast Pichia pastoris using the host promoter of the alcohol oxidase-1 gene. Cell membranes exhibited specific and saturable binding of the opioid antagonist [3H]diprenorphine (Kd = 0.2 nM and Bmax = 400 fmol/mg protein or 800 sites/cell). Competition studies with non-selective, and mu-, delta- and kappa-selective opioid agonists and antagonists revealed a typical mu-opioid receptor binding profile, suggesting proper folding of the protein in yeast membranes.

Early alterations in the brown adipose tissue adenylate cyclase system of pre-obese Zucker rat fa/fa pups: decreased G-proteins and beta 3-adrenoceptor activities.

This study was undertaken to determine whether receptor and non-receptor components of the adenylate cyclase (AC) cascade were altered in brown adipose tissue (BAT) of 14-day-old pre-obese (fa/fa) rats, before endocrine status is strongly modified by fa gene expression. Activity of the AC catalytic subunit did not differ between the two genotypes. In fa/fa rats compared with control Fa/fa rats, there was a 50% decrease in the activity of alpha Gs (stimulated by NaF or guanosine 5'-[gamma-thio]triphosphate) but no change in protein content (Western blotting). alpha Gi function, assessed by the inhibitory action of low concentrations of guanosine 5'-[beta gamma-imido]triphosphate upon 10(-4) M forskolin-stimulated AC activity, was equally low in both genotypes. Analysis of dose-response curves for different beta-agonists revealed that (i) both the basal and the maximally stimulated activity of AC were 2-fold lower in fa/fa rats than in Fa/fa rats; (ii) BRL37344 and CGP12177 (beta 3 agonists) were less potent in fa/fa than in Fa/fa rats (Kact. multiplied by 2); (iii) noradrenaline and isoprenaline (Iso), at the low-affinity site (beta 3-AR), were less potent in fa/fa than in Fa/fa pups (Kact. increased by 30 and 20% respectively). At the high-affinity site (mainly beta 1) these two agonists were more potent in fa/fa than in Fa/fa rats (Kact. decreased by 40 and 80% respectively). In good agreement with the latter result, the beta 1-adrenergic receptor (beta 1-AR)-selective antagonist CGP20712A had more effect on the Iso-stimulated AC activity in pre-obese than in lean pups (2-fold decreased in IC50). Binding experiments with [3H]CGP12177 show that in BAT of suckling rats, beta 3-ARs represent 80% of the total beta-ARs. Bmax values for the two sites were not affected by the genotype, although the beta 3-AR mRNA concentration in BAT (quantitative reverse-transcriptase PCR) was 3-fold lower in fa/fa rats than in Fa/fa pups. In conclusion, these results provide evidence for alterations in beta 1- and beta 3-AR signalling in BAT of 14-day-old suckling pre-obese Zucker rats with a decreased activity of alpha Gs. The impaired AC responsiveness to catecholamines might be a primary contributor to the development of this genetic obesity.

Transcriptional modulation by n-butyric acid of beta 1-, beta 2-, and beta 3-adrenergic receptor balance in 3T3-F442A adipocytes.

3T3-F442A adipocytes, which express major beta 3-adrenergic receptors (beta 3-AR) (90%) and minor beta 1-AR (< 10%) and beta 2-AR (< 1%) populations, were used to investigate regulation by n-butyric acid of beta-AR subtype expression. Following butyrate treatment, EC50 values of beta 1- and beta 2-selective agonists, dobutamine and fenoterol, were decreased, whereas that of the beta 3-selective agonist BRL37344 was increased. Direct binding and competition of (-)-[125I]iodocyanopindolol binding by selective beta 1- and beta 2-AR antagonists, CGP20712A and ICI118551, and by the beta 3-AR agonist, BRL37344, revealed that both beta 1- and beta 2-AR were increased in butyrate-treated adipocytes, whereas beta 3-AR almost totally disappeared. In control adipocytes, beta 1-, beta 2-, and beta 3-AR transcripts (quantitated by a polymerase chain reaction assay) represented 6.5, 0.5, and 93% of total beta-AR mRNA, respectively. In butyrate-exposed cells, proportions of beta-AR proteins and mRNAs were, respectively, 87 and 94% for beta 1 and 9 and 1% for beta 2-AR. beta 3-ARs were barely detectable in binding assays and accounted for 4.5% of beta-AR transcripts. Variations of beta-AR protein and mRNA levels were accompanied by parallel changes in the transcription rates of the corresponding genes. The differential regulation of the three beta-ARs by n-butyric acid, a dietary factor produced from colonic fermentation, may have significant nutritional and energetic consequences.

Evidence for a functional beta 3-adrenoceptor in man.

1. The existence of a functional beta 3-adrenoceptor in man was investigated by studying the lipolytic action of selective beta-adrenoceptor agents in isolated white omental and subcutaneous fat cells. 2. The non-selective beta 1/beta 2-adrenoceptor antagonist, CGP 12177 was lipolytic in both omental and subcutaneous fat cells. The intrinsic activity relative to isoprenaline was greater in omental than in subcutaneous cells. 3. Addition of the beta 2-adrenoceptor antagonist, ICI 118,551 and the beta 1-adrenoceptor antagonist CGP 20712A in combination or the non-selective beta-adrenoceptor antagonist propranolol alone (all 10(-7) M), induced a rightward shift of the dose-response curves for isoprenaline- and BRL37344-stimulated lipolysis of about 4 and 2 log-units, respectively. However, the antagonists did not alter lipolysis induced by CGP12177. 4. Several concentrations of beta-adrenoceptor antagonists were used to determine the pA2 values by Schild analysis. The values for CGP 20712A and ICI 118,551 (6.63 +/- 0.20 and 6.25 +/- 0.12) as antagonists of the lipolytic effects of CGP 12177 were over 2 units lower than the pA2 value for CGP 20712A against the response to the selective beta 1-agonist dobutamine (8.58 +/- 0.23) and the pA2 value for ICI 118,551 against the response to the selective beta 2-agonist terbutaline (9.15 +/- 0.26). 5. beta 3-Adrenoceptor mRNA expression, investigated with a polymerase chain reaction assay, was demonstrated in both types of adipocytes in the same cell preparations that had a lipolytic response to CGP 12177. 6. In conclusion, human white fat cells express an atypical beta-adrenoceptor in addition to beta 1- and beta 2-adrenoceptors. This receptor is stimulated more selectively by the beta1/beta2-antagonist CGP 12177 than by BRL 37344 and is poorly sensitive to blockade by selective beta1- and beta2-antagonists. On the basis of the pharmacological properties and the mRNA analyses, we suggest that this atypical receptor corresponds to the beta 3-adrenoceptor subtype.

The promoter and intron/exon structure of the human and mouse beta 3-adrenergic-receptor genes.

Transcription-start sites for the mouse and human beta 3-adrenergic-receptor mRNA have been localized in a region comprised between 150 and 200 nucleotides 5' from the ATG translation-start codon. Motifs potentially implicated in heterologous regulation of beta 3-adrenergic-receptor expression by glucocorticoids and by beta-adrenergic agonists have been identified upstream from these cap sites. In mouse, a second mRNA initiation region is postulated to exist further upstream. Comparison of the nucleotide sequences of the 3' end of the human and mouse beta 3-adrenergic-receptor genes to those of the corresponding cDNA revealed that in contrast to beta 1 and beta 2 adrenergic receptors, the beta 3-adrenergic-receptor genes comprise several exons. A large exon (1.4 kb) encodes the first 402 and 388 amino-acid residues of the human and mouse beta 3 adrenergic receptor, respectively. In man, a second exon (700 bp) contains the sequence coding for the six carboxy-terminal residues of the receptor and the entire mRNA 3' untranslated region. In mouse, a second exon (68 bp) codes for the 12 carboxy-terminal residues of the receptor and a third exon contains the beta 3-adrenergic-receptor mRNA 3' untranslated region. The use of alternate acceptor splice sites generates two forms of exon 3 (600 bp and 700 bp), yielding two beta 3-adrenergic-receptor transcripts which are differentially expressed in white and brown adipose tissues. Human beta 3-adrenergic-receptor transcripts with different 3' untranslated regions are produced by continuation of transcription beyond termination signals. Together, our results suggest that utilization of alternate promoters and/or 3' untranslated regions may allow tissue-specific regulation of beta 3-adrenergic-receptors expression.

The human beta 3-adrenergic receptor is resistant to short term agonist-promoted desensitization.

The human beta 3-adrenergic receptor (beta 3AR) lacks most of the structural determinants that, in the beta 2AR, contribute to agonist-induced receptor desensitization. To evaluate the effect of these structural differences on the beta 3AR desensitization profile, the human beta 2- and beta 3AR were stably expressed in Chinese hamster fibroblasts (CHW) and murine Ltk- cells (L cells). Incubation of CHW-beta 2 or L-beta 2 cells with 10 microM isoproterenol for 30 min induced a decrease in the maximal agonist-stimulated adenylyl cyclase activity and a cAMP-dependent reduction in the potency of isoproterenol to stimulate the receptor. In addition, this pretreatment impaired the formation of the high affinity heterotrimeric agonist-receptor-guanine nucleotide-binding protein complex and induced the sequestration of approximately 30% of the beta 2AR away from the cell surface. In contrast, similar treatment of CHW-beta 3 and L-beta 3 cells did not affect the maximal receptor-stimulated adenylyl cyclase activity, nor did it induce any significant sequestration of the beta 3AR. In fact, only a modest cAMP-independent decrease in the potency of isoproterenol to stimulate the receptor could be observed after isoproterenol treatment. The rapid desensitization pattern of a chimeric beta 3AR, in which the third cytoplasmic loop and the carboxyl-terminal tail were exchanged with those of the beta 2AR (which include potential phosphorylation sites and other possible molecular determinants of desensitization), was found to be intermediate between those of the two original receptor subtypes. These results demonstrate that (i) the beta 3AR is less prone than the beta 2AR to undergo rapid agonist-promoted desensitization and, (ii) in addition to the phosphorylation sites located in the third cytoplasmic loop and the carboxyl-terminal tail of the beta 2AR, other molecular determinants contribute to short term desensitization.

Tissue distribution of beta 3-adrenergic receptor mRNA in man.

Expression of mRNA for beta 1-, beta 2-, and beta 3-adrenergic receptors (beta 1-, beta 2-, and beta 3-AR) was investigated in human tissues. beta 1- and beta 2-AR mRNA distribution correlated with that of the cognate receptors established by pharmacological studies. beta 3-AR transcripts were abundant in infant perirenal brown adipose tissue, characterized by the presence of uncoupling protein (UCP) mRNA. In adult whole adipose tissues, beta 3-AR mRNA levels were high in deep deposits such as perirenal and omental, and lower in subcutaneous. In these deposits, UCP mRNA levels paralleled those of beta 3-AR. However, isolated omental and subcutaneous adipose cells, enriched in white adipocytes, expressed beta 3-AR but no UCP transcripts. beta 3-AR mRNA was highly expressed in gallbladder, and to a much lower extent in colon, independently of UCP mRNA. Quadriceps or abdominal muscles, heart, liver, lung, kidney, thyroid, and lymphocytes did not express intrinsic beta 3-AR mRNA. This study demonstrates that substantial amounts of brown adipocytes exist throughout life in adipose deposits, which are generally classified as white. These deposits are the main sites of beta 3-AR expression, which also occurs in gallbladder and colon. beta 3-AR may thus be involved in the control of lipid metabolism, possibly from fat assimilation in the digestive tract, to triglyceride storage and mobilization in adipose tissues.

Inhibition by dexamethasone of beta 3-adrenergic receptor responsiveness in 3T3-F442A adipocytes. Evidence for a transcriptional mechanism.

Modulation of beta 3-adrenergic receptor (beta 3AR) expression by dexamethasone was investigated in the murine 3T3-F442A adipocytic cell line. In untreated cells, a major population of binding sites (62,000-114,000 sites/cell) of low affinity for (-)-[3H] CGP12177 and (-)-[125I]iodocyanopindolol (corresponding to the beta 3AR subtype) was present along with a minor population (6,500-8,000 sites/cell) of sites of high affinity for the radioligands (corresponding to a mixture of the beta 1 and beta 2AR subtypes). Long-term exposure of the cells to 250 nM dexamethasone led to a sharp decrease in beta 3AR density (less than 5,000 sites/cell) which paralleled a diminished potency of the beta 3AR-selective agonists BRL37344 and CGP12177 to stimulate the production of intracellular cAMP. Analysis of RNA by polymerase chain reaction and nuclear run-on assays indicated that dexamethasone inhibited the synthesis of beta 3AR mRNA, resulting in 4-8-fold decrease in the steady-state levels of this mRNA. The down-regulation of beta 3AR protein and cellular mRNA appeared to be mediated by the receptor for glucocorticoids as assessed by the antagonistic action of the anti-glucocorticoid RU38486.

The human beta 3-adrenergic receptor: relationship with atypical receptors.

Atypical beta-adrenergic receptors (beta AR), different from beta 1 and beta 2ARs, have been suggested to modulate energy expenditure. We have characterized a gene coding for a third human beta AR, beta 3AR, whose sequence is 402 amino acids long and is 50.7% and 45.5% homologous to that of the human beta 1 and beta 2AR, respectively. The KD of [125I]-iodocyanopindolol for beta 3AR is 10-fold higher than for beta 1 or beta 2AR. The receptor has an apparent molecular weight of 65,000. Agonists for the beta 3AR induce cyclic AMP accumulation. Among 11 beta antagonists tested, only ICI118551 and CGP20712A, previously classified as, respectively, beta 1 and beta 2 selective, inhibit this effect. The beta 1 and beta 2 antagonists pindolol, oxprenolol, and CGP12177 are agonists of the beta 3AR. The potency order of beta agonists at beta 3 sites correlates with that for stimulation of lipolysis in rat fat tissues. Moreover, because beta 3AR mRNA was detected in rodent adipose tissues, liver, and muscle, we propose that the beta 3AR participates to the control by catecholamines of energy expenditure.

Molecular characterization of the mouse beta 3-adrenergic receptor: relationship with the atypical receptor of adipocytes.

The gene encoding the murine beta 3-adrenergic receptor (beta 3AR) has been isolated. It translates into a polypeptide of 388 amino acid residues which shows 82% overall homology with the human beta 3AR. In Southern blot experiments, a probe derived from the murine beta 3AR gene hybridizes to a unique restriction fragment in the murine and human genomes. In both species, the beta 3AR gene is located on chromosome 8, in regions (8A2----8A4 in mouse, and 8p11----8p12 in man) which are conserved between mouse and man. The pharmacological profile of the mouse beta 3AR strongly resembles that of the human beta 3AR. It is characterized by a low affinity toward the radiolabelled beta-adrenergic antagonist [125I]Iodocyanopindolol and a low efficiency of other antagonists such as propranolol, ICI 118551 or CGP 20712A to inhibit cAMP production induced by isoproterenol. Another salient feature shared by the murine and the human beta 3ARs is the very potent effect of the lipolytic compound BRL 37344 on cAMP accumulation and the partial agonistic effect of the beta 1- and beta 2-adrenergic antagonists CGP 12177A, oxprenolol and pindolol. These properties are very close to those ascribed to the atypical beta AR of rodent adipocytes. In addition, Northern blot analyses indicate that the beta 3AR gene is mainly expressed in mouse brown and white adipose tissues, suggesting that the murine beta 3AR described here is the atypical beta AR involved in the control of energy expenditure in fat tissue.