Address correspondence and reprint requests to Carla R. O. Carvalho MD, Departamento de Fisiologia e Biofísica, ICB, USP, Cidade Universitária, São Paulo, SP, Brazil, Caixa Postal: 66208, CEP: 05389–970. E-mail: email@example.com
Melatonin is the pineal hormone that acts via a pertussis toxin-sensitive G-protein to inhibit adenylate cyclase. However, the intracellular signalling effects of melatonin are not completely understood. Melatonin receptors are mainly present in the suprachiasmatic nucleus (SCN) and pars tuberalis of both humans and rats. The SCN directly controls, amongst other mechanisms, the circadian rhythm of plasma glucose concentration. In this study, using immunoprecipitation and immunoblotting, we show that melatonin induces rapid tyrosine phosphorylation and activation of the insulin receptor β-subunit tyrosine kinase (IR) in the rat hypothalamic suprachiasmatic region. Upon IR activation, tyrosine phosphorylation of IRS-1 was detected. In addition, melatonin induced IRS-1/PI(3)-kinase and IRS-1/SHP-2 associations and downstream AKT serine phosphorylation and MAPK (mitogen-activated protein kinase) phosphorylation, respectively. These results not only indicate a new signal transduction pathway for melatonin, but also a potential cross-talk between melatonin and insulin.
sodium dodecyl sulfate – polyacrylamide gel electrophoresis
SH2-containing phosphotyrosine phosphatase
Melatonin (5-methoxy-N-acetyltryptamine) is produced and secreted by the pineal gland in a circadian fashion. The production of melatonin occurs at night and the duration of its secretion is proportional to the length of the dark period. Melatonin, thus, acts as a neuroendocrine transducer of circadian and seasonal photoperiodic information. In mammals, the physiological functions of melatonin are, among other mechanisms, regulation of the circadian clock in the hypothalamic suprachiasmatic nucleus (SCN), seasonal reproduction, and inhibition of dopamine release from the retina (Morgan et al. 1994; Vanecek 1998). Experiments with pinealectomized animals demonstrated that melatonin also plays a role in energy expenditure, body mass regulation, and insulin peripheral action and secretion (Margraf and Lynch 1993; Lima et al. 1998; La Fleur et al. 1999; Picinato et al. 2002). It has been demonstrated that the biological effects of melatonin are mediated, at least in part, by specific high-affinity G protein-coupled receptors (GPCRs). In mammals, two distinct high-affinity membrane-bound receptors, the MT1 and MT2 receptors, have been described. These receptors are encoded by separate genes (Reppert et al. 1994, 1995, 1996). Both melatonin receptors inhibit adenylyl cyclase via pertussis toxin-sensitive Gi proteins (Yung et al. 1995; Barret et al. 1999). The MT1 receptor is localized in the hypothalamic SCN and hypophyseal pars tuberalis (Sugden et al. 1999; revised by Dubocovich et al. 2003). MT2 is localized in the SCN and retina (Reppert et al. 1995; Willians et al. 1995; Hunt et al. 2001).
There is a consistent body of evidence demonstrating that insulin acts in the hypothalamus to suppress feeding behaviour (Woods et al. 1979; Brüning et al. 2000). Insulin receptors (IR) display distinct patterns of expression in the hypothalamus, pituitary and olfactory bulb (Baskin et al. 1994; Folli et al. 1994). IR is a protein tyrosine kinase that catalyzes the phosphorylation of several intracellular substrates, including the insulin receptor substrate (IRS) proteins and the src-homology-2/α collagen related protein, Shc (reviewed by Saltiel and Kahn 2001). The IRS proteins, once phosphorylated at distinct tyrosine residues, act as docking sites for proteins containing Src homology 2 (SH2) domains, including phosphatidylinositol-3-OH kinase [PI(3)K] and the SH2-containing phosphotyrosine phosphatase, SHP-2 (reviewed by Saltiel and Kahn 2001). A downstream substrate of PI(3)K activity is the serine/threonine protein kinase B or AKT (Coffer et al. 1998; Lawlor and Alessi 2001). Upon insulin receptor tyrosine kinase activation and autophosphorylation, there is also recruitment of Shc protein and the adaptor molecule, Grb2, leading to activation of the extracellular signal-regulated protein kinase (p42/44 MAP kinase, ERK) pathway (Skolnik et al. 1993; De Fea and Roth 1997).
In this study, we evaluated the ability of melatonin to induce the tyrosine phosphorylation of IRS-1, as well as IRS-1/PI(3)K and IRS−1/SHP2 associations, and the phosphorylation of AKT/PKB and p42/44MAPK in the rat hypothalamus in vivo. We also identified a kinase that was activated by melatonin, the insulin receptor, and assessed a possible cross-talk between the melatonin and insulin signalling pathways. The present data reveal a new signal transduction pathway for melatonin and demonstrate, for the first time, that the insulin receptor is co-activated by a GPCR-binding hormone.
Materials and methods
The reagents and apparatus for sodium dodecyl sulfate – polyacrylamide gel electrophoresis (SDS–PAGE) and the nitrocellulose membranes were from Bio-Rad (Richmond, CA, USA). Tris, phenylmethylsulfonylfluoride (PMSF), aprotinin, dithiothreitol, and melatonin were from Sigma-Aldrich (St Louis, MO, USA). Sodium amobarbital and human recombinant insulin (Humulin R) were purchased from Eli Lilly Co. (Indianapolis, IN, USA). The melatonin receptors antagonists, MT1/MT2 non-selective luzindole and the MT2-seletive 4-phenyl-2-propionamidotetraline (4P-PDOT), were obtained from Tocris Cookson (Ballwin, MO, USA). Anti-IR, anti-IRS-1, and anti-phosphotyrosine antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Anti-PI3-kinase was from Upstate Biotechnology (Lake Place, NY, USA). Phospho-AKT (Ser473) and Phospho-p42/44 MAPK (Thr202/Tyr204) antibodies were from New England BioLabs (Beverly, MA, USA). The enhanced chemiluminescence reagent kit, ECL, and the protein A Sepharose 6MB were from Amersham-Pharmacia Biotech (Buckinghamshire, UK).
Animals and surgical procedures
Adult male Wistar rats (250–300 g) were used in all experiments in accordance with the guidelines of the Brazilian College for Animal Experimentation (COBEA). After an overnight fast, the rats were anaesthetized with sodium thiopental (5 mg/100 g body weight, i.p.) and positioned on a David Kopf stereotaxic apparatus using head and chin holders. The lateral, anteroposterior and dorsoventral coordinates of the lateral ventricle were obtained from Paxinos and Watson (1986). A cannula was positioned and fixed to the skull with cranioplastic cement. The rats were studied 6 days after cannula implantation into the lateral ventricle.
Anaesthetized 6-h fasted rats (sodium thiopental 5 mg/100 g body weight, i.p.) received an acute infusion of 5 μL of saline containing, or not, insulin (60 μg) or melatonin (100 ng) through the cannula. Following the infusion, and after the indicated time in the figure legend, animals were decapitated by guillotine, the skulls were immediately opened, and the suprachiasmatic region of the hypothalamus was rapidly excised under a surgical microscope and was immediately homogenized as described below. Histological analysis of the remaining brain showed that the removed area included the posterior part of the medial pre-optic area, the suprachiasmatic nucleus and the most anterior and ventral portions of the anterior hypothalamic nucleus. The extent of the lesion never exceeded the basal retrochiasmatic area.
Western blot analysis
Hypothalami from three different animals were treated as pool samples. The removed tissue was immediately homogenized in approximately 100 μL of ice-cold solubilization buffer containing 100 mm Tris, pH 7.6, 1% Triton X-100, 0.01 mg/mL aprotinin, 2 mm PMSF, 10 mm Na3VO4, 10 mm NaF, 10 mm Na4P2O7, and 10 mm EDTA, using a Polytron PTA 2100 generator operated at maximum speed for 30 s. Insoluble material was removed by centrifugation for 30 min at 12 800 g in a Beckman 70.1 Ti rotor (Palo Alto, CA, USA) at 4°C. The supernatants of the samples were used for immunoprecipitation with anti-phosphotyrosine, or anti-IR, or anti-IRS-1, and protein A-Sepharose 6MB. The precipitated proteins were treated with Laemmli sample buffer (Laemmli 1970) and run on SDS–PAGE (8% bis-acrylamide) in a Bio-Rad miniature lab gel apparatus (Mini-Protean, Bio-Rad). To determine the phosphorylation status of AKT and p42/44MAPK, the solubilized proteins from hypothalami samples were treated with Laemmli sample buffer containing 100 mm dithiothreitol and heated in a boiling water bath for 4 min. Similar sized aliquots (80 μg protein) were subjected to SDS–PAGE (8% bis-acrylamide). Mr standards were myosin (210 kDa), β-galactosidase (110 kDa), BSA (85 kDa) and ovalbumin (49 kDa).
Electrotransfer of proteins from the gel to nitrocellulose was performed for 90 min at 120 V (constant) in the miniature transfer apparatus (Mini-Protean). To reduce non-specific protein binding to the nitrocellulose, the filter was pre-incubated overnight at 4°C in blocking buffer (5% non-fat dry milk, 10 mm Tris, 150 mm NaCl and 0.02% Tween 20). The nitrocellulose blots were incubated for 4 h at 22°C with the specific antibodies described in the figure legends diluted in blocking buffer (3% non-fat dry milk). To visualize the autoradiogram, commercial enhanced chemiluminescence reagents exposed to photographic film were used. Quantitative analysis of the blots was done using Scion Image software (Frederick, MD, USA).
IR kinase assay
Insulin receptor tyrosine kinase activity was measured by autophosphorylation. A low dose of melatonin was infused into the lateral ventricle in order to stimulate IR autophosphorylation and partial receptor activation. The insulin receptor was then immunoprecipitated and allowed to autophosphorylate in vitro in the presence of exogenous ATP (15 μm). Tyrosine autophosphorylation was measured by immunoblotting with anti-phosphotyrosine antibody.
Results are presented as direct comparisons of bands present in the autoradiographs from saline or melatonin injected rats. Band intensities were quantified by optical densitometry using the Scion Image software (Frederick, MD, USA). Student's t-test for unpaired data was employed.
Melatonin-induced tyrosine phosphorylation of IRS-1 and insulin receptor β-subunit
Figure 1 illustrates a typical immunoblot with antibody against phosphotyrosine-containing proteins from samples previously immunoprecipitated with anti-phosphotyrosine antibody. Figure 1(a) shows that melatonin induces tyrosine phosphorylation of two distinct bands, an upper broad band migrating at 180–140 kDa and a lower band migrating between 90 and 100 kDa. The maximal tyrosine phosphorylation occurred at 5 min after melatonin infusion and decreased dramatically thereafter. Parallel experiments were performed in which the rats received an infusion of insulin as a positive control for induced tyrosine phosphorylation (Carvalheira et al. 2001) (Fig. 1b).
By re-incubating the membranes with specific antibodies, we detected that most of the 180–140 kDa band corresponded to the insulin receptor substrate-1, IRS-1. The 90–100-kDa band corresponded to the 95 kDa β subunit of the insulin receptor (IR) (data not shown).
Melatonin-induced IRS-1 association with PI (3) kinase and with SHP-2
In order to confirm that the upper band contains at least the protein named IRS-1, and that the lower band contains the β-subunit of the insulin receptor, some excised hypothalami samples from saline or acutely melatonin-treated rats were immunoprecipitated with anti-IR or anti-IRS-1 antibodies, followed by immunoblotting with anti-phosphotyrosine antibody (Figs 2a and 3a, respectively).
In the hypothalami samples previously immunoprecipitated with anti-insulin receptor antibody and subsequently immunoblotted with anti-phosphotyrosine antibody, melatonin was shown to induce an approximately 3-fold increase (p < 0.02) in insulin receptor tyrosine phosphorylation above basal (Fig. 2a), with no difference in the amount of immunoprecipitated insulin receptor β-subunit (Fig. 2b).
In samples from hypothalami previously immunoprecipitated with anti-IRS-1 antibody and immunoblotted with anti-phosphotyrosine antibody, there was a 10-fold increase (p < 0.05) in the melatonin-stimulated IRS-1 phosphorylation (Fig. 3a), without any change in the amount of IRS-1 immunoprecipitated, demonstrated by subsequent immunoblotting of the same nitrocellulose sheets with anti-IRS-1 antibody (Fig. 3b). Previously published data demonstrate a co-immunoprecipitation between IRS-1 and some SH2-containing proteins, such as the 85-kDa subunit of PI(3)K and SHP2, a phosphotyrosine phosphatase with approximately 78 kDa (reviewed by Saltiel and Kahn 2001). When the same blots were subsequently incubated with antibodies directed against the 85-kDa subunit of PI(3)K, there was a faint band in the basal state (saline infusion). After melatonin infusion, the intensity of this band displayed a significant (49 ± 9%, p < 0.05) increase compared with the basal state (Fig. 3c).
The incubation of the same blots with anti-SHP2 antibody demonstrated that melatonin induced a 53 ± 8% increase (p < 0.05) in the IRS-1/SHP−2 association (Fig. 3d). These results are consistent with the formation of a stable association between IRS-1 and each of the demonstrated proteins, PI(3)K and SHP-2.
Melatonin dose–response curve for tyrosine phosphorylation of insulin receptor β-subunit
Figure 4 demonstrates a dose–response experiment. After 5 min of 5 μL infusion containing 10 pg, 1 ng or 100 ng melatonin into the lateral ventricle, the hypothalami from different animals were excised and the soluble proteins were precipitated with the anti-insulin receptor antibody. Melatonin-stimulated insulin receptor tyrosine phosphorylation was dose-dependent. The phosphorylated proteins were detectable after injection of as little as 10 pg melatonin and maximal stimulation occurred when 1 ng was infused.
Effect of melatonin on AKT and MAP kinase phosphorylation
Melatonin infusion induced a 3-fold increase in AKT serine phosphorylation at 10 min, and then returned to basal levels after 15 min (Fig. 5a). This result was confirmed in three distinct experiments, as demonstrated in the representative figure, and the bar graphs indicate the densitometry analysis.
Figure 5(b) shows that melatonin induces increased phosphorylation of p42MAP kinase at 10 min, which remains phosphorylated above basal level for 15 min after melatonin infusion.
To determine whether the melatonin-induced insulin receptor phosphorylation was due to the induction of its tyrosine kinase activity, we performed an in vitro kinase assay based on the dose–response results (Fig. 4). After the infusion of 10 pg melatonin in vivo, the hypothalami were excised and submitted to immunoprecipitation with anti-insulin receptor antibody and protein A. The immunoprecipitated proteins were incubated with exogenous ATP (15 μm) to allow autophosphorylation to occur if the insulin receptor kinase activity was activated by melatonin intracerebroventricular (i.c.v.) infusion.
Immunoblotting of these assays with anti-phosphotyrosine antibody demonstrated that there was a clear increase in IR autophosphorylation (Fig. 6) of approximately 3-fold when compared with basal plus ATP.
Effect of the melatonin receptor inhibitors, luzindole and 4P-PDOT, on melatonin-induced insulin receptor tyrosyl phosphorylation
In order to analyze whether the effect of melatonin detected on insulin receptor autophosphorylation was due to melatonin membrane receptors, MT1 and MT2, experiments using luzindole, a non-selective melatonin receptor antagonist, and 4P-PDOT, a selective MT2 receptor antagonist were performed.
When the rats were pre-treated for 10 min with 10 μm luzindole or 300 nm 4P-PDOT, there were marked reductions in the melatonin-induced tyrosine phosphorylation of the insulin receptor in extracts immunoprecipitated with anti-insulin receptor antibody (Fig. 7).
In the present study, the action of melatonin upon various tyrosine kinase pathways was examined in the hypothalamus. Our results show that melatonin induces a rapid MT1/MT2 membrane receptor-dependent tyrosine phosphorylation and activation of the insulin receptor β-subunit tyrosine kinase (IR).
Upon IR activation, tyrosine phosphorylation of IRS-1 was detected. In addition, melatonin induces IRS-1/PI(3)K and IRS-1/SHP-2 associations and downstream AKT serine phosphorylation and p42MAPK (also known as ERK-2) phosphorylation, respectively. The present report is the first to demonstrate that melatonin stimulates tyrosine phosphorylation of insulin receptor and IRS-1. These data suggest that IRS-1 phosphorylation may play a role as a converging target in insulin- and melatonin-stimulated signalling pathways and also as a probable link for both hormones in the control of body weight (Le Gouic et al. 1996; Rasmussen et al. 1999) and carbohydrate metabolism (Alquier et al. 2003).
The major lipid product of PI(3)K activity is phosphatidylinositol 3,4,5-triphosphate (PIP3). PIP3 has binding affinity for a sequence called the pleckstrin homology (PH) domain. Thus, PH-domain-containing proteins are localized to membrane-associate signalling complexes (Coffer et al. 1998; Lawlor and Alessi 2001). Some targets of PI 3-kinase lipid products are AKT/PKB and its upstream activator, phosphoinositide-dependent kinase (PDK). AKT/PKB activates various enzymes involved in cell growth and in the inhibition of apoptosis, in addition to other functions. These data, taken together with our results of melatonin-induced AKT phosphorylation, suggest that AKT may participate in a novel mechanism by which melatonin acts as a putative modulator of feeding behaviour associated to seasonal variations of body weight and metabolism (Bartness and Wade 1985; Rosenthal et al. 1987).
Evidence from distinct source indicates that SHP-2 activity has a positive role in the activation of MAP kinase members and in mitogenic signalling by insulin, EGF, and interleukin-6 (Xiao et al. 1994; Hausdorff et al. 1995; Deb et al. 1998; Cunnick et al. 2000). We showed that melatonin is able to induce tyrosine phosphorylation of IRS-1 and its simultaneous association with SHP-2 and a later tyrosine/threonine phosphorylation of p42MAP kinase in the hypothalamus of intact rats. In this regard, our data may be of physiological importance, as MAPK activation is related to light-induced clock entrainment (Butcher et al. 2002).
The implications of intracellular interactions between melatonin and insulin signalling systems are interesting. In addition to body weight and food intake modulation, insulin is able to promote sharp decreases in elevated plasma FFA levels in fasted rats through a direct effect on the CNS (Coimbra and Migliorini 1986). Furthermore, there is a growing body of evidence indicating that melatonin, through its MT1 and MT2 receptors, is involved in the neuronal mechanism regulating circadian rhythms, in particular, those of the glucose blood level (La Fleur et al. 1999, 2001).
The concentrations used for Luzindole and 4P-PDOT are in the range for non-specific blockage of MT1 and MT2, in the case of Luzindole and specific blockage of MT2 receptor in the case of 4P-PDOT (Hunt et al. 2001). Since the melatonin-induced IR tyrosine phosphorylation was blocked, at same magnitude, in both cases, it is possible to conclude that the melatonin receptor involved is, most probably, of the MT2 type.
The molecular mechanisms by which the melatonin receptor, which lacks intrinsic tyrosine kinase activity, couples to tyrosine phosphorylation events are not known. MT2 (and MT1), through Gi protein activation, could control the tyrosine kinase capacity of the IR, by modulating the activity of specific phosphotyrosine phosphatases that respond to the intracellular cAMP levels. This possibility is supported by recent evidence obtained with transgenic mice, which over express protein Gi. The Gi-transgenic mice have an enhanced insulin-induced IR and IRS-1 tyrosine phosphorylation mediated by the suppression of protein-tyrosine phosphatase-1B activity and expression (Tao et al. 2001). However, further studies will be required to confirm this hypothesis.
In summary, our study demonstrates the existence of an additional intracellular signalling mechanism stimulated by melatonin in the rat hypothalamus (Fig. 8). Intracellular cross-talk between melatonin and insulin signalling may have a role in the intracellular mechanism controlling body weight, feeding behaviour and blood glucose circadian rhythm.
We thank Ms Luciene M Ribeiro and Mr Luiz Janeri for technical assistance. We also gratefully acknowledge helpful discussions with Drs N. Canteras and L. R. Britto. Grants from Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) and Conselho Nacional de Pesquisa (CNPq) supported this work.