Insulin-Induced Ca2+ Entry in Hepatocytes Is Important for PI 3-Kinase Activation, but Not for Insulin Receptor and IRS-1 Tyrosi

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Biochimica et Biophysica Acta 1495 (2000) 14^23 www.elsevier.com/locate/bba

Insulin-induced Ca2 entry in hepatocytes is important for PI 3-kinase activation, but not for insulin receptor and IRS-1 tyrosine phosphorylation

Kenza Benzeroual a;b, Sanjay K. Pandey c;f , Ashok K. Srivastava c;d;f , Ge¨rald van de Werve a;e;g, Pierre S. Haddad a;b;c;*

a Groupe de Recherche en Transport Membranaire, Universite¨ de Montreäl, Montreal Que., Canada b De¨partement de Pharmacologie, Universite¨ de Montreäl, Montreal, Que., Canada c De¨partement de Physiologie, Universite¨ de Montreäl, Montreal, Que., Canada d De¨partement de Me¨decine, Universite¨ de Montreäl, Montreal, Que., Canada e De¨partement de Nutrition, Universite¨ de Montreäl, Montreal, Que., Canada f Centre Hospitalier de l'Universite¨ de Montreäl, Campus Hoªtel-Dieu, Montreal, Que., Canada g Centre Hospitalier de l'Universite¨ de Montreäl, Campus Notre-Dame, Montreal, Que., Canada

Received 2 June 1999; received in revised form 26 October 1999; accepted 28 October 1999

Abstract

Insulin produces an influx of Ca2 into isolated rat hepatocyte couplets that is important to couple its tyrosine kinase receptor to MAPK activity (Benzeroual et al., Am. J. Physiol. 272, (1997) G1425^G1432. In the present study, we have examined the implication of Ca2 in the phosphorylation state of the insulin receptor (IR) L-subunit and of insulin receptor substrate-1 (IRS-1), as well as in the stimulation of PI 3-kinase activity in cultured hepatocytes. External Ca2 chelation (EGTA 4 mM) or administration of Ca2 channel inhibitors gadolinium 50 WM or nickel 500 WM inhibited insulin-induced PI 3-kinase activation by 85, 50 and 50%, respectively, whereas 200 WM verapamil was without effect. In contrast, the insulin- induced tyrosine phosphorylation of IR L-subunit and of IRS-1 was not affected by any of the experimental conditions. Our data demonstrate that the stimulation of PI 3-kinase activity by the activated insulin receptor, but not the phosphorylation of IR L-subunit and IRS-1, requires an influx of Ca2.Ca2 thus appears to play an important role as a second messenger in insulin signaling in liver cells. ß 2000 Elsevier Science B.V. All rights reserved.

Keywords: Insulin; Hepatocyte; Phosphatidylinositol-3-kinase; Calcium; Insulin receptor; Insulin receptor substrate-1

1. Introduction ety of biological functions, including stimulation of amino acid and glucose transport, modulation of en- At the cellular level, insulin promotes a wide vari- zyme activities and of gene expression, and enhance- ment of DNA synthesis [1]. Through these events, insulin acts as a central regulator in glucose, lipid and protein metabolism in liver and other peripheral * Corresponding author. De¨partement de Pharmacologie, Uni- versite¨ de Montreal, P.O. Box 6128, Downtown Station, Mon- tissues in response to environmental changes. tre¨al, Que¨. H3C 3J7, Canada. Fax: +1-514-343-2291; Signaling through the insulin receptor is a complex E-mail: [email protected] process that involves the activation of protein ki-

BBAMCR 14570 29-12-99 K. Benzeroual et al. / Biochimica et Biophysica Acta 1495 (2000) 14^23 15 nases and phosphatases, principally in at least two The question of the implication of Ca2 in other major transduction systems: one involving mitogen- pathways of insulin action in the liver thus arose. activated protein kinase (MAP kinase) and another Therefore, the aim of the present work was to exam- involving phosphatidylinositol 3-kinase (PI 3-kinase). ine the implication of Ca2 in the other major branch In most cells, insulin initiates its e¡ects by interacting of the insulin signal transduction pathway, namely with the K-subunit of its receptor with subsequent PI 3-kinase activation. To gain further insight on activation of the tyrosine kinase activity of the L- the site of action of the divalent cation, we also ex- subunit followed by receptor autophosphorylation amined the e¡ect of Ca2 on the phosphorylation [1,2]. Activation of the insulin receptor (IR) tyrosine state of the IR L-subunit and of IRS-1. kinase also leads to tyrosyl phosphorylation of a number of intracellular proteins, the most prominent being the cytosolic insulin receptor substrate-1 (IRS- 2. Materials and methods 1) [3]. Considerable evidence demonstrates that IR tyrosine kinase activity and associated IRS-1 phos- 2.1. Materials phorylation are essential for many of the complex cellular responses to insulin [1,4,5]. Gadolinium chloride hexahydrate was purchased IRS-1 acts as an intermediate docking protein be- from Aldrich (Milwaukee, WI). Polyclonal antibody tween the IR and cytoplasmic proteins containing against the p85 regulatory subunit of PI 3-kinase was SH2 domains [3]. For instance, when IRS-1 is tyro- purchased from Santa Cruz Biotechnology, (Santa sine phosphorylated, it can bind and activate PI Cruz, CA); it is directed against an epitope corre- 3-kinase in several cells in vivo and in vitro, including sponding to amino acids 333^430 of the protein. the liver [3,6^9]. This enzyme is composed of a 85- Anti-phosphotyrosine, anti-human insulin receptor kDa regulatory subunit [10] and a 110-kDa catalytic L-subunit and anti-rat carboxyterminal IRS-1 anti- subunit that phosphorylates phosphatidylinositol bodies were purchased from Upstate Biotechnology (PI) as well as its 4 and 4,5 phosphorylated deriva- (Lake Placid, NY). Goat anti-mouse antibodies con- tives in the D-3 position of the inositol ring [11]. PI jugated to alkaline phosphatase were from Bio-Rad 3-kinase plays a key role in mediating several bio- (Mississauga, ON). Silica gel-60 plates were pur- logical e¡ects of insulin, such as glucose transport chased from Merck (Rahway, NJ). Protein A Se- [12], glycogen synthesis [13], and transcriptional reg- pharose beads were obtained from Pharmacia Bio- ulation of several genes [14,15]. Thus, IR, IRS-1, and tech (Mississauga, ON). Bovine insulin, bovine PI 3-kinase represent three of the earliest steps in one serum albumin (BSA fraction V), leupeptin, aproti- branch of the insulin signaling cascade. nin, pepstatin A, as well as other chemicals were In the liver, insulin is also known to promote cat- purchased from Sigma-Aldrich (Oakville, ON). ion in£ux and this was found to be important for the hormone's action [16]. Moreover, we have recently 2.2. Animals demonstrated that insulin induces an increase in cytosolic Ca2 in primary rat hepatocytes. This in- Male Sprague^Dawley rats weighing 150^175 g crease was dependent on the presence of external were ordered from Charles River Breeding Labora- Ca2 but was independent of e¡ects on insulin bind- tories (St.-Constant, QC) and allowed free access to ing [17]. It also had a pharmacological sensitivity Purina rat chow and tap water. Animals were treated reminiscent of non-selective cation channels present in accordance with the guidelines of the Canadian on the hepatocyte plasma membrane [18,19]. Finally, Council for the Care of Animals and all experimental conditions interfering with the e¡ect of insulin on protocols were accepted by the university ethics com- hepatocellular Ca2 were also found to inhibit the mittee. stimulation of MAPK activity by the hormone. We concluded that insulin triggers Ca2 in£ux into liver 2.3. Isolation and incubation of hepatocytes cells and this e¡ect is important for the stimulation of the MAPK signaling cascade [17]. Rats were anesthetized with an intraperitoneal in-

BBAMCR 14570 29-12-99 16 K. Benzeroual et al. / Biochimica et Biophysica Acta 1495 (2000) 14^23 jection of sodium pentobarbital (50 mg/kg; Centre de HEPES, pH 7.5, 150 mM NaCl, 0.1% Triton X- Distribution de Me¨dicaments Ve¨te¨rinaires, St.-Hya- 100, 10% Glycerol, 10 Wg/ml leupeptin, 1 mM cinthe, QC) prior to surgery. Hepatocytes were iso- PMSF), once with 0.5 M LiCl/0.1 M Tris pH 7.5, lated by collagenase digestion of the liver as previ- once with H2O and ¢nally with 0.1 M NaCl/1 mM ously described [20]. Cells were puri¢ed by EDTA/20 mM Tris pH 7.5. Beads were then resus- centrifugation on a Percoll gradient. Viability was pended in 50 Wl of kinase bu¡er (20 mM Tris-HCl, determined by Trypan blue exclusion and exceeded pH 7.5, 10 mM NaCl, 0.5 mM EGTA). Then 1 Wlof 85%. Freshly isolated hepatocytes were suspended in a 20 mg/ml concentration of pure L-K-phosphatidyl- William's E medium containing 25 mM bicarbonate inositol (PI) (bought from Sigma) dissolved in DMSO and 1% BSA, pH 7.4, at 37³C. Cells were then seeded was added to the reaction mixture to make micelles onto 35-mm plastic Petri dishes previously coated of PI. The mixture was vortexed and preincubated at with rat tail collagen. Thereafter, cells were incu- room temperature for 10 min. Subsequently, [Q-32P]- bated for 60 min at 37³C in a humidi¢ed atmosphere ATP (2 mCi/ml) (10 WCi per assay) and MgCl2 of 5% CO2^95% O2. Unattached cells were then re- (10 mM) were added to start the reaction and the moved, medium was replaced with Krebs-Henseleit mixture was incubated for 2^3 min at room temper- bu¡er, and incubation was prolonged for an addi- ature. The reaction was stopped by addition of 150 Wl tional 30 min. These conditions are similar to £uo- of CHCl3 :CH3OH: HCL (100:200:2) and 100 Wl rescence experiments where insulin was found to in- of CHCl3 was added to separate the phases. The crease intracellular calcium [17]. inferior organic phase was taken and washed with one volume of CH3OH:HCl (1:1). The lipid samples 2.4. PI 3-kinase assay were concentrated in vacuo and spotted onto a silica gel 60 plate which was later developed in CHCl3 : The assay was performed essentially as described CH3OH:NH4OH:H2O (43:38:5:7). The phosphory- by Fukui and Hanafusa [21] with minor modi¢ca- lated lipids were visualized by autoradiography and tions [22]. Isolated hepatocytes were cultured as de- the area corresponding to each spot was quanti¢ed scribed above. Cells were challenged at 37³C with by using a phophorimager (Molecular Dynamics). A Krebs^Henseleit bu¡er containing insulin or its ve- similar method has been used by other investigators hicle. To assess the e¡ects of 4 mM EGTA, 200 WM to demonstrate insulin-stimulated PI 3-kinase activity verapamil, 500 WM nickel or 50 WM gadolinium, in CHO cells [23] and adipocytes [24]. these agents were administered 15 min before and throughout the insulin challenge. The incubation 2.5. Insulin receptor L-subunit tyrosine was stopped by two rapid washes with ice-cold phos- phosphorylation phate-bu¡ered saline. The 35-mm Petri dishes were then quickly frozen on liquid nitrogen and stored at The cells were stimulated as described for PI 380³C pending the enzyme assay. The cells were 3-kinase assay and were lysed on ice in 200 Wlof lysed in bu¡er A (25 mM Tris-HCl, pH 7.5, 25 bu¡er A. The lysates were clari¢ed by centrifugation mM NaCl, 1 mM Na orthovanadate, 10 mM Na for 12 min at 10 000Ug and then incubated over- £uoride, 10 mM Na pyrophosphate, 20 nM okadaic night at 4³C with 5 Wl of anti-human IR L-subunit acid, 0.5 mM EDTA, 1 mM phenylmethylsulfonyl antibody preadsorbed to protein A-Sepharose beads. £uoride (PMSF), 10 Wg/ml aprotinin, 1% Triton X- The immunoprecipitate was collected with protein A- 100 and 0.1% SDS). The cells were scraped and cen- Sepharose and washed three times, once with 50 mM trifuged at 12 000Ug for 12 min. Equal amount of HEPES bu¡er, pH 7.5, containing 0.1% Triton X- lysate proteins were incubated with 5 Wl of p85-anti- 100 and 0.1% SDS and twice with the same bu¡er body for 2 h at 4³C. Protein A-Sepharose beads were without SDS. The immunoprecipitate was solubilized then added and the resulting mixture was incubated by boiling in 2Usample bu¡er for 10 min. The solu- for an additional 2 h at 4³C. The beads were centri- bilized fraction was submitted to electrophoresis on fuged in a microtube for 15 s and the pellet was 10% SDS-polyacrylamide gels and blotted with anti- washed three times with HNTG medium (20 mM phosphotyrosine antibody (1:1000). The blots were

BBAMCR 14570 29-12-99 K. Benzeroual et al. / Biochimica et Biophysica Acta 1495 (2000) 14^23 17 developed by using goat-anti-mouse IgG conjugated to alkaline phosphatase (1:3000) [25].

2.6. IRS-1 tyrosine phosphorylation

Cells were stimulated and lysed as described above. Equal amounts of protein from cell lysates were added to the anti-IRS-1 antibody-protein A-Se- pharose complex and incubation continued overnight at 4³C. The immunoprecipitate was collected by centrifugation in a microtube for 15 s and the pellet was washed three times with ice-cold PBS. The beads were boiled in 3ULaemmli's sample bu¡er, submitted to electrophoresis on 10% SDS-polyacrylamide gels, transferred to PVDF membranes, blotted with Fig. 1. Time course of stimulation of PI 3-kinase activity by insulin. Isolated rat hepatocytes were incubated in the presence of anti-phosphotyrosine antibody (1:1000) and detected 10 nM insulin for 1, 3, 5, 10, and 15 min. Cells were lysed and by goat-anti-mouse IgG conjugated to alkaline phos- the protein lysates were immunoprecipitated with p85-antibody phatase (1:3000) [22]. as described in Section 2. The immune complex was assessed for PI 3-kinase activity using phosphatidylinositol as substrate 2.7. Statistical analysis as described in Section 2. Data points are the mean þ S.E.M. of eight separate experiments and are expressed in relation to respective baseline values obtained for each experiment (di¡erent Results are expressed as means þ S.E.M. Statistical hepatocyte preparation). signi¢cance between treatment groups was evaluated by one-way analysis of variance. Di¡erences between 3.2. Dose-dependence of PI 3-kinase activity group means were considered signi¢cant at P 6 0.05. Similarly, the concentration dependence of the insulin action on PI 3-kinase activity in cultured rat 3. Results hepatocytes had not been reported. Therefore, primary hepatocytes were stimulated with increasing 3.1. Time-dependence of PI 3-kinase activity concentrations of insulin (0.1^100 nM) for 3 min. As shown in Fig. 2, an increase in PI 3-kinase activ- We ¢rst evaluated the time course of PI 3-kinase ity was detectable after exposure to 0.1 nM insulin activity in response to insulin in primary cultures of (1.5-fold stimulation above basal values) and optimal rat hepatocytes, as it had not been directly examined stimulation was observed with 10 nM insulin (4-fold in this model previously. Fig. 1 presents the results of stimulation), as seen for several other insulin-depen- experiments where hepatocytes were stimulated with dent actions in such preparations [26,27]. When the 10 nM insulin for the indicated times. PI 3-kinase data were ¢tted for a standard dose^response rela- was then immunoprecipitated by using antibodies tionship, a half-maximal e¡ective dose of 1.1 þ 0.5 to the p85 subunit of the enzyme. Immune complexes nM insulin was obtained which falls within the phys- were subjected to PI 3-kinase assay as detailed in the iological range [28]. Thus, insulin-induced stimula- Section 2. As expected, insulin caused a time-depen- tion of PI 3-kinase in isolated cultured hepatocytes dent increase in PI 3-kinase activity. The e¡ect of was dose-dependent as previously observed in whole insulin was evident as early as 1 min and maximal liver [29] and in isolated adipocytes [24]. at 3 min. Thereafter, PI 3-kinase activity remained elevated for at least 15 min after exposure to the 3.3. E¡ect of Ca2+ on insulin-induced PI 3-kinase hormone. The duration of insulin's e¡ect on PI activity 3-kinase in primary hepatocytes was similar to that reported in CHO cells [23] and in adipocytes [24]. Since interference with Ca2 movements was pre-

BBAMCR 14570 29-12-99 18 K. Benzeroual et al. / Biochimica et Biophysica Acta 1495 (2000) 14^23 viously shown to prevent insulin-stimulated MAPK activity [17], we were interested to explore if Ca2 was also implicated in other insulin-stimulated pathways. As presented above, an optimal increase in hepatocyte PI 3-kinase activity was observed once the cells were stimulated with 10 nM insulin for 3 min (Figs. 1 and 2). We thus used these conditions to examine the e¡ects of external Ca2 chelation and of Ca2 channel inhibition. As shown in Fig. 3, chelation of extracellular Ca2 with EGTA inhibited insulin-induced activation of PI 3-kinase by 85%. We also tested the e¡ect of nickel, a divalent cation blocker of Ca2 channels [30,31], and of gadolinium, a trivalent ion of the lanthanide series previously shown to inhibit non-selective cation channels in hepatoma cells [32] and to interfere with Ca2 entry caused by internal Ca2 store depletion in cultured hepatocytes [33,34]. At respective doses of 500 WM and 50 WM, both Ni2 and Gd3 inhibited insulin- induced PI 3-kinase activation by 50% (Fig. 3). In contrast, verapamil, an inhibitor of voltage-activated Ca2 channels [35], did not signi¢cantly a¡ect the Fig. 3. E¡ects of extracellular calcium chelation (EGTA) and of stimulation of enzyme activity by insulin. Control calcium channel inhibitors on insulin-induced PI 3-kinase activ- experiment also demonstrated that neither external ity. Primary rat hepatocytes were pretreated for 15 min with or without EGTA or the speci¢ed Ca2 channel inhibitors at the indicated concentrations before stimulation with 10 nM insulin for 3 min. Top: cells lysates were prepared and proteins were immunoprecipitated with anti-p85-antibody, the immunocom- plex was assayed for PI 3-kinase activity using PI as substrate and [32P]ATP. The labeled phospholipids were resolved by thin layer chromatography then exposed to X-ray ¢lm. An autora- diogram of a representative experiment is shown. Arrow head indicates the position of the phosphorylated phosphoinositide products (PIP). Bottom: PIP's formed in the above experiment were quanti¢ed with a phosphorimager. Results are presented as the mean þ S.E.M. of eight experiments and are expressed in relation to baseline values obtained in respective controls. Dif- ferences between treatment groups were determined by analysis of variance. *Signi¢cantly di¡erent from control, 2Signi¢cantly di¡erent from insulin, P 6 0.002.

Ca2 chelation nor any of the pharmacological agents a¡ected basal PI 3-kinase activity (results Fig. 2. Dose-dependent e¡ect of insulin on PI 3-kinase activity. not illustrated). The pharmacological pro¢le of the Isolated hepatocytes were incubated with various concentrations PI 3-kinase response is similar to that obtained pre- of insulin (0, 0.1, 1,10, 100 nM) for 3 min at 37³C. PI 3-kinase viously for insulin-induced stimulation of MAPK ac- activity was measured as described in Section 2. The data 2 shown are the mean þ S.E.M. of 6 separate experiments and are tivity [17] and thus suggests that Ca in£ux may expressed in relation to respective baseline values obtained for also play an important role in this other major sig- each experiment (di¡erent hepatocyte preparation). naling pathway in liver cells [29].

BBAMCR 14570 29-12-99 K. Benzeroual et al. / Biochimica et Biophysica Acta 1495 (2000) 14^23 19

3.4. E¡ect of Ca2+ on insulin-induced tyrosine phosphorylation of IR L-subunit and of IRS-1

Moving upstream in the insulin signaling cascade, we next evaluated the implication of Ca2 movements in the insulin-induced phosphorylation of both IRS-1 and IR L-subunit. Hepatocytes were again stimulated for 3 min with 10 nM insulin. As shown in Fig. 4, insulin induced a stimulation of IRS-1 tyrosine phosphorylation as expected. There was no statistical signi¢cant di¡erence in the tyrosine phosphorylation of IRS-1 in groups pretreated with EGTA, Ni2,Gd3 or verapamil as compared to those stimulated with insulin alone. Fig. 5 presents the results obtained for IR L-sub-

Fig. 5. E¡ects of external Ca2 chelation and of Ca2 channel inhibitors on insulin-induced tyrosine phosphorylation of insulin receptor. The cells were treated as described in Fig. 4. The cells were lysed and subjected to immunoprecipitation with anti-human antibody to IR L-subunit followed by immunoblotting with anti-phosphotyrosine antibody as described in Section 2. (A) The immunoblot is representative of one of six independent experiments with similar results. (B) The bands were quanti¢ed as described in Fig. 4. Results are expressed as the mean þ S.E.M. of six experiments.

unit tyrosine phosphorylation. As expected, stimulation of hepatocytes with 10 nM insulin for 3 min yielded a signi¢cant increase in the amount of tyrosine-phosphorylated IR L-subunit. However as ob- Fig. 4. E¡ects of external Ca2 chelation and of Ca2 channel served with IRS-1, chelation of external Ca2 or pre- inhibitors on insulin receptor substrate-1 tyrosine phosphoryla- treatment with Ni2,Gd3 or verapamil did not tion. Primary rat hepatocytes were pretreated for 15 min with the various agents at the indicated concentration before stimu- signi¢cantly in£uence this action of insulin as ascer- lation with 10 nM insulin for 3 min. The cells were lysed and tained by one-way ANOVA. then subjected to immunoprecipitation with an antibody to IRS-1 followed by immunoblotting with antiphosphotyrosine antibody. (A) Representative immunoblot of eight experiments 4. Discussion with similar results. (B) The tyrosine phosphorylation of IRS-1 bands were quanti¢ed by densitometric analysis using NIH im- 2 ager. Results are expressed as the mean þ S.E.M. of eight ex- The implication of intracellular Ca as a mediator periments. in insulin action was originally proposed by Clausen

BBAMCR 14570 29-12-99 20 K. Benzeroual et al. / Biochimica et Biophysica Acta 1495 (2000) 14^23 and colleagues in 1974 [36]. Insulin was later shown direct immune-complex enzyme assay, that PI 3-ki- to increase intracellular Ca2 in normal human adi- nase activity is increased by insulin in primary hepa- pocytes [37] and in vascular smooth muscle cells [38], tocytes, as seen previously for isolated rat adipocytes but not in a skeletal muscle cell line [39]. In liver [24]. PI 3-kinase activity was also detected in the cells, it was shown that the insulin-dependent stim- basal state. Time course data and insulin dose^re- ulation of phosphodiesterases, of succinate oxidation sponse relationship show that hepatocyte PI 3-kinase and of protein synthesis was inhibited by the lack of responds rapidly to physiologically relevant concen- external Ca2 [40,41]. In 1985, Strunecka presented trations of the hormone. indirect evidence that insulin elevated steady-state We next evaluated the implication of Ca2 ions in Ca2 in digitonin-permeabilized rat hepatocyte sus- the insulin-dependent stimulation of PI 3-kinase ac- pensions [42]. Our laboratory recently demonstrated tivity in isolated cultured rat hepatocytes. The results that insulin induces an in£ux of extracellular Ca2 demonstrate that conditions previously shown to in- into primary cultures of rat hepatocytes [17]. We also hibit insulin-induced Ca2 in£ux, namely EGTA, found that interference with insulin-induced Ca2 Ni2 and Gd3, also blocked the stimulation of PI movements, using external Ca2 chelation or Ca2 3-kinase by the hormone, whereas verapamil was in- channel inhibitors, blocked the hormone's ability to e¡ective. This pharmacological pro¢le is identical to stimulate MAPK activity [17]. These data are in line the one we recently observed for the insulin-depen- with recent evidence showing that MAPK activity in dent stimulation of MAPK activity [17]. This rat hepatocytes can be modulated by Ca2-depen- strongly suggests that Ca2 in£ux also plays an im- dent mechanisms [43] as seen in other epithelial cells portant role in PI 3-kinase activation, another major [44]. Thus, our results strongly indicated that insulin- signaling pathway of insulin in liver cells [29]. induced Ca2 in£ux was a crucial component for at We also evaluated the implication of Ca2 move- least part of the insulin signaling cascade in rat liver ments in the insulin-induced phosphorylation of IR cells. We thus sought to determine if other compo- L-subunit and of IRS-1, two early steps in the insulin nents of this signaling cascade were also Ca2-de- signaling cascade that have been de¢ned at the mo- pendent. lecular level and that are essential for mediating sev- The early events occurring after insulin binding eral biological responses to insulin. In contrast to PI have been rather well documented in the liver of 3-kinase, insulin-dependent tyrosine phosphorylation intact animals [45^47]. IRS-1 has been shown to be of IR L-subunit and of IRS-1 was una¡ected by ex- a major and direct substrate of the activated insulin ternal Ca2 chelation or by the Ca2-channel block- receptor both in vivo and in vitro [9,48]. IRS-1 has ers Ni2 and Gd3. One interpretation of our data is multiple potential tyrosine phosphorylation sites thus that insulin-induced Ca2 in£ux may intervene which, once phosphorylated, serve as intracellular in the insulin signaling cascade between IRS-1 and binding sites for proteins containing SH2 domains downstream elements, such as PI 3-kinase and [3,6,9,49,50]. For instance, tyrosine phosphorylation MAPK. of IRS-1 leads to its association with the enzyme PI Indeed, it was shown that PI 3-kinase activity may 3-kinase [3,8,9,51] resulting in an increase in total PI be modulated by Ca2 [55] through a Ca2-depen- 3-kinase activity in several cell types [3,8,9,23,24,50]. dent interaction of calmodulin directly with the p110 Studies based on the use of the inhibitors wortman- catalytic subunit of the enzyme [56], as was also seen nin and LY294002 reported that PI 3-kinase is in- with the SH2 domain of the p85 regulatory subunit volved in the stimulation of glucose uptake and acti- [57]. Moreover, Ca2-dependent conformational vation of glycogen synthase induced by insulin in changes of the N-terminal SH2 domain of p85 were either cultured hepatocytes or in the whole liver, re- reported to be important for PI 3-kinase binding to spectively [52^54]. Thus, the activation of PI 3-kinase K-PDGF receptor [58] and Ca2 ions were found to by insulin in liver cells is physiologically relevant and directly activate puri¢ed PI 3-kinases in bovine thy- we assessed the possible implication of Ca2 in this mus [59]. insulin action. A number of studies also point to a relationship The results presented herein con¢rm, by using a between PI 3-kinase and MAPK pathways. For in-

BBAMCR 14570 29-12-99 K. Benzeroual et al. / Biochimica et Biophysica Acta 1495 (2000) 14^23 21 stance, wortmannin was shown to prevent the stim- since Ni2 inhibition of this channel abolished the ulation of MAPK isoform ERK-1 by insulin in a e¡ect of insulin on DNA synthesis [67]. primary hepatocyte model similar to our own, sug- In conclusion, our study demonstrates that inter- gesting that PI 3-kinase activity may lie upstream of ference of insulin-induced Ca2 in£ux into primary MAPK in liver cells [60]. In addition, PI 3-kinase can rat hepatocytes prevents the activation of PI 3-kinase form a complex with Ras [61] while expression of a by the hormone, but not the associated tyrosine constitutively active p110 catalytic subunit of PI phosphorylation of insulin receptor L-subunit or of 3-kinase can accentuate Ras-dependent cellular pro- IRS-1. Hence, Ca2 ions may intervene in the insulin cesses [62], including MAPK activation. Hence, it is signaling cascade downstream of IRS-1 or at an in- possible that the insulin-induced in£ux of Ca2 ob- dependent site upstream of both PI 3-kinase and served previously by our laboratory [17] stimulates MAPK in hepatocytes. Our studies are within the PI 3-kinase activity which, in turn, triggers the physiological framework since both PI 3-kinase and MAPK cascade. MAPK activities were shown to be important for the An interesting alternative explanation for our re- stimulation of glycogen synthesis by insulin in liver sults would implicate recently identi¢ed Ca2-de- cells [54]. Our results thus support the notion that pendent tyrosine kinases of the focal adhesion kinase insulin-induced Ca2 movements are a crucial part family, such as Pyk2 [63], which are known to medi- of the physiological action of this hormone in the ate Ca2-dependent MAPK activation in certain cell liver. types [64]. Pyk2 or Pyk2-like kinases appear to be activated early in receptor signaling cascades. For instance, Ca2-dependent tyrosine phosphorylation Acknowledgements of SHC by Pyk2 has been reported [65] and downstream activation of both PI 3-kinase and MAPK This study was supported by grants from the Med- could thus ensue. Hence, inhibition of insulin-in- ical Research Council of Canada (Grant MA-11298), duced Ca2 in£ux by EGTA or channel blockers from the Canadian Diabetes Association and from could prevent Pyk2 activation. Finally, although the Fonds pour la Formation des chercheurs et l'aide the phosphorylation of the insulin receptor and of a© la Recherche (F.C.A.R.; Grant 97-ER-2116). K.B. IRS-1 are not a¡ected by our experimental modula- was the recipient of a studentship from the Groupe tion of Ca2 in£ux, our results do not rule out that de Recherche en Transport Membranaire made pos- prevention of Ca2 in£ux may interfere with the for- sible by infrastructure grants from the F.C.A.R. and mation of complexes between IRS-1 and other e¡ec- the Universite¨ de Montre¨al. P.H. is a scholar of the tors, such as GRB2 or the p85 regulatory subunit of Fonds de la recherche en sante¨ du Que¨bec. The au- PI 3-kinase. Further experiments will be necessary to thors wish to express their gratitude to Ms. Elisabeth elucidate the exact site of action of Ca2 in£ux in the Peres for excellent assistance with graphic art. insulin signal transduction cascade. The means by which insulin triggers Ca2 in£ux into liver cells is also unclear. The pharmacological sensitivity of this in£ux and of insulin-stimulated PI References 3-kinase and MAPK activities is compatible with [1] O.M. Rosen, Science 237 (1987) 1452^1458. that of non-selective cation channels present on hep- [2] M. Kasuga, F.A. Karlsson, C.R. Kahn, Science 215 (1982) atocytes [18]. It is thus possible that these channels 185^187. mediate the e¡ect of insulin on hepatocellular Ca2, [3] X.J. Sun, P. Rothenberg, C.R. Kahn, J.M. Backer, E. Araki, as supported by our preliminary observations using P.A. Wilden, D.A. Cahill, B.J. Goldstein, M.F. White, Na- the patch clamp technique [66]. Coupling between ture 352 (1991) 73^77. the insulin receptor and cation channels is not un- [4] H. Maegawa, J.M. Olefsky, S. Thies, D. Boyd, A. Ullrich, D.A. McClaim, J. Biol. Chem. 263 (1988) 12629^12637. precedented since insulin was shown to activate a 10- [5] Y. Ebina, E. Araki, M. Taira, F. Shimada, M. Morri, C.S. pS calcium-permeable channel in CHO-IR cells [67]. Craik, K. Siddle, S.B. Pierce, R.A. Roth, W.J. Rutter, Proc. Moreover, this coupling was physiologically relevant Natl. Acad. Sci. USA 84 (1987) 704^708.

BBAMCR 14570 29-12-99 22 K. Benzeroual et al. / Biochimica et Biophysica Acta 1495 (2000) 14^23

[6] B.E. Lavan, M.R. Kuhne, C.W. Garner, D. Anderson, M. [31] M.H. Nathanson, A. Gautam, R. Bruck, C.M. Isales, J.L. Reedijk, T. Pawson, G.E. Lienhard, J. Biol. Chem. 267 Boyer, Hepatology 15 (1992) 107^116. (1992) 11631^11636. [32] C. Bear, C. Li, Am. J. Physiol. 261 (1991) C1018^C1024. [7] X.-J. Sun, M. Miralpeix, M.G.J. Myers, E. Glasheen, J.M. [33] L.A. Berven, B.P. Hughes, G.J. Barritt, Biochem. J. 299 Backer, C.R. Kahn, M.F. White, J. Biol. Chem. 267 (1992) (1994) 399^407. 22662^22672. [34] K.C. Fernando, G.J. Barritt, Biochim. Biophys. Acta 1268 [8] Y.R. Hadari, E. Tzahar, O. Nadiv, P. Rothenberg, C.T. (1995) 97^106. Roberts Jr., D. LeRoith, Y. Yarden, Y. Zick, J. Biol. [35] W.A. Catterall, J. Striessnig, Trends Pharmacol. Sci. 13 Chem. 267 (1992) 17483^17486. (1992) 256^262. [9] J.M. Backer, M.G. Myers Jr., S.E. Shoelson, D.J. Chin, X.J. [36] T. Clausen, J. Elbrink, B.R. Martin, Acta Endocrinol. 191 Sun, M. Miralpeix, P. Hu, B. Margolis, E.Y. Skolnik, J. (1974) 137^143. Schlessinger, M.F. White, EMBO J. 11 (1992) 3469^3479. [37] B. Draznin, K.E. Sussman, R.H. Eckel, M. Kao, T. Yost, [10] L.C. Cantley, K.R. Auger, C.L. Carpenter, B. Duckworth, N.A. Sherman, J. Clin. Invest. 82 (1988) 1848^1852. A. Graziani, R. Kapeller, S. Solto¡, Cell 64 (1991) 281^302. [38] R.M. Touyz, E.L. Schi¡rin, Hypertension 23 (1994) 931^ [11] I.D. Hiles, M. Otsu, S. Volinia, M.J. Fry, I. Gout, R. 935. Dhand, G. Panayotou, F. Ruiz-larrea, N.F. Thompson [39] A. Klip, G. Li, W.J. Logan, Am. J. Physiol. 247 (1984) Totty, Cell 70 (1992) 419^429. E297^E304. [12] B. Cheatham, C.J. Vlahos, L. Cheatham, L. Wang, J. Blenis, [40] K.S. Rogers, R.A. Menon, C. Mohan, P.J. Geiger, S.P. Bess- C.R. Kahn, Mol. Cell. Biol. 14 (1994) 4902^4911. man, Biochem. Med. Metab. Biol. 45 (1991) 344^349. [13] P.R. Shepherd, B.T. Nave, K. Siddle, Biochem. J. 305 (1995) [41] F. Irvine, A.V. Wallace, S.R. Sarawak, M.D. Houslay, Bio- 25^28. chem. J. 293 (1993) 249^253. [14] R.A. Gabbay, C. Sutherland, L. Gnudi, B.B. Kahn, R.M. [42] A. Strunecka, Gen. Physiol. Biophys. 4 (1985) 505^516. O'Brien, D.K. Granne, J.S. Flier, J. Biol. Chem. 271 (1996) [43] A. Romanelli, G. Van de Werve, Metabolism 46 (1997) 548^ 1890^1897. 555. [15] H. Osawa, C. Sutherland, R.B. Robey, R.L. Printz, D.K. [44] T.-S.O. Chao, K.L. Byron, K.-M. Lee, M. Villereal, M.R. Granne, J. Biol. Chem. 271 (1996) 16690^16694. Rosner, J. Biol. Chem. 267 (1992) 19876^19883. [16] L. Agius, M. Peak, G. Beresford, M. Al-Habori, T.H. Tho- [45] M.J. Saad, F. Folli, C.R. Kahn, J. Clin. Invest. 92 (1993) mas, Biochem. Soc. Trans. 22 (1994) 516^522. 2065^2072. [17] K. Benzeroual, G. Van de Werve, S. Meloche, L. Mathe¨,A. [46] K. Tobe, T. Kadowaki, K. Hara, Y. Gotoh, H. Kosako, S. Romanelli, P. Haddad, Am. J. Physiol. 272 (1997) G1425^ Matsuda, H. Tamemoto, K. Ueki, Y. Akanuma, E. Nishida, G1432. Y. Yazaki, J. Biol. Chem. 267 (1992) 21089^21097. [18] S.D. Lidofsky, Hepatology 22 (1995) 306A (Abstract). [47] Y. Ito, S.I. Takahashi, T. Hidaka, T. Noguchi, J. Endocri- [19] S.D. Lidofsky, M.H. Xie, A. Sostman, B.F. Scharschmidt, nol. 154 (1997) 267^273. J.G. Fitz, J. Biol. Chem. 268 (1993) 14632^14636. [48] P. Rothenberg, W.S. Lane, A. Karasik, J. Backer, M.F. [20] J. Haimovici, J.S. Beck, D. Vallerand, C. Molla-Hosseini, P. White, C.R. Kahn, J. Biol. Chem. 266 (1991) 8302^8311. Haddad, Am. J. Physiol. 267 (1994) G364^G370. [49] M.F. White, C.R. Kahn, J. Biol. Chem. 269 (1994) 1^4. [21] Y. Fukui, H. Hanafusa, Mol. Cell. Biol. 9 (1989) 1651^1658. [50] K. Yonesawa, H. Ueda, K. Hara, K. Nishida, A. Ando, A. [22] S.K. Pandey, M.-B. Anand-Srivastava, A.K. Srivastava, Bio- Chavanieu, H. Matsuba, K. Shii, K. Yokono, Y. Fukui, B. chemistry 37 (1998) 7006^7014. Calas, F. Grigorescu, R. Dhand, I. Gout, M. Otsu, M.D. [23] N.B. Ruderman, R. Kapeller, M.F. White, L.C. Cantley, Water¢eld, M. Kasuga, J. Biol. Chem. 267 (1992) 25958^ Proc. Natl. Acad. Sci. USA 87 (1990) 1411^1415. 25966. [24] K.L. Kelly, N.B. Ruderman, K.S. Chen, J. Biol. Chem. 267 [51] R. Kapeller, K.S. Chen, M. Yoakin, B.S. Scha¡hausen, J.M. (1992) 3423^3428. Backer, M.F. White, L.C. Cantley, N.B. Ruderman, Mol. [25] F. D'Onofrio, M.Q. Le, J.L. Chiasson, A.K. Srivastava, Endocrinol. 5 (1991) 769^777. FEBS Lett. 340 (1994) 269^275. [52] Y. Le Marchand-Brustel, N. Gautier, M. Cormont, E. Van [26] P.B. Lynedjian, D. Jotterand, T. Nouspiket, J. Biol. Chem. 9 Obberghen, Endocrinology 136 (1995) 3566^3570. (1989) 1651^1989. [53] T. Tsakiridis, H.E. McDowell, T. Walker, C.P. Downes, [27] O. Morin, M. Fehlmann, P. Freychet, Mol. Cell. Endocrinol. H.S. Hundal, M. Vranic, A. Klip, Endocrinology 136 (10) 25 (1982) 339^352. (1995) 4314^4322. [28] Karam, J.H. (1995) in Basic and Clinical Pharmacology [54] J. Carlsen, K. Christiansen, J. Vinten, Cell. Signal. 9 (1997) (Katzung, B.G., ed.), pp. 637^654, Prentice Hall Canada, 447^450. toronto. [55] S.S. Singh, A. Chauhan, H. Brockerho¡, V.P. Chauhan, Life [29] F. Folli, M.J.A. Saad, J.M. Backer, C.R. Kahn, J. Biol. Sci. 57 (1995) 685^694. Chem. 267 (1992) 22171^22177. [56] R. Fischer, J. Julsgart, M.W. Berchtolt, FEBS Lett. 425 [30] G.E.N. Kass, J. Llopis, S.C. Chow, S.K. Duddy, S. Orre- (1999) 175^177. nius, J. Biol. Chem. 265 (1990) 17486^17492. [57] J.L. Joyal, D.J. Burks, S. Pons, W.F. Matter, C.J. Vlahos,

BBAMCR 14570 29-12-99 K. Benzeroual et al. / Biochimica et Biophysica Acta 1495 (2000) 14^23 23

M.F. White, D.B. Sacks, J. Biol. Chem. 272 (1997) 28183^ [63] S. Lev, H. Moreno, R. Martinez, P. Canoll, E. Peles, J.M. 28186. Musacchio, G.D. Plowman, B. Rudy, J. Schlessinger, Nature [58] D. Mahadevan, N. Thanki, P. McPhie, J.F. Beeler, J.-C. Yu, 376 (1995) 737^745. A. Wlodawer, M.A. Heidaran, Biochemistry 33 (1994) 746^ [64] L.M. Luttrell, Y. Daaka, R.J. Lefkowitz, Curr. Opin. Cell 754. Biol. 11 (1999) 177^183. [59] F. Shibasaki, Y. Homma, T. Takenawa, J. Biol. Chem. 266 [65] D.D. Schlaepfer, C.R. Hauck, D.J. Sieg, Prog. Biophys. (1991) 8108^8114. Mol. Biol. 71 (1999) 435^478. [60] C.J. Band, B.I. Posner, J. Biol. Chem. 272 (1997) 138^145. [66] Mathe¨, L., Vallerand, D. and Haddad, P. (1997) Hepatology [61] P. Rodriguez-Viciana, P.H. Warne, R. Dhand, B. Vanhae- 26, 263A (Abstract). serbroeck, I. Gout, M.J. Fry, M.D. Water¢eld, J. Down- [67] L. Nie, M. Kanzaki, H. Shibata, I. Kojima, Endocrinology ward, Nature 370 (1994) 527^532. 139 (1998) 179^188. [62] Q. Hu, A. Klippel, A.J. Muslin, W.J. Fantl, L.T. William, Science 268 (1995) 100^102.

BBAMCR 14570 29-12-99