GIPC participates in G protein signaling downstream of insulin-like growth factor 1 receptor.

Several recent studies have demonstrated that insulin-like growth factor (IGF)-1-induced mitogen-activated protein kinase (MAP kinase) activation is abolished by pertussis toxin, suggesting that trimeric G proteins of the G(i) class are novel cellular targets of the IGF-1 signaling pathway. We report here that the intracellular domain of the Xenopus IGF-1 receptor is capable of binding to the Xenopus homolog of mammalian GIPC, a PDZ domain-containing protein previously identified as a binding partner of G(i)-specific GAP (RGS-GAIP). Binding of xGIPC to xIGF-1 receptor is independent of the kinase activity of the receptor and appears to require the PDZ domain of xGIPC. Injection of two C-terminal truncation mutants that retained the PDZ domain blocked IGF-1-induced Xenopus MAP kinase activation and oocyte maturation. While full-length xGIPC injection did not significantly alter insulin response, it greatly enhanced human RGS-GAIP in stimulating the insulin response in frog oocytes. This represents the first demonstration that GIPC x RGS-GAIP complex acts positively in IGF-1 receptor signal transduction.

Insulin-like growth factor 1 (IGF-1) 1 exerts its biological roles by activating the intrinsic protein tyrosine kinase activity of the IGF-1 receptor. The activated IGF-1 receptor autophosphorylates its cytoplasmic domain and phosphorylates insulin receptor substrate 1 (IRS-1) and many other protein substrates. Phosphorylation of these protein substrates leads to changes in multiple intracellular signaling pathways including the Ras-Raf-MAP kinase pathway and the phosphatidylinositol 3-kinase (PI 3-kinase)/Akt pathway (1). However, these same signaling pathways are also similarly regulated by insulin/ insulin receptor and other growth factors such as epidermal growth factor and platelet-derived growth factor. Clearly, simple tissue-and time-specific expressions of the receptors and/or components of these signaling pathways do not fully explain the distinct biological activities of these various growth factors.
Several recent studies have demonstrated that IGF-1 receptors activate heterotrimeric G proteins in some cell types. Luttrell et al. (2) first demonstrated that IGF-1-induced MAP kinase activation in rat 1 fibroblasts was inhibited by either treating cells with pertussis toxin or transfecting them with a G␤␥ scavenger (␤ARK-CT), suggesting the involvement of a pertussis toxin-sensitive G protein (i.e. of the G i class). This study further suggests that the released G␤␥ subunits, rather than the activated GTP-bound G␣ subunit, is responsible for IGF-1-induced MAP kinase activation. Similar results have since been obtained using human intestinal smooth muscle cells (3), 3T3-L1 mouse pre-adipose cells (4), and rat cerebellar granule neurons (5). These latter studies further demonstrated that the IGF-1 receptor forms a complex with a G protein ␣ subunit of the G i class with its associated G␤ subunit (and presumably G␥ subunit) (4,5). Kuemmerle and Murthy (3) were able to demonstrate that IGF-1 selectively activated G␣ i2 but not G␣ i1 , G␣ i3 , or G␣ q in the same cells. As heterotrimeric G proteins are known to be associated with, and activated by, heptahelical G protein-coupled receptors, it remains unclear how IGF-1/IGF-1 receptors activate the G i proteins.
We have been studying IGF-1 receptor signaling in the induction of oocyte maturation in Xenopus laevis (6,7). We have used the cytoplasmic portion of the xIGF-1 receptor in a yeast two-hybrid screen in an attempt to isolate potential downstream signaling proteins. In this paper, we report the identification of the Xenopus homolog of GIPC as an xIGF-1 receptorbinding protein and provide evidence supporting a functional role for xGIPC in the regulation of G protein signaling downstream of the xIGF-1 receptor in the induction of oocyte maturation.

EXPERIMENTAL PROCEDURES
Animal and Oocyte Manipulation-All procedures involving live oocytes were carried out in a room maintained at between 18 and 20°C. Sexually mature, oocyte-positive X. laevis were purchased from NASCO and maintained according to local animal care guidelines. The frogs were injected with pregnant mare serum gonadotropin (Sigma, 50 IU/ frog) 3-10 days before operations. A fragment of ovary was removed surgically under hypothermia. Stage VI oocytes were manually defolliculated and stored in oocyte incubation medium OR2 (82.5 mM NaCl, 2.5 mM KCl, 1 mM CaCl 2 , 1 mM MgCl 2 , 1 mM Na 2 HPO 4 , 5 mM Hepes, pH 7.8).
Cloning and cDNA Manipulation-The nucleotide sequence encoding the cytoplasmic domain (amino acids 958 -1358) of the Xenopus IGF-1R (7) was PCR-amplified using the following primers: forward primer, 5Ј-TATG AAT TCT AAG AAG AGA AAC AGC AAC C-3Ј; reverse primer, 5Ј-TAT GAA TTC ACT GAT ACA GCG GG-3Ј. The amplified cDNA was digested with EcoRI, treated with the Klenow fragment, and ligated into the pAS2 (CLONTECH) vector that had been digested with BamHI and treated with the Klenow fragment. The resulting clone, designated pAS-xIGF-1R cyto , expressed a fusion protein between the GAL4 DNA-binding domain and the xIGF-1R cytoplasmic domain. The kinase-deficient mutant of xIGF-1R (xIGF-1RKA cyto ) in pAS2 was constructed by the two-step PCR procedure (8) using the same forward and reverse primers (above) in combination with the following internal primers changing the catalytically essential Lys-1029 to Ala: forward primer, 5Ј-A GTT GCC ATA GCG ACG GTC AAC  G-3Ј; reverse primer, 5Ј-C GTT GAC CGT CGC TAT GGC AAC TTT  C-3Ј. To create xIGF-1R cyto and xIGF-1RKA cyto for transfection or  mRNA synthesis, the same PCR products were digested with EcoRI and  ligated into pCS2ϩMyc previously treated with EcoRI. Upon sequence analysis, YA 5-2, the clone identified in the yeast two-hybrid screen, was found to have its open reading frame (ORF) (xGIPC) inserted in a reverse orientation relative to the GAL4 activation domain (AD). To clone xGIPC in-frame with the GAL4AD, a 5Ј PCR primer was designed to amplify the coding region (5Ј-TAT GAA TTC ATG CCT CTG GGA TTG CGC GTA AAG-3Ј). This primer, along with a pGAD10 (two-hybrid vector)-specific primer, was used to amplify the ORF. The PCR product was digested with EcoRI and ligated into the pGAD10 previously digested with EcoRI. To create an xGIPC for transfection and mRNA synthesis, the same PCR product was digested with EcoRI and treated with Klenow before ligation into pCS2ϩHA previously digested with XbaI and treated with Klenow.
pCS2ϩHA was modified from pCS2ϩ (9) as follows. Two complementary oligonucleotides were made to code for the hemagglutinin (HA) epitope (YPYDVPDYA) with a translation initiation codon and cohesive BamHI ends. The sequences of the two oligonucleotides were as follows: forward, 5Ј-GAT CCA CCA TGT ACC CAT ACG ATG TTC CAG ATT ACG CTT CCA TG-3Ј; reverse, 5Ј-GAT CCA TGG AAG CGT AAT CTG GAA CAT CGT ATG GGT ACA TGG TG-3Ј. The oligonucleotides were annealed to each other and then directly ligated into pCS2ϩ vector previously digested with BamHI.
Subclones of xGIPC for mapping the binding region of xGIPC and for injection into oocytes were similarly generated by PCR using primers containing an EcoRI site. The PCR fragments were ligated into either pGAD10 vector or pCS2ϩHA vector. The resulting clones generated were fusion proteins with the GAL4AD or a HA tag, respectively. For the sake of brevity, only the amino acids comprising the various constructs are indicated in the figures. To reconstitute full-length xGIPC, we PCR-amplified xGIPC-(1-320) with a reversed primer encoding 11 extra amino acids (GAIGDAKQGRF) derived from the sequence of the Xenopus expressed sequence tag clone (10).
The human GAIP coding sequence was PCR-amplified from a human brain cDNA library (a gift of J. K. Ngsee) using the following primers: forward primer, 5Ј-TAT GAA TTC A AT GCC CAC CCC GCA TGA GG-3Ј; reverse primer, 5Ј-TAT GAA TTC TGT GCT GCT GGG GGC GGC C-3Ј. The amplified sequence was digested with EcoRI and inserted at the EcoRI site of pCS2ϩ or pCS2ϩMT (9), resulting in the expression of an untagged or Myc-tagged version of human GAIP. The identity of the sequence was confirmed by DNA sequence determination.
Yeast Two-hybrid Analysis-These procedures are essentially the same as those described in the Yeast Protocol Handbook provided by CLONTECH. Saccharomyces cerevisiae strain Y190 was co-transformed with the bait plasmid (pAS-xIGF-1R cyto ) and a X. laevis oocyte cDNA library constructed in the pGAD10 vector (CLONTECH). Transformants were plated on 150-mm synthetic dropout medium (Yeast Protocol Handbook, CLONTECH)/His Ϫ /Leu Ϫ /Trp Ϫ /3-amino-1,2,4-triazole (50 mM) plates and incubated at 30°C until colonies reached 1-2 mm. ␤-Galactosidase assays for the detection of potential positive clones were performed according to the protocols supplied by CLONTECH. Potential positive clones were grown in liquid synthetic dropout medium/Leu Ϫ , and the pGAD10 library plasmid was isolated and transformed into Escherichia coli strain MH6 by electroporation. The transformants were plated on M9 media (11 mM Na 2 HPO 4 , 22 mM KH 2 PO 4 , 8.5 mM NaCl, 19 mM NH 4 Cl, 1 mM MgSO 4 , 100 M CaCl 2 , 2 mg/ml thiamine, 20 mg/liter uracil, 100 g/ml ampicillin) and incubated at 37°C for 20 h. The isolated plasmid was used in co-transformation assays with the original bait plated on synthetic dropout medium/Leu Ϫ / Trp Ϫ plates. Colonies were tested for filter-based ␤-galactosidase activity. Yeast protein extraction for immunoblotting was carried out using the urea/SDS method supplied by CLONTECH.
Polyclonal Antibodies against xGIPC-An internal HindIII fragment of xGIPC, encoding amino acids 216 -305, was excised and inserted into pGEX-KT (11). Glutathione S-transferase fusion proteins were induced and purified by binding to glutathione-agarose beads. Immunization of rabbits was carried out according to Harlow and Lane (12). Immune sera were used without further processing.
Co-immunoprecipitation Experiments-COS-7 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 50 units/ml penicillin-streptomycin, and 2 ml/liter Fungizone (Invitrogen) at 37°C in a humidified atmosphere containing 5% CO 2 . Transient transfection of COS-7 cells was accomplished using Lipo-fectAMINE (Invitrogen). Briefly, cells in a six-well dish at 70% confluence were washed once with PBS and incubated with 1 ml of OPTI-MEM medium containing 5 l of LipofectAMINE reagent and 0.25 g of each plasmid for 18 h. Following the transfection, the medium was changed to Dulbecco's modified Eagle's medium with 10% fetal calf serum with antibiotics. The cells were allowed to grow for a further 24 -30 h and lysed for co-immunoprecipitation or kinase assays.
Transiently transfected COS-7 cells were washed with PBS and then lysed with 200 l of ice-cold PBS-Lysis buffer (13) (10 mM phosphate buffer, pH 7.5, 150 mM NaCl, 1% Triton X-100, 100 M phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 100 M sodium orthovanadate). The lysate was cleared by a 10-min centrifugation at 15,000 ϫ g. 150 l of the cleared lysate was added to 450 l of PBS-Lysis buffer containing 10 l of anti-Myc-or anti-HA-Sepharose beads (14). Lysates were incubated at 4°C with rocking for 2 h. Immunocomplexes were washed five times with PBS-Lysis buffer; after washing, 20 l of 2ϫ Laemmli sample buffer was added, and the samples were separated using SDS-PAGE. The proteins were transferred to a nitrocellulose membrane for Western blotting.
Co-immunoprecipitation experiments using frog oocyte extracts were performed essentially the same way. Oocytes were injected with the various mRNA and incubated overnight. Extracts were prepared in PBS-Lysis buffer (10 l/oocyte) and subjected to immunoprecipitation (200 -400 l of extract with 5 l of either preimmune serum or anti-xGIPC immune serum).
Protein Kinase Assays-Protein kinase assays were performed on immunoprecipitated xIGF-1R cyto or xIGF-1RKA cyto . Briefly, immunoprecipitation procedures were carried out as above with the exception that the last wash step was done with kinase buffer (50 mM Hepes, pH 6.9, 1 mM EDTA, 0.1% Triton X-100). After the final wash, 23 l of kinase buffer were added, and the reaction was started with the addition of 10 Ci of [␥-32 P]ATP, 1 M ATP, and 5 mM MnCl 2 . The kinase reaction was carried out at room temperature for 30 min and was stopped with the addition of 2ϫ Laemmli sample buffer.

RESULTS
We inserted the cytoplasmic domain of xIGF-1 receptor (7) into the bait vector pAS2. The resultant plasmid (pAS-xIGF- 1R cyto ) encodes an HA epitope tag followed by the DNA-binding domain of GAL4 and then by the xIGF-1 receptor cytoplasmic domain. We also constructed similar bait (pAS-xIGF-1RKA cyto ) with a single point mutation changing the catalytically essential lysine 1029 to alanine. Transforming yeast with each plasmid resulted in expression of the fusion protein with the anticipated size as indicated by both anti-xIGF-1 receptor blots and anti-HA blots (Fig. 1). As expected, pAS-xIGF-1R cyto was phosphorylated on tyrosine, whereas tyrosine phosphorylation of pAS-xIGF-1RKA cyto was undetectable.
Using pAS-xIGF-1R cyto as bait, we screened an oocyte cDNA library constructed in the pGAD10 vector (encoding the GAL4AD) (by CLONTECH). One particular clone, YA5-2, caught our attention since it contained an ORF highly homologous to a mammalian gene, GIPC, previously identified as a binding partner for a G i -specific GTPase-activating protein (RGS-GAIP or GAIP) (16). We thought that perhaps GIPC might provide a functional link between IGF-1 receptor and G i protein signaling. However, the ORF of 320 amino acids, which included a candidate translation start site but lacked a translation termination codon, was inserted in pGAD10 with a reverse orientation. Therefore xGIPC could not have been made as part of a GAL4AD fusion protein. This clone was therefore initially discarded as "false positive" until one of us (R. A. B.) found, in searching the literature, that the ADH1 termination sequences actually contain functional promoter sequences capable of initiating transcription opposite to the ADH promoter in pGAD10 (17). Consequently xGIPC could have been expressed in yeast from the promoter contained within the ADH1 termination sequences as a nonfusion protein.
The cloned xGIPC was sequenced from both directions, and the sequences assembled and compared with GIPC from several other species (Fig. 2). From these comparisons, it seemed that YA 5-2 only contained partial xGIPC coding sequences lacking the extreme C terminus of perhaps as few as 11 amino acids. However, repeated attempts to amplify the missing C terminus by either 3Ј-rapid amplification of cDNA ends or anchored PCR were unsuccessful (data not shown).
We decided to carry out binding experiments in transiently transfected COS-7 cells following an unsuccessful attempt to demonstrate co-immunoprecipitation of endogenous xIGF-1 receptor and xGIPC in frog oocytes (see below). To facilitate these experiments, we subcloned xIGF-1R cyto or xIGF-1RKA cyto into pCS2ϩMyc (9) and xGIPC or its truncation mutants (Fig. 3A) into pCS2ϩHA. Transient transfection of Myc-xIGF-1R cyto or Myc-xIGF-1RKA cyto resulted in similar levels of the two proteins, recognized by anti-Myc (Fig. 4A, lower panel). Immunoprecipitation followed by an in vitro kinase assay confirmed that Myc-xIGF-1RKA cyto was indeed deficient in catalytic activity (Fig. 4A, upper panel). Co-transfection of each plasmid with HA-xGIPC followed by co-immunoprecipitation experiments clearly indicated that xGIPC was able to bind both the kinase-active and kinase-deficient cytoplasmic domain of the xIGF-1 receptor (Fig. 4B, lanes 7 and 8). Lanes 5 and 6 in Fig.  4B represent HA immunoprecipitation from extracts derived from cells expressing only the Myc-tagged proteins, indicating that the co-immunoprecipitation observed in Fig. 4B, lanes 7 and 8, were not due to the overexpressed Myc fusion proteins being "pulled down" by anti-HA antibodies nonspecifically.
We wished to determine whether overexpression of xGIPC-(1-320) and the various truncation mutants had any effect on IGF-1 receptor signaling in frog oocytes. mRNA encoding green fluorescent protein (18) (used as a control), HA-xGIPC-(1-320), or HA-xGIPC PDZ were injected into immature oocytes. Following overnight incubation to allow translation and accumulation of the corresponding proteins, oocytes were incubated with insulin (200 or 500 nM). We have previously shown that these concentrations of insulin activate the endogenous xIGF-1 receptor (7,19) and ultimately cause oocyte maturation, indicated by, among other criteria, xMAP kinase activation and GVBD. Fig. 5A shows that injection of the PDZ domain mRNA had no significant effect (compared with similar amount of green fluorescent protein mRNA or water) on insulin-induced GVBD. In contrast, injection of xGIPC-(1-320) mRNA severely reduced the ability of insulin to induce GVBD. Fig. 5B shows a typical experiment where we analyzed xMAP kinase phosphorylation (indicative of activation (20, 21)) following scoring oocytes for GVBD (Fig. 5A). Again, xGIPC-(1-320), but not the PDZ domain in isolation, significantly reduced the ability of insulin to activate xMAP kinase phosphorylation (activation). Immunoblotting using anti-HA antibodies demonstrated that both xGIPC-(1-320) and HA-xGIPC PDZ were expressed and accumulated in oocytes (Fig. 5A, inset). Similarly, injection of mRNA encoding either the N terminus or the C terminus of xGIPC did not significantly change insulin-induced xMAP kinase phosphorylation (Fig. 5C) or GVBD (not shown).
A recently deposited Xenopus expressed sequence tag clone (GenBank TM accession no. AW646188) (10) apparently contains the missing 3Ј-end of our xGIPC sequence. This clone contains 240 bp of coding sequence that is identical to the corresponding region of xGIPC (with the exception, of course, of the 11 extra codons that are missing from our clone). Fulllength xGIPC was therefore generated by PCR amplification of xGIPC-(1-320) with a 3Ј-primer that contained 11 extra codons (GAIGDAKQGRF). To our surprise, the full-length xGIPC, unlike xGIPC-(1-320), did not inhibit insulin-induced oocyte maturation assayed by three different criteria: GVBD (Fig. 6A), MPF activation (Fig. 6B), and xMAP kinase phosphorylation (Fig. 6C). In contrast a C-terminal deletion mutant (xGIPC N-PDZ), which was capable of binding xIGF-1 receptor cytoplasmic domain in the yeast assays (Fig. 3B), significantly reduced insulin-induced oocyte maturation. Immunoblotting indicated that xGIPC and xGIPC N-PDZ were expressed at similar levels in the respectively injected oocytes (Fig. 6A, inset). The inhibitory effect of xGIPC N-PDZ was reminiscent of that of xGIPC-(1-320) (Fig. 5, A and B).
To further investigate the mechanism of xGIPC in insulin signaling, we explored the possible functional interaction between xGIPC and GAIP in frog oocytes. Human GAIP (hGAIP) cDNA was PCR-amplified and expressed with or without an N-terminal Myc tag. Our preliminary injection experiments indicated that hGAIP significantly accelerated insulin-induced GVBD (not shown). We used a concentration of insulin (50 nM) that was not sufficient to induce oocyte maturation in most batches of frog oocytes. Under these conditions, injection of hGAIP mRNA (either untagged or Myc-tagged) resulted in oocyte maturation in a significant percentage of oocytes (Fig. 7,  A and B). Although xGIPC alone had little effect, it greatly enhanced the ability of hGAIP in stimulating insulin-induced oocyte maturation (Fig. 7, A and B). To confirm that xGIPC bound hGAIP in frog oocytes, we performed co-immunoprecipitation experiments. Oocytes were injected with Myc-hGAIP alone or together with HA-xGIPC, HA-xGIPC-(1-320), or HA-xGIPC N-PDZ . Immunoprecipitation with antibodies against xGIPC pulled down both endogenous xGIPC (not shown) and all three forms of mRNA-derived xGIPC (Fig. 7C, lower panel). Co-immunoprecipitation of Myc-hGAIP with the endogenous xGIPC was evident (Fig. 7C, lane 5 compared with lane 1, which represents preimmune control). The amounts of Myc-hGAIP co-precipitated in xGIPC immunoprecipitations increased significantly in all three groups of oocytes that had been injected with the various xGIPC mRNAs (Fig. 7C, lanes  6 -8 as compared with lane 5). These results clearly indicated the ability of all three forms of xGIPC to bind Myc-hGAIP. As would be expected, nonspecific binding of Myc-GAIP in preimmune immunoprecipitations (Fig. 7C, lanes 1-4) did not change regardless of whether oocytes had received the xGIPC mRNA injection.  5. xGIPC-(1-320) blocks insulin-induced oocyte maturation. A, oocytes were injected with mRNA for green fluorescent protein (GFP) (as a control), xGIPC-(1-320), or xGIPC PDZ domain. After 8 h of incubation the oocytes were stimulated with 0.2 M insulin. Oocytes were scored for GVBD following an overnight incubation with insulin. Shown are the means (with standard deviations) of three to five experiments with the total number of oocytes indicated above each bar. The inset shows an anti-HA immunoblot indicating protein expression in mRNA-injected oocytes. Cont, control. B, oocyte were treated similarly as in A with the exception that two different concentrations of insulin were used. Following an overnight incubation with insulin, extracts were prepared and subjected to immunoblotting with anti-xMAP kinase antibodies. The upper band and the lower band represent phosphorylated (or active) and unphosphorylated (or inactive) Xenopus MAP kinase, respectively. C, oocytes injected with water or mRNA encoding xGIPC-(1-320), the N terminus (N), or the C terminus (C) were treated with 0.2 M insulin overnight. Oocyte extracts were prepared and analyzed by immunoblotting with anti-xMAP kinase antibodies.

DISCUSSION
The original xGIPC-(1-320) clone (YA 5-2) isolated in the yeast two-hybrid screen was inserted in the pGAD10 vector with a reverse orientation relative to that of GAL4AD. This unusual result meant that xGIPC-(1-320) was not made as a GAL4AD fusion protein or driven by the ADH promoter. Instead it was transcribed from an intrinsic promoter lying within the ADH1 termination sequences (17) and translated as a nonfusion protein. Under this circumstance, xGIPC-(1-320) should bind the bait (GAL4DBD-xIGF-1R cyto ) but should also function as a "transcription activator" capable of replacing the GAL4AD in activating the Gal1-driven reporters in the twohybrid assay. This is not surprising given that more than 1% of random bacterial genomic fragments demonstrate transcription activator function when fused in frame with the yeast GAL4 DNA-binding domain (22). The common feature of these transcription activators is the relative abundance of acidic amino acids (22). Indeed, xGIPC-(1-320) is an acidic protein (overall pI of 6.16) with a particularly acidic C terminus (pI of 4.71 over the last 114 amino acids). A second possibility was for xGIPC-(1-320) to simultaneously bind the bait and a GAL4AD fusion protein derived from the nonsense direction of xGIPC-(1-320).
A recent study (23) indicates that the PDZ domain of mam-malian GIPC is sufficient to mediate interaction to TrkA (receptor for nerve growth factor). Lou et al. (23) have also defined the juxtamembrane domain of TrkA as the likely docking site FIG. 6. xGIPC N-PDZ blocked insulin-induced oocyte maturation. A, oocytes were injected with water (Control) or mRNA for xGIPC N-PDZ or full-length xGIPC. After 8 h of incubation the oocytes were stimulated with 0.2 M insulin. Oocytes were scored for GVBD following an overnight incubation with insulin. Shown are the means of two experiments with the total number of oocytes indicated above each bar. The inset shows an anti-HA immunoblot indicating expression of the corresponding protein in mRNA-injected oocytes. B, oocytes injected with water or mRNA were incubated with the indicated concentrations of insulin for 9 h. MPF extracts were prepared and subjected to in vitro kinase assays using histone H1 as a substrate. C, the same extracts from B were immunoblotted with anti-xMAP kinase antibodies.
FIG. 7. xGIPC synergized with hGAIP in stimulating insulin response in frog oocytes. A, oocytes were injected with water (Control) or the indicated mRNA. After 8 h of incubation the oocytes were stimulated with 0.05 M insulin. Oocytes were scored for GVBD following an overnight incubation with insulin. Shown are means of three independent experiments with the total number of oocytes indicated above each bar. B, following GVBD scoring, MPF extracts were prepared and subjected to in vitro kinase assays using histone H1 as a substrate (upper panel) or immunoblotted with anti-xMAP kinase antibodies (lower panel). Shown is a representative of three independent experiments. C, oocytes were first injected with mRNA for Myc-hGAIP followed by injection with mRNA for the indicated xGIPC forms. Following an overnight incubation, extracts were prepared and immunoprecipitated with either preimmune serum or anti-xGIPC serum. The immunoprecipitations (IPs) were immunoblotted with anti-Myc antibodies (upper panel). The blots were reprobed with anti-HA antibodies (lower panel) following stripping of anti-Myc-antibodies. Shown is a representative of two independent experiments. for the GIPC PDZ domain. Limited, but noticeable, sequence homology was indeed found between the juxtamembrane domains of TrkA and xIGF-1 receptor (not shown). We have not yet determined whether this region of the xIGF-1 receptor is similarly required for interacting with xGIPC. Our data also indicate that both xGIPC-(1-320) and xGIPC N-PDZ are capable of binding to xIGF-1 receptor, consistent with the notion that the PDZ domain is also involved in binding the intracellular domain of xIGF-1 receptor. However, the PDZ domain alone did not suffice to bind xIGF-1 receptor either in yeast two-hybrid assays (Fig. 3B) or in COS cells (not shown). Further studies will be required to clarify the nature of this apparent difference.
Clearly the most interesting implication of our data lies in the possible link between IGF-1 receptor signaling and G protein functions during oocyte maturation. High levels of cAMP are thought to be responsible for maintaining oocyte meiosis arrest (24). Progesterone, the natural maturation-inducing ovarian hormone, binds to the cytoplasmic progesterone receptor (xPR) (25,26) and triggers a reduction of cAMP by inhibiting oocyte adenylyl cyclase (27,28). Insulin and IGF-1 can also induce maturation in vitro (29) by activating the endogenous IGF-1 receptor (7). Forskolin, a potent adenylyl cyclase activator, similarly blocks both progesterone-induced and insulininduced oocyte maturation (30,31), suggesting that a reduction of cAMP is a necessary event in both cases. It has been reported that insulin is able to stimulate oocyte cAMP phosphodiesterase activity (32). A recent study has provided a possible link between the insulin/IGF-1 receptor and the activation of oocyte phosphodiesterase. Andersen et al. (33) reported that injection of protein kinase B/Akt results in activation of cAMP phosphodiesterase (PDE3) and induction of oocyte maturation. Protein kinase B/Akt is a serine/threonine kinase activated downstream of PI 3-kinase. We (6) and others (34) have previously shown that insulin/IGF-1 receptor activates PI 3-kinase in frog oocytes. Therefore, one signaling pathway may consist of IGF-1 receptor-activated PI 3-kinase-protein kinase B/Akt, which directly or indirectly activates oocyte PDE3 (33) and ultimately results in the reduction of oocyte cAMP.
We propose here that following ligand binding, xIGF-1 receptor activates an oocyte G i in addition to the PI 3-kinase/Akt/ PDE3 pathway. The co-operation of the two signaling pathways ensures the required cAMP reduction by inhibiting oocyte adenylyl cyclase (via G i ␣-GTP) and activating oocyte PDE3 (via protein kinase B/Akt). The function of xGIPC is likely to promote GAIP-mediated G i ␣-GTP hydrolysis that timely terminates the G protein signal and recycles the resulting G i ␣-GDP for another round of G protein activation. It is interesting to note that several groups have recently demonstrated that IGF-1 receptor forms a complex with trimeric G i proteins (4,5). Whether the interaction between IGF-1 receptor and G i is direct or mediated by other proteins, it is conceivable that xGIPC could function to link GAIP to the receptor complex. A receptor complex containing both the signal initiator (IGF-1 receptor) and signal terminator (GAIP) may be a unique feature to trimeric G protein signaling downstream of receptor protein tyrosine kinases (such as IGF-1 receptor and TrkA). The two C-terminally truncated xGIPC mutants, xGIPC N-PDZ or xGIPC-(1-320), may act in a dominant-negative fashion over the endogenous xGIPC. The fact that both behaved similarly to full-length xGIPC in binding xIGF-1R (Fig. 3B) and hGAIP (Fig. 7C) would seem to rule out the simple explanation that the mutants fail to link GAIP to the IGF-1 receptor. The reasons for the apparently dominant-negative effect of these mutants therefore remain unknown.
How might the complex of xGIPC and RGS-GAIP stimulate IGF-1 receptor signaling in frog oocytes? Two nonmutually exclusive possibilities are considered here. First, overexpression of xGIPC and hGAIP may increase the IGF-1 receptorassociated GTPase activity of hGAIP toward endogenous G i -GTP and hence accelerate recycling of G i -GDP for further rounds of activation by the IGF-1 receptor. This clearly would have the net effect of increased G i activity, resulting in greater inhibition of oocyte adenylyl cyclase and reduction of cAMP. Second, the complex of xGIPC and hGAIP may function downstream of xIGF-1 receptor, independently of G i , to activate a maturation-promoting signaling pathway.