(Received for publication, April 23, 1997)
From the Centre de Recherche sur l'Endocrinologie Moléculaire et le Développement du CNRS, UPR 1511, 92190 Meudon, France and the § INSERM U145, Avenue de Valombrose, 06107 Nice Cedex 2, France
Insulin receptor substrate-1 (IRS-1) and Shc are two proteins implicated in intracellular signal transduction. They are activated by an increasing number of extracellular signals, mediated by receptor tyrosine kinases, cytokine receptors, and G protein-coupled receptors. In this study we demonstrate that Shc interacts directly with IRS-1, using the yeast two-hybrid system and an in vitro interaction assay. Deletion analysis of the proteins to map the domains implicated in this interaction shows that the phosphotyrosine binding domain of Shc binds to the region of IRS-1 comprising amino acids 583-661. An in vitro association assay, performed with or without activation of tyrosine kinases, gives evidence that tyrosine phosphorylation of IRS-1 and Shc drastically improves the interaction. Site-directed mutagenesis on IRS-1 583-693 shows that the asparagine, but not the tyrosine residue of the N625GDY628motif domain, is implicated in the IRS-1-Shc-phosphotyrosine binding interaction. Mutation of another tyrosine residue, Tyr608, also induced a 40% decrease in the interaction. This study, describing a phosphotyrosine-dependent interaction between IRS-1 and Shc, suggests that this association might be important in signal transduction.
Intracellular signal transduction by growth factors involves a cascade of phosphorylation and dephosphorylation events, beginning by the autophosphorylation and activation of a receptor tyrosine kinase. Among the cohort of intracellular proteins activated by growth factors, two families of proteins, IRSs1 and Shc, are of particular interest. Shc proteins are molecular adaptors characterized by the presence of different interaction domains: two phosphotyrosine interaction domains, a C terminus SH2 domain and an N terminus PTB domain, allowing interaction with receptor tyrosine kinases, and a central collagen homologous (CH1) region (1-3). Shc is known to be implicated in the activation of the Ras signaling pathway via the association of the adaptor Grb2 with phosphotyrosine residues of the CH1 region (4). However the CH1 region can also bind other proteins, like adaptins or Src homology 3 domain-containing proteins (5, 6). This implies that in addition to the activation of the Ras pathway, Shc-related proteins can be involved in other cellular functions (7). IRS proteins IRS-1 and IRS-2 are considered to be docking proteins; their multiple tyrosine phosphorylated residues are potential binding sites for SH2-containing proteins, initiating various signaling pathways (8, 9). Different classes of factors are known to bind to IRSs: molecular adaptors, like Grb2 and Nck, and proteins with catalytic activity, like the tyrosine phosphatase Syp, the tyrosine kinase p59fyn, or the phosphatidylinositol 3-kinase via its p85 regulatory subunit (8, 9).
Shc can be one of the molecular adaptors interacting with IRS-1. In support of this hypothesis, it can be noticed that SH2 containing proteins like Grb2, Nck, p85, and Syp can either bind directly to activated growth factor tyrosine kinase receptors or bind to the docking protein IRS-1 (10-15). Actually, a co-immunoprecipitation of IRS-1 and Shc has been reported in Shc-transfected fibroblasts, after insulin or insulin-like growth factor-1 stimulation (16). This result has not been confirmed by further studies, probably due to the low ability of Shc to coprecipitate with relevant proteins, like the insulin receptor. Indeed, an association between the insulin receptor and Shc, or the insulin receptor and IRS-2, has never been reported in cells, using co-immunoprecipitation assays (17, 18) but was only shown using more sensitive techniques like the two-hybrid system (19-22).
The aim of the present study was then to further analyze the interaction between Shc and IRS-1, using two sensitive systems that allow the detection of transient or weak protein-protein interactions: the yeast two-hybrid system and an association assay between GST fusions and in vitro translated proteins. This study shows a direct interaction between the two proteins IRS-1 and Shc and delineates the minimal interaction domains of these proteins. These experiments demonstrate also that the IRS-1-Shc interaction is phosphorylation-dependent, suggesting that it plays a role in growth factor signaling.
The yeast strains Y190 (MATa, leu2-3, 112, ura3-52, trp1-901, his3-200, ade2-101, gal4D, gal80D, URA3::GAL1-lacZ, LYS2::GAL-HIS3, cyhr) and Y526 (MATa, ura3-52, his3-200, ade2-101, lys2-801, trp1-901, leu2-3, 112, canr, gal4-542, gal80-538, URA3::GAL1-lacZ) were used. Yeast expression plasmids pAS1-CYH2 (pAS2) and pACTII were provided by S. Elledge (Houston, TX) (23, 24), and pGBT9 was provided by S. Fields (New York, NY). pAS2 and pGBT9 contain Gal4 DNA binding domain (Gal4BD) and were used with the yeast strains Y190 and Y526 containing lacZ gene downstream the Gal4 binding sequence. pACTII contains the Gal4 activation domain (Gal4 AD). Rat IRS-1 and IR cDNAs were kindly provided by M. F. White and C. R. Kahn. Human Shc p52 and IR cDNAs were gifts from P. G. Pelicci and E. Clauser, respectively. Synthetic defined dropout yeast (DO) media lacking the appropriate amino acids were obtained from BIO 101 (La Jolla, CA). Oligonucleotides were purchased at the Institut Pasteur (Paris, France). All chemicals were from Sigma France, and enzymes were from New England Biolabs (Beverly, MA).
Plasmid ConstructionsSeveral subdomains of IRS-1 cDNA
were generated by restriction digestion: IRS-1 215-1231
(XmnI) and IRS-1 583-693 (XbaI/NcoI). Other subdomains, IRS-1 215-583, IRS-1 583-661, and IRS-1 633-693, were generated by polymerase chain reaction using pfu DNA polymerase (Stratagene, La Jolla, CA), introducing a BamHI site to each
end of the cDNA fragment to allow the in-frame insertion into pAS2. Site-directed mutagenesis was made in the pAS2 IRS-1 583-693
construct, using the Quick Change site-directed mutagenesis kit
(Stratagene, La Jolla, CA). Single mutations were tyrosine 608 changed
into alanine (Y608A), tyrosine 628 changed into phenylalanine
(Y628F), tyrosine 658 changed into alanine (Y658A), and asparagine 625 changed into alanine (N625A). Double mutants were Y608A/N625A, N625A/Y628F, and Y608A/Y628F. For pBSSK IRS-1, pBSSK IRS-1 1-708, and
pBSSK IRS-1 691-1231 constructs, cDNA fragments were generated by
polymerase chain reaction. The BglI/DraI fragment
of the rat IR cDNA, corresponding to the intracellular domain of
the -subunit, was inserted in pGBT9. Shc constructs (Shc 1-473, Shc
1-230, Shc 234-473) cloned into pACTII were previously described
(20). Shc 1-473, Shc 1-230, and the intracellular domain of the human IR were generated by polymerase chain reaction and cloned into the pGEX
3X (Pharmacia Biotech Inc.). All constructs were verified by DNA
sequence analysis.
Yeast strains were transformed by the lithium acetate
method of Gietz (25). Cotransformants were selected on plates lacking tryptophan and leucine (DO-WL) to select for the presence of pAS2 or
pGBT9 and pACTII, respectively. Three colonies of each cotransformant were patched on DO-WL plates and then replicated on Whatman No. 40 paper, laid on a DO-WL plate. After 24 h, a -galactosidase assay was performed by a color filter assay using
5-bromo-4-chloro-3-indolyl-
-D-galactopyranoside as
described (26). The blue coloration of patches reflects the occurrence
of interaction between the two hybrid proteins. For quantitative
studies of
-galactosidase activity, a solution assay was performed
on eight independent colonies from two independent transformation
experiments. Yeast were grown in DO-WL galactose medium, and substrate
was O-nitrophenyl
-D-galactopyranoside (26).
Yeast protein extracts were performed with glass beads according to the manufacturer's protocol (CLONTECH, Palo Alto, CA). Extracts were subjected to SDS-PAGE, and proteins were blotted to nitrocellulose. Hybrid proteins were detected using anti-Gal4 AD and anti-Gal4 BD antibodies (CLONTECH), and the ECL detection kit (Amersham).
In Vitro Association ExperimentsGST, GST-Shc, GST-Shc 1-230, and GST-IR were produced in bacteria according to the manufacturer's protocol (Pharmacia). Induced bacteria were lysed by sonication in PLC lysis buffer (50 mM HEPES, pH 7.5, 150 mM NaCl, 10% glycerol, 1% Triton X-100, 1.5 mM MgCl2, 1 mM EGTA, 10 µg/ml aprotinin, 10 µg/ml leupeptin, 100 mM NaF, 10 mM sodium pyrophosphate, 1 mM phenylmethylsulfonyl fluoride). Clarified lysates were incubated with glutathione-Sepharose beads for 1 h at 4 °C in PLC lysis buffer. Beads were then washed three times with HNTG buffer (20 mM HEPES, pH 7.5, 150 mM NaCl, 10% glycerol, 0.1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride). IRS-1 1-1231, IRS-1 1-708, and IRS-1 691-1231 were transcribed and translated in vitro using [35S]methionine and the TNT-coupled reticulocyte lysate system (Promega, Madison, WI). A 15-µl aliquot of the translation mix was added to 1 µg of immobilized GST and GST fusion proteins. To allow tyrosine phosphorylation, the proteins were preincubated with PIPES buffer (20 mM PIPES, pH 7.0, 10 mM MnCl2, 0.1 mM sodium vanadate, 10 µg/ml aprotinin) at 30 °C for 30 min to activate the reticulocyte lysate tyrosine kinases as described in Ref. 27. 1 ml of Nonidet P-40/deoxycholate RIPA buffer (1% Nonidet P-40, 0.25% deoxycholate, 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EGTA, 1 mM NaF, 10 µg/ml aprotinin, 1 mM phenylmethylsulfonyl fluoride, 1 mM pepstatin A) was added, and binding was allowed to occur at 4 °C for 1 h. Beads were washed three times with 1 ml of Nonidet P-40/deoxycholate RIPA buffer, and bound proteins were eluted by boiling in SDS sample buffer. The resulting proteins were electrophoresed on 10% SDS-PAGE, and bound translation products were detected by autoradiography.
An interaction was
observed between the two proteins IRS-1 and Shc using two techniques.
In the two-hybrid system, IRS-1 215-1231 fused to the Gal4 BD was
shown to interact with full-length Shc fused to Gal4 AD (Fig.
1A). As assessed by the quantitative
-galactosidase assay, this interaction was of the same order of
magnitude as the interactions between the insulin receptor and Shc (19,
20) and SNF1 and SNF4 (28), measured in the same experiment (Fig. 1A). An in vitro association experiment was
performed using immobilized GST fusion proteins and in vitro
translated IRS-1 labeled with [35S]methionine. As
shown in Fig. 1B, IRS-1 was not retained by GST but was
retained by GST-Shc and by GST-IR used as a positive control. These
results give clear evidence that the two proteins IRS-1 and Shc can
associate.
To determine the domains of the two proteins implicated in the
interaction, we have generated a series of deletion mutants of Gal4
BD-IRS-1 and Gal4 AD-Shc fusion proteins (Fig. 2). The expression of the different hybrid proteins was verified on Western blot, as shown on Fig. 2. It can be noticed that the Shc 234-473 has a
faster migration mobility than what could be expected given its size,
in comparison with Shc 1-230. The explanation for this difference
remains unclear but could be due to differences in the amino acid
composition of the two proteins. The expression of the reporter gene
lacZ was tested for each cotransformant studied (Table
I). The interaction between IRS-1 and Shc was observed with the PTB domain of Shc (Shc 1-230) but not with the CH-SH2 domain
(Shc 234-473). The minimal domain of IRS-1 able to interact with Shc
in this system was the fragment containing the amino acids 583-661.
The other domains either did not interact (IRS-1 215-583) or could not
be tested in the two-hybrid system due to an important background
(IRS-1 693-1231).
|
IRS-1 and
Shc fusion proteins expressed in yeast display a low tyrosine
phosphorylation, as detected in Western blot using an
antiphosphotyrosine antibody (data not shown). To determine the
importance of the phosphorylation state of the proteins for the
interaction observed, we have tested in vitro association of
IRS-1 and Shc, stimulating or not stimulating the activity of tyrosine
kinases present in the reticulocyte lysate (27). Full-length IRS-1,
IRS-1 1-708, and IRS-1 691-1231 were transcribed and translated
in vitro in presence of [35S]methionine. The
radioactive translation products were incubated with GST fusions. This
incubation was performed with or without prior phosphorylation of the
two proteins, as described under "Experimental Procedures." The
phosphorylation state of the proteins was verified by
antiphosphotyrosine Western blot (data not shown). Identical amounts of
the various GST fusions were used, as assessed by Coomassie Blue
staining of the gels before autoradiography (data not shown). As shown
in Fig. 3A, IRS-1 binds to GST-Shc and
GST-Shc 1-230 and also to GST-IR in absence of induction of phosphorylation (lanes 3-5). However, a stronger
interaction was observed when both partners were phosphorylated, as
assessed by the amount of labeled IRS-1 detected on the gel
(lanes 7-9). This experiment suggests that the IRS-1-Shc
interaction, like the IR-IRS-1 interaction, is
phosphorylation-dependent. Such an interaction with GST-Shc and
GST-Shc 1-230 was also observed with the N-terminal domain but not
with the C-terminal domain of IRS-1 (Fig. 3B). These results
show that the N-terminal domain of IRS-1 and not the C-terminal domain
can associate with the Shc PTB domain in a
phosphotyrosine-dependent manner.
Site-directed Mutagenesis on IRS-1 Tyrosine Residues
We have
mutated the three tyrosine residues present in IRS-1 583-661 to
determine their relative importance in the IRS-1-Shc interaction (Fig.
4A). Remarkably, Tyr628 is in a
NGDY motif, resembling the NPXY motif recognized by PTB domains (2, 3). The proline residue of the binding consensus sequence
is not conserved in the IRS-1 583-693 domain, but it has been shown
that it is not critical for the PTB interaction (29, 30). In addition,
all three tyrosine residues are into a YMXM motif.
Site-directed mutagenesis was performed on the Gal4 BD IRS-1 583-693
domain. We have generated four single mutants: Y608A, Y628F, Y658A, and
N625A, and three double mutants: Y608A/N625A, N625A/Y628F, and
Y608A/Y628F. Their expression was confirmed by Western blot (result not
shown). The interaction between IRS-1 mutants and Shc 1-230 was
quantified by a -galactosidase solution assay, and results were
expressed as a percentage of the maximum interaction measured with wild
type IRS-1 583-693 (70 ± 8 units of Miller), as shown in Fig.
4B. IRS-1-Shc interaction was significantly decreased (by
40%) by mutation of Tyr608 but not altered by mutation of
Tyr658. Surprisingly, in the NGDY motif, mutation of
Tyr628 was without effect, whereas mutation of
Asn625 induced a significant alteration of the interaction.
In addition, combination of two mutations did not further decrease the
modification induced by a single mutation, even for the Y608A/N625A
double mutant. These results show that although they are not the only ones, Tyr608 and Asn625 are important residues
for the interaction between IRS-1 and Shc.
In the present study we have shown an interaction between the PTB domain of Shc and a 78-amino acid-long domain of IRS-1. This interaction was demonstrated using two different methods: the yeast two-hybrid system and an in vitro association assay. It is thus likely that the IRS-1-Shc interaction is a direct interaction. If another protein was to be implicated in this association, it should have been present in adequate amounts in both the yeast nucleus and the rabbit reticulocyte lysate.
We show that the PTB domain of Shc and not the SH2 domain is implicated in the association with IRS-1. Then, IRS-1, known to recruit proteins through their SH2 domain, can also bind other factors using a different recognition system. Shc is now considered to be an adaptor protein, with a modular organization (for review see Ref. 7). The identification of proteins able to bind the different domains of Shc are important to shed light on the other roles that Shc might play in addition to Ras activation.
PTB domains were first defined as specialized domains interacting with
a tyrosine residue in a NPXpY motif (2, 3). However, recent
studies have shown that the binding motif recognized by a PTB domain
cannot be restricted to the NPXY motif. For example, the
interaction between APP and the PTB domains of X11 and FE65 is
mediated by the YENPTY motif of
APP (31). Mutation of the first
tyrosine residue of this motif suppresses both interactions, whereas
mutation of the last tyrosine residue has no effect. Furthermore, mutation of the asparagine residue (N) specifically alters the interaction with X11 and not with FE65 (31). Similarly, in the present
study, IRS-1-Shc interaction is significantly altered by mutation of
the first tyrosine and the asparagine residues of the
Y(X)16NGDY motif of IRS-1 583-661, but is
unaffected by mutation of the last tyrosine residue of this motif.
The minimum domain of IRS-1 interacting with Shc, IRS-1 583-661, includes the tyrosine residue Tyr608, which is implicated in the interaction, as shown by mutagenesis experiments. Tyr608 has already been described as binding site for the p85 subunit of phosphatidylinositol 3-kinase (10). Thus, p85 and Shc might compete for the same binding site. A competition for binding to the same residue of IRS-1 has already been described; Tyr895 can bind either Grb2 or p59fyn (32). This suggests that IRS-1 signaling complexes are not homogenous and that different elements in the complex may influence the signal (32). IRS-1 583-661 is also included in the corresponding domain of IRS-2 that mediates an interaction with the insulin receptor (21, 22). It should be interesting to know if the PTB domain of Shc is also able to bind to IRS-2.
In the in vitro association assay, in absence of stimulation of tyrosine phosphorylation, we can observe a basal interaction between IRS-1 and Shc and also between IRS-1 and the insulin receptor. Knowing that IR-IRS-1 interaction is dependent on tyrosine phosphorylation, the meaning of this basal association is unclear (no basal tyrosine phosphorylation was detectable on Western blot in these conditions; data not shown). On the other hand, mutations of the tyrosine residues of the IRS-1 583-693 domain decrease but never suppress the interaction with Shc, which suggests a phosphotyrosine-independent component in this interaction. Supporting this, it has recently been shown that PTB domains can bind to nonphosphorylated proteins (31, 33-35). On the other hand, an IRS-1 molecule in which 18 potential tyrosine phosphorylation sites were replaced by phenylalanine, failed to undergo tyrosine phosphorylation or mediate activation of some effectors but retained the ability to stimulate insulin-mediated mitogenesis at high insulin concentration (36). This suggests that IRS-1 contains phosphotyrosine-independent elements that are involved in mitogenic signals, potentially the phosphotyrosine-independent component of the interaction of Shc with IRS-1.
Tyrosine phosphorylation drastically improves the IRS-1-Shc association, to the same extend as the IR-Shc association used as control, suggesting that the IRS-1-Shc interaction is implicated in growth factor signal transduction. Furthermore, mutation of Tyr608 of IRS-1 583-693 significantly altered the IRS-1-Shc interaction. However, there is a discrepancy between the partial decrease induced by the tyrosine mutations on IRS-1 583-693, and the striking stimulatory effect of tyrosine phosphorylation on the in vitro interaction. One explanation could be that tyrosine residues located outside of IRS-1 583-693 could also be implicated in the phosphotyrosine-dependent interaction observed in vitro with a larger N-terminal domain. In addition, tyrosine phosphorylation occurring in yeast might be insufficient to induce interactions similar to those observed after tyrosine kinase stimulation in vitro.
Recently, structural studies of Shc and IRS-1 PTB domains have revealed that they are members of the pleckstrin homology domain family (37, 38). Pleckstrin homology domains are found in a large variety of proteins, and they are likely to be involved in the localization of proteins in the membrane proximity (for review see Ref. 39). Intracellular localization of the proteins implicated in signal transduction is likely to be important for their recruitment after receptor stimulation. For example, it has been shown that distinct pools of Grb2-Sos are recruited after epidermal growth factor or insulin stimulation (40). Furthermore, Shc proteins are located in the central perinuclear area in unstimulated cells and are redistributed to the peripheral cytosol upon growth factor stimulation (41). It is possible that IRS-1 and Shc could be located in subcellular pools allowing or not allowing their interaction, depending on the activation state of the cell. However, IRS-1-Shc interaction could also allow the localization of the two factors in the same cell compartment.
In summary, we have shown that IRS-1 and Shc interact in vitro and in the yeast two-hybrid system. This association is dependent on tyrosine phosphorylation and thus might be important in intracellular signal transduction. The docking protein IRS-1 is known to interact with an increasing number of factors through their SH2 domains. The present work is the first report of the binding of an adaptor on IRS-1 using a non-SH2 domain. The recruitment of different sets of factors and the formation of various signaling complexes may be a mechanism to induce specific signals. Considering this, the interaction of Shc with IRS-1 could add some degree of specificity in the signal transduced by insulin and growth factors receptors.
We thank M. F. White and C. R. Kahn (Boston, MA) for rat IRS-1 and IR cDNAs, P. G. Pelicci (Milan, Italy) for human Shc p52 cDNA, and E. Clauser (INSERM U36, Paris, France) for human IR cDNA. C. Fleury is gratefully acknowledged for help in site-directed mutagenesis, and A. Leturque and S. Hauguel de Mouzon are thanked for critical reading of the manuscript.