(Received for publication, August 9, 1996, and in revised form, November 20, 1996)
From the Departamento de Biologia Molecular,
Universidad Autonoma de Madrid, Facultad de Ciencias, 28049 Madrid,
Spain, the § Dipartimento di Farmacologia Sperimentale,
Università degli Studi di Napoli "Federico II," 80131 Napoli,
Italy, and INSERM U.344, Endocrinologie Moléculaire,
Faculté de Médecine Necker, 75730 Paris Cedex 15, France
Prolactin (PRL) has been demonstrated to induce
tyrosine phosphorylation and activation of the cytoplasmic tyrosine
kinase JAK2. The present study represents an initial effort to identify the phosphorylation repertoire of the PRL receptor (PRLR). For this
purpose we have modified the rat PRLR cDNA to encode an additional N-terminal epitope specifically designed to allow the rapid
purification of the PRLR and associated proteins from transfected
cells. The Flag-tagged PRLR was stably expressed in the human 293 cell
line. PRL induced tyrosine phosphorylation of proteins of 85, 95, and 185 kDa from 10 to 30 min after PRL stimulation. Immunoblot analysis of
immunoprecipitation indicates that p85 corresponds to the 85-kDa regulatory subunit of phosphatidylinositol (PI)-3 kinase, p95 to PRLR,
and p185 to insulin receptor substrate 1 (IRS-1). Both PI-3
kinase and
IRS-1 appear to associate with PRLR in a PRL-dependent manner. These results thus indicate that kinases other than JAK2, namely PI-3
kinase, are activated by PRL.
PRL1 binding to its cell surface receptor initiates a series of molecular interactions that ultimately determines the specific physiological response. Following PRL stimulation, PRL receptors are tyrosyl-phosphorylated; recent efforts to identify signal transducers activated by the PRL receptor have demonstrated that PRLR associates with and activates two cytoplasmic tyrosine kinases of the Janus T-K family (1-4), although JAK2 appears to be the major kinase involved in most responses.
A number of signaling molecules form stable complexes with other
tyrosyl-phosphorylated receptors via an SH2 domain, including insulin
receptor. The rationale for sharing common intracellular pathways
between IR and PRLR is further substantiated by the insulin-like effect
of growth hormone and to a lesser extent of PRL in a variety of cell
types (5-7). These include 1) increased glucose-stimulated insulin
secretion and decreased threshold of glucose stimulation (8), 2)
increased insulin synthesis (9), 3) increased -cell proliferation
(10), and, more recently, it has been proposed that lactogenic hormones
are primarily responsible for the enhanced islet function observed
during pregnancy (11). To facilitate detection of the interaction
repertoire and the phosphorylation repertoire of the PRLR, we have
modified the rat PRLR cDNA to encode an additional N-terminal
epitope specifically designed to allow the rapid purification of the
PRLR and associated proteins from transfected cells. Our results
clearly show that PRLR associates with insulin receptor substrate 1 and
PI-3
kinase. Upon PRL stimulation, both association with PRLR and
tyrosyl phosphorylation of these two proteins are activated.
Ovine PRL was a gift from the
National Hormone and Pituitary/NIDDK program (Baltimore, MD). The
antiphosphotyrosine antibody (PY), monoclonal IgG2 bk
antiserum to the 85-kDa subunit of PI-3
kinase, and rabbit polyclonal
antibody to IRS-1 (anti-rat C-terminal) were purchased from Upstate
Biotechnology, Inc. Anti-Flag monoclonal antibody M2 and M1 affinity
gel are products of IBI-Kodak.
The 293 fibroblast cells were grown in DMEM nut F12 medium containing 10% fetal calf serum. Several hours before transfection, cells were plated in a rich medium (two-thirds DMEM nut F12, one-third DMEM, 4.5 g/liter glucose) containing 10% fetal calf serum. Cells were incubated in the absence of serum overnight prior to hormone stimulation using serum-free DMEM/F12 (12).
Construction of the Flag PRLR (FPRLR) Expression PlasmidThe rat PRLR cDNA in the expression vector pcDNA3 was modified to encode an additional Flag epitope, Asp-Tyr-Lys-Asp-Asp-Asp-Asp-Lys- between the signal peptide and the N terminus of the receptor via a recombinant polymerase chain reaction strategy (13). The Flag PRLR plasmid was used to transfect 293 cells. G-418 resistant cell lines overexpressing the FPRLR were selected for PRL binding. For purification studies, a clonal 293 FPRLR cell line was used.
Immunoprecipitation and Western BlottingConfluent 293 cells, stably transfected with the Flag-tagged PRLR, were incubated in
serum-free medium overnight. The cells were incubated at 37 °C, 5%
CO2 atmosphere. Cells were stimulated or not with oPRL (400 ng/ml) for 15 min. After stimulation, cells were rapidly washed with
ice-cold phosphate-buffered saline and scraped in lysis buffer: 10 mM Tris-HCl, pH 7.5, 5 mM EDTA, 150 mM NaCl, 30 mM sodium pyrophosphate, 50 mM sodium fluoride, 1 mM
Na3VO4, 10% glycerol, 0.5% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, aprotinin, leupeptin, and
pepstatin A (10 µg/ml) on ice. Lysed cells were centrifuged at 14,000 rpm in a Sorvall centrifuge at 4 °C for 15 min. Lysates from control
or stimulated cells were pooled, and 1.5 ml of supernatants were used
for each immunoprecipitation. Immunocomplexes were collected using
Protein A-Sepharose (Pharmacia Biotech. Inc.) and using the appropriate
antibody. Incubation was carried out overnight at 4 °C. Samples were
centrifuged and washed 3 times with lysis buffer, boiled 5 min in a
mixture of 20% glycerol, 10% -mercaptoethanol, 4.6% SDS, and 125 mM Tris, pH 6.8. Immunoprecipitated proteins were separated
by SDS-polyacrylamide gel electrophoresis on a 7.5% gradient
polyacrylamide. Proteins were transferred on a polyvinylidene
difluoride transfer membrane (PolyscreenTM, DuPont NEN)
using a semidry transfer cell (Trans-Blot SD, Bio-Rad). Blots were
incubated with the appropriate antibody and visualized by ECL detection
(Amersham). To reprobe the blot with another antibody, the blot was
rinsed and incubated with stripping buffer (65 mM Tris-HCl,
pH 6.8, 2% SDS, and 100 mM
-mercaptoethanol).
Confluent 293 cells, stably transfected with the Flag-tagged PRLR, were incubated overnight in serum-free conditions and stimulated or not with oPRL (400 ng/ml). Cells were washed twice with ice-cold phosphate-buffered saline and scraped in lysis buffer (described above). After 30 min on ice, cell lysates were centrifuged at 42,000 rpm for 1 h at 4 °C. Supernatants were collected and diluted 1:5 with buffer B (50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 0.1% Triton X-100, 3% (v/v) glycerol, 0.1 mM phenylmethylsulfonyl fluoride, and 1 mM CaCl2) and loaded on an Anti-Flag M1 affinity column (IBI-Kodak), washed with 30 ml of buffer B, and eluted with 6 ml of buffer C (Buffer B containing 2 mM EDTA without CaCl2). Eluted samples were concentrated until an appropriate volume was obtained using Centriprep and Microcon tubes (Amicon) and boiled for 5 min in loading buffer (composition described above). Immunoblot analysis was performed as described previously (2).
In Vitro Phosphatidylinositol 3Immunocomplexes from 293-FPRLR cells immobilized on
protein A-Sepharose were washed 3 times with lysis buffer, once with
0.5 M LiCl, 100 mM Tris-HCl, pH 7.4, and once
with 100 mM Tris-HCl, pH 7.4, 100 mM NaCl, and
1 mM EDTA. The PI-3 kinase activity was assayed according
to the method described by Auger et al. (14) in a final
volume of a 50-µl reaction mixture containing 20 mM
Hepes, pH 7.4, 2 mM sodium orthovanadate, 5 mM
MgCl2, 50 µM ATP, 5 µCi of
[
-32P]ATP, and 2 mg/ml phosphoinositides. The reaction
was initiated by adding a MgCl2-ATP mixture and was stopped
after 20 min at 25 °C, by the addition of 100 µl of 1 M HCl. The lipids were extracted with 200 µl of
chloroform/methanol (1:1), and they were resolved on thin layer
chromatography TLC60 (Merck). The lipid phosphorylation was visualized
by autoradiography.
Understanding the PRL signaling pathway requires identification of
the interaction repertoire and the phosphorylation repertoire of the
PRLR. A recent report has suggested that GH, interferon-, and
leukemia inhibitory factor stimulate tyrosyl phosphorylation of IRS-1
and its association with PI-3
kinase (15) which provides a
physiological basis for several of the insulin-like metabolic effects
of GH. PRL receptor belongs to the cytokine superfamily of receptors
and activates JAK2 in response to ligand binding; however, only limited
numbers of reports are concerned with insulin-related effects of PRL.
For example, the association between PRL and insulin-like growth factor
has been described in several targets tissues (5, 6). To identify the
PRLR interaction repertoire, we have designed a method using an
epitope-modified PRLR and Ca2+-dependent
immunoaffinity chromatography to purify the PRLR and associated
proteins. Analysis of the FPRLR proteins product in a transfected 293 cell line expressing the FPRL receptor was carried out. Binding
experiments with 125I-labeled PRL demonstrated 1 × 105 binding sites per 293 cell; the association constant
(Ka = 3 nM
1) was similar
to the wild type PRLR. The FPRLR was purified by an anti-Flag (M1)
Sepharose chromatography column. Antiphosphotyrosine immunoblot of
purified FPRLR treated without (
) or with (+) 18 nM oPRL
(Fig. 1) revealed that tyrosyl phosphoproteins of 185 and 85 kDa specifically co-purified with PRL-treated FPRLR. The 95-kDa
band representing the PRLR showed low basal levels of tyrosyl phosphorylation in the absence of PRL. A time course of stimulation by
PRL (Fig. 2A) indicated that PRLRs are
phosphorylated from 10 to 30 min after PRL stimulation, and no tyrosyl
phosphorylation is observed at 60 min. In addition to the 95-kDa FPRLR,
the immunoblot (Fig. 2B) with anti-Flag antibody revealed an
82-kDa product of C-terminal cleavage from PRLR which is present even
in the absence of PRL. Interestingly, this product is not
tyrosyl-phosphorylated (Fig. 2A). At 60 min, the PRLR is
found in degradation product as a result of lysosomal degradation
following internalization. In addition to the PRLR, a 185-kDa protein
is tyrosyl-phosphorylated 20 min after stimulation, while the 85-kDa
protein is activated 10 and 20 min after PRL stimulation. The
identification of p85 as PI-3
kinase was obtained after
immunoprecipitation with anti PI-3
kinase and immunoblot with
antiphosphotyrosine (Fig. 2C); the same pattern of tyrosine
phosphorylation upon PRL stimulation was obtained. Immunoblot analysis
of purified FPRLR using immunoprecipitation with anti-Flag antibody
revealed that p85 corresponds to the p85 subunit of PI-3
kinase and
that it displays a PRL-dependent association with FPRLR
(Fig. 3A). Furthermore, p185 was shown to
correspond to IRS-1 and was also associated with FPRLR in a
PRL-dependent manner (Fig. 3B). The
PRL-dependent association of both PI-3
kinase and IRS-1
with FPRLR was obtained (Fig. 3, A and B,
lanes 3 and 4) after immunoprecipitation with
anti-Flag. PRL also induced the association of IRS-1 with PI-3
kinase
as shown by immunoprecipitation with anti-IRS-1 (Fig. 3A,
lanes 1 and 2) or anti-PI-3
kinase (Fig.
3B, lanes 5 and 6). Both of them were
tyrosine-phosphorylated upon PRL stimulation (Fig. 3, A and
B, lanes 7 and 8) following immunoblotting with anti-PI-3
kinase (Fig. 3A) and
anti-IRS-1 (Fig. 3B).
293-FPRLR cells were maximally stimulated or not with oPRL (400 ng/ml),
and FPRLR, IRS-1, and PI-3 kinase were immunoprecipitated from these
cells. The immunocomplexes were assayed for in vitro PI-3
kinase activity. A representative experiment is shown in Fig.
4, where the presence of PI-3
kinase activity in the
immunocomplexes of FPRLR and IRS-1 was evidenced by the presence of
PI(3)P and PI(3,4)P2 labeled with radioactive phosphate.
The PI-3
kinase activity was significantly increased when cells were
stimulated with PRL compared with nonstimulated. This increase in PI-3
kinase activity seems to be due to a higher amount of PI-3
kinase
associated to PRLR induced by PRL, since the total PI-3
kinase
immunoprecipitated activity was not modified. These data fit well with
Western blot results (Fig. 3A), where PRL stimulation
results in PI-3
kinase association to PRL receptor.
A major pathway for signal transduction has been described for the PRLR, implicating the JAK2 protein kinase. Ligand binding to PRLR activates this tyrosine kinase (2, 4) which appears to be constitutively associated with the receptor. A cytoplasmic proximal region of the receptor is required for JAK2 association, more precisely the Box 1 and the adjacent residues upstream of Box 2 (12). Multiple members of the cytokine receptor family can activate JAK2; some have also been shown to stimulate tyrosyl phosphorylation of IRS-1 (15-17). For example, GHR and IL9-R associated JAK2 are able to tyrosine-phosphorylate IRS-1 upon ligand binding (15, 17). In this report, we describe a PRL-dependent association with FPRLR and tyrosine phosphorylation of IRS-1, albeit the type of interaction between JAK2 and IRS-1 after PRL induction is unknown. A possible direct association would make JAK2 an obvious candidate for the tyrosine kinase responsible for the PRL-dependent tyrosine phosphorylation of IRS-1. The possibility of an auxiliary molecule common to the cytokine receptor family that could bind JAK2 and induce phosphorylation of IRS-1 cannot be excluded. IRS-1 may interact directly or indirectly with the PRLR; cytoplasmic regions involved in this association have to be determined. In the case of GHR (15), none of the tyrosines of the cytoplasmic domain appear to be necessary for IRS-1 tyrosyl phosphorylation indicating that IRS-1 interacts with other as yet unidentified proteins. In conclusion, it appears that signaling through IRS-1 may be common to multiple members of this family that activate JAK2.
IRS-1 has been implicated as intermediate between insulin receptor and
several signal molecules. It provides binding sites for SH2 domains of
PI-3 kinase subunit p85. In our study, we detected an association of
these two molecules together and with the Flag PRLR. Activation of this
enzyme can lead to proliferation and regulation of the cell cycle,
glucose uptake, and vesicular trafficking of proteins (18). The
signaling pathway involving PI-3
kinase could explain the
insulin-related effects of PRL. Further experiments are necessary to
determine tyrosine-phosphorylated regions implicated in the interaction
of the SH2 domains of p85 and either the activated PRLR or a
phosphorylated protein intermediate after stimulation by PRL. The
membrane proximal domain of PRLR contains a YSMM sequence which is a
putative binding site for the SH2 domain of p85 (19).
These experiments revealed a new repertoire of signaling molecules of the PRLR, involving different pathways. Interconnection between these pathways has to be defined more precisely.
We thank J. P. Garcia-Ruiz for the critical review of the manuscript and Claudine Coridun for typing the manuscript.