| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
Department of Medicine, Divisions of Endocrinology and Metabolism (L.L., J.J., S.J.F.) and Rheumatology (T.Z.) and the Department of Cell Biology (L.L., S.J.F.), University of Alabama, and the Veterans Affairs Medical Center (J.J., S.J.F.), Birmingham, Alabama 35294; the Laboratory of Cellular and Molecular Biology, National Cancer Institute, National Institutes of Health (J.H.P.), Bethesda, Maryland 20892; and Metabolex, Inc. (T.A.G.), Hayward, California 94545
Address all correspondence and requests for reprints to: Dr. Stuart J. Frank, University of Alabama, Room 756, Developmental Research and Endocrinology Branch, University of Alabama Station, Birmingham, Alabama 35294. E-mail: frank{at}endo.dom.uab.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
GH-induced GHR dimerization rapidly promotes activation of the GHR-associated cytoplasmic tyrosine kinase, JAK2, which, in turn, initiates a number of tyrosine phosphorylation-dependent signaling pathways (8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20). Molecular mapping studies have revealed a complicated pattern of interactions, GH and/or phosphotyrosine dependent and independent, among the GHR, JAK2, SHC, STAT (signal transducer and activator of transcripton) molecules, and Src Homology 2-containing tyrosine phosphatase (SHP-2), among others (9, 14, 15, 21, 22, 23, 24, 25, 26, 27, 28, 29). Of particular potential physiological interest, GH has recently been shown to cause tyrosine phosphorylation of the so-called insulin receptor substrate proteins, IRS-1 and -2 (30, 31, 32, 33). The IRS molecules are large cytosolic phosphoproteins that by virtue of induced multiple tyrosine phosphorylation serve as docking proteins involved in proliferative and metabolic signaling by insulin, interleukin-4 (IL-4), and other cytokines and peptide hormones (34, 35).
The principal physical interactions between the insulin and IL-4 receptors and IRS proteins are believed to be both ligand and phosphotyrosine dependent. A phosphotyrosine residue in NPXY motifs present in each of these receptors, for example, is thought to mediate their interaction with IRS-1 (36). Regions in the amino-terminus of IRS-1, referred to as the phosphotyrosine-binding (PTB) and Shc and IRS-1 NPXY-binding (SAIN) domains, have been demonstrated to directly interact with the activated insulin receptor (IR) NPXY motif, whereas the pleckstrin homology (PH) domain at the extreme amino-terminus of IRS-1 is also critical in allowing efficient IR-IRS-1 functional coupling (37, 38, 39). Once tyrosine phosphorylated (by the IR kinase domain, in the case of insulin stimulation), IRS proteins can dock various SH2-containing signaling proteins, such as Grb-2, SHP-2, and the 85-kDa regulatory subunit (p85) of phosphatidylinositol 3-kinase (PI-3K) (40).
In this study, we explore the association of IRS-1 with the GHR/JAK2 complex. We observe that JAK2 can specifically interact with the same amino-terminal IRS-1 regions (PH, PTB, and SAIN domains) implicated as important in coupling to the IR. However, the nature of the IRS-1-JAK2 interaction appears to differ substantially from that of the IRS-1-IR, in that it is detectable in the absence of GH stimulation and requires neither IRS-1 nor JAK2 to be tyrosine phosphorylated. We further demonstrate by reconstitution into GHR- and IRS-deficient 32D cells that IRS-1 enhances GH-induced proliferation and mitogen-activated protein (MAP) kinase activation.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Cells, cell culture, and generation of transient and stable
transfectants
COS-7 cells were maintained as described previously (21). NIH
3T3-F442A cells (41), provided by Drs. H. Green (Harvard University,
Boston, MA) and C. Carter-Su (University of Michigan), were
cultured in DMEM (4.5 g/liter glucose; Cellgro, Inc.)
supplemented with 10% calf serum (Biofluids, Rockville,
MD), 50 mg/ml gentamicin sulfate, 100 U/ml penicillin, and 100 mg/ml
streptomycin (all from Biofluids), as previously described
(29). The generation of 32D-rGHR cells has also been previously
described (29). In brief, factor-dependent murine promonocytic 32D
cells, provided by Dr. A. Kraft (University of Colorado, Denver,
CO), were cotransfected by electroporation with the rabbit (r) GHR
complementary DNA (cDNA) in the pSX eukaryotic expression vector and
with a vector (pRc/CMV, Invitrogen, San Diego, CA) that
encodes the neomycin resistance marker. Stably transfected cells were
selected by growth in G418 (0.8 mg/ml; Life Technologies).
Pools of 32D cells stably coexpressing the rGHR and either rat IRS-1 (32D-rGHR-IRS-1) or no IRS-1 (histidinol resistance vector only, referred to as 32D-rGHR) were similarly generated by electroporation of 32D-rGHR cells (2 x 107/ml in complete medium; 250 V, 960 µF; in a GenePulser, Bio-Rad electroporator, Bio-Rad Laboratories, Inc., Richmond, CA) with the pSX-driven IRS-1 cDNA (see below) and a vector (pCMV) that encodes the histidinol resistance marker. Coselection of the rGHR and IRS-1 protein was carried out in G418 and histidinol (2 mM; Sigma Chemical Co.).
[125I]hGH binding assays were performed as previously
described (21, 29), with modifications. In brief, serum-starved
32D-rGHR and 32D-rGHR-IRS-1 cells (10 million cells/ml in binding
buffer; as described below) were incubated in duplicate for 16 h
at 4 C in the presence of 175,000 cpm [125I]hGH
(New England Nuclear-DuPont, Wilmington, DE; SA, 85–130
µCi/µg). Identical duplicate samples were incubated with 5 µg/ml
unlabeled hGH to determine nonspecific binding. Cells were washed twice
with cold binding buffer and solubilized in 1 ml 1% SDS-0.1
N NaOH, and the entire lysate was subjected to
-counting. To determine specific binding, the mean number of
nonspecifically bound counts per min (in the presence of excess
unlabeled hGH) was subtracted from the mean number of total bound
counts per min (in the absence of excess unlabeled hGH; the range in
each case varied <5% from the mean). Specific binding was expressed
as a fraction of the total radioactivity added per sample.
Transient expression of the human IR in COS-7 cells was accomplished as previously described (24). Briefly, COS-7 cells grown in 100 x 20-mm tissue culture dishes were transfected at 60–80% confluence by the calcium phosphate precipitation method (42) with 20 µg of the hIR cDNA (43) in the Rldn expression vector (a gift from Dr. D. McClain, University of Mississippi, Jackson, MS).
Antibodies
Anti-JAK2 peptide serum (directed at residues 758–776 of murine
JAK2; used for immunoblotting) (44), anti-IRS-1 immunoaffinity-purified
rabbit antibody (directed at residues 1221–1235 of rat IRS-1) (34),
anti-IRS-2 protein A-purified rabbit antibody (directed at residues
976-1094 of mouse IRS-2) (35), anti-MAP kinase (anti-MAPK)
affinity-purified rabbit antibody (directed at residues 333–367 of rat
ERK1; recognizes both ERK1 and ERK2), and 4G10 mouse monoclonal
antiphosphotyrosine (APT) antibody were purchased from Upstate Biotechnology, Inc. (Lake Placid, NY). Antihuman IR-ß chain
affinity-purified rabbit antibody and mouse monoclonal
anti-glutathione-S-transferase (anti-GST) were purchased
from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA).
Anti-activated MAPK affinity-purified rabbit antibody (recognizing the
dually phosphorylated Thr183 and Tyr185
residues that correspond to the active forms of ERK1 and ERK2) was
purchased from Promega Corp. (Madison, WI). Anti-JAK1
mouse monoclonal antibody was purchased from Transduction Laboratories, Inc. (Lexington, KY). The anti-JAK2 serum used for
immunoprecipitation (raised in rabbits against a GST fusion protein
incorporating residues 746-1129 of the murine JAK2) (45) and the
anti-GHRcyt rabbit serum (directed at the residue 317–620
region of the human GHR cytoplasmic domain) (46) have been previously
described.
Plasmid construction
The pSX plasmid [a gift from Dr. J. Bonifacino, NIH
(Bethesda, MD), and Dr. K. Arai, DNAX (Palo Alto, CA)], which drives
eukaryotic protein expression from the SR
promoter (47), and the
preparation of pSX rGHR have been previously described (29). The
isolation and preparation of rat IRS-1 cDNA have been described (34).
IRS-1 cDNA was ligated into the pSX plasmid at the SacI and
DraI/SmaI sites. The pCMV-his histidinol
resistance plasmid has been described previously (48).
The generation of cDNA plasmids encoding GST/IRS-1(PH) (including rat IRS-1 residues 21–203), GST/IRS-1(PH+PTB) (residues 21–400), GST/IRS-1(PTB+SAIN) (residues 108–516), and GST/IRS-1(pre-C) (residues 516–896) has been described (37). To generate the plasmid encoding GST/IRS-1(C-ter), the rat cDNA region encoding residues 899-1235 was ligated into the pGEX 2TRS plasmid (46) using the EcoRI and KpnI sites. Correct assembly of this cDNA fragment was verified by dideoxy sequencing and by specific immunoblotting of the bacterially expressed fusion protein using the anti-IRS-1 C-terminal antibody. The plasmid encoding GST/hGHR-(271–389) has been previously described (21).
Cell stimulation, protein extraction, and GST fusion binding
assays
Serum starvation of 3T3-F442A and COS-7 cells was accomplished
by substitution of 0.5% (wt/vol) BSA (fraction V, Boehringer Mannheim, Indianapolis, IN) for serum in the respective culture
medium for 16–20 h before experiments. 32D-rGHR and 32D-rGHR-IRS-1
were serum starved similarly, but for only 4–6 h before stimulation.
Unless otherwise noted, hGH was used at a final concentration of 500
ng/ml in binding buffer [consisting of 25 mM Tris-HCl (pH
7.4), 120 mM NaCl, 5 mM KCl, 1.2 mM
MgCl2, 0.1% (wt/vol) BSA, and 1 mM dextrose],
and stimulations were performed at 37 C for 15 min. Details of the hGH
treatment protocol have been described previously (21). 32D-rGHR and
32D-rGHR-IRS-1 were collected after stimulation with GH by
centrifugation (800 x g for 1 min at 4 C) and
aspiration of the binding buffer. 3T3-F442A cells (after GH
stimulation) and hIR-transfected COS-7 cells (after stimulation with
porcine insulin (Sigma Chemical Co.), 100 ng/ml in PBS for
10 min at 37 C) were washed once with ice-cold PBS containing 0.4
mM sodium orthovanadate (PBS-vanadate) and then harvested
by scraping in PBS-vanadate; pelleted cells were collected by brief
centrifugation. For each cell type, pelleted cells were solubilized for
15 min at 4 C in fusion lysis buffer [FLB; 1%, vol/vol; Triton X-100,
150 mM NaCl, 10% (vol/vol) glycerol, 50 mM
Tris-HCl (pH 8.0), 100 mM NaF, 2 mM EDTA, 1
mM phenylmethylsulfonylfluoride, 1 mM sodium
orthovanadate, 10 mM benzamidine, and 10 µg/ml
aprotinin], as indicated. After centrifugation at 15,000 x
g for 15 min at 4 C, the detergent extracts were subjected
to either immunoprecipitation or affinity precipitation or were
directly electrophoresed, as indicated below.
For affinity precipitation of extracts with GST fusion proteins, fusion protein induction and affinity purification on glutathione-agarose beads (Pharmacia Biotech, Piscataway, NJ) were performed as described previously (21). After glutathione-agarose affinity purification, the amount of each full-length fusion protein purified was estimated by Coomassie staining of eluted full-length fusions in comparison to BSA standards after resolution by SDS-PAGE. Roughly 2 µg of each of the indicated fusions bound to glutathione-agarose beads were incubated with FLB extract from either 3T3-F442A cells or transfected COS-7 cells. After incubation for 2 h at 4 C, the beads were washed extensively with FLB and eluted in reduced Laemmli SDS sample buffer. Ninety percent of the eluate was resolved by SDS-PAGE on 7% gels and immunoblotted sequentially with antibodies as indicated. For verification of similar loading of the fusion proteins on the beads, 5% of the eluate was resolved on a 10% gel and immunoblotted with anti-GST antibody. Where indicated, fractions of the cell extracts that were subjected to glutathione-agarose binding were also analyzed by immunoprecipitation and immunoblotting, as described below, to normalize for the content of the protein being affinity precipitated.
Immunoprecipitation, electrophoresis, and immunoblotting
For immunoprecipitation, cell lysates were incubated with the
appropriate antibody for 90 min at 4 C. Protein-A Sepharose was added
for an additional 30 min, and immune complexes were washed five times
with FLB buffer. Laemmli sample buffer eluates were resolved by
SDS-PAGE and immunoblotted as indicated.
Resolution of proteins under reduced conditions by SDS-PAGE, Western transfer of proteins, and blocking of Hybond-ECL membrane (Amersham, Arlington Heights, IL) with 2% BSA were performed as previously described (14, 21). Membranes were immunoblotted with 1 µg/ml or the indicated dilutions of antibodies against IRS-1, IRS-2, IR ß-chain, 4G10 (1:2,500), GST (1:1,000), MAPK (1 µg/ml), activated MAPK (1:20,000), anti-GHRcyt (1:2,000) or anti-JAK2 peptide antiserum (1:2,000), and horseradish peroxidase-conjugated antimouse or antirabbit secondary antibodies (1:2,000). Detection by enhanced chemiluminescence detection reagents (all from Amersham) and stripping and reprobing of blots were accomplished according to the manufacturer’s suggestions.
Densitometric analysis
Densitometry of enhanced chemiluminescence (ECL) immunoblots was
performed using a solid state video camera (Sony 77, Sony Corp.)
and a 28-mm MicroNikkor lens over a light box of variable intensity
(Northern Light Precision 890, Imaging Research, Inc.,
Toronto, Canada). Quantification was performed using a Macintosh
II-based image analysis program (Image 1.61, developed by W. S.
Rasband, Research Services Branch, NIMH, Bethesda, MD). Basal and
GH-induced MAPK activities were estimated for 32D-rGHR and
32D-rGHR-IRS-1 by normalizing the relative total ERK (ERK1 plus ERK2)
densitometric signal of each sample’s anti-activated MAPK blot by that
of the stripped and reprobed anti-MAPK blot (that is, normalizing the
activated ERK1 and ERK2 for ERK1 and ERK2 abundance).
Proliferation assays
Proliferation of 32D-rGHR and 32D-rGHR-IRS-1 was assayed in
three ways. The CellTiter 96 Aqueous NonRadioactive Cell Proliferation
Assay (Promega Corp.) kit was used according to the
manufacturer’s suggestions. In brief, serum-deprived cells
(25,000/well·0.2 ml) were incubated in 96-well plates in serum-free
medium supplemented with hGH (0–1000 ng/ml) for 24 h at 37 C.
Four hours before the end of the incubation, a tetrazolium compound
(3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxoyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium,
inner salt), referred to as MTS, and an electron-coupling reagent
(phenazine methosulfate) were added; conversion of the MTS into an
aqueous soluble formazan product (a process indicative of cell
viability and correlated in degree to cell number) was monitored at
OD490 in an enzyme-linked immunosorbent assay microplate
reader (Thermo Max, Molecular Devices, Menlo Park, CA).
The data are presented as the fold increase (±SE of
triplicates for each condition) promoted by the indicated GH
concentration vs. that for no GH addition for each cell
line.
Similarly, as indicated in the text, proliferation data generated as described above were verified by manual cell counts. Serum-deprived cells were incubated in triplicate in 24-well plates (400,000/well·0.5 ml) in serum-free medium supplemented with various concentrations of hGH for 24 h. Cells were then resuspended, and trypan blue-excluding cells were counted with a hemacytometer.
Finally, proliferation was also assayed flow cytometrically using the Cell Census Plus system (Sigma Chemical Co.), according to the manufacturer’s instructions and as described previously (49, 50). In this assay, the fluorescent aliphatic reporter molecule, PKH26, is rapidly incorporated and stably retained in the plasma membrane. With each cell division, the dye is equally distributed into the membranes of daughter cells; thus, the number of cell divisions can be monitored by measurement of the PKH26 fluorescence per cell with a flow cytometer. Serum-starved 32D-rGHR and 32D-rGHR-IRS-1 (2.5 million cells) were washed with serum-free RPMI 1640 medium and labeled in 0.5 ml of a solution containing PKH26 (2 µM) for 3 min at room temperature. The labeling reaction was terminated by washing three times in 0.5% BSA-RPMI solution. After removal of an aliquot of each sample for initial fluorescence determination by flow cytometry, the remaining cells were incubated in serum-free medium supplemented with 10 ng/ml hGH for 48 h at 37 C. PKH26 staining was then determined again. PKH26 staining at each time point was determined flow cytometrically (FACS Calibur (Becton Dickinson and Co., Mountain View, CA) by excitation at 488 nm and measurement of emission with a standard filter set-up for phycoerythrin. Raw histogram data acquisition was deconvoluted into peaks using ModFit (Verity Software), and the proliferation index was calculated using Cell Proliferation Model software (Sigma Chemical Co.). The proliferation index is the ratio of the total number of cells at 48 h to the calculated number of cells in the parental generation.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
To address this issue, 3T3-F442A cells (a highly GH-responsive
murine preadipocyte line) were exposed to GH (500 ng/ml) or vehicle for
15 min before harvesting and detergent lysis. Detergent-soluble cell
extracts were then immunoprecipitated with antisera directed at the GHR
cytoplasmic domain, JAK2, IRS-1, IRS-2, and JAK1, as indicated (Fig. 1
). Precipitated proteins were resolved
by SDS-PAGE and immunoblotted with the monoclonal APT antibody 4G10. As
we have previously shown (29), anti-GHRcyt specifically
precipitated both tyrosine-phosphorylated GHR and JAK2 after GH
treatment of these cells. Tyrosine-phosphorylated JAK2 was specifically
precipitated by anti-JAK2, as expected (29), but was not significantly
detected in anti-IRS-1, anti-IRS-2, or anti-JAK1 precipitates. The
identities of the GHR, JAK2, IRS-1, IRS-2, and JAK1 in each direct
precipitate were verified by stripping and reprobing of the membrane
with each antibody (data not shown).
|
JAK2 specifically associates in vitro with amino-terminal regions
of IRS-1
As the stability of protein associations can be adversely affected
under conditions of coimmunoprecipitation, it can be difficult to
characterize weak associations using this technique. To analyze further
the association of JAK2 with IRS-1, we performed in vitro
affinity precipitation experiments in which immobilized GST fusion
proteins incorporating regions of IRS-1 were incubated with detergent
extracts from 3T3-F442A cells that had been treated with or without GH
before lysis.
As diagrammed in Fig. 2, IRS-1 contains
three important modules in its amino-terminal half, the PH, PTB, and
SAIN domains, which have been implicated in physical and functional
coupling to the IR (37, 38, 39, 51). The remainder of the molecule (which
we have arbitrarily divided into the pre-C-terminus and C-terminus
regions, as indicated) contains several critical tyrosine-containing
motifs that, when phosphorylated, form the actual docking sites for
SH2-containing signaling molecules (reviewed in Ref. 50). The GST/IRS-1
fusion proteins employed in this study span all but the
amino-terminal 20 residues of the IRS-1 molecule and in some cases
overlap in certain regions. They are designated GST/IRS-1(PH)
(including IRS-1 residues 21–203), GST/IRS-1(PH+PTB) (residues
21–400), GST/IRS-1(PTB+SAIN) (residues 108–516), GST/IRS-1(pre-C)
(residues 516–896), and GST/IRS-1(C-ter) (residues 899-1235). As a
positive control for binding of JAK2 in this assay system, we used a
previously described GST fusion protein, GST/hGHR-(271–389), which
incorporates the proximal one third of the human GHR cytoplasmic domain
and specifically associates with JAK2 (21). Anti-GST immunoblotting of
these glutathione-agarose-purified fusion proteins verified their
migration at the expected Mr in SDS-PAGE (not shown).
|
. Anti-JAK2 blotting of
precipitated proteins revealed a lack of association of JAK2 extracted
from vehicle- or GH-treated cells with GST alone, GST/IRS-1(pre-C), or
GST/IRS-1(C-ter) (lanes 1, 2, 7–10, 17, and 18). Importantly, these
results validate the specificity of the assay system in that not only
GST-coated glutathione-agarose beads, but also beads coated with GST
coupled to IRS-1 fragments of substantial size, yielded the same
negative result. As expected, GST/hGHR-(271–389) precipitated ample
JAK2 from extracts of cells treated with or without GH before lysis
(lanes 11 and 12), consistent with our previous findings (21).
Similarly, immunoprecipitated JAK2 was detected in each extract (lanes
13 and 14).
|
as well as multiple other such
experiments indicate that the most significant binding of JAK2, when
normalized for the abundance of anti-GST-blottable fusion protein in
the precipitate, was found with the GST/IRS-1(PTB+SAIN) fusion.
To assess the adequacy of GH stimulation in the experiments shown in
Fig. 3, we stripped and reprobed the anti-JAK2 blots with APT antibody
(lower panel). As expected, GH induced significant tyrosine
phosphorylation of specifically immunoprecipitated JAK2 (lane 14), and
the JAK2 affinity precipitated by GST/hGHR-(271–389) included the
GH-induced tyrosine phosphorylated JAK2 (lane 12). Notably,
GST/IRS-1(PH+PTB), GST/IRS-1(PTB+SAIN), and, to a lesser extent,
GST/IRS-1(PH) also precipitated the tyrosine-phosphorylated form of
JAK2 from GH-stimulated cell extracts (lanes 4, 6, and 16). Again, the
specificity of this assay is indicated by the lack of
tyrosine-phosphorylated JAK2 significantly detectable even by the
highly sensitive 4G10 antibody in the GST, GST/IRS-1(pre-C), and
GST/IRS-1-(C-ter) precipitates of extracts from GH-stimulated cells.
As expected, the GST/IRS-1 fusion proteins expressed in
Escherichia coli, were not tyrosine phosphorylated (as
determined by APT immunoblotting in experiments not shown). Thus, we
conclude from the experiments in Fig. 3 that, although amino-terminal
IRS-1 regions can interact with tyrosine-phosphorylated JAK2, the
IRS-1-JAK2 interaction does not require GH stimulation and tyrosine
phosphorylation of either IRS-1 or JAK2.
Our observation of a lack of ligand and phosphotyrosine requirement for
the IRS-1-JAK2 interaction in this assay system led us to test the
IRS-1-IR interaction in a similar system. Since the PTB and SAIN
regions of IRS-1 have been implicated as mediating interaction with the
IR, we assessed the ability of GST, GST/IRS-1(PH), GST/IRS-1(pre-C),
and GST/IRS-1(PTB+SAIN) fusions to precipitate the IR from extracts of
vehicle- and insulin-stimulated COS-7 cells transiently expressing the
IR. As shown in Fig. 4, A (anti-IR blot)
and B (APT blot), insulin-induced tyrosine-phosphorylated IR was
specifically precipitated by GST/IRS-1(PTB+SAIN), but not by the
negative control proteins, GST, GST/IRS-1(PH), and GST/IRS-1(pre-C).
This is in concert with previous data indicating that the interaction
of the IRS-1 PTB and SAIN regions with the IR is greatly increased by
insulin-induced IR tyrosine phosphorylation (37). Anti-IR
immunoblotting of equal aliquots of unprecipitated cell extract
verified that similar amounts of IR were expressed in both unstimulated
and insulin-treated cells (not shown). Thus, although PTB and/or SAIN
domain-containing fusions were best at precipitating both JAK2 and the
IR, the mechanism of IRS-1’s interaction with JAK2 differs
fundamentally from that with IR with regard to its lack of ligand
and/or phosphotyrosine dependence.
|
cDNAs encoding IRS-1 and the rGHR were subcloned into the pSX eukaryotic expression vector and stably transfected into 32D cells, as described in Materials and Methods. rGHR-expressing stably transfected pools that coexpressed either no IRS-1 or IRS-1 were selected both negatively (survival in histidinol- and G418-containing medium by virtue of cotransfected histidinol and G418 resistance markers) and positively (by growth in bovine GH-containing FCS-supplemented medium in the absence of IL-3) (29). The resultant transfectant cell lines were designated 32D-rGHR and 32D-rGHR-IRS-1.
To document expression of the transfected IRS-1 molecule, detergent
extracts of vehicle- and GH-stimulated 32D-rGHR and 32D-rGHR-IRS-1
cells were resolved by SDS-PAGE and immunoblotted sequentially with
anti-IRS-1 (Fig. 5A, left
panel) and APT (Fig. 5A, right panel) antibodies. As
expected, based on previous results (52), 32D-rGHR had no detectable
IRS-1 protein. The transfected IRS-1 in 32D-rGHR-IRS-1 cells was easily
detectable by blotting; the dominant immunoreactive form of this
protein migrated at the expected Mr of 160–170 kDa, but
several less abundant lower Mr forms, presumably resulting
from proteolysis, were also observed. APT immunoblotting revealed that
transfected IRS-1 was significantly tyrosine phosphorylated, even in
the absence of GH stimulation (Fig. 5A, right panel). This
degree of basal tyrosine phosphorylation may be a reflection of the
abundant level of IRS-1 expression in 32D-rGHR-IRS-1 (although we note
that substantial basal tyrosine phosphorylation of endogenous IRS-1 was
also observed in 3T3-F442A cells; Fig. 1). In response to GH
stimulation, a modest increase in IRS-1 tyrosine phosphorylation was
detected in 32D-rGHR-IRS-1. In contrast to the significant basal and
only modestly GH-enhanced levels of tyrosine phosphorylation of
transfected IRS-1, specific immunoprecipitation from 32D-rGHR and
32D-rGHR-IRS-1 indicated that both cells displayed similar abundance of
endogenous JAK2 (Fig. 5B, left panel) and that JAK2 tyrosine
phosphorylation was similarly and significantly GH inducible in both
cell lines (Fig. 5B, right panel). [We also note that the
time course of GH-induced JAK2 tyrosine phosphorylation was similar in
the two cell lines (data not shown).] Consistent with this similarity
in GH-induced JAK2 activation, the abundance of transfected rGHR
present in each cell was also quite comparable, as assessed by both
anti-GHRcyt immunoblotting of detergent extracts (Fig. 5C, left panel) and [125I]hGH cell surface binding
(Fig. 5C, right panel). [As previously observed (29),
nontransfected 32D cells express no GHRs, as assessed by either of
these two assays.]
|
). Equal aliquots of 32D-rGHR and
32D-rGHR-IRS-1 cells were incubated in serum-free medium in the
presence of various concentrations of recombinant hGH for 24 h.
The number of viable cells present after this incubation period
relative to the number present after incubation with serum-free medium
only was assessed by a well validated nonradioactive cell proliferation
assay based on the cellular conversion of a tetrazolium salt into a
formazan product that can be detected by an enzyme-linked immunosorbent
assay plate reader, as described in Materials and Methods.
As we previously observed (29), 32D-rGHR cells increased their number
in a GH concentration-dependent fashion (maximal response of roughly
2-fold in comparison to no added GH; ED50, 
0.1 ng/ml of
hGH) in this assay. Interestingly, 32D-rGHR-IRS-1 cells exhibited a
dramatic enhancement of GH-induced increase in cell number; although
the ED50 for these cells was quite similar to that for
32D-rGHR, the maximal response was significantly augmented (roughly
6-fold over no added GH) in the cells expressing WT IRS-1 compared to
that in cells expressing no IRS-1. This same pattern of enhanced
GH-induced proliferation in cells expressing IRS-1 was observed in
separate experiments in which cell viability was assessed by standard
cell counts of trypan blue-excluding cells (not shown).
|
, fluorescence data acquired immediately after
cell labeling (T0) indicated that each cell was equally
labeled by PKH26. After 48 h of incubation in the presence of hGH
(10 ng/ml) (T48), 32D-rGHR-IRS-1 exhibited markedly
enhanced proliferation potential relative to 32D-rGHR; 66% of cells
underwent at least one cell division, compared with 10% of 32D-rGHR
cells. Thus, the proliferative index for 32D-rGHR-IRS-1 (1.62) was
substantially greater than that of 32D-rGHR (1.07). In concert with the
findings in Fig. 6A, these results argue that GH-induced proliferation
is enhanced by expression of IRS-1 and suggest that increased
mitogenesis is a factor that contributes to this IRS-1 enhancement of
the GH-induced increase in the number of viable cells. It is possible,
however, that a decrease in cell death may also contribute to this
IRS-1 effect.
IRS-1 has been implicated as being involved in activation of the MAP
kinase pathway (40) and might thereby be related to growth
factor-induced mitogenesis. We, therefore, tested whether 32D-rGHR
manifested GH-induced MAP kinase activation and, if so, whether IRS-1
had effects on the magnitude of this activation. As a proxy for
enzymatic activation, we monitored GH-induced MAP kinase
phosphorylation by immunoblotting cell extracts with a state-specific
anti-phospho-MAP kinase antibody. This antibody specifically recognizes
the phosphorylated threonine 183 and tyrosine 185 residues in the MAP
kinase molecule that correlate to its enzymatic activation. Treatment
of cells with GH for 15 min resulted in MAP kinase phosphorylation in
both 32D-rGHR and 32D-rGHR-IRS-1 (Fig. 7A), indicating that MAP kinase
activation by GH is not entirely IRS-1 (or -2)-dependent. However,
expression of IRS-1 significantly augmented the level of GH-induced MAP
kinase phosphorylation compared to that in 32D-rGHR. Stripping and
reprobing of this blot with anti-MAP kinase (Fig. 7B) confirmed the
presence of similar amounts of MAP kinase in each extract. The pooled
data from the experiment shown in Fig. 7, A and B, and two other
separate experiments are displayed graphically in Fig. 7C and indicate
the reproducibility and significance of these results. Further
stripping and reprobing with APT antibody also confirmed, as expected
given the anti-phospho-MAP kinase blotting result, that GH-induced
tyrosine phosphorylation of MAP kinase in both cell lines, but showed
that it was greatly augmented only in the 32D-rGHR-IRS-1 line (not
shown). Further, other experiments (not shown) indicated that the
relative time course of GH-induced MAP kinase phosphorylation was
similar in both cell types. Therefore, we conclude that, similar to the
findings for GH-induced proliferation, the degree of MAP kinase
activation induced by GH is greatly augmented by the presence of
IRS-1.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In this study, we confirm that GH stimulation of the 3T3-F442A preadipocyte leads to enhanced tyrosine phosphorylation of IRS-1 and that a tyrosine phosphoprotein that specifically coimmunoprecipitates with JAK2 is likely to be IRS-1. Importantly, our in vitro affinity precipitation studies extend our understanding of the JAK2-IRS-1 interaction in two ways. First, they provide clear evidence that IRS-1’s ability to interact with JAK2 in such an assay is reliant on amino-terminal regions of the IRS-1 molecule. Second, since the GST/IRS-1 fusion proteins are bacterially expressed and are therefore not tyrosine phosphorylated (this has been verified by APT immunoblotting; not shown), and JAK2’s ability to associate with the IRS-1 amino-terminal region-containing fusion proteins does not depend on its tyrosine phosphorylation, we conclude that this IRS-1-JAK2 interaction is nonphosphotyrosine dependent.
Although these findings do not preclude an element of phosphotyrosine
dependence to the IRS-1-JAK2 interaction as it naturally occurs within
cells, we note that GST/IRS-1-IR in vitro affinity
precipitation experiments using the same fusion proteins (Fig. 4 in
this report and Ref. 37) faithfully reflect the IR phosphotyrosine
dependence of the IR-IRS-1 interaction seen in intact cells and in the
yeast two-hybrid system (37). Therefore, our data are consistent with
the idea that a significant element of the IRS-1-JAK2 interaction in
cells may be phosphotyrosine independent and that IRS-1 and JAK2 might
be associated even in the absence of GH stimulation. The finding that
JAK2 extracted from cells previously exposed to GH is less precipitable
by the relevant GST/IRS-1 fusion proteins than JAK2 from unstimulated
cells raises the possibility that GH treatment changes the affinity of
endogenous IRS-1 and/or other cellular proteins for JAK2; in this
respect, an element(s) of the IRS-1-JAK2 interaction may indeed be GH
dependent. This issue requires further study. Similarly, we do not yet
know whether the IRS-1-JAK2 interaction is direct or is mediated by
other proteins present in cells or cell extracts. Further, based on the
inability of us and others to reliably coimmunoprecipitate
immunoblottable IRS-1 with JAK2, we believe that whether direct or
indirect, the association of IRS-1 and JAK2, although specific, is
weak.
Our finding that the PTB domain of IRS-1, in combination with the PH or SAIN domain, is the most avid region of IRS-1 for association with JAK2 is particularly interesting. The IRS-1 PTB and SAIN domains are also the regions of the molecule that most avidly interact with the tyrosine-phosphorylated IR. PTB domains were originally described as regions within certain molecules that bind peptides phosphorylated on tyrosine residues. PTB domains were distinguished from SH2 domains in that the PTB domain binding specificity is determined by residues that lie amino-terminal, rather than carboxyl-terminal, to the phosphotyrosine residue (56, 57, 58, 59). More recently, however, there have been descriptions of important interactions between PTB domains and motifs similar to NPXY without a requirement for tyrosine phosphorylation within the motif (60) and between the PTB domain and an Asparagine Proline Leucine Histidine sequence (61). These examples of nonphosphotyrosine-dependent PTB domain-mediated interactions have led some investigators to refer to the PTB domain as a general protein interaction domain (60). Similarly, there are also a number of examples of nonphosphotyrosine-dependent interactions involving SH2 domains, although in these instances the target motifs are not yet mapped (62, 63, 64, 65). Our findings raise the possibility that the IRS-1 PTB domain might serve as both a phosphotyrosine-binding domain (e.g. in its interaction with the IR) and a protein interaction domain (e.g. in its interaction, whether direct or indirect, with JAK2). Such an arrangement raises the further possibility that such mechanistic heterogeneity in the interaction of IRS-1 with the insulin and GH (and perhaps other cytokine) signaling pathways may underlie the complexity of the physiological interactions between these hormones.
Given the structural similarity between PH and PTB domains (66), it is
notable that we also observe specific interaction between the IRS-1 PH
domain and JAK2, although this interaction appears less pronounced than
that of the PTB domain and JAK2. Others have similarly observed a
specific nonphosphotyrosine-dependent in vitro interaction
between IRS and JAK family members; TYK2 extracted from U-266 cells
interacted with a GST/IRS-2 PH domain-containing fusion protein
independently of prior stimulation of the cells with interferon-
(67). Although the IRS-1 PH domain has been shown to be quite important
in functionally coupling IRS-1 to IR signaling, a physical interaction
between the IRS-1 PH domain and the IR has not been observed. Thus, it
is possible that the use of IRS proteins in cytokine receptor/JAK
family signaling may differ from that in IR signaling in this aspect as
well. We note the recent findings of Gaul et al. that JAK2
may directly interact with the IR and be tyrosine phosphorylated in
response to insulin (68); we as yet have no information as to whether
our observed PH- or PTB/SAIN-mediated interactions between IRS-1 and
JAK2 have an impact on the degree to which insulin treatment leads to
IR-induced IRS-1 association and tyrosine phosphorylation.
Although GH could, in principle, employ IRS-1 in signaling several of its various physiological actions (insulinomimesis, insulin antagonism, proliferation, etc.), we opted for relative simplicity in assaying whether IRS-1 influenced GH-induced proliferation in the 32D cell. Although this is not a cell in which GH signaling normally has a role, it is a convenient factor-dependent line that lacks IRS-1 and -2, and we have already characterized various aspects of GH-induced biochemical activation in these cells reconstituted with the GHR (29). The finding that 32D-rGHR exhibited GH-dependent proliferation indicates that IRS-1 (or IRS-2) is not required for the GHR to couple to mitogenic and/or antiapoptotic signaling pathways in this cell. Using several assays of cell proliferation, however, we observe that expression of IRS-1 confers enhanced GH-induced proliferation. Similar effects of IRS-1 expression on growth factor- and cytokine-induced proliferation have been documented in 32D and other cell systems (35, 51, 69). For IL-4, at least part of the augmentation in cytokine-induced proliferation conferred by IRS-1 expression in 32D cells is thought to be due to protection from apoptosis (54).
Although we do not yet know the exact mechanism(s) involved in IRS-1’s enhancement of GH-induced proliferation, we note that GH-induced MAP kinase activation is significantly augmented in 32D-rGHR-IRS-1 cells in comparison to 32D-rGHR cells. We are actively pursuing whether this IRS-1-mediated increased MAP kinase activation relates to the enhanced inclusion of particular Ras pathway activators in the GH-activated GHR/JAK2 complex and/or whether IRS-1 facilitates GH-induced access to other pathways leading to MAP kinase activation. It is tempting to speculate that the coupling of GH to pathways related to glucose transport and metabolism might similarly be modulated by IRS-1. Surprisingly, insulin-mediated glucose transport in 3T3-L1 adipocytes was, however, recently shown to be unaffected by overexpression of the IRS-1 PTB or SAIN domains despite the inhibition by these IRS-1 fragments of insulin-induced tyrosine phosphorylation of endogenous IRS-1 and SHC, IRS-1-associated PI-3K activation, p70s6k activation, and MAP kinase phosphorylation (70). In the same study, insulin-induced cell cycle progression of HIRcB fibroblasts was also inhibited by overexpression of these IRS-1 domains (70). Although we have yet to test the role of IRS-1 in GH-induced metabolic signaling, these findings and others indicating that GH-induced glucose transport in 3T3-L1 adipocytes can be independent of PI-3K (71) compel us to consider the possibility that IRS-1 may be linked to the GH signaling system primarily to affect proliferative, rather than metabolic, effects of GH.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received November 3, 1998.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
, and leukemia
inhibitory factor promoted tyrosyl phosphorylationn receptor
substrate-1. J Biol Chem 270:14685–14692
