(Received for publication, September 11, 1996, and in revised form, November 26, 1996)
From the Research Division, Joslin Diabetes Center,
and Department of Medicine, Brigham and Women's Hospital, Harvard
Medical School, Boston, Massachusetts 02215 and the
§ Istituto di Clinica Medica, Endocrinologia e Malattie
Metaboliche, University of Bari School of Medicine,
Bari, Italy 70124
The stimulation of
phosphatidylinositol (PI) 3-kinase by insulin-like growth factor I
(IGF-I) in L6 cultured skeletal muscle cells is inhibited by the
glucocorticoid dexamethasone. The objective of this study was to
investigate the mechanism of dexamethasone action by determining its
effects on the expression of the p85 and p85
regulatory subunit
isoforms of PI 3-kinase, their coupling with the p110 catalytic
subunit, and their association with insulin receptor substrate 1 (IRS-1) in response to IGF-I stimulation. Dexamethasone induced a 300%
increase in p85
protein content in the L6 cultured myoblast cell
line, whereas it increased p110 content by only 38% and had no effect
on p85
. The increase in p85
protein was associated with a
coordinate increase in p85
mRNA. Stimulation with IGF-I induced
the association of p85
and p85
with IRS-1, and this was
accompanied by increased amounts of the p110 catalytic subunit and
markedly increased PI 3-kinase activity in IRS-1 immunoprecipitates. In
cells treated with dexamethasone, greater amounts of p85
and lower
amounts of p85
, respectively, were found in IRS-1
immunoprecipitates, such that the
/
ratio was markedly higher
than in control cells. In spite of the increase in both total and
IRS-1-associated p85
following dexamethasone treatment,
IRS-1-associated p110 catalytic subunit and PI 3-kinase activity were
decreased by approximately 50%. Thus, dexamethasone induces a specific
increase in expression of the p85
regulatory subunit that is not
associated with a coordinate increase in the p110 catalytic subunit of
PI 3-kinase. As a consequence, in dexamethasone-treated cells, p85
that is not coupled with p110 competes with both p85
·p110 and
p85
·p110 complexes for association with IRS-1, leading to increased p85
but decreased p85
, p110, and PI 3-kinase activity in IRS-1 immunoprecipitates.
Growth factor activation of transmembrane tyrosine kinase
receptors results in rapid recruitment of phosphatidylinositol
(PI)1 3-kinase activity to
tyrosine-phosphorylated proteins. In intact cells, PI 3-kinase
catalyzes the phosphorylation of PI 4,5-bisphosphate (PI-4,5-P2) at the 3-position of the inositol ring, thus
leading to elevation in intracellular PI 3,4,5-trisphosphate
(PI-3,4,5-P3) (reviewed in Ref. 1). Unlike the products of
PI kinases in the classical PI cycle (PI-4-P and
PI-4,5-P2), 3-phosphorylated phosphoinositides are not
cleaved by phospholipase C-
(2, 3), and it has been suggested that
they may serve as intracellular second messengers for yet unidentified
in vivo targets (1, 4). Extensive experimental evidence has
established a key role for PI 3-kinase in the signal transduction
mechanisms of a number of peptide growth factors, including epidermal
growth factor, platelet-derived growth factor (PDGF), insulin, and
insulin-like growth factor-I (IGF-I). PI 3-kinase has thus been
implicated in the regulation of multiple general and specialized
cellular processes, including membrane ruffling (5), receptor
endocytosis (6), mitogenesis (7, 8), cell differentiation (9), and
insulin stimulation of glucose transport (10-12) and glycogen synthesis (13, 14).
Mammalian PI 3-kinase is a heterodimer composed of an 85-kDa (p85)
regulatory subunit and a 110-kDa (p110) catalytic subunit (15, 16). Two
distinct and closely related 85-kDa protein isoforms, p85 and
p85
, have been cloned and shown to be the products of separate genes
(17-20). Both of these p85 proteins have the capacity to form stable
high affinity complexes with the p110 component of PI 3-kinase (21,
22). The two p85 isoforms have a multidomain structure, containing two
SH2 (src homology 2) domains, one SH3 domain, and
a region with significant sequence similarity to a GTPase-activating
protein domain of the product of the breakpoint cluster region gene
(23). The presence of several functional domains suggests that the p85
proteins may have multiple interactive and regulatory roles. At
present, functional differences between the p85
and p85
protein
isoforms have not been established.
Two forms of p110 also have been cloned, one from bovine brain
designated p110 (20), and a second human variant designated p110
(24). Expression studies have demonstrated that both p110 proteins have
intrinsic PI 3-kinase activity and can associate with the p85 component
in intact cells (20, 24). The domains in p85 and p110 required for
subunit interaction have been identified and mapped to an amino acid
sequence between the two SH2 domains of p85 and an
NH2-terminal amino acid sequence of p110, respectively (20,
25-27). These studies have established a structural model of the PI
3-kinase complex, in which the p110 subunit contains catalytic activity
and is tightly associated with the p85 subunit, which acts as an
adaptor and/or regulatory subunit. Integrity of the p85·p110 complex
appears to be necessary for p110 catalytic activity (26). Thus,
overexpression of the p85 subunit or a portion of the p85 protein, such
as an intact p85 SH2 domain, through transfection or microinjection of
cells results in inhibition of PI 3-kinase activation and cell
signaling (28, 29). The physiological occurrence of selective
up-regulation of p85 expression as a mechanism of inhibition of PI
3-kinase activity has not been investigated.
The best characterized mode of PI 3-kinase activation in response to
peptide growth factors involves changes induced in the p85 protein upon
binding to certain phosphorylated tyrosine residues, which are then
transmitted to the associated p110 catalytic subunit and cause its
activation. In the case of the epidermal growth factor and PDGF
receptors, the p85 protein binds directly to phosphorylated tyrosines
in the receptor molecule through its SH2 domains (reviewed in Ref. 1).
In the case of the receptors for insulin and IGF-I, only a limited
fraction of the total cell PI 3-kinase associates directly with the
receptor, while most of PI 3-kinase interacts with specific
tyrosine-phosphorylation sites of the receptor substrates IRS-1 and
IRS-2 (30-32). Binding of p85 to tyrosine-phosphorylated IRS-1 or
phosphorylated IRS-1-related YMXM peptide sequences results in
increased catalytic activity of the PI 3-kinase complex (33, 34).
Tyr-608 and Tyr-939 of IRS-1 appear to be the predominant sites for
interaction with the amino-terminal SH2 domain of p85
(35). Although
these studies have provided important insight into the mechanism of PI
3-kinase activation in response to insulin and IGF-I, limited
information is available on the relative abundance of p85
and p85
within the IRS-1·PI 3-kinase complex and the coupling of
IRS-1-associated p85 isoforms with catalytically active p110
subunits.
In a previous report (36), we described the inhibition of IGF-I
activated PI 3-kinase activity by the glucocorticoid dexamethasone in
the L6 skeletal muscle cell line. The objective of this study was to
investigate the effects of dexamethasone on the expression of p85
and p85
isoforms of PI 3-kinase in L6 cells, their coupling with the
p110 catalytic subunit, and their association with IRS-1 in response to
IGF-I stimulation. We show that L6 myoblasts express both p85
and
p85
and that the expression of p85
, but not that of p85
, is
specifically increased by dexamethasone. The increase in the cellular
pool of p85
correlates with an increased amount of this protein
associated with IRS-1 after IGF-I stimulation. However, despite the
increase in IRS-1-associated p85
, both p110 and PI 3-kinase activity
associated with IRS-1 are diminished. These data support the concept
that a substantial fraction of p85
is "free" (i.e.
not coupled with p110) in dexamethasone-treated cells. We propose that
the selective increase in p85
expression may represent a novel
physiological mechanism leading to inhibition of PI 3-kinase activity
by glucocorticoids.
Culture medium and donor calf
bovine serum were from Life Technologies, Inc. Recombinant human
insulin and recombinant human IGF-I were gifts from Eli Lilly
(Indianapolis, IN). PDGF-BB was purchased from Genzyme (Boston, MA).
Protein A-Sepharose was from Pierce (Rockford, IL). Reagents for
SDS-PAGE and the Bradford protein assay were purchased from Bio-Rad.
125I-Labeled protein A and Tran35S-label were
from ICN, [-35S]dATP was from Amersham, and
[
-32P]ATP was from DuPont NEN. Oligonucleotides were
purchased from Bio-Synthesis (Lewisville, TX). Moloney murine leukemia
virus-H
Reverse Transcriptase was from Life Technologies,
Inc. Ampli-Taq polymerase and other reagents for PCR
amplification were from Perkin Elmer. Dexamethasone and other chemicals
were from Sigma.
Polyclonal phosphotyrosine (anti-PY) antibody was prepared in rabbits
by injection of phosphotyrosine polymerized by
1-ethyl-3(3-dimethyl-aminopropyl)carbodiimide with alanine, threonine,
and keyhole limpet hemocyanin, and purified by affinity chromatography
on a phosphotyramine-Sepharose column as described previously (37).
Polyclonal anti-peptide antibodies against a
Tyr-Ala-Ser-Ile-Asn-Phe-Gln-Lys-Gln-Pro-Glu-Asp-Arg-Gln peptide,
corresponding to the last 14 amino acids in the COOH-terminal region of
rat IRS-1 (38), were generated and purified on specific peptide
affinity columns as described previously (36, 37). Polyclonal anti-p85
(anti-p85) antibodies that recognize both and
isoforms of p85
were purchased from Upstate Biotechnology Inc. Monoclonal antibodies
specific to p85
(anti-p85
) were from Transduction Laboratories.
Monoclonal antibodies to the p110 subunit of PI 3-kinase (anti-p110)
were a generous gift of Dr. L. T. Williams (University of California,
San Francisco, CA).
The line of L6 rat skeletal muscle cells has
been described previously (36). Cells were cultured in MEM supplemented
with 10% donor calf bovine serum, 2 mM glutamine, and
nonessential amino acids in a 5% CO2 atmosphere at
37 °C. Cells were plated in MEM containing 10% donor calf bovine
serum in 150-mm culture dishes. On day 4 after plating, when the cells
were 50-60% confluent, the medium was replaced with serum-free MEM
containing 0.5% bovine serum albumin, and cells were incubated in the
presence or absence of 1 µM dexamethasone (concentration
confirmed by absorbance at 242 nm using a molar absorbance coefficient
of 1.5 × 104 M1
cm
1) for the indicated times. All experiments were
carried out with undifferentiated myoblasts.
L6 cell
monolayers were rinsed twice with ice-cold phosphate-buffered saline
and solubilized directly on the tissue culture plates with 4 M guanidinium isothiocyanate, 0.1 M Tris-HCl,
pH 7.5, 0.66% N-laurylsarcosine, and 5%
-mercaptoethanol. Total cellular RNA was isolated by low temperature
4 M guanidinium isothiocyanate-phenol-chloroform extraction, followed by cold ethanol precipitation (39), and quantitated by spectrophotometry at 260 nm. In all samples, intact ribosomal RNA bands were visualized after electrophoresis. To quantitate p85
and p85
mRNA content in L6 cells, a polymerase chain reaction (PCR) amplification method was used with oligonucleotide primers complementary to sequences that are identical or highly homologous in p85
and p85
mRNA but flank regions of different sizes, such that amplified cDNA fragments from the two p85 isoforms could be separated by polyacrylamide gel electrophoresis (40). Specific
first strand cDNA copies of p85
and p85
mRNAs were synthesized using 100 units of Moloney murine leukemia
virus-H
Reverse Transcriptase in a 20-µl reaction
volume containing 3-5 µg of total cell RNA and 0.75 µM
reverse primer 5
-GTACAGGTTGTAGGGCTC-3
at 42 °C. This primer is
complementary to a nucleotide sequence that is identical in bovine
p85
and p85
(nucleotides 2064-2047 and 2046-2029 in p85
and
p85
cDNA sequences, respectively) and in bovine and mouse p85
(17, 19). The reverse transcription products were combined with 2.5 units of Ampli-Taq, 1 × PCR buffer, 100 µM dNTPs, 4 mM MgCl2, and 0.15 µM oligonucleotide primers in a 100-µl final reaction
volume for PCR amplification. The sense primer for PCR amplification
was a (1:1) mixture of the oligonucleotides 5
-GACAAACGCATGAACAG-3
and
5
-GACAAGCGCATGAACAG-3
, corresponding to nucleotides 1678-1694 and
1657-1693 of bovine p85
and bovine p85
cDNAs, respectively.
The oligonucleotide sequence 5
-GACAAACGCATGAACAG-3
is identical in
bovine and mouse p85
(17, 19). The antisense primer was the same one
used for reverse transcription. Based on p85
and p85
cDNA
sequences, PCR amplification with these primers should generate
cDNA products of 386 and 389 bases for p85
and p85
,
respectively. 30 cycles of PCR were performed with denaturation at
95 °C for 1 min, annealing at 50 °C for 1 min, and extension at
72 °C for 1 min. PCR products were 35S-labeled by using
a sense primer phosphorylated with T4 polynucleotide kinase (Life
Technologies, Inc.) in the presence of [35S]ATP.
Amplified cDNA fragments were precipitated with ethanol, resolved
on a 6% denaturing polyacrylamide gel, and visualized by
autoradiography. Scanning optical densitometry (Molecular Dynamics, Sunnyvale, CA) was performed to quantify the relative amounts of
amplified cDNA species.
After incubation in serum-free medium without or with dexamethasone for the indicated times, cells were left untreated or stimulated with IGF-I (100 nM) for 10 min, washed once with ice-cold phosphate-buffered saline containing 100 µM sodium orthovanadate and twice with 20 mM Tris-HCl (pH 7.6) containing 137 mM NaCl, 1 mM MgCl2, 1 mM CaCl2, and 100 µM sodium orthovanadate (Buffer A). The cells were lysed in Buffer A (1 ml/150-mm dish) containing 1% Nonidet P-40, 10% glycerol, 10 mM sodium pyrophosphate, 10 mM sodium fluoride, 2 mM EDTA, 2 mM phenylmethylsulfonyl fluoride, 4 µg/ml leupeptin, and 2 mM sodium orthovanadate (lysis buffer). Insoluble material was removed by centrifugation at 12,000 × g for 10 min at 4 °C, the protein concentration in the resulting supernatant was determined with the Bradford dye binding assay (41), and the final protein concentration was adjusted to 2 mg/ml with lysis buffer.
Immunoprecipitation was carried out by incubation of the cell lysate overnight at 4 °C with anti-IRS-1 antibodies and adsorption of resulting immune complexes to protein A-Sepharose beads for 2 h at 4 °C. The pelleted beads were washed successively in phosphate-buffered saline containing 1% Nonidet P-40 and 100 µM sodium orthovanadate (three times), 100 mM Tris-HCl, pH 7.6, containing 500 mM LiCl and 100 µM sodium orthovanadate (three times), and 10 mM Tris-HCl, pH 7.6, containing 100 mM NaCl, 1 mM EDTA, and 100 µM sodium orthovanadate (twice).
PI 3-Kinase AssayThe washed immunoprecipitates were
resuspended in 50 µl of 10 mM Tris-HCl, pH 7.6, containing 100 mM NaCl, 1 mM EDTA, and 100 µM sodium orthovanadate, and then combined with 10 µl
of 100 mM MgCl2 and 10 µl of 2 µg/µl
phosphatidylinositol (Avanti) that had been sonicated in 10 mM Tris-HCl, pH 7.6, containing 1 mM EGTA. The
PI 3-kinase reaction was started by the addition of 10 µl of 10 mM ATP containing 30 µCi of [-32P]ATP.
After 10 min at 22 °C with constant vortexing, the reaction was
stopped by the addition of 20 µl of 8 N HCl and 160 µl
of chloroform/methanol (1:1, v/v). The mixture was vigorously mixed and
centrifuged to separate the phases, and the lower organic phase was
removed and applied to a silica gel TLC plate precoated with 1%
potassium oxalate (Merck, Darmstadt, Germany). The plates were
developed in chloroform/methanol/water/ammonia (60:47:11.3:2, v/v), as
described previously (36). The PI-3 product was identified by its
comigration with a PI-4 standard and quantitated by scanning densitometry.
For identification and quantitation of
specific proteins, cell lysates were prepared with 1% Nonidet P-40 as
described above and analyzed by immunoblotting either directly or
following immunoprecipitation with anti-IRS-1, anti-p85, or anti-p110
antibodies, as indicated. Immunoprecipitates or total cell lysates for
immunoblotting were boiled in Laemmli buffer with 100 mM
dithiothreitol for 4 min, electrophoresed on 8% SDS-polyacrylamide
gels, and transferred by electroblotting onto nitrocellulose sheets
(Schleicher & Schuell) as described previously (37). The transfer
buffer used was 10 mM Tris, 192 mM glycine,
20% methanol (v/v), and 0.02% SDS, and blotting was carried out at 80 volts for 2.5 h. The membranes were incubated with anti-IRS-1,
anti-p85, anti-p85, or anti-p110 antibodies as appropriate under
conditions reported previously (37). Specific protein bands on
autoradiographic images were quantified by scanning optical
densitometry, and the data were expressed using arbitrary units
normalized to control values within each gel.
Subconfluent cell monolayers in 100-mm dishes were washed
twice in phosphate-buffered saline and incubated in serum-free, methionine-free MEM containing 0.1 mCi/ml Tran35S-label at
37 °C for 20 h. Cell lysates then were prepared as described
above and precleared by mixing for 16 h at 4 °C with 1 µl of
normal goat serum (pre-adsorbed to 30 µl settled volume of protein
A-Sepharose) for each ml of lysate. Anti-IRS-1 antibody was added (5 µg adsorbed for 2 h at 4 °C to 30 µl of protein A-Sepharose for each ml of precleared supernatant) and mixed for 4 h at
4 °C. The Sepharose beads were washed as described previously,
boiled in Laemmli buffer with 100 mM dithiothreitol for 4 min, and the solubilized proteins electrophoresed on 8%
SDS-polyacrylamide gels. Gels were prepared for fluorography using
AmplifyTM (Amersham Int., Little Chalfont, United Kingdom),
dried at 80 °C under vacuum, and exposed to Kodak XAR-5 film at
80 °C for 6-36 h.
Data are presented as mean ± S.E. Statistical analysis was performed by paired and unpaired Student's t tests as appropriate.
Two isoforms of the p85 regulatory subunit of PI
3-kinase have been described, p85 and p85
, that are expressed in
several, although not all, cell types (16, 19). Previous work has
demonstrated that p85 is expressed in rat skeletal muscle, but the
isoform pattern in this tissue is not known (42). To investigate the expression of p85 isoforms in rat L6 skeletal muscle cells, we employed
polyclonal antibodies to the full-length 85-kDa subunit of p85
that
detect both p85
and p85
(anti-p85) and a monoclonal antibody
raised against a 19-kDa fragment corresponding to the COOH terminus of
p85
that specifically recognizes p85
(anti-p85
). As shown in
Fig. 1A, two distinct protein species of 85 and 87 kDa were detected by direct immunoblotting with the
isoform-nonspecific anti-p85 antibodies. The apparent molecular weights
of these two protein species are similar to the reported size of p85
and p85
expressed in a reticulocyte lysate expression system (19).
Immunoprecipitation of Nonidet P-40 extracts of L6 cells with anti-p85
prior to immunoblotting produced similar results, although the ratio of
immunoreactive p85
to immunoreactive p85
was considerably greater
as compared to direct immunoblotting (Fig. 1A), likely
reflecting the relatively higher affinity of anti-p85 antibodies for
native as compared to denatured p85
. Immunoblotting of these
immunoprecipitates with the
isoform-specific antibody anti-p85
confirmed the identity of the lower 85-kDa protein as p85
(Fig.
1B). These data indicate that both p85
and p85
are
expressed in rat L6 myoblasts, migrating as 85- and 87-kDa proteins,
respectively, on SDS-polyacrylamide gels.
Regulation of p85
To
investigate the effects of the glucocorticoid dexamethasone on the
cellular content of p85 and p85
, serum-starved L6 myoblasts were
incubated in the presence or absence of 1 µM
dexamethasone for different times, and the total cell content of p85
and p85
was determined by immunoblotting with the polyclonal
anti-p85 antibody. The content of p85
and p85
did not change in
control L6 myoblasts over the 72-h study period (Fig.
2A). In cells treated with 1 µM
dexamethasone, the amount of p85
progressively increased in a
time-dependent manner (Fig. 2A). After 72 h
of treatment, p85
content was approximately 4-fold higher in
dexamethasone-treated L6 myoblasts compared to control (Fig.
2B; p < 0.05). By contrast, dexamethasone
did not alter the amount of p85
(Fig. 2, A and B), indicating that the effect on the
isoform of p85 was
specific.
To determine whether the increase in the p85 regulatory subunit of
PI 3-kinase was associated with a coordinate increase in the catalytic
p110 subunit, similar L6 myoblast preparations were analyzed by
immunoblotting with anti-p110, a specific monoclonal antibody raised
against mouse p110 (26). As shown in Fig. 3A, one protein band of 110 kDa was recognized by anti-p110 in myoblast extracts. Dexamethasone treatment for 72 h increased p110 content by 38% (Fig. 3B; p < 0.05). However, the
magnitude of this response was much less than the 300% increase in
p85
(Fig. 2A), indicating that the effects of
dexamethasone on the regulatory subunit p85
and the catalytic
subunit p110, respectively, were not coordinate.
To assess whether augmentation of p85 versus p85
by
dexamethasone could involve a selective increase in p85
mRNA,
p85
and p85
mRNA levels were measured by quantitative PCR.
For this purpose, total RNA was extracted from cells incubated in the
presence or absence of 1 µM dexamethasone for 48 h
and subjected to reverse transcription and PCR amplification. The
p85
and p85
mRNAs were co-amplified in the same PCR reaction
using primers complementary to identical or highly homologous
nucleotide sequences in the p85
and p85
genes that flank
intervening fragments of different length. 35S-dATP was
included in the PCR mixture during the reaction and resulting
35S-labeled PCR products were resolved on a denaturing
acrylamide gel and visualized by autoradiography. The correct size of
the PCR fragments was confirmed by comparison with a sequencing
reaction run on the same gel (not shown). A representative
autoradiogram of the PCR products is illustrated in Fig.
4A. The products of 386 and 389 bases
correspond to the PCR products obtained by amplification of p85
and
p85
mRNA, respectively. With this method, the p85
PCR product
was predominant over the p85
PCR product, likely reflecting higher
p85
than p85
mRNA levels in control L6 myoblasts (ratio of
10:1, with a similar number of 35S-labeled adenylic acid
molecules in either PCR fragment). Dexamethasone treatment did not
alter p85
mRNA levels, but induced a 3-fold increase in the
amount of p85
mRNA compared to control (Fig. 4B).
These results are consistent with the observed effects of dexamethasone
on p85
and p85
protein levels, indicating that the specific
augmentation of p85
protein content in L6 myoblasts can be explained
by increased levels of p85
mRNA.
PI 3-Kinase Activity Associated with IRS-1
The mechanism of
PI 3-kinase activation in L6 cells stimulated with IGF-I involves
association of the p85 subunit to tyrosine-phosphorylated residues of
the substrate IRS-1 and subsequent activation of the lipid kinase
intrinsic to the p110 subunit. Although IGF-I also increases the amount
of PI 3-kinase activity recoverable in anti-IGF-I receptor
immunoprecipitates, receptor-associated activity represents a minor
fraction compared to IRS-1-associated PI 3-kinase activity in L6
myoblasts (36). Since dexamethasone treatment induced a marked increase
in the cellular content of p85 not paralleled by similar changes in
either p85
or p110, the cellular pool of p85
should be in excess
relative to other PI 3-kinase components. The consequences of the
cellular excess of p85
on activation of IRS-1-associated PI 3-kinase
were studied by measuring PI 3-kinase activity in IRS-1
immunoprecipitates. Antibody to the carboxyl terminus of rat IRS-1
precipitated 70-80% of immunoreactive IRS-1 from total cell lysates
of L6 myoblasts (not shown). In control cells, IGF-I stimulation
induced a rapid and marked (42-fold) increase in the amount of PI
3-kinase activity assayed in vitro with
[32P]ATP and phosphatidylinositol in reconstituted IRS-1
immunoprecipitates (Fig. 5A). In cells
treated with 1 µM dexamethasone for 72 h, there was
no difference in the basal level of PI 3-kinase activity associated
with IRS-1, but IGF-1 stimulation induced only a 23-fold increase in
IRS-1-associated PI 3-kinase activity (Fig. 5A). Thus, in
comparison with control cells, dexamethasone reduced by 45% the level
of IGF-I-stimulated PI 3-kinase activity associated with IRS-1 (Fig.
5B).
Tyrosine Phosphorylation and Cellular Content of IRS-1
The
reduced levels of IRS-1-associated PI 3-kinase activity despite the
marked increase in p85 protein content in dexamethasone-treated L6
myoblasts could be explained by several mechanisms, including decreased
tyrosine phosphorylation and/or cellular content of IRS-1 protein,
impaired association between p85 proteins and tyrosine-phosphorylated IRS-1, altered "coupling" between p85 and the catalytic subunit p110, or a combination of these factors. To investigate the tyrosine phosphorylation and cellular content of IRS-1, Nonidet P-40 extracts from control and dexamethasone-treated L6 myoblasts were resolved on
7% polyacrylamide gels, transferred to nitrocellulose, and subjected
to immunoblotting with anti-PY or anti-IRS-1 antibodies. The effects of
dexamethasone on IGF-I stimulation of IRS-1 tyrosine phosphorylation
are illustrated in the autoradiograph in Fig. 6A and quantitated in the bar
graph in Fig. 6B. Maximum IRS-1 tyrosine
phosphorylation was modestly but significantly decreased in
dexamethasone-treated compared with control cells (25% decrease, p < 0.05). In addition, the amount of IRS-1 protein
was markedly reduced to 35% of control levels (p < 0.05) by dexamethasone (Fig. 6C), as reported previously
(36). The decrease in IRS-1 protein levels occurred more rapidly than
the increase in p85
protein levels, such that IRS-1 protein was
reduced to 40% of control after only 12 h of exposure to
dexamethasone (not shown). The greater decrease in protein content than
tyrosine phosphorylation of IRS-1 resulted in an approximately 2-fold
increase in the ratio of tyrosine-phosphorylated IRS-1 to total IRS-1
in L6 myoblasts treated with dexamethasone (Fig. 6C).
Association of PI 3-Kinase Subunits with IRS-1
The reduced
level of tyrosine-phosphorylated IRS-1 in L6 myoblasts treated with
dexamethasone could result in decreased association of this protein
with p85, leading to impaired activation of PI 3-kinase. To investigate
this possibility, the association of IRS-1 with PI 3-kinase regulatory
and catalytic subunits was studied both with metabolic labeling
experiments in which anti-IRS-1 immunoprecipitates were isolated from
[35S]methionine-labeled cells (Fig. 7),
and with immunoblotting analysis of IRS-1 immunoprecipitates using
antibodies specific to p85, p85
, or p110 (Figs. 8
and 9).
For metabolic labeling studies, L6 myoblasts were labeled with
[35S]methionine in the absence or presence of 1 µM dexamethasone for 20 h. Under these experimental
conditions, steady-state levels of protein labeling were achieved (not
shown). IRS-1 immunoprecipitates were then obtained as described under
"Experimental Procedures" and subjected to SDS-PAGE followed by
fluorography. As compared with the basal state, IGF-I stimulation
induced the association of three protein species of 85, 87, and 110 kDa, respectively, with IRS-1 in control cells (Fig. 7,
first and second lanes). These protein species
were found to be the only detectable 35S-labeled proteins
in the molecular mass range from 40 to 230 kDa undergoing increased
association with IRS-1 upon IGF-I stimulation. Based on their
electrophoretic mobilities, the 85-, 87-, and 110-kDa proteins are
likely to represent the p85, p85
, and p110 subunits of PI
3-kinase. In consideration of the similar number of methionine and
cysteine residues contained in p85 isoforms (9 methionines and 6 cysteines in p85
, 8 methionines and 6 cysteines in p85
) (19), the
stoichiometry of p85
and p85
in the IRS-1 complex upon IGF-I
stimulation appears to be 1:1.7 (Fig. 7, second lane), indicating that more p85
than p85
associates with IRS-1 after stimulation with IGF-I in L6 myoblasts. The 110-kDa band induced by
IGF-I appears as a single protein species and shows higher labeling
compared to the p85 proteins (Fig. 7, second lane), likely reflecting the greater number of methionine and cysteine residues in
p110 compared to p85 (a total of 72 in bovine p110) (20) and the
potential association of multiple p110 molecules with IRS-1 through
distinct p85 proteins (19). IGF-I stimulation also induced the
association of 35S-labeled p85
, p85
, and p110 with
IRS-1 in dexamethasone-treated L6 myoblasts (Fig. 7, third
and fourth lanes). However, the amounts of IRS-1 associated
35S-labeled p85
and 35S-labeled p85
in
IRS-1 immunoprecipitates were approximately 200 and 70%, respectively,
in L6 myoblasts treated with dexamethasone compared to control cells,
such that the p85
/p85
ratio was increased. In spite of the
increase in IRS-1 associated p85
, the amount of
35S-labeled p110 was found to be decreased by 30% in cells
treated with dexamethasone (Fig. 7, fourth lane). Thus, the
increase in IRS-1 associated p85
was not associated with a
coordinate increase in IRS-1 associated p110.
The levels of IRS-1 associated p85 and p85
in control and
dexamethasone-treated L6 myoblasts also were determined by
immunoblotting with anti-p85 antibody (Fig. 8A, left
panel). To specifically detect p85
, similar immunoprecipitates
were also analyzed with a monoclonal antibody to p85
(Fig. 8A,
right panel). Quantitation of the two p85 isoforms in IRS-1
immunoprecipitates is presented in Table I. Small
amounts of both p85
and p85
were associated with IRS-1 in the
basal state, and the amount of immunoreactive p85
was approximately
2-fold greater than the amount of p85
(
/
ratio = 0.5).
IGF-I stimulation increased the amount of p85
and p85
associated
with IRS-1 by 9.4- and 5.5-fold, respectively, resulting in a slight
increase in the
/
ratio to 0.85. Although this experiment does
not provide direct quantitation of the absolute amounts of p85
and
p85
, it should be noted that the
/
ratio detected by direct
immunoblotting with anti-p85 antibody was considerably greater than the
/
ratio in IRS-1 immunoprecipitates (compare Figs. 1A
and 8A). This indicates that a greater proportion of cellular p85
was engaged in complexes with IRS-1 than the comparable fraction of cellular p85
. As shown in Fig. 8A and Table
I, in dexamethasone-treated L6 myoblasts, a considerably greater amount of p85
and a moderately lower amount of p85
were associated with
IRS-1 in the absence of IGF-I stimulation (430 and 78% of control
levels, respectively, p < 0.05), such that the
/
ratio was increased to 2.69. IGF-I stimulation increased the amounts of
both p85
and p85
in IRS-1 immunoprecipitates of
dexamethasone-treated cells (4.4- and 4.1-fold, respectively). However,
in comparison with IGF-I stimulated cells not treated with
dexamethasone, IRS-1-associated p85
was increased and p85
was
decreased (205 and 60% of control levels, respectively,
p < 0.05). Thus, the
/
ratio in IGF-I stimulated
cells was further increased to 2.88 (Table I). These results indicate
that IRS-1 is complexed to greater amounts of p85
and reduced
amounts of p85
both in the basal state and after stimulation with
IGF-I in dexamethasone-treated compared with control L6 myoblasts. It
should be noted that complex formation between IRS-1 and p85
was
increased in dexamethasone-treated L6 myoblasts, even though the levels
of total and tyrosine-phosphorylated IRS-1 in the immunoprecipitates
analyzed in these experiments were decreased by 68 ± 15% and
27 ± 12%, respectively (Fig. 8B, p < 0.5), in agreement with the results presented in Fig. 6. These findings
were confirmed by immunoprecipitation with antibodies to p85
and
subsequent immunoblotting with anti-IRS-1 or anti-PY antibodies, as
illustrated in Fig. 8C. Taken together, these results indicate that considerably greater amounts of p85
and moderately lower amounts of p85
, respectively, were associated with IRS-1 in L6
myoblasts treated with dexamethasone, in agreement with the results
from metabolic labeling studies (Fig. 7). Thus, the increase in total
p85
observed in L6 myoblasts treated with dexamethasone correlates
with an increase in IRS-1-associated p85
. This indicates that the
inhibition of PI 3-kinase activation cannot be attributed to reduced
complex formation between IRS-1 and p85 proteins.
|
Decreased IRS-1-associated PI 3-kinase activity (Fig. 5) in spite of
the marked increase in IRS-1-associated p85 in L6 myoblasts treated
with dexamethasone raises the possibility that a significant amount of
the p85
in IRS-1 immunoprecipitates may not exist as a complex with
the p110 catalytic subunit of PI 3-kinase. To directly measure the
amount of the p110 subunit associated with IRS-1, proteins from control
or dexamethasone-treated L6 myoblasts were subjected to
immunoprecipitation with IRS-1 antibody followed by immunoblotting with
monoclonal antibody specific for p110 (26) (Fig. 9A). In the
absence of IGF-I stimulation little p110 was detectable in IRS-1
immunoprecipitates from control or dexamethasone-treated cells.
Stimulation of control cells with 100 nM IGF-I for 10 min increased the amount of p110 associated with IRS-1 by 3.9-fold (Fig. 9,
A and B). In dexamethasone-treated L6 myoblasts,
although the amount of p110 in IRS-1 immunoprecipitates also increased upon IGF-I stimulation, the amount of p110 associated with IRS-1 was
decreased by 41% in comparison with control cells (Fig. 9, A and B, p < 0.05). These
results are similar to the results presented in Fig. 7, showing reduced
levels of 35S-labeled p110 associated with IRS-1 upon IGF-I
stimulation in dexamethasone-treated L6 myoblasts. If related to the
level of tyrosine-phosphorylated IRS-1 in IRS-1 immunoprecipitates
(Fig. 8B), IRS-1 associated p110 was still decreased by 22%
in dexamethasone-treated cells compared to control. This correlates
with the finding of a 25% decrease in the level of IRS-1 associated PI
3-kinase activity corrected for the level of tyrosine-phosphorylated
IRS-1 in response to dexamethasone treatment. These results support the
concept that PI 3-kinase inhibition by dexamethasone cannot be entirely explained by decreased tyrosine phosphorylation of IRS-1. The reduced
complex formation between IRS-1 and p110 in dexamethasone-treated cells
was confirmed by carrying out the reverse experiment with p110
immunoprecipitation followed by anti-IRS-1 immunoblotting (Fig.
9C). These observations indicate that IRS-1 is complexed with p85
molecules that are not coupled stoichiometrically with p110
in L6 myoblasts treated with dexamethasone. The decreased amounts of
the p110 catalytic subunit in IRS-1 immunoprecipitates from L6
myoblasts treated with dexamethasone correlate with and can account for
the observed decrease in the levels of PI 3-kinase activity.
This study demonstrates that both the cellular amounts of the
various subunits constituting the PI 3-kinase enzyme complex and PI
3-kinase subunit association with tyrosine-phosphorylated IRS-1 are
differentially regulated by the glucocorticoid dexamethasone in
undifferentiated L6 skeletal muscle cells. Dexamethasone markedly increased p85 in L6 myoblasts, but did not alter the levels of p85
and induced only a modest increase in cellular p110 content. Under these conditions, a greater amount of p85
and reduced amounts of both p85
and p110 were recruited to the IGF-I receptor substrate IRS-1 upon hormone stimulation. In addition, the activity of PI 3-kinase measured in IRS-1 immune complexes was significantly decreased
by dexamethasone, likely reflecting the reduced amounts of
IRS-1-associated p110 catalytic subunit.
Glucocorticoids have been reported to increase the amount of p85
protein in rat skeletal muscle (42) and in F442A adipocytes (43). In
these previous studies, the p85 isoform pattern was not determined and,
therefore, it is not known whether the effects of dexamethasone were
isoform-specific. In L6 myoblasts, the increase in p85 protein
content was associated with an increase in p85
mRNA and no
change in p85
mRNA, suggesting that dexamethasone may act by
specifically increasing expression of the p85
gene. The p85
and
p85
isoforms possess 62% overall identity at the amino acid level
and 58% nucleotide identity and, thus, are thought to be encoded by
two distinct but related genes (19). Information on p85 gene structure
is very limited at present and, in future studies, it will be important
to identify the gene regulatory elements that dictate the tissue
distribution of the two p85 isoforms as well as the glucocorticoid
responsiveness limited to p85
. Our data would indicate that
glucocorticoid response elements may be identified exclusively in the
p85
gene (and not in the p85
gene).
The selective regulation of the isoform of p85 by dexamethasone
supports the concept that p85
and p85
may have distinct biological roles. Although the related p85
and p85
regulatory subunits both have been shown to form stable complexes with the catalytic p110 component of PI 3-kinase (21, 22), functional differences of p85
as compared to p85
previously have been
reported. Studies conducted in T-lymphocytes have demonstrated that the two p85 isoforms have a different phosphorylation pattern upon T-cell
activation (44). Following activation of the CD3 antigen complex in
T-cells, rapid serine phosphorylation of p85
was observed, whereas
phosphorylation of p85
was unchanged. In addition, the catalytic
subunit p110 was shown to undergo rapid threonine phosphorylation when
associated with p85
but not with p85
. It has recently been reported that much larger stimulation of PI 3-kinase is found in p85
compared to p85
immunoprecipitates upon insulin stimulation of CHO-T
cells, even though both p85
and p85
associate with IRS-1(45).
This has led to the suggestion that insulin causes recruitment of both
p85
and p85
regulatory subunits to IRS-1 signaling complexes, but
the activity of IRS-1-associated PI 3-kinase is stimulated only in the
p85
·p110 complex, with little or no stimulation of the
p85
·p110·PI 3-kinase complex. The results in the current study
demonstrating specific augmentation of p85
protein content by
dexamethasone indicate for the first time that the control of p85
and p85
expression represents an additional level of differential
regulation of the two p85 isoforms.
IGF-I stimulation of L6 myoblasts induced a severalfold increase in the
association of both p85 and p85
isoforms with IRS-1 immune
complexes. Both p85 isoforms have reportedly been shown to associate
with IRS-1 signaling complexes upon insulin stimulation in COS-1 cells
transiently transfected with the insulin receptor and in CHO-T cells
stably overexpressing the insulin receptor (45). In the latter study,
the
/
ratio in IRS-1 immune complexes generally reflected the
/
ratio in the total cell lysate, indicating no preferential
recruitment of a given p85 isoform to IRS-1 (45). By contrast, our
results indicate that p85
may preferentially associate with IRS-1 in
signaling complexes, since the association of p85
with IRS-1 immune
complexes was greater quantitatively than that of p85
in response to
IGF-I stimulation (Figs. 7 and 8). It is possible that association of
p85
and p85
with IRS-1 signaling complexes may be modulated in a
cell context specific manner and differ in cell lines expressing high
levels of insulin receptors in which hyperphosphorylation of IRS-1 may
occur. Interestingly, the amount of p85
in anti-phosphotyrosine
immunoprecipitates from PDGF-stimulated L6 myoblasts was significantly
higher than the amount of p85
and thus the
/
ratio was much
greater than in IRS-1 immune complexes.2
This suggests that association of p85
and p85
with
tyrosine-phosphorylated proteins in L6 skeletal muscle cells may differ
depending on the specific protein target (i.e. IRS-1
versus the PDGF receptor).
Ample experimental evidence has established the concept of IRS-1 acting
as a "docking protein" capable of simultaneously binding multiple
protein components through specific amino acid motifs containing
phosphorylated tyrosine residues. GST fusions of the NH2-terminal SH2 domain of p85 were found to bind
strongly to Tyr608-Met-Pro-Met and
Tyr939-Met-Asn-Met and, to a lesser extent, to
Tyr461-Ile-Cys-Met and Tyr987-Met-Tyr-Met in
IRS-1 (35). At present, it is not known whether the same sites serve
also for the two SH2 domains present in p85
. Although both p85
and p85
were shown to bind simultaneously to a single IRS-1 molecule
in COS-1 cells transiently transfected with the insulin receptor, this
was not the case in CHO-T cells stably expressing the insulin receptor
(45). In the PDGF receptor, Tyr751 is a common binding site
for both Nck and p85
, indicating that SH2 domains of different
signaling molecules can compete for binding to the same phosphorylated
tyrosine motif (46). In L6 myoblasts treated with dexamethasone, the
increase in IRS-1-bound p85
was associated with a coordinate
decrease in IRS-1-associated p85
, which could potentially be
explained by p85
and p85
competing for the same binding site(s)
on tyrosine-phosphorylated IRS-1.
Dexamethasone induced a cellular excess of p85 and a greater number
of IRS-1·p85
complexes in L6 myoblasts. That the pool of
IRS-1-associated p85
was largely composed of regulatory subunit lacking the p110 catalytic subunit is indicated by the detection of
reduced amounts of IRS-1-associated p110 (demonstrated by both p110
immunoblotting and metabolic labeling experiments) and a corresponding
decrease in PI 3-kinase activity. Although there was an increase in
p110 protein in total cell lysates following treatment of L6 myoblasts
with dexamethasone, this was much less pronounced than the increase in
p85
protein (38 versus 300%, respectively). In
interpreting these results, it is important to note that two p110
proteins (p110
and p110
) have been identified (20, 24). We have
used a monoclonal antibody raised against bovine p110 that recognizes
only the p110
isoform. Therefore, it is not known whether or not
dexamethasone affects cellular levels of p110
and what fraction of
IRS-1-associated p85
in dexamethasone-treated cells was coupled with
p110
. However, the decrease in IRS-1-associated p110
and
IRS-1-associated PI 3-kinase activity in response to dexamethasone were
coordinate. In addition, a single 35S-labeled protein of
110 kDa, possibly representing both p110
and p110
, was detectable
in IRS-1 immunoprecipitates upon IGF-I stimulation. This band was less
intense in L6 myoblasts treated with dexamethasone, confirming that the
total amount of IRS-1-associated catalytic p110 subunit was
decreased.
A p85 subunit lacking a p110 subunit may have the capacity to bind to
the specific tyrosine-phosphorylated motif in target proteins without
localizing catalytic activity in the protein-protein signaling complex.
Generation of non-coupled (i.e. monomeric or free) p85
can be achieved experimentally by transfection and overexpression of
the p85
cDNA in mammalian cells. When this experiment was
performed in 293 cells overexpressing the PDGF receptor, overexpression
of p85
was shown to completely abrogate activation of PDGF
receptor-associated PI 3-kinase activity (28). In addition,
microinjection of the p85
NH2-terminal SH2 domain into
rat 1 fibroblasts overexpressing the insulin receptor was shown to
inhibit insulin- and IGF-I-induced DNA synthesis by competing with
endogenous PI 3-kinase for binding to IRS-1 (29). These studies support
the concept that cellular overexpression of the p85
regulatory
subunit of the PI 3-kinase complex can lead to inhibition of PI
3-kinase activity and impairment of cell signaling through activation
of this enzyme. Accordingly, the excess of free p85
induced by
dexamethasone in L6 myoblasts may compete with both p85
·p110 and
p85
·p110 complexes for binding to IRS-1. If there are functional
specificities for IRS-1-bound p85
and p85
, respectively, this may
have distinct effects by disrupting specific signaling responses not
only through IRS-1·p85
·p110 complexes, but also through
IRS-1·p85
·p110 complexes.
There is evidence that free p85 binds to tyrosine-phosphorylated
proteins with greater avidity than the intact p85·p110 complex (19,
29). Thus, preferential complex formation between IRS-1 and p85
may
have occurred in L6 myoblasts treated with dexamethasone. It recently
has been suggested that in situ concentrations of PI
3,4,5-P3 synthesized locally by the catalytic subunit of PI 3-kinase may bind to the SH2 domain of p85 and dissociate PI 3-kinase from the tyrosine phosphoprotein (47). The lack of PI
3,4,5-P3 production by the catalytically inactive p85
monomer induced by dexamethasone could provide a mechanism conferring
greater affinity to this protein for binding to tyrosine-phosphorylated IRS-1 compared to p85
·p110 and p85
·p110 complexes. Such a
mechanism could also lead to displacement of IRS-1-bound p85
·p110
by free p85
.
Activation of PI 3-kinase by receptor and nonreceptor protein tyrosine
kinases has been implicated in a broad spectrum of cellular responses,
including mitogenesis (1, 7, 8), chemotaxis (48, 49), membrane ruffling
(5), activation of p70 S6 kinase (50), insulin-dependent
GLUT-4 translocation (51), glycogen synthesis (13, 14), activation of
integrins in platelets (52), histamine release (53), receptor
down-regulation (6), and inhibition of apoptosis (54). All of these
biological responses require a functionally active PI 3-kinase enzyme
complex that reflects the coordinate expression of both p85 and p110 PI 3-kinase subunits in the cell. The generation of excess p85
regulatory subunit relative to the other components of the PI 3-kinase
complex along with inhibition of IRS-1-associated PI 3-kinase activity in response to dexamethasone demonstrated in this study suggests a
novel mechanism of PI 3-kinase regulation. This mechanism may not be
limited to muscle cells and may occur in other cell types under
physiological circumstances and/or in disease states characterized by
high levels of endogenous cortisol or exogenously administered glucocorticoids.
We acknowledge Karen TenDyke for excellent technical support.