understanding the specific actions of GH on growth and me- tabolism may be found in these other nuclear phosphoproteins; elucidation of their functions may ...
THEJOURNALOF BIOI~W~ICAL CHEMISTRY
Val. 269, No. 11, Issue of March 18,pp. 7874-7878, 1994 Printed in U.S.A.
0 1994 by The American Society for Biochemistry and Molecular Biology, Inc.
Rapid Changes in Nuclear Protein Tyrosine Phosphorylation after Growth Hormone Treatment in Vivo IDENTIFICATION OF PHOSPHORYLATED MITOGEN-ACTIVATED PROTEIN KINASE AND STAT91” (Received for publication, December 8, 1993)
Ann M. GronowskiS and PeterRotweins From the Department of Biochemistry and Molecular Biophysics a n d the Department of Medicine, Washington University School of Medicine, Saint Louis, Missouri 63110
Growth hormone (GH) plays a centralrole in regulating growth and intermediarymetabolism in vertebrates, although the mechanisms by which GH initiates these actions are largely unknown. The GH receptor, a member of the cytokine receptor superfamily, does not demonstrate homology with any known tyrosine kinases. However, addition of GH to cells in vitro has been shown to stimulate tyrosine phosphorylation of various intracellular proteins including mitogen-activated protein kinases ( M A P kinases) and the newly described Janus kinase, JAK2. Subsequent steps in GH-mediated signal transduction have not been delineated. In the present study, we have examined early events in GH action in vivo. Hypophysectomized juvenile male rats were treated with GH for 15,30,or 60 min. Rat liver whole cell and nuclear extracts were prepared and analyzed via SDS-polyacrylamidegel electrophoresis and Western blotting techniques. GH rapidly stimulated the tyrosine phosphorylation of at least 8 nuclear proteins of 205,91, 83,80,65,53,44, and 42 kDa, and caused the dephosphorylation of a single “149-kDa protein. Using specificantibodies, we have identified three of these nuclear phosphoproteins as 42- and 44-kDa MAP kinases, and as STAT91, a 91-kDa component of the interferon-stimulated gene factor-3protein complex. Oneconsequence of the activation of STAT91 in the nucleus is the appearance of GH-stimulatedDNA binding activity, as assessed by gel-mobility shift assay using an oligonucleotidecontaining a c-sis-inducible element from the c-fos promoter. Theseresults show that nuclear proteintyrosine phosphorylation is a prominent early event in GH action in vivo and demonstrate a linkbetween GH-stimulated signal transduction and targetgene expression. Growth hormone (GH)l is essential for normal growth, development, and metabolism (1). Although these effects of GH
* The costs of publication of this article were defrayedin part by the payment of page charges. This article must therefore be hereby marked “aduertisement”in accordance with 18 U.S.C.Section1734solelyto indicate this fact. This work was supported in part by National Institutes of Health (NIH) Grant 5-R01-DK37449(to P. R.). Oligonucleotides were synthesized at the Washington University Protein Chemistry Facility under support of the Washington University Diabetes Research and Training Center (NIH Grant DK20579). t Supported by NIH Training Grant DK07120. 0 To whom correspondence shouldbe addressed: Washington University School of Medicine, Box 8231, 660 S. Euclid Ave., Saint Louis, MO 63110. Tel.: 314-362-2703; Fax: 314-362-7183. The abbreviations used are: GH, growth hormone; IFN, interferon; EGF, epidermal growth factor; PDGF, platelet-derived growth factor; CSF-1, colony-stimulatingfactor-1; hypox, hypophysectomized;PMSF, phenylmethylsulfonyl fluoride; DTT, dithiothreitol; PAGE, polyacrylamide gel electrophoresis;PVDF, polyvinylidenefluoride; MAP kinase, mitogen-activated protein kinase; ERK, extracellular signal-regulated protein kinase; ISGF3, interferon-stimulated gene factor-3; GAS, IFN-
have been known for many years, the mechanisms of GH action remain unclear. Recent studies have demonstrated that activation of t h e GH receptor by ligand binding leads to the rapid tyrosine phosphorylation of multiple intracellular proteins(26). These alterations in protein phosphorylation may result from association of the activated, dimerized GH receptor ( 7 ) with the non-receptor protein kinase JAK2 (81, a member of the Janus kinase family (9). Ultimately, in order to promote growth and to influence development, GH action must lead to the altered expression of specific target genes. Insulin-like growth factor-I is a major mediator of G H s actions on somatic growth (10,11), and transcription of this gene is induced within 30 min of GH treatment in vivo (12). Transcription of early response GH (13,14), genes c-fos and c j u n a r ealso stimulated rapidly by as are genes encoding the serine protease inhibitors Spi 2.1 and
2.3 (15, 16). Our laboratoryis interested in the mechanisms by which GH alters target gene expression. In this report we have focused on early steps in GH-mediated signal transduction by examining changes in hepatic nuclear protein phosphorylation in response to GH treatment. We show that GH rapidly induces the tyrosine phosphorylation of at least eight nuclear proteins and causes the dephosphorylation of one additional nuclear protein. We have identified three of these phosphoproteins as t h e 42and 44-kDa MAP kinases (ERK1 and ERK2) (171, a n d as t h e 91-kDa component of the interferon-regulated ISGF3 complex, STAT91 (18, 19). Sincethe phosphorylation of STAT91 leads to inducible DNA binding activity,these findings suggesta role for hormone-activated protein tyrosine phosphorylation in GH-induced gene regulation. EXPERIMENTALPROCEDURES Materials-The following antibodies were obtained from commercial sources: anti-phosphotyrosine (monoclonal IgG2bk) from Upstate Biotechnology Inc. (Lake Placid, N Y ) ; anti-phosphotyrosine-coupledagarose (monoclonal IgG2b)and monoclonal anti-ISGF3 from Transduction Laboratories (Lexington, K Y ) ; anti-MAE’ kinase antibody (monoclonal 2033, isotype IgG,) from Zymed Laboratories, Inc. (South San Francisco, CAI; anti-mouse horseradish peroxidase conjugate from Sigma. Enhanced chemiluminescent (ECL) detection system was purchased from Amersham Corp. PVDF membranes were obtained from Millipore (Bedford, MA). Recombinant hGH was obtained from Eli Lilly Co. (Indianapolis, IN) and Genentech, Inc. (South San Francisco, CAI. a-[32P]dATP(800 Ci/mmol) waspurchased from DuPont NEN. The following oligonucleotides were prepared at the Washington University Protein Chemistry Facility: high affinity c-sis-inducible element (SIE) m67 (201, top strand, 5’-GTCGACATTTCCCGTAAATCGTCGA-3’ and, bottom strand, 5’-TCGACGA””IACGGG-3‘;IFN-y-activated sequence (GAS)sequencefrom the Ly6gene(211, top strand, 5”AATTTATGCATATTCCTGTAAGTGAC-3’and, bottom strand, 5‘-GTCACTTACAGy-activated sequence; Oct,octamer transcription factor; STAT91, signal transducer and activator of transcription 91; SIE, c-sis-inducible element; TBS, Tris-buffered saline; ECL, enhanced chemiluminescence.
7874
GH Induces Qrosine Phosphorylation
of Nuclear Proteins
7875
GAATATG-3'; Oct 1 (22),topstrand, 5'-TI"MAGAGGATCCATG- 7.4, 1mM EDTA, 1mM EGTA, 1% TritonX-100, 0.5% Nonidet P-40, 0.2 CAAATGGACGTACG-3' and bottom strand, 5'"ITTTCGTACGTCC- m Na3V0,, 0.2 mM PMSF. Samples were centrifuged for 4 min at AGTTGCATGGATCCTCT-3'. Other chemicals were reagent grade and 16,000 x g in a microcentrifuge and were washed twice in the same buffer. The beads then were suspended in SampleBuffer a s described were purchased from commercial suppliers. above, boiled for 5 min, centrifuged for 4 min a t 16,000 x g , and the AnimalStudies-Pituitary-intact or hypox maleSprague-Dawley rats, age 7 weeks, were purchasedfrom Harlan Sprague-Dawley (Indi- supernatant was separated by SDS-PAGE (10-12%). Separated proanapolis, IN). Completenessof hypophysectomy was confirmed by lack teins were transblotted and analyzedby Western blotting as described of weight gain over the next 10-14 days. Animals were maintained a t above. Immunoprecipitation experiments wereperformed twice on difthe Washington University Animal Care Facility on a 12-h lighudark ferent nuclear extracts. Gel-mobility Shift Assay-Labeled double-stranded probes were preschedule with free access tofood and water (hypox animals were given water supplemented with 5% sucrose). At 8-9 weeks of age, hypox pared by annealing overlapping complementary single-stranded oligoanimals received a single intraperitoneal injectionof 1.5 pglg recombi- nucleotides and filling in the ends usinga-[32PldATP,unlabeled dCTP, nant hGHor vehicle, and were sacrificed 15,30, or60 min later. Livers dGTP, and dTTP, and the Klenow fragment of E. coli DNA polymerase were excised,weighed, and placed on ice. Portal veins were flushed withI. Probes were gel-purifiedbefore use. Rat liver nuclear extracts (-5 pg ice-coldDulbecco's Ca+2/Mg+-2-freephosphate-bufferedsalinesuppleof protein) were preincubatedfor 15 min on ice in 6.25 m HEPES, pH mented with 1m Na3V04. Liverswere then dissected freeof diaphrag7.6, 20 m KCI, 3 m EDTA, 4% glycerol, 0.25 m Na3V04, 1.25 m matic remnants and extra-hepatic tissues. After removal of - 100 mgof DTT, and 0.125m PMSF, 2 pgof poly(dI.dC), and unlabeled competitor tissue for whole cell extractions, the remainder was usedfor prepara- oligonucleotide when indicated (26). In antibody-inhibition assays, this tion of nuclei. All animal protocols were approved by the Washington initial preincubation was followed by an additional 15-min incubation on ice in thepresence of antibody. The binding reactionswere initiated University Animal Studies Committee. Whole Cell and NuclearExtractPreparation-Unlessotherwise by addition of labeled oligonucleotide (-4-8 x lo4 cpm) and were conspecified, the following procedures were performed a t 4 "C. PMSF and tinued for 30 min at 25 "C in a final volume of 20 pl. Samples were DTT were added to all buffers immediately prior to use. Whole cell resolved by native PAGE (4.5%) in 45 m Tns, 25 m boric acid, 1.25mM protein extracts were made as described (23).Minced liver was homog- EDTA. enized with a Tissumizer (Tekmar, Cincinnati, OH) in 100 "C 1% SDS, 50 m HEPES, pH7.4, 100 mM sodium pyrophosphate, 100mM NaF, 10 RESULTS mM EDTA, 10 mM Na3V04. The homogenate was boiled for 10 min, GH D e a t m eSntti m u l a tNe su c l eParro t eTi ny r o s i n e cooled on ice for 40 min, and centrifuged for 30 s a t 16,000 x g in a effect of GH on liver nuclearprotein microcentrifuge. The supernatant was aliquoted and stored at-80 "C. Phosphorylation-The Nuclei were isolated from livers of normal, hypox, or GH-treated hypox phosphorylation was examined after a single intraperitoneal rats by modifying previously published methods (24). Finely minced injection of recombinant hGH to hypox male rats. Within 15 tissue (10-30 g) was suspended in-30 ml of Buffer A (10m HEPES, min of GH treatment, changes were observed in the tyrosine pH 7.6, 25 mM KCl, 0.15 mM spermine, 0.5 mM spermidine, 1 m EDTA, phosphorylation of severalnuclearproteins(Fig. 1). Most 2 M sucrose, 10%glycerol, 1m DTT, 0.5 mM PMSF, 1 mv Na,V04, and 10 m NaF), andhomogenized with six passesof a motor-driven Potter prominent was the appearance of a 91-kDa band detected by homogenizer. Buffer A wasthen added toa final volume of -85 ml, and Western blotting with an anti-phosphotyrosine monoclonal anthe suspension was divided into three parts and layered onto a 10-ml tibody. Tyrosine phosphorylation of pp91 was evident by 15 cushion of Buffer A. The liver suspension was then centrifuged for 45 min, was maximal by 15-30 min, and was maintained through min at 100,000 x g in an SWZ8 rotor. The pelleted nuclei were washed 60 min. Other phosphoproteins were induced by 15 min after in 10 ml of 10 mM HEPES, pH 7.6,25 mM KCl, 0.15 m spermine, 0.5 nm spermidine, 1 mM EDTA, 1 m EGTA, 20% glycerol, 1 mM Na,V04, 100 GH injection, including205-, 65-,53-, and42-kDa bands, anda -149-kDa protein became dephosphorylated (Fig. 1,A X ) .The mM NaF, 1 mM DTT, 0.5 mM PMSF, and were centrifuged in Sorvall GLC-1 rotor for 5 min at 740 x g. Washed nuclei were then suspended 65-kDa band was diffuse and may in fact represent several in threevolumes of 10 ~ HEPES, l l ~ pH7.6,lOO mM KCI, 3 m MgCI,, 0.1 proteins. By 30 min after GH treatment, several additional mM EDTA, 10%glycerol, 1 m Na3V04, 100 nm NaF, 1 mM DTT, 0.5 mM phosphoproteins appeared including bands at 83, 80, and 44 PMSF, andpassedseveraltimesthrougha 22-gauge blunt-ended kDa(Fig. 1, A X ) . Constitutivelyphosphorylated bands of needle. Nuclei were transferred to a glass vial, and M 4 KC1 was added -105 and -180 kDa also were detected in all experiments. slowly t o a final concentration of 0.4 M under constant stirring. This mixture wasallowed t o stir for a n additional 30 min. The salt-extracted Table I summarizes the time course and extent of changes in proteins were then separated by centrifugation a t 188,000 x g for 30 hepatic nuclear protein tyrosine phosphorylation after GH inmin in a SW50.1rotor. The supernatant was dialyzedtwice for 60 min jection. in 25 mM HEPES, pH 7.6,80 IILM KCI, 0.1 nm EDTA, 10% glycerol, 1mM Identification of 42- and 44-kDa MAP Kinases as GH-stimuNa,VO,, 1 m DTT, 0.5 mM PMSF. Protein concentrations of nuclear lated Qrosine-phosphorylated Proteins-GH has been shown to and whole cell extracts were quantified using the methodof Bradford stimulate the tyrosine phosphorylation of MAP kinases in cul(25). tured cells (3-5), but this effect has not been examined i n uiuo. Western Blotting-Nuclear or whole cell extracts were diluted in SampleBuffer (0.05 M Tris base, 4.28 SDS, 0.4 M glycine, 100 p~ Furthermore, MAP kinases have been found to be both phosNa3V04, 10 p~ leupeptin, 0.2 trypsin inhibitor units of aprotinin, 2 0 8 phorylated and translocated into thenucleus after stimulation glycerol, 0.08%bromphenol blue, 5% P-mercaptoethanol, 100 mM DTT) with mitogens suchas EGF (27). Forthese reasons, we asked if and boiled for 5 min. Equivalent amounts of protein were separatedby the 42- and 44-kDa nuclear proteins that were phosphorylated SDS-PAGE (7.5-12%), and transferred to PVDF membranes with a Trans-Blot SD semi-dry transfer cell (Bio-Rad) in 48 mM Tris, 39 mM in response to GH were in fact MAP kinases. Western blotting glycine, 20% methanol. Membranes were blocked overnight in Tris- using an antibody that recognizes both the 42- and 44-kDa buffered saline (TBS; 10 m Tris, pH 8.0, 150 mM NaCl) with 0.05% MAP kinases (ERK1 and ERK2) indicated that GH stimulated Tween 20 and 0.25% gelatin (TBSTG), incubated with primary antibod- the appearance of both isoforms in the nucleus (Fig. 2 A ) , with ies in TBSTG for 2 h, washed twice for 15 min with TBSTG and incu- a time course identical to that seen using an anti-phosphotybated with secondary antibodiesfor 1 h in TBSTG. After an additional rosine antibody (see Fig. 1B). The more abundant 42-kDa prowashing step (three times for 20 min in TBS with 0.05% Tween 201, antibody binding was visualized using an ECL detection system.Before tein was detected at the earliest timepoint examined, 15 min, while p44 was not seen consistently until 30 min. In order to reuse, membranes were stripped in 2% SDS, 100 m P-mercaptoethanol, 50 m Tris, pH 6.8, for 30 min at 55 "C, followed by thorough assess whether the appearance of these proteins in the nucleus washing with TBS. Membranes were then reblocked overnight with was a result of GH-stimulated synthesisof MAP kinases or was TBSTG and probed as described above. Experiments were performed secondary to hormone-induced nuclear translocation, Western two to five times using whole cell and nuclear extracts prepared from blotting was performed in parallel on whole cell liver extracts. different groups of hypox and GH-treated rats. As seen in Fig. 2B, GH treatment did not alter levels of immuImmunoprecipitation-Equivalent quantities (100 pg) of rat liver increase in p42 nuclear proteins were incubated with 125 pg of anti-phosphotyrosine- nodetectable MAP kinases, suggesting that the coupled agarose beads for 1 h a t 4 "C in 150 m NaCl, 10 m Tris, pH and p44 seen in nuclear extracts was caused by GH-activated
GH Induces Qrosine Phosphorylation of Nuclear Proteins
7876
0.
A. x
GH-treated
x
0
(rnin)
0
a
' 15
-
FIG.1. GH-induced changes in hepatic nuclear protein tyrosine phosphorylation. Liver nuclear proteins were extracted from normal rats from and hypox rats treated with GHfor 15,30, or 60min.Proteinswereseparatedusing SDS-PAGE, transblotted onto PVDF membranes,and probed with a monoclonal anti-phosphotyrosineantibody. Antibody binding was detected using ECL Westerndetectionsystem. A, proteins separated using a 12% gel, Western blot exposed to x-ray film for5 s; B , the same membrane exposed for 1min; C, proteins separated using a 7.5% gel, Western blot exposed to x-ray filmfor 4 min. Molecular weight markers are indicated.
30 6 0
a I
' ppl49-
- v
PP205 PPI49 ++++ PP91 ++++ -++++ ~ ~ 8 3 PP80 + PP65 ++ + PP53 PP44 PP42
-46
pp53PP44-
- 30
pp42--
A
s pp 205
60min
+
++
+++
++++ ++++
+++ +++ +++ ++ + ++
++
++ +
' -92 kDa
z
-46
5*-
30
-
'
15
3
E0
30 60 ' z
200 kDa
A. Nuclear
GH
-
-
GH-treated (rnin)
CL
30min
-
-
60
!
- 69
x 0
15 min
++++
30
C.
TABLE I
Hypox
(rnln)
15
- 69
Summary of GH-stimulatedchangesintyrosinephosphorylation of rat liver nuclear proteins Symbols depict the relative intensityof bands, with - indicating not detected, + indicating visible, and ++ to ++++ indicating increasing intensity. -kDa
'
4 I
-92 kDa
GH-treated
-
+++
++ +++
+
x
X I" '
GH-treated (min) 15 30 6 0
'
Normal
8.WholeCell
+ ++++ +++
+ + +
+
translocation of these proteins. (Whole cell extracts of serumtreated, undifferentiated 3T3 F442A cells were included as a positive control for this experiment(Fig. 2B ).). Further confirmation that the GH-stimulated, phosphorylated 42- and 44kDa proteins were MAP kinases was provided by reprobing anti-phosphotyrosine Western blots of nuclear protein extracts with anti-MAP kinase antibodies.2 In sum, these resultsshow that 42- and 44-kDa MAP kinases are rapidlyphosphorylated on tyrosine residues and are translocated to the nucleus in response to in vivo GH treatment. Activation of STAT91 by GH Deatment--Recent studies have shown that the91- and 84-kDa components of the IFN-a-regulated ISGF3complex become tyrosine-phosphorylated and are translocated to the nucleusresponse in to a variety of agents in addition to IFN-a (281, including IFN-y, EGF, PDGF, CSF-1, interleukin-6, and other cytokines (20, 26, 29, 30). In order to determine if the 91-kDa nuclear protein that becomes tyrosinephosphorylated in response to GH wasIFN-inducible STAT91 (191, Western blotting wasperformed using a monoclonal anti~
A. M. Gronowski and P. Rotwein, unpublished observations.
p44p42-
-.
. - II
-46
FIG.2. Appearance of 42- and 44-kDaMAP kinases in nuclear extracts afterGH treatment. Protein extracts, prepared in parallel, from nuclei (A) or from whole cells ( B )were separated on 12% SDSPAGE, transblotted to PVDF, and probed with a monoclonal anti-MAP kinase antibody. Whole cell extracts of serum-treated, undifferentiated 3T3 F442A cells were includeda s a control ( B ) .Antibody binding was detected using a n ECL Westem detection system. Western blots were exposed to x-ray film for1 min ( A ) or 5 min ( B ) .
body that recognizes both the 84- and 91-kDa proteins (Fig. 3A, left panel in parallel with an anti-phosphotyrosine Western blot (Fig. 3A, rightpanel). As shown in Fig. 3A (leftpanel1, after GH treatment levels of both p91 and p84 ISGF3 proteins increased in nuclear extracts within15 min of GH injection, and reached a maximum by 30 min. As also indicated in thisfigure, both p91 and p84 are present at low levels in liver nuclear extracts from normally growingrats and from untreated hypox animals. A comparison of the part of the blot incubated with anti-ISGF3 (left panel) with the portion probed with an antiphosphotyrosineantibody (rightpanel)demonstratesthat STAT91 migrates at the same relative mobility as the prominent 91-kDa phosphoprotein. By contrast, in the experiment depicted in the right panel of Fig. 3A, the p84 component of the ISGF3 complex is not seen, indicating either that the84-kDa ISGF3 protein is phosphorylated to a lesser degree than
GH Induces Drosine Phosphorylation of Nuclear Proteins
7877
A. Nucleor
1-
aISGF3
aPY
8.Whole Cell x
GH-treated
Z
p91-
- 9 2 kDa
p84'
- 69
FIG.4. GH-induced DNA-protein interactions. Nuclear protein extracts were prepared from hypox or GH treated rats.A time courseof GH-induced protein bindingto a high affinitySIE oligonucleotide(20) is x GH-treated shown in thegel-mobility shift experiment inpanel A. Specificity studg (min) ies are shown inpanel B. Competitor experiments were performed as 2 . r 30 60 ' I 15 described under "Experimental Procedures," using a n unlabeled GAS oligonucleotide (21),a n Oct 1 (22) competitor, and monoclonal antibodies directed against ISGF3 and MAP kinases. Specific DNA-protein antibody complexes are markedA, B , and C . Note that the anti-ISGF3 inhibits formation of complex C . Free probe is visible at the bottom of I IP-aPY panel B , but was electrophoresedoff the gel pictured inpanel A. Dried Western-a ISGF3 gels were exposed to x-ray film at -80 "C for 45 min (A) or 3 h ( B). FIG.3. Appearance of ISGFS proteins in nuclear extracts after GH treatment. Protein extracts, prepared in parallel, from nuclei(A or whole cells ( B )were separatedon 10% SDS-PAGE, transblotted, and liver nuclear extracts were incubated with a 32P-labeled high probed with monoclonal antibodies against ISGF3(A, left panel and B ) or anti-phosphotyrosine (A, right panel). Left and right panels ofA are affinity SIE oligonucleotide derived from the c-fos gene promoter (20) for 30 min at 25 "C, followed by PAGE and autorafrom the samegel and were separated after transblotting ontoPVDF a membrane. In C , proteins from nuclear extracts were immunoprecipi- diography. As shown in Fig. 4 A , GH rapidly (within 15 min) tated using anti-phosphotyrosine antibodies. Immunoprecipitated pro- induced the appearanceof three DNA-protein complexes, A, B, teins were separated using 10% SDS-PAGE, transblotted, and probed and C, that were similar in mobility to those seen in other with monoclonal anti-ISGF3 antibody. Antibody binding was detected sytems inresponse to EGF andIFN-y treatment (20,261. Alow using a n ECL Western detection system. Blots were exposed to x-ray level of complex formation was found using liver nuclear exfilm for 10 s (A) or 1 min ( B and C ) . C. Nucleor-IP
~
tracts from hypox and normal2rats. Fig. 4B demonstrates that formation of these DNA-protein complexes was inhibited by STAT91 or that it is recognized poorly by the anti-phosphotyrosine antibody (also see Fig. Ut). Based on the low level of excess unlabeled SIE competitor and to a lesser degree by an detectable tyrosinephosphorylation of the 84-kDa ISGF3 com- unlabeled GAS oligonucleotide derived from the Ly6 gene (211, ponent in response to GH, it is likely that GH-inducible pp83 but not by a n Oct 1 competitor (22). The Ly6-GAS oligonucleand pp80 (see Fig. 1and Table I) arecomposed of other nuclear otide also causeda gel-mobility shift; however, complex formaproteins. In orderto assess whether the increase in abundancetion did not change in response to GH.2 As shown also in Fig. of p91 and p84 was caused by GH-stimulated synthesis of 4L3, formation of the fastest migrating gel-shifted band was ISGF3 or was secondary to GH-induced nuclear translocation, inhibited by addition of anti-ISGF3 antibody (20) butwas not Western blotting was performed on whole cell liver extracts blocked by another, irrelevant, antibody. The results of these that were prepared in parallel with nuclear protein extracts. As experiments demonstrate that GH treatment rapidly induces illustrated inFig. 3B, there wasno change inlevels of p84 or of DNA binding activity toward an SIEoligonucleotide, and that p91 until 60 min after GH treatment, indicating that the in- a t least one inducible protein component is STAT91. crease detected by 15 and 30 min in nuclear extracts was secDISCUSSION ondary to translocationof these proteins into thenucleus. The studies presented hereshow that in vivo GH treatment In order toconfirm that the ISGF3 proteinsdetected in the nucleus weretyrosine-phosphorylated, an immunoprecipita- leads to rapid changes in nuclear protein tyrosine phosphoGH stimulates thephostion experiment was performed using monoclonal anti-phos- rylation. Our results demonstrate that phorylation of at least eight proteins in liver nuclei ranging in photyrosine antibodies coupled to agarose.Immunoprecipitatednuclearproteins were separated on SDS-PAGE, and size from 42 to 205 kDa, with predominant tyrosine-phosphoWestern blots were probed with anti-ISGF3 antibody. The re- rylated proteins of -91, 83, and 80 kDa. GH treatment also sults, shown in Fig. 3C, indicate thatboth the 91- and 84-kDa induces the dephosphorylation of a 149-kDa protein. We have ISGF3proteins became tyrosine-phosphorylated after GH further identified three of these proteins as the p42 and p44 MAP kinases andSTAT91, a 91-kDa cytoplasmic proteinthat is treatment.Further confirmation thattheGH-stimulated, phosphorylated 91-kDa protein was STAT91 was provided.by tyrosine-phosphorylated and translocated to the nucleus in rereprobing anti-phosphotyrosine Western blots of nuclear pro- sponse to multiple signaling molecules (20, 26, 28-30). Our studies also reveal that an additional nuclear action of GH is tein extracts with anti-ISGF3 antibodies2. These results all indicate that the 91-kDa GH-inducible nuclear phosphoprotein the induction of DNA binding activity as assessed by gel-mobility shift assay, using as probe a modified SIE oligonucleotide is STAT91. To determine if GH stimulates the DNA binding ability of from the c-fos gene (20). Inhibition of one of the threeinducible STAT91, gel-mobility shift assayswere performed (Fig. 4). Rat DNA-protein complexes by anti-ISGF3 antibody indicates that
7878
GH Induces Tyrosine Phosphorylation of Nuclear Proteins
STAT91 comprises at least part of this GH-regulated DNA binding activity. Studies using cultured cells have shown that GH rapidly stimulates thephosphorylation and activation of MAP kinases (3-5). Our results demonstrateadditionally that GH treatment enhances the tyrosine phosphorylation of at least two MAP kinases in vivo and induces their translocation to the nucleus. Both the 42-kDa (ERK2) and 44-kDa (ERK1)MAP kinase isoforms are regulated by GH; however, the more abundant 42kDa isoform is phosphorylated and translocated more rapidly and more extensively after GH injection. Nuclear translocation of MAP kinases also has been demonstrated inresponse to EGF and tophorbol esters (27). At present, therole of MAP kinases in GH action is not known. However, as these enzymes phosphorylate ribosomal 56 kinase (pp9OrSk),c-myc, and c-jun (171, they may be involved in GH-induced activation of early response genes. STAT91 was discovered originally as a 91-kDa component of ISGF3, a transcription complex activated by IFN-a and containing 113-, 91-, 84-, and 48-kDa proteins (28). The three largest proteins in this complex are tyrosine-phosphorylated and translocated to the nucleus in response to IFN-a (28). In the nucleus these proteins interact with the 48-kDa component and then bind with high affinity to the interferon-stimulated response element (28). In contrast, IFN-yonly induces phosphorylation of the 91-kDa (and 84-kDa) protein, which is then translocated to the nucleus and binds to the IFN-y-activated sequence (GAS) (29). Recent reports have demonstrated that STAT91 also is activated in response to other growth factors and cytokines including interleukin-6, PDGF, CSF-1, and EGF (20, 26, 30), suggesting that this protein is involved in a common, Ras-independent (30) signal transduction pathway. Our results show that STAT91 is rapidly tyrosine-phosphorylated, translocated into the nucleus, and activated in response to a single GH injection. It is interesting tonote that thereceptors for GH, IFN-y, and IFN-a all interact with members of the Janus tyrosine kinase family (8,311, suggestinga potential link between activation of Janus kinases and stimulation of the STAT91 signaling pathway. By contrast, actions of the EGF, PDGF, and CSF-1 receptors have intrinsic tyrosine kinase activity (18).As with MAP kinases, the role of STAT91 in GH action is not yet known. STAT91 has been shown t o bind to a sequence that isfound in thepromoter regions of several IFNy-responsive genes (21, 29), and in thec-sis-inducible element GH-activated of the c-fos gene (20).Our data demonstrate that STAT91 is able to interact witha high affinity oligonucleotide containing the SIE consensus sequence, although STAT91 appears to comprise only part of the GH-inducible DNA binding activity. Therefore, STAT91 may provide a pathway, separate from MAP kinases, for induction of early response genesby GH. The identities of the other GH-induced nuclear phosphoproteins detected in this study are unknown. Proteins of similar electrophoretic mobilities have been shown to become tyrosinephosphorylated after GH treatment of 3T3-F442A fibroblasts
(97,90, and 75 kDa) and IM9 lymphocytes (93 kDa), butnone of these proteins have been localized to the nucleus (3, 6). Because neither theMAP kinase nor the STAT91 signal transduction pathway isexclusive to GH, it islikely that othernovel signaling molecules are activated by GH treatment. A key to understanding the specific actions of GH on growth and metabolism may be found in these othernuclear phosphoproteins; elucidation of their functions may help reveal the mechanisms by which the pleiotropic effects of GH are regulated. Acknowledgments-We thank Dr. Michael Thomas for critical review of the manuscript,andweacknowledge his help and that of Paula Kellerman in isolating liver nuclei and with gel-shifi assays.
REFERENCES 1. Kelly, P. A., Djiane, J., Postel-Vinay, M.-C., andEderly, M. (1991)Endocr. Reu. 12,235-251 2. Campbell, G. S., Christian, L. J., and Carter-Su,C. (1993)J.Biol. Chem. 268, 7427-7434 3. Campbell, G. S., Pang, L., Miyasaka, T., Saltiel, A. R., and Carter-Su,C. (1992) J. Biol. Chem. 267, 6074-6080 4. Moller, C., Hansson,A.,Enberg, B., h b i e , P. E., and Norsted,G. (1992)J. Biol. Chem. 267,23403-23408 5. Winston, L. A., and Bertics, P. J. (1992)J. Biol. Chem. 267, 47474751 6. Silva, C.M., Day, R. N., Weber, M. J., and Thorner, M. 0. (1993)Endocrinology 133,2307-2312 7. Cunningham, B. C., Ultsch, M., De Vos,A. M., Mulkerrin, M. G.,Clauser, K. R., and Wells, J. A. (19911Science 253,821-825 8. Argetsinger, L. S., Campbell, G. S.,Yang,X., Witthuhn, B.A., Silvennoinen, O., Ihle, J. N., and Carter-Su, C. (1993)Cell 74, 237-244 9. Harpur, A. G.,Andres, A.-C.,Ziemiecki, A.,Aston, R. R., andWilks, A. F. (1992) Oncogene 7, 1347-1353 10. Froesch, E. R., Schmid, C., Schwander, J., and Zapf, J. (1985)Annu. Rev. Physiol. 47, 443467 11. Daughaday, W. H., and Rotwein, P. (1989)Endocr. Reu. 10,68-91 12. Bichell, D. P., Kikuchi, K., and Rotwein, P. (1992)Mol. Endocrinol. 6, 18991908 13. Gurland, G.,Ashcom, G., Cochran, B. H., and Schwartz,J.(1990)Endocrinology 127, 3187-3195 14. Slootweg, M. C., de Groot, R. P., Herrmann-Erlee, M. P. M., Koornneef, I., Kruijer, W., and Kramer, Y. M. (19911J. Mol. Endocrinol. 6, 179-188 15. Yoon, J. B., Berry, S. A,, Seelig,S., and Towle, H. C . (1990)J. Biol. Chem. 265, 19947-19954 16. Paquereau, L., Vilarem, M. J., Rossi, V., Rouayrenc, J. F., and Le Cam, A. (19921Eur. J. Biochem. 209, 1053-1061 17. Thomas, G. (1992)Cell 68,3-6 18. Montminy, M. (1993)Science 261, 1694-1695 19. Shaui, K., Stark, G. R., Ken; I. M., and Darnell, J. E. (1993)Science 261, 1744-1746 20. Sadowski, H. B., Shaui, K., Darnell, J.E., Jr., and Gilman,M.2. (1993)Science 261, 1739-1744 21. Strehlow, I., Seegert, D., Frick, C., Bange, F.-C., Schindler, C., Bottger, E. C., and Decker, T. (1993)J. Biol. Chem. 268, 16590-16595 22. Rosenfeld, M. G. (1991)Genes & Deu. 5,897-907 23. Folli, F., Saad, M. J. A., Backer, J. M., and Kahn, C. R. (1992)J . Biol. Chem. 267,22171-22177 24. Sladek, F. M., Zhong, W., Lai, E., and Darnell, J. E. (1990)Genes & Deu. 4, 2353-2365 25. Bradford, M. M. (1976)Anal. Biochem. 72,248-254 26. Ruff-Jamison, S., Chen, K., and Cohen, S . (1993)Science 261, 1733-1736 27. Lamy, F., Wilkin, F., Baptist, M., Posada, J., Roger, P. P., and Dumont, J. E. (1993)J . Biol. Chem. 268, 8398-8401 28. Schindler, C., Shuai,K., Prezioso, V. R., and Darnell, J. E., Jr. (1992)Science 257,809-813 29. Shaui, K., Schindler, C., Prezioso, V. R., and Darnell, J. E., Jr. (1992)Science 268.1808-1812 30. Silvennoinen, O., Schindler, C., Schlessinger, J.,and L e v y , D.E. (1993)Science 261, 1736-1739 31. Hunter, T. (1993)Nature 366,114-116 ~~
~
I
~~
~
