Proc. Natl. Acad. Sci. USAVol. 90, pp. 8429-8433, September 1993Medical Sciences

Structure of the murine Jak2 protein-tyrosine kinase and its role ininterleukin 3 signal transductionOLLI SILVENNOINEN*t, BRUCE A. WITTHUHN*, FREDERICK W. QUELLE*, JOHN L. CLEVELAND**,TAOLIN YI*, AND JAMES N. IHLE*:t§*Department of Biochemistry, St. Jude Children's Research Hospital, 332 North Lauderdale Street, Memphis, TN 38105; and *Department of Biochemistry,University of Tennessee, Memphis, TN 38163

Communicated by George R. Stark, May 18, 1993

ABSTRACT Interleukin 3 (IL-3) regulates the prolifera-tion and differentiation of hematopoietic cells. Although the1L-3 receptor chains lack kinase catalytic domins, IL-3 in-duces tyrosine phosphorylation of cellular proteins. To inves-tigate the potential role of the JAK family of protein-tyrosinekinases in EL-3 signal transduction, we have obtained full-length cDNA clones for murine Jakl and Jak2 protein-tyrosinekcinas and prepared antiserum against the predicted proteins.Using antisera against Jak2, we demonstrate that 1L-3 stimu-lation results in the rapid and specific tyrosine phosphorylationof Jak2 and activates its in vitro kinase activity.

Hematopoiesis is regulated through the interaction ofgrowthfactors with their cognate receptors (1, 2). Interleukin 3 (IL-3)supports the proliferation and differentiation of early pro-genitors and cells committed to several myeloid lineages (3).The receptor for IL-3 is composed of an a subunit of 60-70kDa and a ,B subunit of 130-140 kDa (4). Both subunitscontain the extracellular conserved motifs found in the cy-tokine receptor superfamily. The cytoplasmic domains sharelimited similarity with other cytokine receptors and lackdetectable catalytic domains. However, considerable evi-dence suggests that signal transduction involves tyrosinephosphorylation (1, 4). Thus there has been considerableinterest in identifying a protein-tyrosine kinase that associ-ates with the receptor and is activated by IL-3 binding.Polymerase chain reactions (PCRs) have been done with

degenerative oligonucleotides to conserved tyrosine kinasedomains to identify potential kinases in myeloid cells (5). Byusing this approach and Northern blot analysis, IL-3-dependent cells express (6) the genes for a limited number oftyrosine kinases including Lyn, Tec, c-Fes, Jaki, and Jak2.Lyn kinase has been implicated in IL-3 signaling (7). How-ever, we have not detected an effect of IL-3 on Lyn kinaseactivity or on its tyrosine phosphorylation (unpublisheddata). We have also not detected any tyrosine phosphoryla-tion or activation ofTec or c-Fes kinase activity (unpublisheddata). Therefore, we focused on assessing the role of murineJakl and Jak2 genes in IL-3 signal transduction.The JAK (Janus kinase; alternatively referred to as just

another kinase) family of kinases was detected in PCRamplification of tyrosine kinase domains in hematopoieticcells (8). The complete structure ofthe humanJAKI gene hasbeen reported (9) and, recently, a partial sequence of themurine Jak2 gene has been reported (10). Independently, athird member (tyk2) was isolated by homology screening (11).The family is characterized by two kinase domains. TheC-terminal domain has all the hallmarks of protein kinasesand the second domain has the hallmarks of a protein kinasebut differs significantly from both the protein-tyrosine and-serine/threonine kinases. There are no SH2 and SH3 do-

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

mains that characterize many tyrosine kinases. There isextensive similarity in the N-terminal region among the JAKfamily members (10).A role for the tyk2 gene in signal transduction was estab-

lished in studies examining interferon a (IFN-a) (12). Byusing a genetic approach, the tyk2 gene was cloned by itsability to reconstitute the cellular response to IFN-a inmutant cells. It is hypothesized that the kinase activity ofTyk2 is activated after IFN-a binding and is responsible forthe phosphorylation ofthe 113- and 91/84-kDa proteins oftheIFN-stimulated gene factor 3a complex (13, 14). This com-plex associates with the IFN-stimulated gene factor 3y pro-tein and activates gene expression by binding to the inter-feron-stimulated response element.

Jak2 is involved in the response to erythropoietin (EPO)(15). EPO stimulation induces tyrosine phosphorylation ofJak2 and activates its in vitro autophosphorylation activity.Using a series ofEPO receptor mutants, the induction ofJak2tyrosine phosphorylation was found to correlate with theinduction of biological responses. Jak2 was also shown tophysically associate with the membrane-proximal cytoplas-mic region of the EPO receptor that is required for biologicalactivity. Here we report the complete structure ofthe murineJak2 gene. We demonstrate that Jak2 is rapidly tyrosinephosphorylated and its kinase activity is activated in re-sponse to IL-3.

MATERIALS AND METHODSIsolation of Murine Jak2 Clones. PCR, with degenerative

oligonucleotides corresponding to the conserved protein-tyrosine kinase domain, was used to amplify cDNAs frommurine bone-marrow-derived monocytes (16). The Jak2cDNA clone was used to screen myeloid cDNA libraries(17-19). The longest cDNAs were subcloned and the nucle-otide sequence was determined (20).

Cells and Cell Culture. The cell lines used have beendescribed (21) and were maintained in IL-3 (25 units/ml).Mouse bone-marrow-derived monocytes were grown as de-scribed (17).Generation of Antibodies. Synthetic peptides correspond-

ing to the hinge region between domains 1 and 2 (aa 758-776)were coupled to keyhole limpet hemocyanin by gluteralde-hyde and used for immunization of rabbits. A syntheticpeptide to Jakl (aa 785-804) was used for competitionstudies.

Abbreviations: IL-3, interleukin 3; IFN-a, interferon a; EPO, eryth-ropoietin.tPresent address: Department of Pharmacology, New York Univer-sity Medical Center, School of Medicine, 550 First Avenue, NewYork, NY 10016.1To whom reprint requests should be addressed.¶The sequence reported in this paper has been deposited in theGenBank data base (accession no. L16956).

8429

8430 Medical Sciences: Silvennoinen et al.

CCCCGCAACAAGATCGCAACTOT T?CCCTCCC 33

CAGAAGAACAGCCCTTTTTTTCCCTCCCGCGAAGCCCAATGTTCTCAAAAAAGCTCTAG 93

ATGAATCGCCTGCCCTTACAATCACAGAAATOOAGGCAACCTCCACATCTCCTGTACAT 1S30 N A C L T N T t A T S T S P V H 20

CAGAATGCTGATATTCCTGGAAGTGCTAATTCTGTGAACCAGATAGAGCCAGTCCTTCAA 213Q P G D I P G S A N S V X Q I t P V L Q 40

CTGTATCT0TACCATTCTCTTGGGCCCAA GAAGGCAGAGTATCTGAAGTTTCCAAGT0GA 273V Y L r H S L G Q A G Y L K * P S G 60

CGCTATGCTTCAGAAOAAATTTGTGT0GCTGCTTCTAAAGCTTGTGGTATTACGCCTCTG 333t Y V A I C V A A S K A C C I T P V 80

TATCATASATATGTTTCCGTTAATOAGTGAAACCGAAAGCATCTGGTACCCACCCAATCAT 393Y N N A L N S t T R I W Y P N H 100

?TCTTCCACATAGACGAGTCAACCAGOCATCACATACTCTACAGCATAAGGCTCTAC?TC 453V F H I 0 S T R H 0 L Y R I R F Y 120

CCTCATTOGTACTOTAGTO?CAGCACCAGAACCTACAGATACGCAGTGTCCCGGOGGGC S13P N W Y C S 0 S S R T Y R Y C V S a G A 140

> cscCAAGCTCCTCTGCTTGAT0ACTTTTCATGTCTTACCTTVVV0CVCAGTGGC00CAT0AT 573

A P L L D D V N S Y L F A Q w R N 0 160S P

TTTGTTCAcOGATGOATAAAAGTACCTG?GACTCATCAAACTCAGGAAGAOVOVCTTC00 633f V H C W I K V P V T H t Q Y C L 0 180

ATVGCGGT0TTAOACATCATGAGAATACGCAAGOACAAACACCAGACTCCACTOOC=TC 693N A V L D N R I A K t K D Q T P L A V 200

TATAACTCTGTCAOCTACAAGACATTCTTACCAAAGTCcCTTCGAGCOAAcATCCAAGAC 753Y N S V S Y K T f L P K C V R A K I Q D 220

TATCACATTTTAACCCGGAAGCGAATCAGGcACAGATTTCGCAGATTCATTCACCAATTC 613N I L T R K R I R aA R R I Q Q 240

AGTCAAT0VAAAGCCACTCCCAGGAACCTAAAACTTAAGTATCTTATAAACCTOGAAACC 673S Q C A t A R L K L K Y L I L T 260

CVCCAGVCTGCCTTCrACACAGAACAGTTTVAATAAAGAATCT0CAAGA0VCTVCA 933L Q S A Y t Q V K S A C P S 250

GGTGAG0AOATTV0ACAVVAAC0AAACOGTOGAATTCAOV00VCAAGA 993

033e VAT I I I T GS G G I Q W S a 300

GCGAAACATAAGGAAAOOA0ACACT0ACAOAACAG0ACOTACA0TTATATTGT0ATTTC 1053C H S 3 T L T Q 0 V Q L Y C D 320

tCCTGATATTAVTTCATTCAGTAVAATT CAACAaccAT COAA AAAAGTAGA 1113P D I I D V S I Q A N Q a C S a S a 340

9 9Ca C99a9 9 C

ATTGTAACT0TCCATAAACAACATG0TAAAGrTTCGACATAGAACrTAGCTCATTAAAA 1173S V t V H K Q D 0 x V L I L S S L K 360V

t

AAGCC'tTCTCATTCGTCTCATTAATTGACGGGCATTACACACTAACTGCGGAVCGCCAC 1233A L S V S L I 0 G Y Y R L T A 0 A N 380

t t C t a tCATTACCTCTrCCAAAGkGGTCGCTCCCCCAGCTGTGCTCGAGAACATACACACCAACTCC 1293H Y L C X V A P P A V L N I N S N C 400

t c a a

CACGCCCAATATCAATGGATTTTGCCATTAGCAAACTAAACAAGGCGGGTAACCACACT 13S3H P t N D A I S K L K K A c N Q T 420

9 a t t c c

GGACTATATOTGCTACGATGCAGCCCTAAGCACTTCAACAAATACTTTCTGACCTTTOCT 1413L Y V L R C S P K o r N K Y r L T A 440

t a

CTTGAGCGAGAAAATGTCATTGAATATAAACACTGTTTGATTACGAAGAATGAGAATOGA 1473V R N V I Y K H C L t T N N G 460

9 t gtGAATACAACCTCAGCOGGACTAAGAGGAACTTCAGTAACCTTAAGGACCTTTTGAATTGC 1533

Y N L S C T K A N S N L K 0 L L N C 480

C t tTACCACATCGAAACTGTCCGCCACACAGTATCATCTTCCAGTTTACCAAATGCTGCCCC 1S93Y Q t V R S 0 5 I I V Q T C C P S00

9 t 9

CCAAAGCCAAAAGATAAATCAAACCTVTCCGTCTTCACAACAAATGGTATTTCTGATGTT 1653

P K P K D K S N L L V r R T N G I S 0 V S20v

C

CACATCTCACCAACATTACAGAGGCATAATAATCTGAATCAAATAGCTGTTCACAAAATC 1713Q I S P T L N N V N Q N V I 540

L

9 a

Proc. Natl. Acad. Sci. USA 90 (1993)

9 9 C

AAAOOOTAAGAAGACAAGTTOGACGATTATGGCAACTVCACAAAACGGAAGTTCTTVVV 1

X G v R R X V G D Y G Q L H X t t V L L

a tAAAGTCCTA0ATAAAOCACATAGGAACTATTCAGAG CVVCVVGAAGCAGA TCoV 11X V L D X a H R P Y S I S f T 9 a a S n

a a

AT0MACTCA.TTCTCACAGCA&VTTTTTTVVCAATTATGGT0VCV0V0TCVT0GA0A0 14

N S Q L S H X H L V L P Y c v c v C e a

t 9

GAOAACATTCTOOTTCAGAAATVTTOAAAAmTTTATCACTOGATACATACCTGAAAAG 2

9 H I L V Q a r v X f G S L D T T L X X

t 9

ssecasaaasatatsalw^auaX C ~~~~~~~~~~~~~2110 X 9 * I 0 S L V R L 0 V a R Q L a V a

c a cAVCATTTTCTA0AA0AAAAATCCCTTATTCATVC0AATV0VOeVCVAAAAATATccT 2n V. r L X t K S L I H C P V C a K H t L

CfTATCA0AGAAGAAGOACA00AGAOOc0AACCCACCTTTCATCAAACTTAGTGATcT 2L I R L * D a R T ¢ 11 P P * I X L S D P

a

c ATTAcATTACAGTACTACC0AACAVCATTCTTCAGAAVACAATACC 2

I S I t V L P D S IL 9 a I P w V p

9 acaCCT0AAVOCATTOAOAATCCTAAAAATCVCAATCTOOCAACAOACAAGTO0ACVOOSp a C I a H P X P L 0 L a T D K W S 7 a

ACCACTCT0V00AOATCTATGGA0GAO.ATAAOCCCCCOAGTCCCVGATTCTCAA 2T T L W * I C S C e D K P L 3 A L D S Q

9

AGAAAGCTGCAGTTCTATCAAGATAAGCATCACTTCCTCCACCCAAGTGCACAGACTTA ;R K L Q r Y D K H Q L P A P K W T t L

CCAAACCTTATAAATAATTGCATCGACTATCACCAGAT?TCAGCCCCCCTTTCAGAGCTA N L I N N C H DVY P D r R P A r A

GtCATCCGTGATCTTAACAGCCTGtTTACTCCAGATTATGAACCACTAACACAAAATGACV I R D L N S L r T PD VY L L T I N D

ATVCTACCAAACATCAGAATACGGTCCCTAGGGTTTTCTGGTGCTTTTGAA0ACAGGCACN L P N N R I G A L C r S C A r C D R D

CCTACACACGTTGAAGACAGACACTTGAAGTTTCTACAGCAGCTTGGCAAAGVTAACTTCP T Q r t R H L K r L Q Q L C K C N

G0GAGTGTGGAGATGTGCCGCTATGACCCCCTCCAGGACMCACTGGCCACGVGCTCGCTG S V N C R Y D P L Q D N T G t V V A

CTOAAGAAACTCCACCACAGCACTGAAGAGCACCCCGAGACTTGAGAGCCAGAGCGACGV K K L Q H SVt t tH L RD r t R I I t

ATCCTGA.AATCCTTGCACCATGACAACATCGTCAAGTACAAGGCAGTGTVCTACACSCCCI L K S L Q H D N I V K Y K C V C Y S A

CGGCGGCCCAACCTVAAGATTAATTATGAATAVTTACCATATGCAACGTTACGAGACTATe R R N L R LINS Y L P Y C S L D Y

CTCCAAAAACATAAAGAACCCATAGATCACAAAAAACTTCTTCAATACACATCTCAGATAL Q K HKK R I D H K K L L Q Y S Q I

TGCAACGGCATGGAATATCTTGCTACAAAAAGGTATATCCACAGCGACCTGGCAACAAGCC K 0 N t Y L C T K R Y IH R D L A T R

AACATATTGGTCOAAAATCAGAACAGGOTTAAAATAGGCAGACTTCGCATAACCAAAGTCN I L V t N t N R V K I C D C L T K V

VV0CCGCAGGACAAAGAATACTACAAAGTAAAGGAGCCAGGCAAAGCCCCATATTCTCGL P Q DKK Y Y t V K t P G t S P I W

TACGCACCT0AATCCTGCAGAGACCCAAGTTTCTGTCGCCTCCACATGTTCGAGCTTTY A P t 5 L T t 5 K P S V A S D V W 5 F

COACTOGTTCTATACCAACVTTCACATACATCGAGAAGACTAAAAGTCCACCCGVCCAAC V V L 2 L Y I K S K S P P V

TTTATGOAVAJTVGCAATGATAAACAACGCCAAATGATTOTGTTCCATTTCATAGAGR I 0 H D K Q C Q I V N L 2

CTACTCAAGCACAACGOAAGATTOCCAAGOCCAGAAGGATVGCCCAGATGAOATTTATOTO

TTCCACACACATCCCATCCOCAACATATAGGCCACCGAA CCCACVCATTCACTCAGACC W I K S 0 T v

ACCAAGCAGACTTCCACAACCACAACAAACTCTCTACCCVVCVCTACACATCCTTAT

1773 C (a) (a) (a)

560 CATGATVCCAGCTACCCAGAAGAAACTGTGACGCCCTCTGCTCAAAGCTTTVCTTC

L833So

l893600

L9S3620

1013640

1073660

11336$0

1193700

t2S3720

1313740

1373760

2433

780

2493

800

2553

820

2613

840

2673

860

2733

680

2793

900

26S3920

2913

940

2973

960

3033

980

3093

1000

3153

1020

32 13

1040

3273

1060

3333

1080

3393

1100

34S31120

3S131129

3573

FIG. 1. (Legend appears at the bottom of the opposite page.)

AGAATCAA-ATTTAlATATTTAATGAAAGTCTTGGCCAAGGTACTTTTACAAAAAT1TTTTR U Z D L I F N * S L C Q C T F T X I

Proc. Natl. Acad. Sci. USA 90 (1993) 8431

In Vitro Translation and Transcription. Full-length Jakl orJak2 cDNAs were inserted into pBSK (Stratagene) to maketranscripts with T3 RNA polymerase. Approximately 3 ,g ofRNA was used in translation reaction mixtures (Stratagene)in the presence of 35S translabel (NEN). The products weredivided equally and either subjected to SDS/PAGE directlyor immunoprecipitated with Jakl or Jak2 antisera. Compe-titions were preformed by incubating peptides (100 pug/ml)with antisera for 1 h at 4°( prior to immunoprecipitation.In Vito Kinase Assays. Immunoprecipitated proteins on

protein A-Sepharose (Pharmacia) were washed with kinasebuffer (50 mM NaCI/5 mM MgCl2/5 mM MnCl2/0.1 mMNa3VO4/10 mM Hepes, pH 7.4) and incubated for 30 min atroom temperature with an equal volume of kinase buffercontaining ([32P]ATP at 0.25 mCi/ml. After extensive wash-ing, proteins were eluted with sample buffer for SDS/PAGEand separated on 7% gels. 32P-containing proteins werevisualized by autoradiography. In vitro-phosphorylated Jak2was isolated from gel slices and the phosphoamino acidcontent was determined (22).

RESULTSThe tyrosine kinases expressed in IL-3-dependent cells wereidentified by PCR using degenerative oligonucleotides cor-responding to the conserved regions of the tyrosine kinasedomain (5). One ofthe most frequently isolated cDNA cloneswas identical to Jak2 (16). Expression analysis indicated thatJak2 was widely expressed and prompted us to obtainfull-length cDNA clones. Screening of myeloid cDNA librar-ies allowed the isolation of several overlapping clones, thelongest (4 kb) of which contained the entire coding region ofJak2. The nucleotide sequence of Jak2 contains an openreading frame of3387 bp and the 5' end has three stop codonsbefore the firstATG (Fig. 1). Although the firstATG does notfulfill the Kozak consensus flanking sequences, it is imme-diately followed by an ATG codon in the typical translationinitiation environment (23). The 5' end does not contain anobvious signal peptide. The compiled size of the 3' untrans-lated region of the Jak2 clones is 0.9 kb, which wouldcorrespond to a 4.4-kb transcript. The Jak2 open readingframe encodes 1129 aa with a calculated molecular mass of130 kDa. Hydrophilicity analysis failed to identify transmem-brane regions. During our studies, a partial sequence ofJak2was published (10), which lacked the first 143 aa. A compar-ison ofthe sequences indicates 71 nt differences in the codingregion, resulting in 9 aa changes (Fig. 1). Our cDNA clonesdid not contain a 7-aa insert at position 711 found in a cDNAclone described by Harpur et al. (10).The murine Jak2 gene is closely related to the human tyk2

and JAKI genes (42% and 43% identities, respectively). Wehave also obtained full-length cDNA clones for the murineJaki gene (data not shown), which has 45.5% identity to Jak2at the nucleotide level in the coding region. The Jak2 proteinhas a 600-aa N terminus that lacks obvious SH2 or SH3domains. After this is a kinase-related domain (domain 2) anda C-terminal kinase domain (domain 1). The C-terminalkinase domain contains the conserved residues in subdo-mains VI-VIII that are characteristically found in tyrosinekinases (24). Subdomain VIII, which may contribute tosubstrate recognition, contains a unique FWY amino acidmotiffound in all Jak family members. Domain 2 begins at aa543 and the structural subdomains can be identified. How-

A B C D

KDa :

215 -

105 -

69 -

43 -

FIG. 2. In vitro translation andimmunoprecipitation of Jak2.Capped Jak2 mRNA was used inin vitro translation reaction mix-tures and products labeled with[35S]methionine. After the reac-tions, Jak2 was immunoprecipi-tated from reaction mixtures in theabsence of competing peptides(lane A) or in the presence of thepeptide to which the antiserumwas raised (lane B), to an irrele-vant peptide (lane C), or to a pep-tide corresponding to the sameregion of Jakl. The immunopre-cipitates were resolved by SDS/PAGE.

ever, clear differences in the amino acid composition andspacing in critical kinase subdomains I, II, VI, and VIII (24)raise the possibility that this domain may have a regulatoryfunction or an unknown substrate specificity.Although the N terminus of the JAK family proteins is less

homologous than the kinase domains (36-39%o vs. 49-56%),there are several stretches of homology. Database searcheswith the N-terminal sequence did not show homology withother proteins. Close comparisons of the region immediatelyN-terminal to the kinase-like domain reveal some similarityto SH2 domains.To biochemically characterize the Jak2 protein, anti-

peptide antisera were prepared against a region that wasunique for Jak2 (aa 758-776). To assess the reactivity,immunoprecipitations were done with in vitro-synthesizedJak2 (Fig. 2). In vitro translation of Jak2 RNA yielded a130-kDa protein (data not shown). This protein was immu-noprecipitated by the Jak2 anti-peptide antiserum (Fig. 2,lane A) but not by an irrelevant antiserum (data not shown).Immunoprecipitation was competed by the homologous pep-tide (lane B) but not by an irrelevant peptide (lane C) or bya peptide that is the homologous region of Jakl (lane D). TheJak2 antiserum did not immunoprecipitate in vitro-synthesized Jakl (data not shown). Lastly, the Jak2 antise-rum immunoprecipitated a 130-kDa protein from in vivo-methionine-labeled cells, which was competed by the homol-ogous peptide (data not shown).

IL-3 stimulation induces tyrosine phosphorylation of sev-eral cellular substrates including the IL-3 receptor subunit(3, 25). We therefore determined whether Jak2 was a sub-strate (Fig. 3). Western blot analysis of cell lysates with amonoclonal antibody against phosphotyrosine (4G10) de-tected the appearance of several proteins after IL-3 stimu-lation including a broad band at 130-140 kDa, a minor bandat 70 kDa, and major bands at 55 kDa, 50 kDa, and 38 kDa.When extracts were immunoprecipitated with Jak2 antise-rum, a 130-kDa protein was readily detected in stimulatedcells. Also of note are proteins of 110 kDa, 70 kDa, and 60kDa that coimmunoprecipitated with Jak2. These have beenconsistently seen in Jak2 immunoprecipitations. Immunopre-cipitation with Jakl antiserum consistently detected a weak

FIG. 1 (on oppositepage). Sequence ofthe murine Jak2 cDNA. The sequence oftheJak2 open reading frame and noncoding regions is shown.Nucleotide and amino acid differences from the published partial Jak2 cDNA sequences (10) are shown. The ATG codons are indicated (*). Thesymbol > above nt 522 designates the 5' end of the reported Jak2 sequence. The symbol A at nt 2226 indicates the location of the 7-aa insertas detected (10). The nucleotides shown in parentheses in the 3' noncoding region were present in the previous studies (10) but not detectedin our studies.

Medical Sciences: Silvennoinen et al.

8432 Medical Sciences: Silvennoinen et al.

Antiserum: TCL aJAK2 aIL3Rf3 aJAK1r- F l r - 1

IL3: - + - + - + +

KDa

215 -

105 -

69 -

43 -

Probed with anti-PTyr

FIG. 3. IL-3 stimulation of the tyrosine phosphorylation of Jak2.DA-3 cells were deprived of growth factors for 14 hr, eitherunstimulated (-) or stimulated with IL-3 (10 units/ml) (+), andlysed, and extracts were incubated with the indicated antisera. Theimmunoprecipitates or total cell lysates (lanes TCL) were resolvedby SDS/PAGE. The proteins were transferred to nitrocellulose andthe blots were probed with the 4G10 monoclonal antibody againstphosphotyrosine. The positions of the migration of standards areindicated on the left. Antisera: aJAK2, anti-JAK2; aIL3R,B, anti-IL-3 receptor , chain; aJAK1, anti-JAK1.

band at 130 kDa, indicating that Jakl may also be a substrate.The inducible tyrosine phosphorylation of the IL-3p chain,which migrates slightly above Jak2, is shown in Fig. 3. Thusthe broad band seen in total cell lysates consists of both Jak2and the IL-313 chain.To further establish Jak2 phosphorylation, cells were stim-

ulated with IL-3 and the phosphotyrosine-containing fractionwas isolated on Sepharose beads containing anti-phosphoty-rosine monoclonal antibody 1G2. Jak2 was readily detectedon Westem blots of the IG2-isolated proteins (Fig. 4A) from

B0 5 10 30 60 120

A TCL IG2-..,

1L3: - + -+

::.: :II

C

FIG. 4. Specificity and kinetics of Jak2 tyrosine phosphorylationafter IL-3 stimulation. Cell extracts were prepared from unstimulatedcells (lanes - and 0) or from cells stimulated with IL-3 at 10 units/mlfor 10 min (+) or stimulated for the indicated times. (A) Thephosphotyrosine-containing protein fraction was isolated by bindingto and elution from lG2-coupled Sepharose beads; the 1G2 mono-clonal antibody against phosphotyrosine is as described (26, 33). Theproteins were resolved by SDS/PAGE and transferred to nitrocel-lulose, and the blots were probed with the Jak2 antiserum. Alterna-tively, the total cell lysate (TCL) was electrophoresed and trans-ferred, and the blots were probed with the Jak2 antiserum. (B and C)The cell extracts were immunoprecipitated with the Jak2 antiserum,the precipitates were resolved by SDS/PAGE and transferred tonitrocellulose, and the blots were probed with the 4G10 monoclonalantibody (B) or the Jak2 antiserum (C).

kDa

N~o *No_~+4 , _q + _ +

r- +- +

215 -

_~ c - ~- Jak2

105 -

70

43-

Autoradiography

** , 4- Jak2

Probed with Anti-Jak2

FIG. 5. IL-3 stimulation activates Jak2 in vitro kinase activity inimmunoprecipitates. DA3 cells were removed from growth factorsand were either unstimulated (-) or stimulated with IL-3 for 10 min(+). Extracts were immunoprecipitated with normal rabbit serum(NRS) or Jak2 antiserum in the absence ofcompeting peptide (aJak2)or in the presence of the peptide (30 ug/ml) to which the antiserumwas raised (aJak2 + Jak2 peptide) or in the presence ofan equivalentamount of the peptide that corresponds to the comparable region ofJakl (aJak2 + Jakl peptide). The immunoprecipitates were used forin vitro kinase assays and the products were resolved by SDS/PAGE, transferred to nitrocellulose, and detected by autoradiogra-phy (Upper). The blots were subsequently probed with the Jak2antiserum (Lower).

stimulated but not unstimulated cells. The kinetics are illus-trated in Fig. 4B. Jak2 tyrosine phosphorylation was maximalat 5 min after IL-3 stimulation and, subsequently, decreased.These kinetics are comparable to the general pattern oftyrosine phosphorylation (26). During this period, there wasno change in the levels of Jak2 as assessed by Western blotanalysis with Jak2 antiserum (Fig. 4B).To assess the effect on Jak2 kinase activity, cells were

stimulated with IL-3 for 10 min, Jak2 was immunoprecipi-tated and in vitro kinase assays were performed. No in vitrokinase activity was detected in immunoprecipitates withnormal serum from unstimulated or stimulated cells (Fig. 5).However, immunoprecipitates with Jak2 antiserum con-tained kinase activity and resulted in the phosphorylation ofa protein of 130 kDa that comigrated with Jak2. In contrast,the 130-kDa band was not detected with unstimulated cells.Phosphoamino acid analysis of the 130-kDa band demon-strated the presence ofpredominantly phosphotyrosine (datanot shown). Interestingly, there were no other major proteinbands phosphorylated in these in vitro reactions. The Jak2peptide blocked precipitation of kinase activity, whereas apeptide to the corresponding region of Jakl had no effect.

DISCUSSIONOur studies provide the complete sequence of the murineJak2 as indicated by three criteria. (i) The coding sequenceinitiates at the same site as the published sequences ofhumantyk2 and JAKI. (ii) The first ATG is preceded by stop codonsin all reading frames. (iii) The size of the compiled cDNAs isconsistent with the observed transcript. Our sequence varies

Proc. Natl. Acad Sci. USA 90 (1993)

Proc. Natl. Acad. Sci. USA 90 (1993) 8433

from the published partial sequence (10). Since a portion ofthe previous sequence was obtained from PCR-amplifiedDNAs, some differences may be PCR artifacts.

IL-3 stimulation induces tyrosine phosphorylation (3, 27).One of the substrates is Jak2 (3). Among the tyrosine kinasesin IL-3-dependent cells, there was a remarkable specificityfor Jak2. In particular, no changes in phosphorylation ofLyn,Tec, or c-Fes have been seen (data not shown). However, weconsistently see a low level of Jakl phosphorylation afterIL-3 stimulation. Therefore, Jakl may share sufficient sim-ilarity to Jak2 to weakly associate with the IL-3 receptorcomplex. Alternatively, since there is considerable sequencehomology between Jakl and Jak2 at the potential autophos-phorylation site, Jakl may be a substrate for Jak2. However,we have not detected activation of Jakl in vitro kinaseactivity (data not shown).

IL-3 stimulation results in the activation of Jak2 in vitrokinase activity. The C-terminal protein-tyrosine kinase do-main ofJak2 contains the characteristic autophosphorylationsite of tyrosine kinases (24).The requirement for IL-3 binding for kinase activity indi-

cates that Jak2 is highly regulated in cells, consistent with amajor role in growth regulation. The primary substrate of thein vitro kinase reactions was Jak2. In particular, there was nodetectable phosphorylation of immunoglobulins nor is eno-lase a substrate for Jak2 (data not shown), indicating that Jak2may have a strict substrate specificity. The requirement forreceptor activation and the substrate specificity may accountfor the inability to demonstrate Jakl protein-tyrosine kinaseactivity in previous studies (9).We have shown (15) that Jak2 is tyrosine-phosphorylated

and activated afterEPO stimulation. Moreover, these studiesdemonstrated that Jak2 physically associates with a mem-brane-proximal region of the cytoplasmic domain ofthe EPOreceptor that is essential for function. However, like the EPOreceptor, the (3 subunit of the IL-3 receptor is rapidlytyrosine-phosphorylated and it can be hypothesized that thisphosphorylation is mediated by Jak2. For the EPO receptor,tyrosine phosphorylation occurs at sites in the cytoplasmicC-terminal end and this region is not required for mitogene-sis.The ability of IL-3 and EPO to activate Jak2 suggests the

possibility that Jak2 may be a component in the signaltransducing pathways of several cytokine receptors. In pre-liminary experiments, granulocyte-macrophage and granu-locyte colony-stimulating factors were found to induce thetyrosine phosphorylation of Jak2. This is consistent withstudies showing that these factors induce comparable pat-terns of tyrosine phosphorylation (3). We have also observedtyrosine phosphorylation of Jak2 in response to IFN-y in amacrophage cell line (unpublished data).The hematopoietic growth factor receptors are members of

a receptor superfamily that includes growth hormone recep-tor, the prolactin receptor, ciliary neurotropic factor recep-tor, and others (28). Moreover, the receptors for IFN havebeen speculated to have evolved from the same progenitor.The recent studies implicating Tyk2 in IFN-a signaling (12)and Jak2 in signaling by IL-3, EPO (15), granulocyte andgranulocyte-macrophage colony-stimulating factor, IFN-y,and growth hormone (29) suggest that the JAK family kinasesare involved in the signal-transducing pathways utilized bythe cytokine/IFN superfamily of receptors. If correct, it isfurther tempting to speculate that the JAK family of kinasesmay regulate gene expression through comparable pathwaysinvolving family members related to the IFN-stimulated genefactor 3a proteins (30, 31) and the IFN-stimulated gene factor

3y-related DNA binding proteins including ICSBP, IRE1,IRF2, and possibly Myb (32).

We thank Linda Snyder, Brenda Simmons, and Cynthia Miller forexcellent technical assistance in these studies. This work wassupported in part by the National Cancer Institute Cancer CenterSupport (CORE) Grant P30 CA21765, by Grants RO1 DK42932(J.N.I.) and DK44158 (J.L.C.) from the National Institute of Diabe-tes and Digestive and Kidney Diseases and by the American Leba-nese Syrian Associated Charities (ALSAC).

1. Metcalf, D. (1989) Nature (London) 339, 27-30.2. Clark, S. C. & Kamen, R. (1987) Science 236, 1229-1237.3. Ihle, J. N. (1992) in Interleukins: Molecular Biology and Im-

munology, ed. Kishimoto, T. (Karger, Basel), pp. 65-106.4. Miyajima, A., Kitamura, T., Harada, N., Yokota, T. & Arai, K.

(1992) Annu. Rev. Immunol. 10, 295-331.5. Wilks, A. F. (1991) Methods Enzymol. 200, 533-546.6. Mano, H., Mano, K., Tang, B., Koehler, M., Yi, T., Gilbert,

D. J., Jenkins, N. A., Copeland, N. G. & Ihle, J. N. (1993)Oncogene 8, 417-424.

7. Torigoe, T., O'Connor, R., Santoli, D. & Reed, J. C. (1992)Blood 80, 617-624.

8. Wilks, A. F. (1989) Proc. Natl. Acad. Sci. USA 86, 1603-1607.9. Wilks, A. F., Harpur, A. G., Kurban, R. R., Ralph, S. J.,

Zurcher, G. & Ziemiecki, A. (1991) Mol. Cell. Biol. 11, 2057-2065.

10. Harpur, A. G., Andres, A. C., Ziemiecki, A., Aston, R. R. &Wilks, A. F. (1992) Oncogene 7, 1347-1353.

11. Firmbach-Kraft, I., Byers, M., Shows, T., Dalla-Favera, R. &Krolewski, J. J. (1990) Oncogene 5, 1329-1336.

12. Velazquez, L., Fellous, M., Stark, G. R. & Pellegrini, S. (1992)Cell 70, 313-322.

13. Fu, X. Y. (1992) Cell 70, 323-335.14. Schindler, C., Shuai, K., Prezioso, V. R. & Darnell, J. E., Jr.

(1992) Science 257, 809-813.15. Witthuhn, B., Quelle, F. W., Silvennoinen, O., Yi, T., Tang,

B., Miura, 0. & Ihle, J. N. (1993) Cell 74, 227-236.16. Wilks, A. F. (1989) Proc. Natl. Acad. Sci. USA 86, 1603-1607.17. Yi, T. L. & Willman, C. L. (1989) Oncogene 4, 1081-1087.18. Morishita, K., Parker, D. S., Mucenski, M. L., Jenkins, N. A.,

Copeland, N. G. & Ihle, J. N. (1988) Cell 54, 831-840.19. Bartholomew, C. & Ihle, J. N. (1991) Mol. Cell. Biol. 11,

1820-1828.20. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl.

Acad. Sci. USA 74, 5463-5467.21. Ihle, J. N. & Askew, D. (1989) Int. J. Cell Cloning 1, 1-30.22. Cooper, J. A., Sefton, B. M. & Hunter, T. (1983) Methods

Enzymol. 9, 387-402.23. Kozak, M. (1987) Nucleic Acids Res. 15, 8125-8148.24. Hanks, S. K., Quinn, A. M. & Hunter, T. (1988) Science 241,

42-52.25. Sorensen, P., Mui, A. L. F. & Krystal, G. (1989)J. Biol. Chem.

264, 19253-19258.26. Isfort, R., Abraham, R., Huhn, R. D., Frackelton, A. R., Jr.,

& Ihle, J. N. (1988) J. Biol. Chem. 263, 19203-19209.27. Ihle, J. N. (1990) in Peptide Growth Factors and Their Recep-

tors, eds. Sporn, M. & Roberts, A. (Springer, New York), pp.541-575.

28. Bazan, J. F. (1992) Science 257, 410-413.29. Artgetsinger, L. S., Campbell, G. S., Yang, X., Witthuhn,

B. A., Silvennoinen, O., Ihle, J. N. & Carter-Su, C. (1993) Cell74, 237-244.

30. Schindler, C., Fu, X.-Y., Improta, T., Aebersold, R. & Damell,J. E., Jr. (1992) Proc. Natl. Acad. Sci. USA 89, 7836-7839.

31. Fu, X.-Y., Schindler, C., Improta, T., Aebersold, R. &Darnell,J. E., Jr. (1992) Proc. Natl. Acad. Sci. USA 89, 7840-7843.

32. Veals, S. A., Schindler, C., Leonard, D., Fu, X.-Y., Aeber-sold, R., Darnell, J. E., Jr., & Levy, D. E. (1992) Mol. Cell.Biol. 12, 3315-3324.

33. Frackelton, A. R., Jr., Ross, A. H. & Eisen, H. N. (1983) Mol.Cell. Biol. 3, 1343-1352.

Medical Sciences: Sflvennoinen et aL