of August 25, 2018.This information is current as

Cytokine ResponsesKinase Activity and Differentially Affects An Unusual Insertion in Jak2 Is Crucial for

Peter C. Heinrich and Serge HaanBehrmann,Sommer, Tanja Nöcker, Catherine Rolvering, Iris

Claude Haan, Daniela C. Kroy, Stefan Wüller, Ulrike

http://www.jimmunol.org/content/182/5/2969doi: 10.4049/jimmunol.0800572

2009; 182:2969-2977; ;J Immunol 

MaterialSupplementary http://www.jimmunol.org/content/suppl/2009/02/18/182.5.2969.DC1

Referenceshttp://www.jimmunol.org/content/182/5/2969.full#ref-list-1

, 15 of which you can access for free at: cites 29 articlesThis article

        average*  

4 weeks from acceptance to publicationFast Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

Submit online. ?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists, Inc. All rights reserved.Copyright © 2009 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on August 25, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

by guest on August 25, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

An Unusual Insertion in Jak2 Is Crucial for Kinase Activityand Differentially Affects Cytokine Responses

Claude Haan,1* Daniela C. Kroy,1† Stefan Wuller,†‡ Ulrike Sommer,† Tanja Nocker,†

Catherine Rolvering,* Iris Behrmann,* Peter C. Heinrich,†§ and Serge Haan2*†

The Janus kinases, Jaks, constitutively associate with the cytoplasmic region of cytokine receptors and play an important role ina multitude of biological processes. Jak2 dysfunction has been implicated in myeloproliferative diseases and leukemia. AlthoughJaks were studied extensively for many years, the molecular mechanism of Jak activation upon cytokine stimulation of cells is stillincompletely understood. In this study, we investigated the importance of an unusual insertion located within the kinase domainin Jak2. We found that the deletion of this insertion, which we named the Jak-specific insertion (JSI), totally abrogates Jak2autophosphorylation. We further point mutated four residues within the JSI that are conserved in all Jak family members. Threeof these mutants showed abrogated or reduced autophosphorylation, whereas the fourth displayed increased autophosphorylation.We found that the phosphorylation state of these mutants is not influenced by other domains of the kinase. Our data furthersuggest that the JSI is not required for the negative regulation of kinase activity by the suppressor of cytokine signaling proteins,SOCS. Most importantly, we show that mutations in this region differentially affect IFN-� and erythropoietin signal transduction.Taken together, the dramatic effects on the phosphorylation status of Jak2 as well as the differential effects on the signaling viadifferent cytokines highlight the importance of this unusual region for the catalytic activity of Jaks. The Journal of Immunology,2009, 182: 2969–2977.

P rotein kinases are involved in many biological pathwaysand deregulation of their kinase activity is observed inmany diseases. In recent years, many reports on the three-

dimensional structures of protein kinases have given valuable in-sights into the mechanisms underlying the regulation of kinaseactivity (1, 2). Although these studies revealed a high degree ofconformational plasticity of protein kinase domains between theinactive and active state, the basic protein kinase fold is highlyconserved.

Jaks are cytoplasmic kinases that constitutively associate withthe cytoplasmic region of cytokine receptors. They activate tran-scription factors of the STAT family and play a pivotal role in avariety of biological processes such as cell growth, differentiation,and survival as well as in hematopoietic and immune responses (3,4). Jaks contain a 4.1/ezrin/radixin/moesin (FERM)3 domain, apotential SH2 domain, a kinase-like (or pseudokinase) domain,and a classical tyrosine kinase domain. Although Jaks have been

studied extensively for many years, their mechanism of activationupon cytokine stimulation is still incompletely understood. Jak2recently came into the focus of interest after a somatic mutation inJak2 (Jak2-V617F) was identified in many patients suffering frommyeloproliferative disorders (5, 6). This mutation leads to a cyto-kine-independent activation of Jak2 and downstream signalingproteins. Jak2 is involved in the signal transduction of many cy-tokines such as erythropoietin (Epo), IFN-�, IL-3, growth hor-mone, thrombopoietin, and others (7–13). Depending on the cyto-kine/cytokine receptor system, signal transduction can involveeither two Jak2 molecules or Jak2 paired with Jak1 or Tyk2, re-spectively. Upon treatment of cells with Epo, the ligand recruitstwo Epo receptor chains, each associated with Jak2, whereasIFN-� stimulation triggers a heteromeric receptor complex com-posed of IFN-�R� and IFN-�R�, which are associated with Jak1and Jak2, respectively. Using a molecular modeling approach, wedetected an unusual insertion in the kinase domain of Jaks locatedin close proximity to the catalytic cleft that we designated Jak-specific insertion (JSI). This region was also highlighted in thepublished crystal structures of the Jak2 and Jak3 kinase domainsand termed “Lip” in the context of Jak2 (14, 15). In the presentstudy, we investigated the importance of this unusual insertion inthe context of Jak2. We find the JSI/Lip to be crucial for kinaseactivity and report that mutations in this region can differentiallyaffect signal transduction via different cytokines.

Materials and MethodsAlignment of kinase domains, structure representation, andmolecular modeling

For the graphic representation of protein structures, the programs PyMOL(16) and Ribbons (17) were used. The Brookhaven Protein Data Bank entrycode for the Jak2 structure is 2b7a. The initial alignment of the kinasedomain sequences was performed by the use of the BLAST (basic localalignment search tool) program (18). Modifications were then introduced tomeet the structural requirements derived from the solved structures. The

*Life Sciences Research Unit, University of Luxembourg, Luxembourg, Luxem-bourg; †Department of Biochemistry and ‡Department of Pediatrics, University Hos-pital of Rheinish-Westphalian Technical University, Aachen, Germany; and §Insi-tute of Biochemistry and Molecular Biology, Albert-Ludwigs University,Freiburg, Germany

Received for publication February 20, 2008. Accepted for publication December23, 2008.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 C.H. and D.C.K. contributed equally to this work.2 Address correspondence and reprint requests to Dr. Serge Haan, Life Sciences Re-search Unit, University of Luxembourg, 162A Avenue de la Faïencerie, L-1511 Lux-embourg, Luxembourg. E-mail address: serge.haan@uni.lu3 Abbreviations used in this paper: FERM, 4.1, ezrin, radixin, moesin; CIS, cytokine-inducible SH2-containing protein; Epo, erythropoietin; eYFP, enhanced yellow flu-orescent protein; JSI, Jak-specific insertion; FRT, Flp recombinase target; HEK, hu-man embryonic kidney; Luc, luciferase; SOCS, suppressor of cytokine signaling; WT,wild type.

Copyright © 2009 by The American Association of Immunologists, Inc. 0022-1767/09/$2.00

The Journal of Immunology

www.jimmunol.org/cgi/doi/10.4049/jimmunol.0800572

by guest on August 25, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

GenPept data bank accession numbers for the sequences used are as fol-lows: P52332 (murine Jak1); NP_032439 (murine Jak2); P29597 (humanTyk2); AAC50950 (human Jak3); NP_511119 (hopscotch); NP_005408(human c-Src); EAW63310 (human fibroblast growth factor receptor); andEAW69042 (human insulin receptor). For the molecular modeling ap-proach, mutations were introduced using the PyMOL (16) program andenergy minimization of the resulting structures was performed under vac-uum conditions with the GROMOS program library (W. F. van Gunsteren;distributed by BIOMOS Biomolecular Software, Laboratory of PhysicalChemistry, University of Groningen, Groningen, The Netherlands). For thegeneration of the Jak2-JSI.mpg movie,4 the program Ribbons (17) wasused for representation of the structure and rendering was performed withPOV-Ray software (Persistence of Vision Raytracer, version 3.6 (2004;www.povray.org); Persistence of Vision).

Cell culture and transfection

�2A cells (human fibrosarcoma cells provided by Dr. I. M. Kerr, CancerResearch UK, London, U.K.), HEK-FRT (where HEK is human embryonickidney and FRT is Flp recombinase target) cells (Flp-In T-REx-293 cells;Invitrogen), and COS-7 cells (simian kidney cells, American Type CultureCollection no. CRL1619) were maintained in DMEM (Invitrogen). Allmedia were supplemented with 10% FCS, 100 mg/L streptomycin, and 60mg/L penicillin. HEK-FRT cell lines stably expressing Jak2 constructs andreconstituted �2A cells were cultured with 250 �g/ml hygromycin. Cellswere grown at 37°C in a water-saturated atmosphere at 5% CO2. Transienttransfection of COS-7 and �2A cells (�1 � 107 cells) was conducted usingFuGENE 6 (Roche) or Superfect (Qiagen) according to the manufacturers’recommendations. HEK-FRT cells stably expressing eYFP-Jak2-WT,eYFP-Jak2-�JSI, eYFP-Jak2-F1061A, eYFP-Jak2-M1062A, eYFP-Jak2-M1064A, and eYFP-Jak2-I1065A as well as �2A-FRT-EpoR cells (whereeYFP is enhanced yellow fluorescent protein and WT is wild type) recon-stituted with Jak2 and the various mutants were generated using the Flp-InT-Rex System from Invitrogen according to the manufacturers’ recommen-dations. The expression of eYFP-tagged Jak2 proteins was induced byincubating the cells with 10 �g/ml doxycycline for 12 h unless otherwisestated. IFN-� and Epo were obtained from Peprotech. Cells were stimu-lated with 1000 U/ml IFN-� or 5–10 �g/ml Epo for the times indicated.

Constructs

Generation of the Jak2 deletion construct �JSI and the F1061A mutant wasperformed using standard PCR techniques. M1062A, M1064A, I1065A,and L87A/M88A were generated using the QuikChange kit (Stratagene)following the manufacturer’s recommendations. The various constructswere transferred to the different vectors pSVL-Jak2, pcDNA5/FRT/TO-Jak2 (19), pcDNA5/FRT/TO-YFP-Jak2, and pEF-eYFP-Jak2-JH1wt (20)using standard techniques. For the generation of pcDNA5/FRT/TO-YFP-Jak2, YFP-Jak2 was subcloned from pEF-YFP-Jak2 (21) into pcDNA5/FRT/TO (Invitrogen). The pcDNA5-HA-EpoR (human) construct was pro-vided by Dr. S. Constantinescu (Ludwig Institute for Cancer Research,Brussels, Belgium). pGV-CIS-promoter-Luc and pEF-FLAG/SOCS-1 orSOCS-3 (where CIS is cytokine-inducible SH2-containing protein, Luc isluciferase, and SOCS is suppressor of cytokine signaling) were provided byDr. A. Yoshimura (Kyushu University, Fukuoka, Japan) and Dr. D. J.Hilton (Walter and Eliza Hall Institute of Medical Research, Melbourne,Australia), respectively. The pGL2-IRF1-tk-Luc construct contains theSTAT1-responsive element of the IRF1 promoter. For reporter gene anal-ysis, a �-galactosidase expression plasmid, pCH110 (Pharmacia), wasused. Detailed information on the cloning procedures is available uponrequest.

Reporter gene analysis

Cells were transiently transfected with 0.25 �g of the Luc reporter-geneconstructs (pGV-CIS-promoter-Luc or pGL2-IRF1-tk-Luc) together withthe respective SOCS expression plasmid or empty vector using Superfectreagent (Qiagen). Jak2 expression was induced for 24 h and subsequentlythe cells were stimulated with the respective cytokine for 24 h. Cell lysisand Luc assays were conducted using the Promega Luc assay system asdescribed by the manufacturer. Luc activity values were normalized totransfection efficiency monitored by cotransfection of a �-galactosidaseexpression vector (0.2 �g of pCH110lacZ; Amersham Biosciences). Ex-periments were performed in triplicate.

Cell lysis, immunoprecipitation, and Western blot analysis

Cell lysis, immunoprecipitations, coimmunoprecipitations, and Westernblot analysis were performed as described previously (22, 23). The fol-lowing Abs were used: anti-GFP (Rockland); anti-Jak2 (Upstate Biotech-nology for immunoprecipitation and Cell Signaling for Western blotting);anti-pYpYJak2 (Cell Signaling); anti-Jak1 (Santa Cruz Biotechnology);anti-FLAG (Stratagene); anti-phosphotyrosine 4G10 (Upstate Biotechnol-ogy); anti-STAT3 (Transduction Laboratories); anti-phospho-Y705STAT3(Cell Signaling); anti-STAT1 (Transduction Laboratories); anti-phospho-Y701STAT1 (Cell Signaling); anti-STAT5 (Santa Cruz Biotechnology);anti-pY694STAT5 (Cell Signaling); and anti-SOCS3 (Santa Cruz Biotech-nology). HRP-conjugated secondary Abs were purchased from Dako-Cytomation. Signals were detected using a self-made ECL solution (24).

Fluorescence microscopy

HEK/FRT-YFP-Jak2 cells were treated with 10 �g/ml doxycycline to in-duce Jak2 expression. Confocal imaging was conducted in a live cell cham-ber of a Zeiss LSM 510 confocal microscope, enabling the investigationunder cell culture conditions (5% CO2 and 37°C). The plasma membranewas visualized after the addition of trypan blue (0.05% in PBS) at �exc �543 nm and �em � 590 nm (where exc is excitation and em is emission).Excitations were achieved by an argon-ion laser (�exc � 458, 488, and 514nm) and a helium/neon laser (�exc � 543 nm). The thickness of opticalsections was �0.6 �m. Twenty-four hours after being seeded on poly-L-lysine-coated glass cover slips, cells were incubated with doxycycline andinvestigated under cell culture conditions 12 h later in phenol red-freemedium.

ResultsJaks contain an unusual insertion within their catalytic domains

In this study, we set out to investigate the functional importance ofan unusual insertion that we detected upon performing sequencealignments of various kinase domains with Jaks (Fig. 1A). Theinsertion is present in all published Jak sequences with the excep-tion of hopscotch, the Drosophila homolog. Interestingly, it is alsoabsent in the pseudokinase domain of the Jaks. Secondary structureanalysis of this 15-aa JSI predicted a small �-helix, and a molec-ular modeling approach suggested the existence of an additionalhelix in the lower lobe of the kinase domain, in close proximity tothe catalytic cleft. The recently published crystal structures of theJak3 and Jak2 kinase domains provided proof for the existence ofsuch a helix, which was termed �H-helix for Jak2 and FG helix inthe case of Jak3 (14, 15). The region encompassing the Jak2 �Hhelix and the preceding 3/10 helix was named the Lip region (14)(Fig. 1A). In Fig. 1, B and C show the location of the so-called�H-helix in Jak2 (highlighted in red). The helix flanks the pre-dicted substrate-binding site, indicating potential importance forkinase activity. To investigate the significance of this region forJak2 kinase activity, we generated a deletion mutant (�JSI) as wellas four point mutations (F1061A, M1062A, M1064A, and I1065A)targeting large conserved hydrophobic residues (Fig. 1C andmovie Jak2-JSI.mpg). For the �JSI deletion, 19 aa between the �Gand �I helices were replaced by a flexible 8-aa GS linker to spanthe distance between the two helices (Fig. 2A). As all of the mu-tated hydrophobic residues are involved in hydrophobic interac-tions with the surrounding residues, we used a molecular modelingapproach to determine how their mutation to alanine would affectthese interactions. Table I gives an overview over the interactionsof the four investigated residues with the surrounding amino acids(based on the crystal structure by Boggon et al. in Ref. 15) andindicates the changes that occur upon their mutation to alanine(based on molecular modeling). The crystal structure shows thatthe first residue that we mutated, Phe1061, is an important structuralresidue that connects the JSI to the lower lobe of the kinase domain(Fig. 1C). Met1064 and Ile1065 connect the JSI helix to the �I�Jloop and the �I helix, respectively. Met1062 is orientated towardthe substrate binding site, the activation loop and the catalytic4 The online version of this article contains supplemental material.

2970 AN UNUSUAL INSERTION IN Jak2 THAT AFFECTS ITS ACTIVITY

by guest on August 25, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

cleft. It also has the fewest contacts with neighboring amino acids(Fig. 1C and Table I). As expected, upon mutation of the JSI res-idues to alanine, hydrophobic contacts with surrounding residuesare lost or reduced for all of the investigated residues (Table I).The residue whose mutation affects contacts with neighboringamino acids least is Met1062. The most important structural rolecan be attributed to Phe1061.

Mutations within the JSI affect Jak2 kinase activity

To monitor mutant Jak2 activity, we first used an overexpressionassay and studied Jak2 autophosphorylation (Fig. 2B). We foundthat the �JSI deletion mutant as well as the F1061A and I1065Amutants failed to show significant autophosphorylation. M1062Aand M1064A displayed autophosphorylation, although to differentextents. Most interestingly, the M1062A mutant showed increased

autophosphorylation compared with WT protein. Mutation ofM1064A resulted in decreased autophosphorylation.

The differential phosphorylation status of the mutants isindependent of contacts with other Jak2 domains

As one of our original assumptions was that the JSI could serve asdocking site for the negatively regulating kinase-like domain (JH2domain), we transiently expressed an isolated kinase domain (JH1domain) of Jak2-WT and mutants (Fig. 2C). An eYFP tag wasadded for detection purposes. Again, the �JSI, F1061A, M1064,and I1065A mutants showed no or reduced phosphorylation,whereas the M1062A mutation led to increased autophosphoryla-tion, indicating that the observed level of phosphorylation is nei-ther due to a regulatory interaction with the JH2 domain nor withthe FERM or SH2 domains.

FIGURE 1. Location of the JSI.A, Sequence alignment of the kinaseand kinase-like domains of Jaks(Jak1, Jak2, Jak3, and Tyk2) with thekinase domains of hopscotch (HOP),c-Src (SRC), fibroblast growth factorreceptor 1 (FGFR), and insulin recep-tor (IR). Highly conserved residuesare depicted in red. Secondary struc-ture elements for Jak2 (14) andamino acid numbers are given. Theactivation loop and the JSI regionare highlighted in green and red, re-spectively. The Jak2 Lip region isindicated in blue. B, Surface repre-sentation of the Jak2 structure(Brookhaven Data Bank entry code2b7a) (14). The location of an ATPanalog, the activation loop (green),the JSI (red), and the potential sub-strate-binding site are highlighted. C,Ribbon representation of the Jak2 ki-nase domain structure. The activationloop (green) and the JSI-helix (red)are highlighted. The residues mutatedin this study are indicated as stickmodels and the helices �I and �J arelabeled. The program Ribbons (17)was used for representation of thestructure. A movie (Jak2-JSI.mpg) ofthe represented structure is providedonline.

2971The Journal of Immunology

by guest on August 25, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Expression of Jak2-M1062A leads to constitutive STATactivation

To further characterize the JSI mutants, we generated stable in-ducible transfectants expressing Jak2-WT or one of the four pointmutants using the Jak2-deficient �2A fibrosarcoma cells in whichwe inserted a FRT site. The different mutants were then insertedinto the FRT site. This system thus avoids clonal differences be-cause of the targeted transgene incorporation. Induction by doxy-cycline enables the expression of controlled levels of Jaks depen-dent on the doxycycline concentration used. Before the insertion ofthe Jak2 constructs, we stably transfected the cells with an EpoRconstruct. After induction of Jak2 expression with doxycycline, weinvestigated the phosphorylation of Jak2, STAT3, and STAT5. Asdepicted in Fig. 3, expression of the M1062A mutant leads to Jak2autophosphorylation as well as to the phosphorylation of STAT3and STAT5, whereas WT Jak2 did not elicit constitutive signaling.Furthermore, none of the other mutants induced signaling events.Treatment with 5 or 10 �g/ml doxycycline led to comparable Jak2expression levels, indicating that maximal expression is reached.

JSI mutants display different cellular localization

Next, we generated HEK cells that stably express Jak2 proteinsunder the control of a doxycycline-inducible promoter. All pro-

teins were N-terminally tagged with eYFP to monitor their local-ization in living cells using confocal laser scanning microscopy.To test the YFP-tagged Jak2 construct, we stably and induciblyexpressed YFP-Jak2 WT or a Jak2-FERM domain mutant (L87A/M88A) designed to abrogate receptor interaction in Jak2-deficient�2A-FRT cells and HEK-FRT cells. As shown in Fig. 4A, theYFP-tagged Jak2 protein is activated in a manner comparable tothat of nontagged Jak2. Also, as described previously (21), YFP-Jak2-WT is localized predominantly at the plasma membrane with

FIGURE 2. Mutations in the JSI affect kinase activity. A, Representa-tion of the different mutants generated in this study. Modified residues areunderlined. B, Comparison of the autophosphorylation elicited by the dif-ferent Jak2 mutants. COS-7 cells were transfected with 2 �g of the variouspSVL-Jak2 constructs. After cell lysis (48 h after transfection), Jak2 wasprecipitated and the proteins were resolved by SDS-PAGE and subjected toWestern blot analysis. Autophosphorylation was monitored using a phos-photyrosine (pY) Ab (4G10), and Jak2 expression was assessed by strip-ping and reprobing the blot with a Jak2 Ab. C, COS-7 cells were trans-fected with a 0.5-�g expression plasmid of Jak2-WT, M1062A, andM1064A, 1.0 �g of I1065A, and 1.5 �g of the F1061A and �JSI eYFP-tagged constructs. The lysates were processed as described above. Auto-phosphorylation was monitored using a phosphotyrosine (pY) Ab (4G10),and Jak2 kinase domain expression was assessed by stripping and reprob-ing the blot with a GFP Ab.

FIGURE 3. Jak2-M1062A is constitutively active. �2A cells stably andinducibly expressing Jak2-WT or mutants were treated with 5 or 10 �g/mldoxycycline (Dox) for 24 h or were left untreated. After lysis of the cells,the proteins were resolved by SDS-PAGE and subjected to Western blotanalysis. Phosphorylation of Jak2, STAT5, and STAT3 was assessed usingphospho-specific Abs. Expression of the proteins was checked by strippingthe blot and reprobing it with Abs directed against the various proteins.Equal loading was assessed using a Fin13 Ab. pY, Phosphotyrosine.

Table I. Effect of JSI mutations on the interactions with neighboringamino acidsa

JSIResidue

ContactedAmino Acid

Nature ofthe Interaction

Effect ofAlanine Mutation

F1061 F1019 Hydrophobic ReducedP1057 Hydrophobic Lost

H-bond Not affectedM1064 Hydrophobic ReducedI1065 Hydrophobic Not affectedL1078 Hydrophobic LostL1081 Hydrophobic LostL1082 Hydrophobic LostL1088 Hydrophobic Lost

M1062 F1019 Hydrophobic LostP1058 H-bond Not affectedI1065 Hydrophobic ReducedN1067 H-bond Not affectedI1074 Hydrophobic Lost

M1064 E1060 Hydrophobic ReducedH-bond Not affected

F1061 Hydrophobic LostL1081 Hydrophobic Not affectedG1086 Hydrophobic LostR1087 Hydrophobic LostL1088 Hydrophobic LostP1089 Hydrophobic Lost

I1065 F1019 Hydrophobic LostF1061 Hydrophobic Lost

H-bond Not affectedM1062 Hydrophobic ReducedI1074 Hydrophobic LostH1077 Hydrophobic LostL1078 Hydrophobic Lost

a The main interaction partners for the JSI residues mutated in this study are given.The effect of a mutation to alanine on the interaction is indicated.

2972 AN UNUSUAL INSERTION IN Jak2 THAT AFFECTS ITS ACTIVITY

by guest on August 25, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

a slight cytoplasmic staining and an exclusion of Jak2 from thenucleus (Fig. 4B). In contrast, Jak2-L87A/M88A is not activated(Fig. 4A) and displays cytoplasmic localization (Fig. 4B).

We then evaluated the localization of the JSI mutants (Fig. 4C)in stable HEK-FRT cells using live cell microscopy. The induc-ibility of Jak2 expression is shown in lanes 1 and 3 (Fig. 4C).Membrane staining with trypan blue is provided in lanes 2, 4, and6 (Fig. 4C). As demonstrated in lane 5 (Fig. 4C), the mutantsM1062A and M1064A showed the same distribution as Jak2-WT.The mutant I1065A also displayed membrane localization butshowed an increased portion of cytoplasmic protein. The effect ofthe F1061A mutant was even more pronounced, as YFP-Jak2-F1061A clearly displayed cytoplasmic localization. Colocalizationwith the membrane staining was hardly observed.

Mutations in the JSI region can differentially affect thesignaling of IFN-� and erythropoietin

To investigate whether mutations in the JSI region affect cytokineresponses, we investigated signaling in response to IFN-� and Epoin reconstituted �2A-FRT cells expressing the different JSI mu-

tants. Fig. 5 illustrates the signal transduction after stimulation ofthe cells with IFN-�. First, we investigated the phosphorylation ofJak2 and Jak1. For this, Jak2 expression was induced with 1 �g/mldoxycycline (a dose for which M1062A shows little constitutivesignaling) and the cells were stimulated with IFN-�. After immu-noprecipitation of Jak2 (Fig. 5A) or Jak1 (Fig. 5B), the phosphor-ylation status of the kinase was monitored. As shown in Fig. 5A,phosphorylation of Jak2 increased in cells expressing Jak2-WT,-M1062A, or -M1064A. However, in cells expressing the I1065Amutant, Jak2 phosphorylation was significantly reduced and cellsexpressing the F1061A mutant did not show Jak2 phosphorylation.These two mutants also affect the basal phosphorylation of Jak2(Figs. 5A and 6A). In contrast, Jak1 phosphorylation was compa-rable in Jak2-WT, -M1062A, -M1064A, and -I1065A cells,whereas it was impaired in cells expressing the F1061A mutant(Fig. 5B). To investigate downstream signaling events, we nextinvestigated the kinetics of STAT1 phosphorylation (Fig. 5C) aswell as its dose-dependent activation (Fig. 5D) upon IFN-� treat-ment. We observed that STAT1 phosphorylation was comparablein cells expressing Jak2-WT, -M1061A, -M1064A, and -I1065A

FIGURE 4. Distribution of thedifferent Jak2 proteins in living cells.A, �2A cells stably and inducibly ex-pressing Jak2-WT, eYFP-Jak2-WTor eYFP-Jak2-L87A/M88A weretreated with 1 �g/ml doxycycline for24 h to induce Jak2 expression andwere subsequently treated with 10ng/ml Epo for the times indicated.Jak2 as well as STAT3 phosphoryla-tion was monitored as described forFig. 3. B, Live cell microscopy ofHEK/FRT-cells stably expressingeYFP-tagged Jak2-WT or Jak2-L87A/M88A. Jak2 localization wasassessed by detecting the eYFP fluo-rescence and membrane staining wasperformed by incubation with trypanblue. Scale: 10 �m. C, HEK/FRT-cells stably and inducibly expressingeYFP-tagged Jak2-WT or JSI mu-tants were monitored by confocal mi-croscopy under cell culture condi-tions. Jak2 was detected using theeYFP-fluorescence, and trypan bluewas used to stain the membrane.Lanes 1 and 2, Overview of nonin-duced cells; lanes 3 and 4, overviewof cells in which Jak2 expression wasinduced with 10 �g/ml doxycycline;lanes 5 and 6, localization of Jak2proteins in cells treated with doxycy-cline; lane 7, merged image of lanes5 and 6.

2973The Journal of Immunology

by guest on August 25, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

but significantly reduced in cells expressing the F1061A mutant. Asimilar effect was visible when we checked STAT1 target genesthat are up-regulated in response to IFN-�. Only the F1061A mu-tant showed a clearly impaired induction of IRF1, GBP2, andSTAT1 protein expression. It must be noted that a slight reductionin IRF1 and GBP2 expression was observed in cells expressing theI1065A mutant. This was also observed in a reporter gene exper-iment using an IRF1 promoter reporter construct (Fig. 1F). Re-porter gene analysis consistently showed an induction of the re-porter in nonstimulated cells expressing the M1062A mutant. Thisindicates that the M1062A mutant induces constitutive signalingthat is not detectable in the Western blot analysis but becomesapparent in a reporter gene analysis integrating the STAT signalsover a long time period.

Next, we studied signal transduction in the stably reconstituted�2A cells upon treatment with Epo (Fig. 6). Again, we firstchecked the phosphorylation status of immunoprecipitated Jak2upon Epo stimulation (Fig. 6A). Although Jak2 phosphorylationupon treatment with Epo is more pronounced in comparison toIFN-� stimulation, the phosphorylation pattern for the differentmutants is comparable with a clear reduction of Jak2 phosphory-lation being observed for the I1065A mutant and a loss of phos-phorylation occurring in cells expressing the F1061A mutant. Asillustrated in Fig. 6B, this translates to downstream STAT5 acti-vation. Whereas cells expressing Jak2-WT, -M1062A, or

-M1064A display comparable phosphorylation of STAT5, theI1065A mutation significantly impairs Epo-mediated activation ofSTAT5. In Jak2-F1061A cells, STAT5 signaling upon Epo stim-ulation is abrogated. To further investigate this phenomenon, weperformed a reporter gene analysis using a STAT5-dependent CISpromoter-reporter construct (Fig. 6C). Again, whereas the reportergene induction in cells reconstituted with the WT protein and thetwo Met to Ala mutants is comparable, expression of I1065A andF1061A dramatically impairs the induction of this STAT5-depen-dent reporter. Of note, a constitutive induction of reporter geneactivity was again apparent in nonstimulated cells expressing theM1062A mutant. Finally, we tested whether the different mutantsare able to bind to the Epo receptor. Fig. 6D shows that all con-structs can be coprecipitated with the EpoR upon induction of theprotein with doxycycline. Jak2-F1061A, however, shows reducedbinding to the EpoR.

JSI mutants are susceptible to negative regulation via SOCSproteins

As the SOCS proteins SOCS1 and SOCS3 are able to inhibitJak2 kinase activity by directly binding to its kinase domain viatheir kinase inhibitory region and/or SH2 domain (25–27), weinvestigated whether JSI mutants could be inhibited by SOCS1and SOCS3. To address this point, we expressed increasing

FIGURE 5. IFN-� signaling in cells stably expressing Jak2-WT or JSI mutants. A and B, Jak2-deficient �2A-EpoR cells stably reconstituted withJak2-WT or the different mutants were treated with 1 �g/ml doxycycline for 24 h to induce Jak2 expression and then stimulated with IFN-� (10 ng/ml)for 15 min. Jak2 (A) or Jak1 (B) were precipitated and their phosphorylation was assessed by Western blot analysis. IP, Immunoprecipitation; pY,phosphotyrosine. C, Reconstituted �2A-EpoR-Jak2 cells were treated with IFN-� (10 ng/ml) for the times indicated. STAT1 activation and up-regulationwas monitored by Western blot analysis. Detection with a Fin13 Ab is provided as loading control. D, �2A-EpoR-Jak2-cells were treated with the indicatedamounts of IFN-� for 1 h and STAT1 activation was monitored. E, �2A-EpoR-Jak2-cells were treated with 10 ng/ml IFN-� for the times indicated andthe expression of IRF1, GBP2, STAT1, and Fin13 (loading control) was monitored. F, �2A-EpoR-Jak2-cells were transfected with a STAT1-dependentIRF1 promoter-reporter construct and Jak2 expression was induced with 1 �g/ml doxycycline. After stimulation of the cells with 10 ng/ml IFN-� for 24 h,reporter gene activity was measured as described in Materials and Methods. Error bars represent the SD of a triplicate experiment.

2974 AN UNUSUAL INSERTION IN Jak2 THAT AFFECTS ITS ACTIVITY

by guest on August 25, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

amounts of SOCS1 (Fig. 7A) or SOCS3 (Fig. 7B) in reconsti-tuted �2A-FRT-EpoR cells expressing the different JSI mutants.Because both SOCS proteins also inhibit Jak1 activity, the spe-cific effect of SOCS1 and SOCS3 on the Jak2 mutants wasstudied after stimulation with Epo using the STAT5-dependentCIS reporter construct. As apparent from Fig. 7, both SOCS1and SOCS3 dose-dependently inhibit STAT5-dependent signal-ing in cells expressing the M1062A, M1064A, and I1065A mu-tants in a manner that is comparable to the inhibition of the WTJak2 protein. As cells expressing the F1061 mutant do not dis-play STAT5 activation, the regulation of this mutant could notbe investigated. Because all JSI mutants that elicit STAT5 ac-tivation are sensitive to negative regulation by SOCS1 andSOCS3, it can be concluded that the JSI region is not involvedin the negative regulation of Jak2 by these proteins.

DiscussionIn this study, we investigated the role of an unusual insertion, theJak-specific insertion or JSI, in the kinase domain of Jak2. Thisinsertion of �15 aa, which is present in all Jak family memberswith the exception of the Drosophila homolog, encompasses an�-helical structure. This is lining the substrate-binding site of thekinase and lies in close proximity to the catalytic cleft of the en-zyme (14). The location of this unusual insertion indicates its po-tential importance for kinase activity.

We performed a mutagenesis study for which we generated aJSI-deletion mutant as well as four point mutants of conservedhydrophobic residues. Our results clearly show that the JSI regionis crucial for Jak2 activity. The mutation of the structurally im-portant residue Phe1061 to alanine leads to the total loss of auto-

phosphorylation capacity upon transient overexpression (Fig. 2B)and also prevents Jak2 phosphorylation upon stimulation withIFN-� or Epo in stably transfected cells (Figs. 5A and 6A). To-gether with the observation that this mutation drastically reducesJak2 membrane localization (Fig. 4C) and displays reduced bind-ing to the EpoR, this shows that the JSI region is important for thestructural integrity of the kinase domain. The lack of membranelocalization also supports previous data on Jak1 showing that amutation in the kinase domain can influence not only the enzy-matic activity but can also affect binding of the kinase to cytokinereceptors (19). The mutants M1062A, M1064A, and I1065A dis-play membrane localization (Fig. 4C) and bind to the EpoR (Fig.6D), showing that they can act at the cell membrane. Anothermutant, M1062A, displays increased autophosphorylation upontransient overexpression compared with WT Jak2 (Fig. 2). Fur-thermore, this mutant induces constitutive STAT activation if in-ducibly expressed in stably reconstituted Jak2-deficient �2A cells(Fig. 3). Constitutive activity is not apparent in the Western blotanalyses in Fig. 5, C–E, and Fig. 6B where lower Jak2 levels wereinduced, but is nevertheless detectable in the reporter gene analy-ses where the signal is acquired over 24 h (Figs. 5F, 6C, and 7).This indicates that the JSI region may not only be of structuralrelevance but may also modulate Jak2 activity or even participatein substrate recognition. The idea makes sense as the JSI helix islining the substrate-binding site of the Jaks (Fig. 1).

As the activity of the kinase domain of Jaks can be modulatedby other domains such as the kinase-like (JH2) domain or even theFERM domain (19, 28), we investigated whether the phosphory-lation pattern we observed could be due to interactions with otherJak domains. Expressing an isolated kinase domain of Jak2-WT

FIGURE 6. Epo signaling in cells stably expressing Jak2-WT or JSI-mutants. A, Jak2-deficient �2A-EpoR cells stably reconstituted with Jak2-WT orthe different mutants were treated with 1 �g/ml doxycycline for 24 h to induce Jak2 expression and were then stimulated with Epo (10 ng/ml) for 15 min.Jak2 was precipitated and its phosphorylation was assessed by Western blot analysis. IP, Immunoprecipitation; pY, phosphotyrosine. B, �2A-EpoR-Jak2-cells were treated with the indicated amount of Epo for 1 h and STAT5 activation was monitored using a phospho-specific Ab, and Jak2, STAT5, and Fin13(loading control) expression was assessed using the respective Abs. pY, Phosphotyrosine. C, �2A-EpoR-Jak2 cells were transfected with a STAT5-dependent CIS promoter-reporter construct and Jak2 expression was induced with 1 �g/ml doxycycline. After stimulation of the cells with 10 ng/ml Epofor 24 h, reporter gene activity was measured as described in Materials and Methods. Error bars represent the SD of a triplicate experiment. D, �2A-EpoR-Jak2-cells were treated with 1 �g/ml doxycycline (Dox) for 24 h to induce the expression of the different Jak2 proteins or were left untreated. Afterlysis of the cells, hemagglutinin (HA)-EpoR was precipitated and coprecipitated Jak2 was analyzed by Western blotting. IP, Immunoprecipitation.

2975The Journal of Immunology

by guest on August 25, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

and Jak2-mutants (Fig. 2C), we observed the same phosphoryla-tion patterns as for the transiently expressed full-length proteins(Fig. 2B). Compared with Jak2-WT, the M1062A mutant dis-played hyperphosphorylation and the other mutants showed re-duced or no phosphorylation. Thus, the observed phosphorylationpattern and especially the observed hyperphosphorylation of theM1062A mutant are not due to impaired interactions with otherdomains of the kinase.

Next, we tested whether the point mutants were still sensitive tonegative regulation by SOCS proteins. Both SOCS1 and SOCS3directly bind to the Jak2 kinase domain (25, 26). As it was pro-posed that the kinase inhibitory region binds as a pseudosubstrateto the kinase domain of Jaks (25), it is conceivable that the JSIregion could be involved in this interaction. We therefore testedthe dose-dependent inhibition of Epo-mediated signaling bySOCS1 and SOCS3 in Jak2-deficient cells reconstituted with WTJak2 or the different mutants. We found the mutants M1062A,M1064A, and I1065A to be still sensitive to SOCS-mediated neg-ative regulation (Fig. 7). The effect on the F1061A mutant couldnot be investigated, because Epo signaling is abrogated in cellsexpressing this mutant. However, as Phe1061 is a structural residuethat is entirely buried within the structure, it cannot be involved insurface contacts with the SOCS proteins. Taken together, thesedata suggest that the JSI region is not involved in the negativeregulation of Jaks by SOCS1 and SOCS3.

Finally, we found that mutations in the JSI can differentiallyaffect IFN-� and Epo signal transduction. In Jak2-deficient cellsreconstituted with different Jak2 constructs, we found that the Jak2mutants M1062A, M1064A, and I1065A induce STAT1 phosphor-ylation to the same extent as WT Jak2 after IFN-� stimulation (Fig.5, C and D). Similarly, up-regulation of the STAT1-dependentproteins IRF1, GBP2, and STAT1 was also unaffected by thesemutations (Fig. 5E). However, the F1061A mutant displayed im-paired STAT1 signaling and reduced STAT1-dependent gene ex-pression upon treatment with IFN-�. In clear contrast, the mutantI1065A showed a strongly decreased STAT5 activation (Fig. 6B)and STAT5-dependent CIS induction (Fig. 6C) after Epo stimula-

tion. The signaling observed in cells expressing the M1062A andM1064A mutants was comparable to that of the signals initiated byWT Jak2 but was totally abrogated in F1061A-expressing cells.These data indicate that the JSI mutants behave differently depend-ing on the triggered receptor system. In a homodimeric receptorcomplex involving two Jak2 molecules, the I1065A mutant ishardly able to induce downstream signaling, whereas it can elicitsignaling in the context of a heterodimeric receptor complex suchas the IFN-� receptor recruiting different Jaks, namely Jak1 andJak2. Similarly, although Jak2-F1061A can to some extent mediatesignaling upon IFN-� stimulation, it cannot initiate Jak2/STAT5signaling after Epo stimulation. In the context of Epo stimulation,the signaling behavior of the mutants reflects the phosphorylationpattern of Jak2 (Fig. 6A). After treatment with IFN-�, however, thesignaling characteristics reflect the phosphorylation pattern of Jak1(Fig. 5B). The reduced Jak2 phosphorylation seen for I1065A, aswell as the absence of phosphorylated Jak2 in the case of F1061A,nevertheless led to the activation of STAT1 and the induction ofSTAT1 target genes (Fig. 5, C–F). This does, however, not implythat Jak2 catalytic activity is not required for IFN-� signaling, asSTAT1 activation in �2A cells reconstituted with a kinase-deadJak2 mutant (Jak2-K882E) is abrogated (19, 29). Thus, althoughphosphorylation of the F1061A mutant cannot be detected, it mustretain some catalytic activity. Taken together, the data show thatmutations in Jak2, and in particular in the Jak2 kinase domain, mayaffect signaling via homomeric receptor complexes more dramat-ically than signals mediated via heteromeric receptors involvingdifferent Jak family members. It also suggests that the bindingand/or activation mechanisms of Jaks significantly differ in thesesettings. Interestingly, it was proposed that the presence of ho-modimeric class I cytokine receptors is needed for Jak2-V617F-mediated transformation (30). This mutant Jak2 protein carryingan activating mutation within the kinase-like domain is found inthe majority of patients suffering from myeloproliferative diseases(5, 6). Overall, the intramolecular interactions between the differ-ent domains of Jaks, but also the contacts between two Jak mol-ecules in receptor dimers, may account for significant differences

FIGURE 7. Sensitivity of JSI-mutants toward SOCS proteins.�2A-EpoR-Jak2-cells were trans-fected with a STAT5-dependent CISpromoter-reporter construct togetherwith increasing amounts of SOCS1(A) or SOCS3 (B). Jak2 expressionwas induced with 1 �g/ml doxycy-cline for 24 h and the cells were stim-ulated with 10 ng/ml Epo for another24 h. Luciferase activity was mea-sured as described in Materials andMethods. Error bars represent the SDof a triplicate experiment.

2976 AN UNUSUAL INSERTION IN Jak2 THAT AFFECTS ITS ACTIVITY

by guest on August 25, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

in the activation mechanism of Jaks in the various cytokine recep-tor contexts.

DisclosuresThe authors have no financial conflict of interest.

References1. Huse, M., and J. Kuriyan. 2002. The conformational plasticity of protein kinases.

Cell 109: 275–282.2. Pellicena, P., and J. Kuriyan. 2006. Protein-protein interactions in the allosteric

regulation of protein kinases. Curr. Opin. Struct. Biol. 16: 702–709.3. Yamaoka, K., P. Saharinen, M. Pesu, V. E. Holt, III, O. Silvennoinen, and J. J.

O’Shea. 2004. The Janus kinases (Jaks). Genome Biol. 5: 253.4. Haan, C., S. Kreis, C. Margue, and I. Behrmann. 2006. Jaks and cytokine recep-

tors–an intimate relationship. Biochem. Pharmacol. 72: 1538–1546.5. Levine, R. L., A. Pardanani, A. Tefferi, and D. G. Gilliland. 2007. Role of JAK2

in the pathogenesis and therapy of myeloproliferative disorders. Nat. Rev. Cancer7: 673–683.

6. Delhommeau, F., D. F. Pisani, C. James, N. Casadevall, S. Constantinescu, andW. Vainchenker. 2006. Oncogenic mechanisms in myeloproliferative disorders.Cell. Mol. Life Sci. 63: 2939–2953.

7. Witthuhn, B. A., F. W. Quelle, O. Silvennoinen, T. Yi, B. Tang, O. Miura, andJ. N. Ihle. 1993. JAK2 associates with the erythropoietin receptor and is tyrosinephosphorylated and activated following stimulation with erythropoietin. Cell 74:227–236.

8. Argetsinger, L. S., G. S. Campbell, X. Yang, B. A. Witthuhn, O. Silvennoinen,J. N. Ihle, and C. Carter-Su. 1993. Identification of JAK2 as a growth hormonereceptor-associated tyrosine kinase. Cell 74: 237–244.

9. Watling, D., D. Guschin, M. Muller, O. Silvennoinen, B. A. Witthuhn,F. W. Quelle, N. C. Rogers, C. Schindler, G. R. Stark, J. N. Ihle, and I. M. Kerr.1993. Complementation by the protein tyrosine kinase JAK2 of a mutant cell linedefective in the interferon-� signal transduction pathway. Nature 366: 166–170.

10. Miyakawa, Y., A. Oda, B. J. Druker, T. Kato, H. Miyazaki, M. Handa, andY. Ikeda. 1995. Recombinant thrombopoietin induces rapid protein tyrosinephosphorylation of Janus kinase 2 and Shc in human blood platelets. Blood 86:23–27.

11. Drachman, J. G., J. D. Griffin, and K. Kaushansky. 1995. The c-Mpl ligand(thrombopoietin) stimulates tyrosine phosphorylation of Jak2, Shc, and c-Mpl.J. Biol. Chem. 270: 4979–4982.

12. Tortolani, P. J., J. A. Johnston, C. M. Bacon, D. W. McVicar, A. Shimosaka,D. Linnekin, D. L. Longo, and J. J. O’Shea. 1995. Thrombopoietin induces ty-rosine phosphorylation and activation of the Janus kinase, JAK2. Blood 85:3444–3451.

13. Silvennoinen, O., B. A. Witthuhn, F. W. Quelle, J. L. Cleveland, T. Yi, andJ. N. Ihle. 1993. Structure of the murine Jak2 protein-tyrosine kinase and its rolein interleukin 3 signal transduction. Proc. Natl. Acad. Sci. USA 90: 8429–8433.

14. Lucet, I. S., E. Fantino, M. Styles, R. Bamert, O. Patel, S. E. Broughton,M. Walter, C. J. Burns, H. Treutlein, A. F. Wilks, and J. Rossjohn. 2006. Thestructural basis of Janus kinase 2 inhibition by a potent and specific pan-Januskinase inhibitor. Blood 107: 176–183.

15. Boggon, T. J., Y. Li, P. W. Manley, and M. J. Eck. 2005. Crystal structure of theJak3 kinase domain in complex with a staurosporine analog. Blood 106:996–1002.

16. DeLano, W. L. 2002. The PyMOL Molecular Graphics System. DeLano Scien-tific, San Carlos, CA.

17. Carson, M. 1991. Ribbons 2.0. J. Appl. Crystallogr. 35: 958–961.18. Altschul, S. F., W. Gish, W. Miller, E. W. Myers, and D. J. Lipman. 1990. Basic

local alignment search tool. J. Mol. Biol. 215: 403–410.19. Haan, S., C. Margue, A. Engrand, C. Rolvering, H. Schmitz-Van de Leur,

P. C. Heinrich, I. Behrmann, and C. Haan. 2008. Dual role of the Jak1 FERM andkinase domains in cytokine receptor binding and in stimulation-dependent Jakactivation. J. Immunol. 180: 998–1007.

20. Sommer, U., C. Schmid, R. M. Sobota, U. Lehmann, N. J. Stevenson,J. A. Johnston, F. Schaper, P. C. Heinrich, and S. Haan. 2005. Mechanisms ofSOCS3 phosphorylation upon interleukin-6 stimulation. Contributions of Src-and receptor-tyrosine kinases. J. Biol. Chem. 280: 31478–31488.

21. Behrmann, I., T. Smyczek, P. C. Heinrich, H. Schmitz-Van de Leur, W. Komyod,B. Giese, G. Muller-Newen, S. Haan, and C. Haan. 2004. Janus kinase (Jak)subcellular localization revisited: the exclusive membrane localization of endog-enous Janus kinase 1 by cytokine receptor interaction uncovers the Jak.receptorcomplex to be equivalent to a receptor tyrosine kinase. J. Biol. Chem. 279:35486–35493.

22. Radtke, S., S. Haan, A. Jorissen, H. M. Hermanns, S. Diefenbach, T. Smyczek,H. Schmitz-Vandeleur, P. C. Heinrich, I. Behrmann, and C. Haan. 2005. The Jak1SH2 domain does not fulfill a classical SH2 function in Jak/STAT signaling butplays a structural role for receptor interaction and up-regulation of receptor sur-face expression. J. Biol. Chem. 280: 25760–25768.

23. Haan, C., H. Is’harc, H. M. Hermanns, H. Schmitz-Van De Leur, I. M. Kerr,P. C. Heinrich, J. Grotzinger, and I. Behrmann. 2001. Mapping of a region withinthe N terminus of Jak1 involved in cytokine receptor interaction. J. Biol. Chem.276: 37451–37458.

24. Haan, C., and I. Behrmann. 2007. A cost effective non-commercial ECL-solutionfor Western blot detections yielding strong signals and low background. J. Im-munol. Methods 318: 11–19.

25. Yasukawa, H., H. Misawa, H. Sakamoto, M. Masuhara, A. Sasaki, T. Wakioka,S. Ohtsuka, T. Imaizumi, T. Matsuda, J. N. Ihle, and A. Yoshimura. 1999. TheJAK-binding protein JAB inhibits Janus tyrosine kinase activity through bindingin the activation loop. EMBO J. 18: 1309–1320.

26. Sasaki, A., H. Yasukawa, A. Susuki, S. Kamizono, T. Syoda, I. Kinjyo,M. Sasaki, J. A. Johnston, and A. Yoshimura. 1999. Cytokine-inducible SH2protein-3 (CIS3/SOCS3) inhibits Janus tyrosine kinase by binding through theN-terminal kinase inhibitory region as well as SH2 domain. Genes Cells 4:339–351.

27. Sasaki, A., H. Yasukawa, T. Shouda, T. Kitamura, I. Dikic, and A. Yoshimura.2000. CIS3/SOCS-3 suppresses erythropoietin (EPO) signaling by binding theEPO receptor and JAK2. J. Biol. Chem. 275: 29338–29347.

28. Saharinen, P., K. Takaluoma, and O. Silvennoinen. 2000. Regulation of the Jak2tyrosine kinase by its pseudokinase domain. Mol. Cell. Biol. 20: 3387–3395.

29. Briscoe, J., N. C. Rogers, B. A. Witthuhn, D. Watling, A. G. Harpur, A. F. Wilks,G. R. Stark, J. N. Ihle, and I. M. Kerr. 1996. Kinase-negative mutants of JAK1can sustain interferon-�-inducible gene expression but not an antiviral state.EMBO J. 15: 799–809.

30. Lu, X., R. Levine, W. Tong, G. Wernig, Y. Pikman, S. Zarnegar, D. G. Gilliland,and H. Lodish. 2005. Expression of a homodimeric type I cytokine receptor isrequired for JAK2V617F-mediated transformation. Proc. Natl. Acad. Sci. USA102: 18962–18967.

2977The Journal of Immunology

by guest on August 25, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from