PHYSIOLOGICAL ROLES OF PROTEIN

INVOLVED IN TNF SIGNAL TRANSDUCTION

Kiaus Peter Hoeflich

A thesis submitted in conformity with the requùements for the degree of Doctor of Philosopby Graduate Department of Medical Biophysics Universiiy of Toronto

O Copyright by Kiaus Peter Hoeflich (2001) National Library Bibiiothèque nationale l*b oi Canada du Canada Acquisitions and Acquisitions et Bibliographie Services services bibliographiques 395 Wellington Street 395. rue WeiKngton Onawa ON KIA ON4 OtlawaON K1AûN4 Canada Canada

The author has granted a non- L'auteur a accordé une licence non exclusive licence aiiowing the exclusive permettant à la National Library of Canada to Bibliothèque nationale du Canada de reproduce, loan, distribute or seii reproduire, prêter, distribuer ou copies of this thesis in microfotm, vendre des copies de cette thèse sous paper or elecironic formats. la forme de microfichelfilm,de reproduction sur papier ou sur format électronique.

The author retains ownership of the L'auteur conserve la propriété du copyright in this thesis. Neither the droit d'auteur qui protège cette thèse. thesis nor substantial extracts fiom it Ni la thèse ni des extraits substantiels may be printed or othewise de celle-ci ne doivent être imprimés reproduced without the author's ou autrement reproduits sans son permission. autorisation. PHYSIOLOGICAL ROLES OF PROTEIN KINASES LNVOLVED iN TNF SIGNAL TRANSDUCTION

Klaus Peter Hoeflich

A thesis submitted in conformity with the requirements for the degree of ktorof Phümphy Graduate Department of Medical Biophysics University of Toronto

O Copyright by Kiaus Peter Hoeflich (2001)

Tumor necrosis factor ('ITJF) is a proinflammacory cytokine essentid for the generation of systemic and local responses to infection, injury and irnrnunologicai challenge. Changes in gene expression by TNF are mediated in part by two transcription factors, NF-KB and AP-I. Understanding OC the TNF-regulated upstrem activators of NF-& and AP-1 is incomplete and therefore potentid therapttic targets remain to be discovered. TNF-induced activation of c-iun, a principle component of the AP- I transcription factor complex, requires activation of the stress-activated protein kinases (SAPKs) downstream of TNF receptor-associated Factor-2 (TRAF2). Here 1 show that ASKl interacts with members of the TRAF famiiy via their TRAF-N and RiNG finger domains, leading to the activation of the SAPK cascade. In untransfected cells, endogenous ASKl npidly associates with TRAF2 in a TNF-dependent manner. While a truncated derivative of T€&4Fî, which inhibits SAPK activation by TNF, blocks TNF-induced ASKl activation, the reciprocal dso holds true and a catalytically-inactive mutant of ASKl is a dominant-negative inhibitor of TNF-and TRAM-induced SAPK activation. Furthemore, protection from TNF-induced ce11 death conferred by an ASKl mutant is dependent upon TRAF2, Hence, ASKl is a common mediator of TRAF-regulated SAPK, and the TRAF2-ASKL connection completes the signalling cascade from TNF to SAPK activation. A novel role for the protein GSK-3p in regulation of NF-& was suggested by the finding that disruption of this in rnice leads to hepatic , a phenotype shared by mutations in TNF regulation of NF-K.,AccordingIy, this phenotype cmbe rescued by inhibiting TNF-ct signalling in the embryos, in the absence of GSK-3P, embryonic fibroblasts exhibit substantialIy reduced TNF-mediated NF-KB transactivation and the cells are highly susceptible to TNF-induced apoptosis. The early steps leading to NF-KB activation (1-K.degradation and NF-& nuclear translocation) are unaffected by loss of GSK-3p,indicating that regdation of NF-KB by GSK-3P occurs at the Ievel of the transcriptionai complex. Lithium treatment, which inhibits GSK-3,also sensitizes wild- type fibroblasts to TNF and inhibits NF-KB transactivm-on. Evaiuation of an inhiibitor known to impede NF-KB hnction, debromohymenialdisine, revealed that it has a potent and direct affect on GSK-3p activity. Together, these findings indicate a noveI role for GSK-3b in facibting NF-KB function. To my parents und Si Tuen:

For yortr faith, love and support throzcghout the cozcrse of my sttrdies Acknowledgements

Thank you, Jim, for everything you have done over the years. I'm sure that not al1 of it is even known to me. I am very gratehl for the many opportunities and direction you have provided, and the interest you took in my growth as a scientist.

Sevenl faculty members have also guided and advised me through various stages of my doctord training. In particular, I am very indebted to Amen Manoukian, Sean Egan, Wen-Chen Yeh, MimgSound Tsao and Linda Penn for their efforts and kindness.

A great many that have worked and studied with me at the OC1 have contributed invahably to my leaming experience. The list is long, but each person is important: Adnan Ali, Madeleine Bonnard, Jennifer Cannons, Jim Dimitroulakos, Brad Doble, Jason Goncalves, Hotvard Gutstein, Jin Jing, Friedemann Kiefer, Lisa Kockecitz, Chao Lu, Juan Luo, Christine Mirtsos, Tmdey Nicklee, Mike Parsons, Liz Rubie, Laurent Ruel, Mike Scheid, Vuk Stambolic, Mark Takahashi, Maude Tessier, and Lee Anne Tibbles,

A special thanks as well to John Jackson and the staffof the PMH library for cheerfiilly tolerating my countless requests for articles, Publications

KIaus P. Hoeflich, Adnan Ali, James R. Woodgett (2001). Glycogen Synthase Kinase-3: Properties, Func tions, and Regdation. In: Chemical Reviews: Protein Phosphorylafion and DepliosplioryIation. (N. Ah, Ed). The American Chemical Society, USA.

Klaus P. Hoefiich, James R. Woodgett (2001). -activated protein kinases and stress. In: Cell and Moleculor Respoiises to Stress, Volume 2. &B. Storey, J.M. Storey, Eds), Elsevier Science, Netherlands.

Klaus P. Hoeflich, Juan Luo, Elizabeth A. Rubie, Ming-Sound Tsao, Ou Jin and James R. Woodgett. Requirement for glycogen synthase kinase-3b in ce11 survivai and NF43 activation. Nature. 2000 Jul6; 406(6791): 86-90,

Klaus P. Hoeflich. Wen-Chen Yeh, Zhengbin Yao. Tak W. Mak and James R. Woodgett Medisltion of TNF receptor-associated factor effector functions by apoptosis signal- regulatingknase-1 (ASKL). Oncogene. 1999 Oct 14; 18(42): 58 14-20,

Jennifer L. Cannons, Klaus P. Hoeflich, James R. Woodgett and Tania H. Watts. Role of stress kinase pathway in signalling via the T ce11 costirnulatory receptor 4- 1BB. Journal of lmmrinology. 1999 Sept 15; 163(6): 7990-8.

Si Tuen Lee, Klaus P. Hoeflich, Gihane W. Wasfy, James R. Woodgett, David W. Andrews, Brian Lebec David W. Hedley and Linda 2. Penn. Bci-2 targeted to the endoplasmic reticuiurn cminhibit apoptosis induced by Myc but not etoposide in Rat-1 fibroblasts. Oncoge~te.1999 June 10; 18(23): 3520-8.

Jim Dimitroulakos, Mark J. Benzaquen, Gihane H. Wasfy, Klaus P. Hoeflich, David W. Hedley, lames R, Woodgett, Mark D. Minden and Linda Z. Penn. An analysis of reactive oxygen intermediates, stress-activated protein kinase and bc1-2 in the apoptotic response of acute myeloid leukemic cet1 Iines to lovastatin. Blood. 1998 Nov 15; 92110): Supplement 1,215. Table of Contents

List of tables ...... X List of figures...... xi... List of appendices...... m11 List of abbreviations...... xiv

Chaptcr 1: Introduction

1.1 Thesis objectives...... 7

1.2 Tumour necrosis factor signalling pathway ...... 3

t 2.1 TRAF3 and RIP activate distinct signalling pathways ...... 6 I 2.2 The NF-KB transcription factor...... 7 1.2.3 Biological functions of IKKP, NEMO and T2K ...... 9

1.3 S tress-activated protein kinases ...... 13

1.3 .1 Historical overview of MAPK pathways ...... 15 1.3.2 The SAPK system ...... 16 1.3.3 Dual-speciflcity protein kinases of the SAPK family ...... 33 1.3.4 Regulation of SAPK by E/lAFKKKs and TNF ...... 29 1.3.5 MAPKKK kinases and RIF ...... 31 1.3.6 The p38 MAPK farniIy ...... 35 1.3.7 Genetic analysis of p38a in mice ...... 37 1.3.8 Surnrnary...... 40

41

1.4.1 isolation and characterizauon of GSK-3 ...... 42 1.4.2 Regulation ofGSK-3 activity ...... 44 1.4.3 GSK-3 homologues ...... 49 1-44 Marnrnalian GSK-3 and Wnt signalling ...... 51 1.4.5 GSK-3 phosphoryIation of B-catenin and APC ...... 56 1.4.6 Regulation ofGSK-3 by other signalling pathways ...... 59 1.4.7 Lithium as an inhibitor of GSK-3 ...... 64 1.4.8 Summary ...... 66

Chapter 2: Mediation of TNF' receptor-associated factor effector functions by apoptosis signai-regulating kinase4 (ASK1)

2.1 Abstract ...... 68

2.2 Attributions ...... 68 Introduction...... Materials and rnethods ...... 2.4.1 cDNA constructs ...... 2.4.2 Ce11 cuiture, transfection and treatment...... --**....-. 2.4.3 Immunoprecipitations and immunoblotting ...... 2.4.4 Kinase assay ...... 2.4.5 Embryonic fibroblast death assay ...... Resuits ...... 2.5.1 ASKI and TEUF proteins associate in vivo ...... 2.52 Mapping of in vivo interaction domains between ASKl and TRAF2...... 2.5.3 TNF-dependent endogenous interaction between TRAF2 and ASK1 ...... 2.5.4 W activation of the SAPK pathway is mediated by ASKl 2.5.5 TRAF2 regulates ASKl activation by RIF ...... ,,.. 2.5.6 Overeqression of dominant-negative ASKl inhibits TNF-mediated apoptosis ody in the presence of TM2..... Discussion ......

Chapter 3: Requirement for glycogen synthase kinase-3$ in ceIl survival and NF-KB activation

Abstract......

Attributions...... ,,., ......

Materials and methods ...... 3.3.1 Reagents ...... 3.3.2 Apoptosis assays ...... 3.3.3 Electrophoretic mobility shift assay ...... 3.3.4 Luciferase Assay ...... Results ......

Discussion......

Chapter 4: Further analysis of NI;-KEi regulation by GSK-3$ and its inhibiton

4.1 Abstract...... 121

4.2 Attributions...... 121 Materiais and methods ......

43.1 Reagents ...... 4.3.2 Assay of GSK-3 activity ...... 4.3.3 Ce11 culture......

Results...... 4.4.1 Mechanistic analysis of NF-& regulation by GSK-3P ...... 4.4.2 GSK-3P does not phosphorylate 1-&-a in vitro ...... 4.4.3 Debromohyrnenialdisine...... e...... ~... 4.4.4 ~enerationof GSK-~@-WFRI*-mice ......

Discussion ...... 4.5.1 Involvement of PKB in NF-KB regulation by GSK-30 ...... 4.5.2 Potential role of the SCF ubi uitin iigase complex ...... 4.5.3 Genetic rescue of the GSK-31 hockout phenotype ......

Chapter 5: Conclusions and future directions

5.1 Apoptosis signal-regulating kinase- 1 ......

5.2 Glycogen synthase kinase-3a ......

5.3 Evolutionary Perspective ......

Chrpter 6: References......

Appendir 1: Activation of endogenous stress-activated protein kinase (SAPK) isoforms by osmotic stress and anisomycin

A1.1 Abstract...... 182

A1 -2 introduction...... 185

Al .3 Materials and methods ...... 185

Al .3.1 cDNA consuucts ...... 185 A1 .3.2 Ce11 culture and treatment ...... 185 A1.3.3 Partial purification of SAPK isofoms ...... 185 AU4Assay of SAPK activity...... 186 A1.3.5 Immunoprecipitations and immunoblotting...... 187 A1.4.1 Endogenous levels of p46 and p54 SAPK isofoms ...... A14.2 Examining the kinetics and amplitude of SAPK activity in U937 cells ...... -...... A1 -4.3 Partial. . purification of p46 and p54 SAPK isofonns by liquid chromatography...... A1 -4.4 Differential activation of endogenous SAPKs in response to sorbitol or anisomycin......

A1 -5 Discussion......

A 1.6 References......

Appendiir 2: Assessrnent of postnatal survival of T~vFRI'/'GsK-~~/-mice ...... -. -. -. -. . . -. .. . . -...... List of Tables

Table 1 Mutant phenotypes of SAPK and p3 8 pathway components.. 21

Table 2 Components of mammalian stress-regulated MAPK signalhg pathways...... -...... 28

Table 3 Protein substrates of GSK-3...... 43

Table 4 GSK-3 is highly conserved throughout evolution...... 48 List of figures

Figure 1 Marnmalian MAPK modules ......

Figure 2 Overview of TNFISAPK signalling......

Figure 3 Role of GSK-3 in the Wnt signahg pathway ......

Figure 4 Role of GSK-3 in phosphatidylinositol3-kinase signalling ...

Figure 5 ASKl associates with TRAF2, WSand W6......

Figure 6 Interaction of ASK 1 with TUF2 mutants ......

Figure 7 Physical association of endogenous -2 and ASKI ......

Figure 8 ASKl is downstream of TRAFs ......

Figure 9 Mechanism of ce11 death inhibition by ASKl ......

Figure 10 ASKl does not inhibit the association of TRAF3 with HIAP1 or HIAP2 ......

Figure 11 Targeted disruption of the GSK-3fi locus ......

Figure 12 Pheno type of GSK-3pLembryos ......

Figure 13 GSK-3P-deficient ceIls show increased sensitivity to TNF-induced apoptosis......

Figure 14 Apoptotic sensitivity of GSK-3f fibroblasts ......

Figure 15 Inhibition/inactivation of GSK-3p reduces TNF-inducible . * NF-KB activation ......

Figure 16 Current mode1 of GSK-38.s involvement in NF-KB Activation......

Figure 17 Effect of RJF-a and IGF-I on GSK-3B activity ......

Figure 18 GST-GSK-3fi does not phosphorylate I-KB-a......

Figure 19 Debromohymenialdisine inhibits GSK-3 ...... ~~~~~~.~ Figure 20 GSK-3fi1TNFR1 mouse breedig strategy ...... 134

Figure 21 ~henoty~eof TNFRI%SK-~~- embryos ...... 136

Figure 22 Detection of endogenous SAPK protein and activity in U937 cells ...... 189

Figure 23 Separation of p46 SAPK and p54 SAPK by anion exchange liquid chromatography ...... 192

Figure 24 Post-natal survival of TNFRI"GSK-3pmice ...... 202 List of Appendices

Appendix 1 Activation of endogenous stress-activated protein kinase (SAPK) isoforms by osmotic stress and anisomycin...... , 182

Appendi. 2 Assessment of postnatai nwivd of WFR I-'-GSK-~~Ymice...... 200 List of abbreviations

AP- 1 activator proteia- 1 MC adenomatous polyposis coli protein Arm annadillo ASKl apoptosis signaI-regulated kinase-1 ATF1 activating transcription factor-2 ATP adenosine 5'-triphosphate bp base-pair cm cyclohe'umide CKII casein kinase41 CREB CAMP-response-element-bindingprotein CSAID cytokine-suppressing anti-idammatory dnig DAK dishevelled-associated kinase DBH de bromo hymenialdisine DEP domain dishevelled, egl-10 and pleckstrin domain DIX domain dishevelled and avin domain DLK dual-leucine zipper-bearing kinase DNA deoxynbonucleic acid DshlDvl dishevelled EDTA ethylenediaminetetraacetic acid EF embryonic fibroblast ELAM endotheliai leukocyte adhesion molecuie EMSA electrophoretic mobility shifi assay ERK extracellular signai-regulating kinase ES embryonic stem FUS3 fusion-3 FRAT fiequently rearranged in advanced T-ce11 lymphomas FZ-2 frizzied-2 GADDI53 growth anest and DNA damage-153 GBP GSK-3 binding protein GCK germinal center kinase GCKR germinal center kinase-related GLK germinal center kinase-like GS glycogen synthase GS K-3 glycogen synthase kinase4 GST giutathione S- GTPase guanine 5'-triphosphatase HA hemagglutinin HD hymenialdisine Hep hemipterous [gG immunoglobulin G 1-K. inhibitor of NF-kappaB W( I-kappaB kinase IL interleukin [LK integrin-linked kinase

xiv IMPase inositol monophosphatase n'TG isopropyl- 1-thoi-beta-D-galactoside nP JNK-interacting protein JNK c-Jun NHz-terminal kinase Da kilo-Dalton KSS 1 kinase suppressor of ssf2 Lck lymphoid ce11 kinase LEF-I leukemia enhancer factor-1 LPS lipopolysaccharide LZK leucine zipper-bearing kinase MAPK mitogen-activated protein kinase MAPKK mitogen-activated protein kinase kinase MAPKAPKZ MAPK-activated protein hase-2 MAPKKK mitogen-activated protein kinase kinase kinase MBP rnyelin basic protein MEF2 myocyte enhancer factor-2 MEK MAPKERK kinase MEKK ME WERK kinase MKK7 MAPK kinase-7 ML'K mixed-lineage kinase mM milii molar MP-1 MEK partner-l NFAT4 nuclear factor of activated T cells NF-& nuclear factor-kappa oct-l octamer-1 PAGE polyacrylamide gel eiectrophoresis PARP poly (ADP-ribose) polymerase PBS phosphate-bufTered saline PDK PU'K-dependent protein kinase PD2 domain PSD-95, discs-large and 20-1 domain P13'K phosphatidylinositol-3-OH kinase PKB PKC PP2A protein phosphatase-2A PRC protein phosphatase-2C RGS regdators of G-protein signahg w ultraviolet MG recombination-activating gene RNA riionucleic acid SM serine, alanine, methionine and proline repeats SAPK stress-activated protein kinase SDS sodium dodecyl sulfate SEKl SAPWERK kinase-1 sgg shagw Shb supemumerary iimbs STE sterile TAKl transforming growth factor-beta-activatedprotein kinase TCF teniary complex factor Tcf T ceii factor TNF tumor necrosis factor TNFR tumor necrosis factor receptor TPL-2 tumor progression locus-2 WD TNFR-associated death domain protein TRAF TNFR-associated factor Tris tns(hydroqmethy~)aminomethane TUNEL terminal deoxynucleotidyl transferase-rnediated dUTP nick end labelling wingless wild-type zeste-white3 Chrpter 1

Introduction 1.1 Thesis objectives

When immune cells encounter noxious stimuli they produce various infiammatory cytokines, including turnour necrosis factor (TNF). Interestingly, while TNF can potently prevent disease (ie: by exhibiting anti-cancer activity), it is also a key pathogenic rnediator of many infectious and infiammatory diseases. For instance, rnisregulation of

TNF contributes to sepsis, meningitis, Lupus, rheumatoid arthntis, Crohn's disease, multiple sclerosis, organ transplant rejection, rnyocardial infarction, and AIDS. In 1998, approxirnately 28 million individuals in the United States suffered iÎom these diseases.

Although information on Tm: and its activities is rapidly accumulating, many important questions remain to be addressed. Firstly, what switches RIF between its helpfÙ1 and deleterious effect on cells? By what mechanisms does TNF function in these interrelated diseases? The answer to these questions lies in understanding the biochemical changes induced by TNF, which is the focus of the research presented here.

Exposure to TNF results in activation of several transcription factors, including

NF-KB and c-Jun (part of the AP-1 transcription factor complex). It is known that TNF- mediated activation of the 1-KB kinases results in NF-KB-regulated gene transcription, while the stress-activated MAP kinases, SAPK and p38 MAPK, contribute to induction of AP-1 activity. However, the pathways by which TNF-receptor ligation causes activation of these transcription factors have not been fuiiy elucidated.

In this thesis, I describe efforts to understand the signaiiing mechanisms that transduce TNF signals iato the nucleus to impact gene expression. Specificaily: a Chapter 2 assesses the biochemical mechanism of apoptosis signal-regulating kinase-

1 (ASK1) activation, a TNF-regulated kinase upsûeam of SAPKs and p38 MAPKs.

Chapters 3 and 4 describe a novel role for glycogen synthase kinase43 (GSK-3B) in

NF-KB-mediated transcription and ce11 survival.

a Appendix 1 describes experiments performed to investigate agonist-induced

activation of specific SAPK isoforms.

The remainder of this chapter introduces the three main players, namely the biology of

RIF. SAPK and GSK-3.

13 Tumour necrosis factor signalling pathway

TNF is involved in a variety of biological activities through its binding to two distinct ceU surface receptors, p55 TNFRl and p75 TNFR2, that are present on most ce11 types

(Baker & Reddy, 1998). TNF binding results in receptor tnmerization and subsequent signailing of T'NF's many activities. A variety of studies, including gene targeting in mice, have demonstrated that TNFRl is responsible for eliciting the majority of these activities (Tartaglia and Goeddel, 1992). in contrast, TNFR2 signais directly in some

Iymphoid celi types and has an au..iary role in modulating TNFRl signalling. Two of TNF's most extensively studied properties, namely NF-KB and SAPK activation, are triggered by TNFR1.

Both TNF receptors are part of a receptor superfamily consisting of more than 20 strucnually related type 1 transmembrane proteins that can be divided into two subgroups, depending on whether theu intracellular region contains an 80 amino acid motif, termed the '*deah domain". The most intensively studied death domain-containing receptors are

TNFRI and Fas: while TNFRl only induces ce11 death under certain circumstances and more often induces transcriptional gene activation, Fas is very efficient in ceil death induction. TNF receptor family members that do not contain death domains are represented by TNFR2. CD40, CD30, CD27, among others, and are involved primarily in gene transcription for ce11 survival, growth, and daerentiation. Neither TNFR possesses intrinsic enzymatic activity and upon ligand binding and receptor aggregation these receptors trigger downstream signalling pathways by recniiting receptor-associated effector molecules. For instance, the 34 kDa protein TNFR-associated death domain protein (TRADD), one of the first identified TNFRl adapter molecules, is recruited to

RSFRl in a RIF-dependent manner (Hsu et al., 1995) and interacts with another adaptor,

TRAFî,(Rothe el al., 1994).

W2is one of six known marnmalian members of the TNE: receptor-associated

factor (TRAF) family, each of which consists of carboxy-terminal TRAF domains,

central zinc fmger repeats, and with the exception of TRAF 1, an amino-terminal RING

finger dornain (Baker & Reddy, 1998; Cao et al., 1996). The carboxy-terminal-170

amino acids of the TRAF domain, the TRAF-C region, is required for receptor binding

and contributes to TRAF oligomerization (Rothe et al., 1994; Cheng et al., 1995). The less conserved TRAF-N domain, preceding TRAF-C, forms a coiled coi[ that mediates oligomerization. Finally, the RING frnger is critical for TRAF2 signaihg to downstrearn effectors. Several proteins, including NF-KB-inducing kinase (NIK)and the cellular inhibitors of apoptosis (c-IAPs), have been reported to interact with al1 or parts of the

TRAF domain of TRAF3 (Arch et al., 1998). While direct interactions of many of these proteins with TRAFs remain to be demonstrated, the surface of the TRAF domain contains many topographical features that could mediate protein-protein interactions

(Park et al., 1999).

TRAFL and TM2were isolated from the TNFR2 signalling complex by biochemical purification and the yeast hvo-hybrid system (Rothe et al., 1994). TRAF3 and TRAFS were identified by virtue of their interactions with several members of the

RIF receptor farnily (Hu et al., 1994; Mosialos et al., 1995). The overexpression of

TRAF4 in breast carcinoma cells resulted in its isolation and cloning (Regnier et al.,

1995). TRAF6 was identified independently as a cornponent of the LIand CD40 signalling pathways (Cao et al., 1996; Ishida et al., 1996). TRAFs are also genetically conserved across other multicellular organisms including Drosophila melamgaster,

Caenorhabdiris elegans, and Dictyostelium discoideum.

In most cases. multiple TRAFs can bind the same TNF farnily receptor, and many ceIls types are known to express different subsets of TRAFs (Arch et al., 1998). Hence, the particular signal that is generated by receptor-mediated TRAF recruitment is dependent on the affinity of different TRAFs for the same receptor and the ce11 type in which the signalling occurs. As the TNF farnily of cytokine ligands are trimeric in nature, the active signalhg complex for TNFRs and TRAF2 are fonned by trimerization. Presumably though, multirneric TRAF complexes cm be assembled even in the absence of receptor ligation, as experiments overexpressing individual TRAFs demonstrate that

TRAFs can directly promote downstream signalling and specific cellular responses (Liu et al., 1996; Takeuchi et al., 1996). In this mamer* transient overexpression of TRAF2, -

5, and -6 cm activate SAPK and transcription factors in the AP-1 and NF-KB families

(Song et al., 1997).

1.2.1 TRAFZ and RLP activate distinct sigualling pat hways

RIP was originally discovered in a yeast two-hybild screen using the cytoplasmic domain of the TNF family receptor, Fas, as bait (Stanger et al., 1995). However, a more precise biochemical analysis reveaIed NP to be recruited in an activation-dependent manner to

TNFRl by virtue of its interaction with TRADD. RIP contains an amino-terminal kinase domain and a carboxy-terminal death domain. Overexpression of RIP deletion mutants revealed the death domain to be pro-apoptotic whereas the region lying between the kinase and death domains, termed the intermediate domain, mediates NF-KB activation

(Hsu et al., 1996a). The carboxy region of NP,just preceding the death domain, was sufficient for triggering potent SAPK activation (Liu et al., 1996). Surprisingly, kinase activity was not required for these fhctions, and although RIP has fonnally been show to be an autophosphorylating serine/iheonine kinase, the bction of this kinase activity remains an enigma.

ResuIts fiom the aforementioned transfection experiments suggest that both

TRAF2 and RiP are involved in TNF-mediated pathways Ieading to activation of both

SAPK and NF-KB. However, data generated in gene targethg experiments have led to a clearer definition of the roles of each of these proteins in TNFRl signalling. Embryonic fibrobiasts derived from either RIP or TRAF2 nuii mice display increased sensitivity to

TNF-induced ce11 death, consistent with both these proteins mediating ad-apoptotic effects (Yeti et al., 1997; Kelliher el al., 1998). Also, TRAFZdeficient ceils have oniy a minor reduction in TNF-induced NF-KB activation but are incapable of activating SAPK

in response to TNF-Conversely, RiP is required for TNF-rnediated activation of NF-&

but not SAPK. RP-deficient rnice die perinatally with evidence of extensive apoptosis in

the Lymphoid and adipose tissue, consistent with an anti-apoptotic role for W.TR,~F?'-

mice also have an unexpectantly severe phenotype, and mutant pups become

progressively runted, are devoid of fat deposirs, and die prernaturely. Taken together,

these data support a mode1 in which TNF-cl activates distinct signdling pathways that

lead to SAPK and NF-@ activation.

1.2.2 The NF-KB transcription factor

The Re1 fmily of proceins represenîs a large group of tram-acting transcription factors

that have been irnplicaeed in development, differentiation and oncogenesis. The founding

mernber of the family v-rel is the oncogene of reticuloendotheliosis virus strain T, which

causes rapid and fata1 leukemia in juvenile birds (for review, see Perkins, 2000). Cellular

counterparts of v-rel have been cloned in numerous other organisms. The most highly

characterized rnember of this family is NF-KB which plays an important role as a

reguiator of the immune response and was f~stdiscovered as a constitutively nuclear

transcription factor in mature B cells that bound to an element in the kappa

immunoglobulin light-chah enhancer. NF-KB is, in fact, a group of binary compIexes of proteins with related promoter-binding and transactivation activities. Al1 members of this farnily contain a 300 arnino acid N-terminai DNA-binding and dimerization domain, known as the Rel-homology domain, and most combinations of NF-KB homo- and heterodimers can be found in vivo. The prototypicd NF-KB complex consists of a p6.5- p50 heterodimer. p65/RelA, RelB and c-Re1 stimulate transcription, whereas ~SO/NF-KBI and p52/NF-KB~serve primariIy to bind DNA. Involvernent of NF-KB in the immune system was confirmed by the discovery that treatment with infiammatory cytokines, such as RIF-a and interleukin- l (IL- l), facilitated its reiease fiom cytoplasmic 1-KB inhibitor proteins which resulted in translocation to the nucleus, binding of DNA, and induction of gene expression. In addition to the regulation of NF-KB activity at the level of subceIlular localization, new information has ernerged regarding the role of NF-KB phosphorylation in transactivation of the p6YRelA subunit and proteolytic processing of p10.5, the precursor for the p50 subunit.

An important physiological functiun for NF-KB was revealed by studies that interfered with NF-KB induced gene expression (for instance, using actinornycin D or cyclohexïmide), Treatment of cells with TNF-a in the presence of such inhibitors potently induces ce11 death (Barkett & Gilmore, 1999). This effect was traced to an anti- apoptotic funçtion for NF-KB. Thus, under normai circumstances, RJF-a induces both pro- and anti-apoptotic pathways. Inhibition of the latter switches the baIance towards ce11 death. The consequences of suppressing this protective role is also strikingly manifested in the developing moue embryo.

Mouse embryos lacking the p65 subunit of NF-KB die during day 13-15 of gestation due to massive apoptosis of their developing hepatocytes (Beg et al., 1995). Embryonic fibrobht cells from these mice have defects in NF-KB activation and heightened sensitivity to TNF-a-mediated apoptosis (Beg & Baltimore, 1996). p65- deficient mice that also lack TNF-GLor TNFRl have been generated to examine whether

TNF-dTNFR1-mediated signalhg is responsible for the Lethai phenotype of p65 knockouts (Doi et al.! 1999; Rosenfeld et al., 2000). These double-knockout mice survive ernbryonic development and gestation is apparently unaffected because newbom mice have normal liver morphology with Little detectable apoptosis. Thus, p6~4-mice die fiom apoptotic signals mediated thraugh WU1during liver development. p65 and p50, the components necessary for NF-KBactivation, are present during fetal liver development in wild-type mice from at Ieast E17, suggesting that the timing of NF-KB activation is not related to the timing of embryonic lethality in p63" mice. Survival of the p65RNFRl double knockouts also suggests that neither p65 nor TNFRl is criticai for hepatic development in mice. However, newbom animais with these deficiencies become sensitive to Section and neutrophilic invasion, leading to damage to the liver and other organs and early postnatal mortality .

1.2.3 Biological functioos of [KKB, NEMO and T2K

NF-KB activation eatails the phosphorylation of the ihiitory I-KB proteins on two specific serine residues. These phosphorytation events trigger I-icB degradation through the ubiquitin/proteosome pathway, consequentiy releasing NF-KB and ailowing its translocation into the nucleus (ManÏatis, 1999). Recentiy, an I-KB kinase activity was identified as a large cornpiex, whkh con& two cataiytic subunits, MKa and IKKP

(Regnier et al. 1997; DiDonato et al- 1997). These kinases are 52% identical and share several structural domains, including an amino-terminai kinase domain, a helix-loop-

helix (HLH) domain and a leucine zipper (LZ) domain. IKKa and IKKP form

heterodimers in vivo, and phosphorylation assays have shown that both kinases can

phosphorylate 1-KB-u on Ser-32 and Ser-36 (Woronicz et al. 1997). The IKKs exist in a

stable complex with a third molecule designated NEMO (NF-KB essential modulator),

which has an important role in cytokine-induced NF-@ activation (Yamaoka et al.,

1998).

Genetic analysis of IKK function through gene targeting in mice have revealed

that despite the high degree of sequence similarity and identicai substrate specificity, the

two catalytic subunits of the tKK complex differ drarnatically in biological function.

While IKKu is required for proper development and differentiation of the epidermis and

other ectodermal derivatives, it is not required for NF-KB induction by proidammatory

stimuli (Hu et al., 1999). In contrat, IKK/~"and NE~~CI'O'embryos display a phenotype

similar to those deficient in the NF-KB subunit p65 (Li et al., l999b; Rudolph et al.,

2000; Beg et al., 19951, and die at rnidgestation due to massive hepatocyte apoptosis.

Consistent with earlier studies indicating that NF-KB protects ceIls against apoptotic

signais, fibroblasts lacking lKKP and NEMO are more sensitive to TNF-mediated ce11

death and cannot activate NF-KB in response to cytokines (Li er al., 1999b; Rudolph er

al., 2000). Endogenous TNF signailing is suggested to be responsible for the observed

ber degeneration, as IKKB/TNFRl double-knockout mice survive until after birth (Li et

al., 1999b).

Charactenzation of the phy siologicai function of a novel TRAFZassociated

kinase, T2K (aiso known as TBKl or NAK) demonstrated that T2K, Iike p65, MKf3 and NEMO,is criticai in protecting embryonic liver fiom apoptosis (Bo~ardet al., 2000). In

response to either RIF-a or IL-1 induction, TzK'*MEFs exhibit normal degradation of I-

KB and NF-KB DNA-binding activity. However, NF-KB-directed transcription is

dramatically reduced. These results demonstrate that, T2K has a unique role in the

activation of NF-KB-directed transcription, apparently independent of 1-KB degradation

and NF-KB DNA binding. Several recent studies have demonstrated that 1-KB

phosphorylation is not the sole means of regulating NF-KB activity and that signal-

induced phosphorylation of the p65 subunit is critical for the induction of NF-KB-

dependent transcription (Zhong et al., 1998; Wang et al., 2000; Sizemore et al., 1999).

WhiIe this inducible phosphorylation of p6j increases the transcriptionai activity of NF-

KB, it does not appear to affect its nuclear translocation or DNA binding activity (Wang

and Baldwin, 1998). Taken together, it is therefore possible that T2K may regulate the

transactivation potential of NF43 through phosphorylation of NF-KB proteins, either

directly or indirectly (Bonnard et al., 2000).

From these gene-targeting studies, it is apparent that mutation of some signalling

components (for instance KKP, NEMO and T2K, but not RIP) that are essenual for

cytokine-mediated NF-KB activation exhibit the tell-tale phenotype of profound liver

degeneration and apoptosis during embryonic deveIopment. Hence, the analysis of

kaockout phenotypes can be used to confiwhich pathway cornponents have a critical

role in the activation of NF-KB by proinfiammatory cytokines and thereby can greatiy

enhance our understanding of the moIecular mechanisms of TNF signal transduction

upstream of NF-KB. 1.3 Stress-activated protein kinases

Full-Iength sequences of more than a hundred MAPKs fiom numerous species have since been reported. Severai MAPK modules have been identified in mammais including the emcellular signai-reguiating kinase (ERK), stress-activated protein kinase (SAPK; or c-

Jun NHz-terminal kinase, JNK), and the p38 group of kinases (Figure l). A number of extensive reviews on MAPK signai transduction have recently been published (Tibbles &

Woodgett, 1999; Widmann et al., 1999). Hence, the goai of this commentary is not to

provide another comprehensive review of the literature, but rather to focus on recent

deveIopments pertaining to TNF-mediated activation of stress-activated MAPKs, their

physiological fùnctions, and to present sorne perspectives. Cellular stresses, Growth factors cytokines

Transcription MEF2 c-Myc c-Jun ATF2 MEF2 factors Sapla p62 TCF NFAT4 CHOP STATs P53 CREB Figure 1. Mammalian MAPK modules

The MAPK module comprises a MAPKKK, MAPKK and a WK.These pathways respond to extracellular signals, including growth and differentiation factors, cellular stress and cytokines. Once activated, MAPKs can phosphorylate a wide variety of proteins, including transcription factors and other kinases. See text for full details. 1.3.1 Historical overview of MAPK pathways

Exposure of cells to either environmentai stress or strong deviations fiom normal conditions initiates complex cascades of stress-inducible transductory that

impact processes such as gene transcription as an adaptive response for cells. The mitogen-activated protein kinase (MAPK) superfamily plays an important role in

transducing signals from the ce11 surface to the nucleus, effecting both the cell's ability to

cope with outside changes as well as cellular coordination in the case of muiticellular

organisms, The term MAPK is most widely used as a general denominator of this farnily

of protein kinases. The MAPK acronym originally described the "microtubule-associated

protein-2 kinase" but evolved into mitogen-activated protein kinase when it was

discovered that the enzyme was induced by a variety of hormones and . Upon

rnolecular cloning of these enzymes, it was realized that they existed in severai classes

that were structurally related but distinctiy regulated. The term MAPK is now commonly

used to denote the entire class of protein-serine kinases that share the following features:

the core functional unit of a MAPK moduie consists of a triad of three kinases that act

sequentiaily; MAPKs are activated via phosphorylatioo on both a threonine and tyrosine

residue by selective upstrearn regulatory kinases; MAPK kinases (MAPKKs or

MAP2Ks). MAPKKs are, in turn, phosphorylated and activated by a group of stnicturaily

related kinases termed MAPK kinase kinases (MAPKKKs or MAP3Ks).

The first MAPK to be cloned was isoIated in a genetic screen of the budding yeast

Saccharomyces cerevisiae (Courchesne et al., 1989; Elion et al., 1990). in yeast ceiis,

mating-specific processes are initiated by the binding of mathg type-specific peptides,

known as cc factor and a factor, to a G protein-coupled pheromone receptor on the ce11 surface (Herskowitz, 1995; Madhani & Fink, 1998). Subsequent signai transduction culminates in a set of physiological responses that prepare celIs for mating, such as arrest

of the cd1 cycle, and aitered ce11 polarity and morphology- Genetic screens identified a

group of "sterile" mutants defective in rnating which were initially grouped into two

categories: deficient in pheromone response or in pheromone production. Additional

components of the pheromone response pathway were identified by a variety of other

approaches, including screens for genetic interactions with the original sterile aiieles. The

approaches that revealed the first MAPKs employed screens for suppressors of

supersensitivity to mating pheromone-induced growth arrest (yielding KSSl, kinase

suppressor oEsst2: Courchesne et al., 1989) and mutations which prevented yeast from

proceeding through mating-induced ce11 fusion (yielding FUS:, fision-3; Elion et al.,

1990). Furtfier genetic anaiysis identified components for five distinct WKpathways

in S. cerevisiae (Herskowitz, 1993). These WKpathways are essential for processes

inchding mating, sporulation, osmoregulation, ce11 wall integrity, starvation and

filamentous growth (Madhani & Fink, 1998; Schaeffer & Weber, 1999).

1.3.2 The SAPK system

Mamrnalian MAPKs have been classified on the basis of two criteria: sequence

homology and differential activation by agonists (Tibbks & Woodgett, 1999; Widmann

et al., 1999). Firstly. the activity of MAPKs is controlled by duai phosphoryiation within

an amino acid sequence known as the activation loop (Canagarajah et al., 1997).

Phosphorylation of the signature motif threonine-X-tyrosine in this L~op(where X is

glutamic acid, proline or glycine for the ERK, SAPK and p38 MAPKs, respectively) is catalyzed by specific MAPKKç and results in a conformational change and a >1000-fold increase in specific activity of the MAPK. In essence, the enzymes are inactive until phosphorylated by theu upstream enzymes. While the ERK class of MAPKs is primarily activated by growth factors and rnitogens, SAPKs and p38 WKsare preferentially induced by a variety of stress signals (Kyriakis er al., 1994). These stimuli include genotoxic agents (irradiation and carcinogens), pathogenic signals (LPS and dsRNA), proinfl ammatory cytokines (turnor necrosis factor (TNF)-a and interleukin (IL)-lp), homeostatic perturbations (in temperature, osmolarity and pH), oxygen tension, intracellular calcium, and other chemical insults (e.g. exposure to arsenite or anisomycin).

As expected, a major point of regulation occurs at the level of the MAPK. Since phosphorylation of both threonine and tyrosine residues is required for MAPK activity, dephosphorylation of either is sufticient for inactivation. This can be achieved through complex regulation by tyrosine-specific phosphatases, serinelthreonine-specific phosphatases or by dual specificity (threonineltyrosine) protein phosphatases (reviewed in Keyse, 2000). It is clear, however, that the duration and magnitude of MAPK activation reflects a balance between the activities of the upstream activating kinases and protein phosphatases.

Three SAPK genes (termed a, $ and y; or JNK2, JNIU, and Ml,respectively) have been cloned (Adler et al., 1992; Derijard er ni., 1994; Kyriakis et al., 1994). Overaii, the family members share 8592% identity and are 42-45% identical within the catalytic domain to the ERK family. The SAPK genes are fierdiversified by alternative mRNA splicing into as many as ten isoforms. Each gene generates 54 kDa and 46 kDa polypeptides, the latter variants arising through the introduction of a 5 bp sequence into the carboxy-terminal region which introduces a premature stop codon. To date, clear functional differences between those 46 kDa and 55 kDa isoforms have not been reported. SAPKu and SAPQ are widely expressed, while SAPKP is selectively expressed in the brain, heart and testis.

The SAPKs were originally identified as the major serinelthreonine kinases

responsible for the phosphoryIation of the c-Jun transcription factor (Gupta et al., 1996;

Aibi et al.. 1993; Kyriakis et al., 1994). c-Jwi dimerizes with members of the Fos, Jun or

activating transcription factor (ATF) family of transcription factors to fom the activator

protein-1 (AP-1) transcription factor cornplex. AP-1 activity is induced by a number of

stressful stimuli and part of this activation cm be attributed to phosphorylation of serines

63 and 73 in the c-Jun transactivation dornain, catalyzed by the SAPKs. Through these

e ffects on AP- 1, SAPKs influence ce Il proliferation and oncogenic transformation.

Additional SAPK targets include: other Jun proteins (JunB and JunD; Kallunki et al.,

1996) and the related activating transcription factor-2 (ATF2; Gupta et al.) 1995); the

ternary complex factor (TCF) subfarnily of ETS-domain transcription factors (Whitmarsh

et al., 1995); tumor suppressor p53 (Fuchs ef al., 1998); Smad3 (Engel et al., 1999);

nuclear factor of activated T cells (NFAT4; Chow et al., 1997); and the basic-helix-loop-

helix transcription factor, Myc (Noguchi et al., 1999)- Whiie some of the aforementioned

SAPK targets still await physiologicd vaiidation, to date, SAPK targets are exclusively

transcription factors. This is in contrast to the ERK and p38 families that phosphorylate

substrates outside of the nucleus as weil as within. SAPK isoforms have varying substrate

affinities and may therefore selectively target transcription factors for distinct biological functions in vivo (Gupta et al., 1996). To address this question, mutant mice lacking each member of the SAPK family have been generated and their role in embryonic development assessed (Table 1).

Although mutant mice with a single deletion of SAPKafJNK2 (Yang et al.,

1998), SAPKfifJMü (Yang et al., 1997b) or SAPKyfJNKl (Dong et al., 1998) are viable without overt structural abnormalities, compound deficiencies of the SAPK tàmily have developmental consequences (Kuan et al., 1999; Sabapathy er al., 1999b). Mice with

SAPKdNK2 and SAPKyiJNKl dual deficiencies, but not other SAPK double mutations, exhibit aberrant brain apoptosis and early embryonic lethality. The most conspicuous feature of E10.5 SAPKdSAPKy (JNKI/JNK2) double mutants is failed closure of the neural folds in the hindbrain region. These embryos display decreased apoptosis in the hindbrain at E9.25 and increased apoptosis in both the hindbrain and

Forebrain regions at E10.5. Loss of three out of four SAPK alleles also affects embryonic development as 25% of SAPK~"-SMK~'{~K~~'JNK~"-)fetuses exhibited exencephaly similar to the double mutant phenotype- By conaast, mice that completely lack SAPKdJNK2 and have oniy one SrlPKy/JlVKl aiieie were obtained in Mendelian ratio and did not display developmental abnormalities. The reason for this difference is not known, although detection of SAPK proteins in brain extracts indicates that gene dosage plays a cntical role in controiling SAPK protein levels. It is possible that a certain threshold of SAPK expression is essential whereby both SAPKcdJNKl and

SAPKyfJNK2 either phosphorylate a cornmon target, or act in parailel to phosphoryIate multiple targets, to induce ceIL death during neural tube c1osu.e. These results reveai a functional diversification of the SAPK family in vivo and a role for SAPKs in mammalian morphogenesis, analogous to the requirement of the SAPK signalling pathway in Drosophila embryos during dorsaI closure (Riesgo-Escovar et al., 1996; Sluss et al., 1996), a morphogenetic process that occurs during mid-embryogenesis.

The preferential expression of SAPKPIJNKI in neural tissue suggests a unique fwiction. Indeed, gene-targeting studies demonstrate that SAPKB/JNIU deficiency, but not SAPKc(IJNK3 or SAPKyLJNKl nuIl mutations, results in increased resistance to kainic acid-induced seizures and apoptosis of hippocampal neurons (Yang et al., 1997b).

Kainate elicits epileptic seinires by direct stimulation of the AMPAkainate class of glutamate receptors and indirectly by increasing the release of excitatory amino acids from nerve terminais. Administration of kainate to wild-type mice induces severe seinires that lasted for 1-2 hours, whereas SAPKF'- (JNK~-)mice show milder symptoms and recover faster, These results are phenocopied by mice with a "knock-in" c-

Jun mutation that eliminates the SAPK phosphorylation sites (Behrens et al., 1999).

While c-Jun appears to be the essential substrate for SAPKfVJNK3 in stress-induced neuronal apoptosis, the mechanism by which these molecules function in excitotoxicity remains to be defined. Together with the observation that SAPK-deficiency causes defects in thymocyte apoptosis (Rincon et al., 1998; Sabapathy et al., 1999a), the data obtained with the SHKB/' (JNK~~)and SAPK~'-s/~PK~-* KI '-JNKT'~) mutant mice provides compelling evidence for roles of SAPKs in apoptotic responses. It wouid therefore be of great interest to use these systems to identify physiologically relevant targets of the SAPK apoptotic pathway. stress-activated protein kinase (INK2/3/1,respectively) p38 MAPK, p38/HOG1, MPK2, Uu2, CSBPIR

MAPWERK kinase 3 MAPWERK kinase 6 SAPWERK kinase 1 (MKK3, -1) MAPKERK kinase 7 (SEK2, JNKK2)

LIWPKKKS ASKlI7, apoptosis signal-regulatingkinase (ASKI = MAPKKKS) DLK dual leucine-zipper bearing kinase (MUK, ZPK) MEKKl-4 MAPKERK kinase kinase (MEKK3 = MTKI) MLK2 rniued-lineage kinase (ML= = MST; MLK3 = SPRK) P AK p21 -activated kinase TAK 1 TGF-activated protein kinase TpE turnor progression locus 2 (Cot)

STE20s GCK germinal center kinase GCKR GCK-related GLK GCK-like kinase HGK HPWGCK-like kinase HPKI hernatopoietic progenitor kinase 1 MST 1 rnamrnalian Ste2O-like protein kinase NESK NIK-like ernbryo specific kinase NIK Nck-interacting kinase TAO II? one thousand and one amino acid protein kiaase 1

Scaffold proteins CE31 Met-Brain 1 IrP 1 JNK-interacting protein 1

Table 2. Components of mammalian stress-regulated MAPK signaling pathways. Recent insight into the mechanism of SAPK in apoptosis has been provided by

Tournier et al. (2000). Murine embryonic fibroblasts (MEFs) were derived from

SAPK(X+SAPK~'-(JNKI+~JNKZ~J embryos and, given that the neuronal-specXc SAPKB

(JNK3) isofonn cannot be detected, these MEFs lack a functional SAPK and represent a

usehl mode1 for studying the SAEX signal transduction pathway. To Merdefme the

requirement for SAPK in apoptosis, wild-type and "SAPK-null" MEFs were exposed to a

variety of ce11 killing agents. SAPK-deficient MEFs were alrnost completely protected

frorn apoptosis induced by UV irradiation, methyt methanesulfonate and anisomycin,

while normal apoptosis was observed by activation of the Fas death-signalling pathway.

Increased survival signailing by the transcription factor NF43 and the protein kinase

PKBIAkt did not account for the resistance of SAPK-nul1 MEFs to UV-induced

apoptosis. SAPK-nul1 MEFs express slightly more p53 than their wild-type courtterparts

but the potential contribution of p53 to the LN resistance of SAPK-nul1 MEFs is dificult

to reconcile, mechanistically. lmponantly, SAPK is not required for the death receptor-

signalling pathway mediated by caspase-8, but is essential for stress-induced apoptosis

utilizing the Apafl, initiator caspase-9, and effector caspase-3 genetic pathway.

Accordingly, mitochondriaI membrane permeability and subsequent cytochrome c release

is also blocked in SAPK-nul1 cells in response to W but not in response to Fas.

Clearly, the molecular mechanism by which SAPKs function in apoptotic signai

transduction and mitochondrial depolarkation is an important, as yet unresolved,

question. An important clue cornes fiom the observation that inhibitors of protein and

inRNA synthesis (cycloheximide and actinomycin D, respectively) do not inhibit W-

induced apoptosis (Tournier et al., 2000). This implies that SAPKs can promote stress- induced killing in a trmscriptiodtranslation-indepeadent mechanism. This could occur by affecting mtmbers of the Bcl-2 farnily of apoptotic regdatory proteins, for exampk. It is possible to reconcile these findings with the data obtained uing the SAPK~~SAPK~~

(JNKI~~K~"-)in vivo apoptosis mode1 in which these SAPK isofoms temporally rnediate both ce11 survival and apoptosis durhg brain deveIopment. Possible scenarios include both separate and çooperative mechanisms where the cellular outcome is detemined by the varying kinetics, isoform selectivity, feedback loops, and autocrine secretions promoted by transcription-dependent and transcription-independent SAPK signalhg events. In this way, a baIancing act between SAPK survival and apoptotic signalhg analogous to that described for TNF-mediated NF-KB tramactivation and caspase processing may explain the duality of effects (Baker & Reddy, 1998).

1.3.3 Dual-specificity protein kinases of the SAPK pathway

Two MAPKKs have been identified as upstream activators of the SAPKs, SEKl

(SAPKIERK kinase-l; ais0 known as kIKK4 or JNKK1; Table 2; Derijard et al., 1995;

Lin er al., 1995; Sanchez et al., 1994) and MKK7 (MAPK kinase-7; also known as SEK2 or JNKK2; Moriguchi et al,, 1997; Tournier et al., 1997; Yao et al., 1997). While the existence of SAPK activators distinct fiom SEKl was suggested by earIy studies using chrornatographically hctionated ceil emacts (Moriguchi et al., 1995) and SEKL- dehient murine ceil lines (Nishina et al., 1997b; Yang et al., 1997a), until recently ody

SEKl had been molecdarly cloned. Thus, information regarding the biological fiuictions of MKK7 is just beginning to emerge. The MKK7 gene consists of 14 exons and alternative splicing leads to the inclusion or excIusion of exons located in the 5' and 3' regions of the gene, resulting in the expression of six MKK7 isoforms that differ in their amino and carboxy termini (Tournier et al., 1999). Cornparison of the activities of the

MKK7 isoforms demonstrates that the MKK7a isofom exhibits lower activity, but a higher level of inducible fold activation, than the corresponding MKK7P and MKK7y isoforms, in response to different upstrearn components of the SAPK signailhg pathway

(Tournier et al., 1999). Although the mouse SEKl and MKK7 genes reside on the sarne chromosome (Tournier et al., 1999; White et al., 1996), it does not appear that this

Linkage is evolutionarily conserved. For instance, in Drosophila the hep (MKK7 homologue; Glise et al., 1995) and D-MKK4 (SEK1 homologue; Han et al., 1998) genes are located on different chromosomes.

The physiological role of SEKl has been extensively studied in mice and SEKI% embryos display defective liver organization and massive hepatocyte apoptosis

(Ganiatsas et al., 1998; Nishina et al., 1999; Yang et al., 1997a). These embryos die between El 1.5-12.5, later in development than the SAPKorlSAPKy double knockout

(Kuan et al., 1999; Sabapathy et al., 1999b). This phenotype can be partially understood by considering the tissue expression of SEKl and MKK7. Although the genes are fairly ubiquitously expressed in mice, SEKl expression is highest (and MKK7 is the lowest) in the liver (Nishina et al., 1999; Yao et al., 1997). Thus, while MKK7 cm perhaps compensate for some SEK1-related functions in embryogenesis there is insufficient

MKK7 present in hepatocytes to rescue SAPK signahg in the iiver. It will be intriguing to know which receptors are responsible for triggering these SEK1-dependent survival signals in hepatocytes during embryogenesis. Of note, c-Jun-deficient mice exhibit a similar phenotype aithough the iiver defects are Iess severe than in SEKI" embryos and livers fiom E12.5 c-hm'- embcyos stiil contain residual hepatocytes (Hilberg et al., 1993;

Johnson et al., 1993). These data led to the belief that the SEKI-SAPK-c-Jun pathway is

required for the anti-apoptotic bction of c-Jun during liver organogenesis.

To further investigate the physiological reievance of the amino-terminal

phosphorylation of c-Jun by SAPKs, mice harboring an allele of cJun with serines 63

and 73 mutated to alanine (referred to as JunAA) were generated (l3ehrens et al.' 1999).

Surprisingly, JunAA homozygotes were obtained at Mendelian frequency, although a

slight but sipifkant reduction in body weight was observed in cornparison to wild-type

adult animais. Histological examination of severai organs, including the liver, revealed no

obvious abnormaiities. This implies that SAPK phosphoryIation of c-Jun is not essential

during hepatogenesis or during developmental regulation of ce11 differentiation and

apoptosis. Of note, SEKI~-C-JU~'-double mutants die very early in embryogenesis

(between E7.5-8.5;the cause of lethality is not known; Nishina et al., 1999). The additive

severity of the double mutant phenotype Mersupports the notion that the SEKIISAPK

module and c-Jun cm fùnction in parailel during deveIopment. However, SAPK is clearly

required for many other functions of c-Jun as JunAA fibroblasts exhibit a defect in

proliferation and reduced transformation by components of the Ras pathway and by

oncogenicfos (Behrens et al., 2000). It is possible that c-Jun phosphorylation by SAPK

cm act as a molecuiar switch that increases the spectnrm of functions of c-Jun (possibly

by regulating the recruitment of distinct CO-activatorcomplexes).

Since SEKI-deficieut mice are inviable, its function has beea furthet studied by

employing SEKI-'- ES cells to complement recombination-activating gene (MG)-

Meficient blastocysts. Recent genetic evidence suggests that SEKL'- thymocytes and peripheral T ceils exhibit increased sensitivity to Fas (CD95) and CD!-mediated apoptosis and are defective in CD28-mediated costimulation for proliferation and IL-2 production (Nishina et al., 1997a; Nishina et al., 1997b). B lymphocyte development is also partially impaired in these reconstituted animals (Nishina et al., 1997a). However, these results and conclusions contrast with those of another study that employed RAG-

2-bIastocyst complementation with a different line of SEK1-targeted ES cells and reported that SEKl is dispensable for the development of both the B and T lineages

(Swat et al., 1998). Aging SEK~'-MG-~-'-chimeric mice fkequently developed lyrnphadenopathy and polyclonal B and T ceIl expansions, indicating that SEKl may be required for maintaining peripheral lymphoid homeostasis (Swat et al., 1998). Further investigation is required to resolve the differences between these two reports.

Extensive characterization of the signal transduction in SEKI" cells has been performed. Ce11 lines lacking SEKl exhibit defective SAPK activation and cJun activation in response to some (anisomycin, heat shock, TNF-a and IL-lp), but not ail

(UV irradiation and sorbitol), cellular stresses (Ganiatsas et al., 1998; Nishina et al.,

1997b; Yang et al., 1997a). Currently, some discrepancy exists as to the precise Ievels of

SAPK activation by these agonists in the absence of SEK1. There also appear to be ceil type-specific effects inasfar as deficits in responses are distinct between ES cells and fibroblasts, for example.

SAPK and p38 are activated with both quantitative and qualitative dserences afler a varïety of stress stimuli, which must reflect a divergence in activating pathways immediateiy upstream of these kinases (Zanke et al., 1996). However, SEKl has been show to phosphorytate p38 in vitro and uiis promiscuity has raised the possibiiity that SAPK and p38 are CO-regulatedby this kinase (Lin et al., 1995). In support of evidence that dominant-negative SEKl specifically acts as an inhibitor of the SAPK signal transduction pathway (Zanke et al., 1996), bioctiemical studies with the homozygous knockout SEKl cells indicates that despite defective SAPK signalling, activation of p38 by a variety of agonists is unaitered (Ganiatsas et al., 1998; Nishina et al., 1997b). While it still remains theoretically possible that the effect of SEKl gene disruption on p38 is hlly complemented by the p38 upstream activators MKK3 and MKK6, these data suggest that SEKl functions as a specific activator of SAPK, and not p38, in vivo. An

MKK6 knockout mouse has not yet been reported, but assaying for any SEK1-dependent p38 activity in a I~~KK~~*MKKS'~background should fially resolve this issue. There is agreement that SAPK does not associate with either MKK3 and MKK6, and it was recently shown that MKK3 disruption has no effect on SAPK activation by W radiation, osmotic shock, K.-1B and TNF-a(Wysk el al., 1999). Mouse homozygaus phenotype Reference

Viable; hypochondroplasia Reimold et al. Nature 379,262-5 (1996) Lethal E 13.5-14.5; impaired hepatogenesis Hilberg et al. Nature 365, 179-18 1 (1993); Johnson Hilberg et al. Genes Dev. 7,1309-13 17 (1993) Viable; impaired spennatogenesis Thepot et al. Development 127,143-53 (2000) Viable; no detectable phenotype Yujiri et al. Science 282, 19 11-19 14 (1998) Lethal El 1; defective cardiovascular Yang et al. Nat Genet. 24,309-3 13 development (2000) Viable; impaired IL-12 production Lu et al. EMBO 1.18, 1845-1857 (1999) Lethal Unpublished (Josef Penninger, personal communication) Lethal E13.5; defective placental Adams et al. Mol Cell6, 109-1 16 development and erythropoiesis (2000); Mudgen et al. Proc Natl Acad Sci USA 97,10454-10459 (2000); Tamura et al Ce11 102,221-23 1 (2000) Viable; defective T ce11 differentiation Yang et al. Immunity 9,575-585 (1998) Viable; reduction in neuronal apoptosis Yang et al. Nature 389,865-870 (1997) Viable; defective T ce11 differentiation Dong et al. Science 282,2092-2095 (1998) Lethal El0.5; dysregulation of apoptosis in Kuan et al. Neuron 22,667-676 (1999) brain Lethal El 1.512.5; abnormal Nishina et al. Development 126,505- hepatogenesis 516 (1999); Ganiatsas et al. Proc Natl Acad Sci USA 95,688 1-6886 (1998); Yang et al. Proc Natl Acad Sci USA 94, 3004-3009 (1997)

Table 1. Mutant phenotypes of SAPK and p38 pathway components. 1.3.4 Regulation of SAPK by MAPKKKs and TNF

A large and diverse array of MAPKKKs has been shown to activate SAPKs when overexpressed in cells (reviewed in Widmann et al., 1999; Table 2). These include the

MEKIERK kinase (MEKK) subgroup, the mixed-lineage kinase (MLK) group, tumor progression locus-2 (T'PL-2, the product of the Cot oncogene), and TGFP-activated protein kinase (TAK1). The MEKK group of MAPKKKs includes MEKK1-4 and apoptosis signal-regulated kinase-1 and -2 (ASKLIZ. or MAPKKK 516) which are

marnmalian orthologues of S. cerevisiae STE11. The MLK group of MAPKKKs, which share significant sequence identity witb both serineithreonine and tyrosine kinases,

includes MLKI-3' dual-leucine zipper-bearing kinase @LK or MUK), and leucine

zipper-bearing kinase (LZK). OP these, MEKK1-3 and Tpl-2 can also activate the ERK

pathway, while TAK1, ASKI, MLKj and MEKK4 have been shown to strongly activate

p38s as well, There are no known MAPKKKs that activate only p38 or p381ERK MAPK

pathways,

The importance of Ras family GTPases in mamrnalian MAPK signal transduction

was first appreciated with the discovery that oncogenic Ras could activate the ERK

pathway. GTPases of the Rho family (Rho, Rac, Cdc42) were originally thought only to

regulate the actin cytoskeleton (Bishop & Hdl, 2000). More recently, however, these

GTPases have been implicated in MAPK signai transduction since constitutively-active

mutants of Racl and Cdc42 can activate SAPK and p38 (Bishop & Hall, 2000; Coso et

al., 1995). Downstrearn targets of Racl and Cdc42 possess a common motif that is

critical for G-protein binding (Burbelo er al., 1995)- This site, the CDC42tRacl

interaction and binding domain (CRiB domain), is present on severai protein kinases and has been described for MLK1-3 (but not DLK or LZK; Nagata et al., 1998). It has also been reported that MEKKl and MEK4 (but not MEKK2 or MEKIU) bind directly to

Cdc42 and Raci (Fanger, Johnson & Johnson, 1997). These interactions between Rho

GTPases and MAPKKKs may contribute to the effects of Rho GTPases on SAPK and p38 activation, however, their regdation is not well defined.

The best characterized mechanism of SAPK activation is by proinfiammatory cytokines of the TNF fmily and recent studies have shown several MAPKKKs to be involved in TNF signa1 transduction (Figure 2). For instance, NF-KB-inducing kinase

(NIK) associates with TRAF2 and other members of the TRAF farnily and mediates activation of NF-@, but not SAPK (Malinin et al., 1997). TAKl is also activated by

TNF, and was demonstrated to associate with TRAF6, but not TRAF2, in an IL-1- dependent manner (Ninomiya-Tsuji et al., 1999). On the other hand, TRAF2-mediated activation of SAPK is inhibited by catalytically-inactive mutants of MEKKl (Yuasa et cd., 1998) and ASKl (Hoeflich et al., 1999; Nishitoh et al., 1998). ASKl is responsive to

TNF treatment in many ceIl types (Ichijo et al., 1997) and TNF signailing to SAPK is mediated by ASKl association with members of the TRAF farnily (Hoefiich et al., 1999;

Liu et ai., 2000; Nishitoh et al., 1998). These interactions require the conserved amino- terminal zinc EUNG and TRAF-N motifs typical to TRAF family members. WhiIe overexpression of a wild-type or activated allele of ASKI induces apoptosis in various ce11 types through mitochondria-dependent caspase activation matai et al., 20001, catalyticaily-inactive ASKl rescues ceii fiom TNF-mediated killing (Ichijo et al., 1997).

This rescue by dominant-negative ASKI is dependent on the presence of endogenous

TRAF2, as determined by comparing TNF-induced apoptosis in TRAF2-deficient and wild-type control fibroblasts (Hoeflich et al., 1999). In addition, through genetic screening for ASK1-binding proteins, the redox-sensing enzyme thioredoxin (Trx) was recently identified as a physiological inhibitor of ASKl (Liu et al., 2000; Saitoh et al.,

1998). Upon treatment of cells with TNF or reactive oxygen species generators such as hydrogen peroxide, Trx appears to be oxidized, ASKl dissociates from Trx and is bound and activated by TRAFî, leading to downstream signalling.

1.3.5 MAPKKK kinases and TNF

Serinelthreonine kinases homologous to yeast STE20, such as germinal center kinase

(GCK), GCK-related (GCKR, also referred to as KHSI) and GCK-like kinase (GLK), have also been shown to be important effectors for TNF signaihg of SAPK activation

(Kyriakis, 1999; Shi & Kehrl, 1997; Yuasa et al., 1998). Expression of antisense DNA constructs of GCKR cm block TNF and TRAF2 activation of SAPK. In addition, expression of full-length TRAF2, but not TRAF2 mutants wherein the RING domain has been deleted, activates GCKR and the SAPKs in vivo. Via their carboxy terminal domains, GCK and GCKR can both associate in vivo with TRAF2 and GCK can also associate in vivo with TRAF6. It is noteworthy that the carboxy region of GCK is also required for binding to MEKKL, thereby constituting an ASKl independent signalling pathway fiom TRAF2 to SAPK. However, the physiotogical effect of this interaction is

Iikely to differ somewhat from that of TRAFZmediated ASKl activation. For instance,

MEKKl and GCWGCKR cmonly activate SAPK, while TNF-a,TRAF2 and ASKL can activate both the SEKl/MKK7-SAPK and MKKI/MKK6-p38 pathways. Thus, ASKl is likely to be a physiological target of TRAF2 in recmiting p38. Moreover, coimmunoprecipitation experiments indicate that GCK and ASKl do not retiably interact in vivo. Also, no apoptotic function has been assigned to GCK. MEKKl has been implicated in the activation of IKB-kinase (LUC)' a component in the anti-apoptotic NF-

KB pathway (Lee et al., 1997), but expression of antisense RNA to GCKR has no effect on the transactivation of NF-KB. Despite a great deal of interest in the GCK group of kinases and the TNFISAPK pathway, there has yet to be a direct dernonstration of the activation of a MAPKKK by a GCK homologue- AIthough these kinases may activate

S.4PK when over-expressed in cells, their regulation of different MAPKKKs has not been demonstrated biochemically or geneticaliy. Taken together, however, one funçtion of TRAFs may be to regulate the interactions between cytokine-activated GCKs and their effectors. TNF-induced ternary complex formation of TRAFL-GCK-MEKK1 wili be of interest to study.

Efforts to elucidate the mechanisms of MAPKKK regulation have been hampered by the fact that al1 marnrnaiian SAPK-activating MAPKKKs identified thus far are constitutively active upon overexpression. Correspondingly, while ASKl can associate with components of the TNFRl cornplex, mere overexpression of ASKl results in its potent activation, overwhelming any endogenous inhibitors present in limiting concentrations. While we still await data from ASKI-deficient cells to determine if

ASKl is selectively required for TNF-induced sustained activation of SAPK and p38, there are already two reports with conflicting resuits regardiig the regulation of SAPK by

TNF-a and [L-1p in ILEXKI* and MEKKI" macrophages and fibroblasts (Xia et al.,

2000; Yujiri et al., 2000). It remains possibk that as yet unidentifïed proteins cm also bind TRAF2 and mediate SAPK activation. Fas-L TNF-a I L-1 4 /'\ 4 Fas TNFRl = 4 TNFR2 IL-1 R = FADD 4 TRAF2 TRAFG

Caspase8 MEKKI ASK1 TAK1

SEKI MKK3

\ SAPK ~38

c-Jun ATF2 CHOP Bcl-2 Bid Bcl-XL I I Figure 2. Ovewiew of TNFISAPK signalling

Inflammation is an important biological response to the exposure of tissues to stress. SAPK and p38 mediate inflammatory signals fiom the tumor necrosis factor (TM) family of cytokines. Central to this, the MAPKKK ASKl associates with TRAF2 and cm be negatively regulated by thioredoxin (Trx). As determincd by targeted gene disruptions in mice, SAPK is required for caspase-9 activation by the mitochondrid pathway. Potentiai targets of SAPK include members of the Bc12 group of apoptotic regulatory proteins. There is evidence to indicate that signalling specificity may be niediated through formation of multi-protein comptexes held together by "scaffold" proteins. The fist example of such a fiamework molecule is tiom the yeast pheromone pathway where the

MAPK FUS3 binds to the STES together with the MAPKK STE7 and h,IAPKKK STE11 (Herskowitz, 1995). Scaffold proteins have now aiso been identilïed in marnrnalian celIs. These include MEK partner-1 (MPI) which interacts with the MAPK

ERKl and MAPKK MEKl (Schaeffer er al., 1998), and the JNK-interacting protein (JIP) group of proteins that bind to SAPK, MKK7 and mixed-lineage protein kinases (Dickens et al., 1997; Yasuda er al., 1999). JIPs have yet to be shown to play a role in TNF receptor-signdiing.

1.3.6 The p38 MAPK family

Mammalian p38 MAPK was fist identified as an LPS-inducible activity in murine peritoneal macrophages (Han et al., 1994). Activation of p38 has traditionally been associated with the stress response and some forms of apoptosis, however, recent studies indicate that a Iarger variety of cellular processes are regulated by p38 (Nebreda &

Porras, 2000). For instance, p38 MAPKs have been proposed to play a physiological role in inflammation and the immune response; inducing differentiation in adipocytes, myoblasts, neurons, chrondrocytes, cardiomyocytes and erythroid cells; and promoting or inhibithg ceii proIiferation and survival in a cell-type specific manner.

Four p38 MAPKs have been cloned that are 60Y&-70% identicai in their amino acid sequences: p3 8cr/Mpk2/CSBP, p3 88, p38yERK6, and p3 86. Substrates of p3 8

WKsinclude pro tein kinases such as MAPKAPKS (MAPK-activated protein kinase- 2) and several transcription factors including MEF2 (myocyte enhancer factor-2),

CHOPIGADD 153 (CEBP homology proteidgrowth arrest and DNA damage-153),

CREB (CAMP-response-element-bindingprotein) and ATF2 (Gupta et al., 1995; Han et al., 1997; Iordanov et al., 1997; Stokoe et al., 1992; Wang & Ron, 1996). Although p38s have overlapping substrate specificity, some targets appear to be preferentially phosphorylated by one or more isoforms (Cohen, 1997). This, together with the observation that the isoforms have distinct tissue expression patterns (Wang et al., 1997), suggests that p38 MAPKs may have both redundant and specific functions. However, the precise biochemical role that each isoform serves in vivo remains unclear.

Studies aimed at understanding the function of p38 have been greatly facilitated by a novel class of pyridinylimidazoles known as cytokine-suppressing anti- inflamrnatory dmgs (CSPLLDs; Le. SB203580) that achieve their effect, at least in part, by inhibition of the a and P isoforms of this kinase (Cuenda et ai., 1995). Since p38y, p386 and many other protein kinases tested appear to be insensitive to CSAiDs, these dmgs have been used extensively to identify substrates and physiological roles of p38a and p38p(Cohen, 1997). However, whether al1 reported CSAlD effects are attributable to p38 inhibition remains to be ciarified. Structural and site-directed mutagenesis studies have recently provided a basis for the selectivity of CSAiDs: the dmgs insert into the

ATP-binding pocket of p38a and bind competitively with ATP (Tong et al., 1997;

Wilson et al., 1997). CSAIDs, however, do not make contact with residues of the ATP- binding pocket that actuaiiy interact with ATP and recent studies have established that threonine-106 of p38a interacts with the 4-fluorophenyl moiety of CSAiDs and plays a criticai roIe in determinhg hgsensitivity- Mutation of this residue to methionine or glutamine, amino acids present at the equivalent position in other MAPKs, or to other residues with bulky side chahs, rnakes p38a and p380 insensitive to CSAIDs (Wilson et al., 1997). Conversely, mutation of this residue to threonine in other MAPK family members (SAPKs, p38y and p386) confers sensitivity to CSAiDs (Eyers el al., 1998;

Gum er al., 1998). Examination of the sequences of protein kinases in the databases reveds that a bulky residue is almost always found at the position equivalent to threonine-106. However, a small number of proteins do have threonine at this position and recent work demonstrating that CSAiDs can aiso modulate the activity of type41

TGF-B receptor, Lck (lymphoid ce11 kinase), Raf-1 MAPKKK, PDK1, and cyclooxygenase (Borsch-Haubold, Pasquet & Watson, 1998; Eyers er al., 1998; Hall-

Jackson er al., 1999; Laii et al., 2000) highlight the potential for CSAID-mediated cellular effects that are independent of p38a and p. Thus, to better understand the biological function of p38 and to whai extent the various p38 isoforms participate in separate physiological processes, four independent groups have recently used homologous recombination to disrupt p38a in mice (Table 1).

1.3.7 Genetic analysis of p38a in mice

Deletion of the p38a gene results in embryonic lethality commencing at E10.5 (Adams et al., 2000; Allen et al., 2000; Mudgett et al., 2000; Tamura et al., 2000). Biochemical assays determinhg the stimulus-induced phosphorylation of p3&c-dependent targets

(MAPKAPIG, ATF2) confimed that the p38a targeting strategy successhlly and specifically abolished signailhg by p38a. Since mice express multiple p38 MAPK family members, this developmental arrest demonstrates that the different enzymes do not perform entirely redundant activities, at least during embryonic development.

Although Allen et al. (2000) noted that mice nul1 for the p38a allele die during embryonic development, no specific phenotype was descriied. Adams et ai (2000) and

Mudgett et al. (2000) observed similar phenotypes. These groups indicated that with time, p38a-deficient embryos become pale and anemic, have deficiencies in vascularization of the embryo and yolk sac, and show varying degrees of growth retardation. The challenge was to iden* the defects that are directly due to loss of p38 function and to distinguish these from secondary defects associated with the loss of viability.

Since a P-gaiactosidase cassette was inserted into the p38a locus, the first clue came fiom studying the expression of p38a during embryonic development (Adams et al., 2000). At E10.5, abundant Iabeling was seen in many regions of the embryo including the hem, branchial arches, limb buds, and somites. High levels of p38a were also found in the extraembryonic tissues, such as the endoderm, mesodem, and the vasculature of the yolk sac and the placenta. Closer examination revealed that while the materna1 part of the placenta was normal, there was a striking reduction in the

Iabyrinthine layer and embryonic blood vessels seemed to be trapped in the superficiai

layers of the placenta and could not intemingle with maternai blood vessels. Histologicai

analysis in one study (Mudgett er al., 2000) desccibed a greatly reduced

spongiotrophoblast layer while the other study (Adams et al., 2000) reported this

structure to be normal. However, fiom the phenotype both groups concluded that

surprisingly, in spite of being broadly expressed in the embryo, p38a appears to be

criticai only for placenta organogenesis. What might be the upstream activators of p38a in this process? A number of proteins has been shown tu be vitai for choriodIantoic placental development including the basic-helix-loop-helix transcription factors MASH-2 and TFEB, nuclear hormone receptor peroxisome proliferator-activated receptor-y, estrogen-receptor-related receptor- p, the von Hippel-Lindau tumor suppressor protein, heat shock protein-90fi, and retinoid

X receptor CY or B (Me, 2000). An interaction benveen p38u and some of these known phyers in placentaI development is conceivable, as p38s previously have been shown to be activated by heat shock, oxidative stress and hormone signalling. Early response proto-oncogenes, such as c-Jun and JunB (Dungy, Siddiqi & Khan, 199 1; Schorpp-

Kistner er al., 1999), have been associated with both proliferation and differentiation events of extra-embryonic tissues. These genes are also expressed in the placenta of human and rodents throughout gestation, suggesting an additional mechanism for p38a in placental deveiopment, Lady, the p38cr-nuil mice phenotype is highly reminiscent of the defective labyrinthine layer of the placenta and cardiovascular malformation that has been recently described for mice lacking MEKIU, an upstream activator of p38 MAPK signailing (Yang et al., 2000).

The fourth study by Tamura et al. (2000) suggested a different analysis and interpretation of the p380i" phenotype. The authors indicate that the p38u-deficiency results in two distinct developmental defects. As described by the other groups, the p38a nul1 embryos are initially challenged by placental hidficiency, but a signifcant portion

(613 1 embryos) rernain viable until much later in development (E16.5) with normal morphology but highly anemic appearance. The basis for the anemic phenotype was traced to a block in erythroid differentiation in fetd liver ceiis and deficiency in Epo gene expression. The reasons for the dirences between the various examples of p38a mutant animais is currently unclear as al1 appear to be true nulls (rather than, for example, hypomorphs).

Despite some discrepancies, these studies demonstrate several important conclusions. Firstly, the phenotypes associated with the p38a nul1 mutations suggest that the function of p38a is, at least partiauy, nonreduadant with other p3 8 MAPK family members. This is quite different than what has been reveaied from knockouts of the

SAPK or ERK family mernbers. None of these mutants revealed a critical role for any of the individual MAPKs in development, and therefore, it has been assumed that this fmily of protein kinases has extensively overlapping iünctions within each subtype. In addition, although previous studies have established that MAPKAPK2 is a substrate of p38cr, the extent to which other kinases may participate in vivo in the activation of

MAPKAPEC.2 remained unclear. Biochemical anaiysis in the p38a knockout studies

Found that UV, anisornycin and sodium arsenite-induced activation of MAPKAPK.2 was completely impaired in p38a4 cells (Adams et al., 2000; Allen et al., 2000). Thus, while the Iack of a comparable phenotype in MKK3-deficient embryos suggests that this kinase is not uniquely required for p38a activation, it can be concluded that MAPKAPK2 is a nonredundant component of the pathway.

1.3.8 Summary

The stress-activated protein kinases are activated by a surprisingly broad array of celIular stress stimuli and current evidence supports a key role for these enzymes in a variety of ceiiuiar functions and developmentai processes. While key signahg intermediates of the cascade mediating activation of c-Jun by proinflammatory cytokines have been identified

(namely ASK 1 and GCKfMEKKl), their involvement is not well understood

mechanistically. tn addition, TRAFs are known to play significant roles in transducing

signals frorn receptors to SAPKslp38s, but their direct targets have not been fuliy

elucidated. Therefore, understanding the physiological roles of such MAPKKKs and

MAPKKK kinases in TNF signaling will fixther disclose the precise regulatory

mechanisms of these kinase cascades.

Glycogen synthase kinase-3 (GSK-3) is a serine-threonine kinase encoded by two

isofonns in mamrnals, termed GSK-3a and GSK-3B (Woodgett, 1990). GSK-3 was

primarily irnplicated in musclt energy storage and metaboiism but since its cloning, a

more generalized role in cellular regdation has emerged, highiighted by the wide array of

substrates controled by this enzyme, including cytopIasmic proteins and nuclear

transcription factors. GSK-3 targets encompass proteins implicated in Alzheimer's

Disease, neurological disorders and cancer. GSK-3 genes are highiy conserved and have

been identified in every eukaryote ïnvestigated to date. Studies of GSK-3 homologues in

various organisrns have revealed physiological roles for the enzyme in differentiation,

ce11 fate determination and in spatial patterning to establish bilateral embryonic

syrnmetry. GSK-3 is already known to play an important role in multiple signal

transductory systems, including the WntiWg and PI3'kinase pathways, which influence proliferation and apoptosis respectively. Here 1wiI1 focus on the basic characteristics of

GSK-3 and discuss recent advances in understanding the invoivement of this unusual enzyme in the development of genetically tractable organisms and in human

pathophysiology.

1.4.1 Isolation and characterization of GSK-3

GSK-3 was orïginaiiy identified as one of five protein kinases that phosphorylate the

rate-limiting enzyme of glycogen synthesis, glycogen synthase (GS; Embi et al., 1980;

Hemmings er al., 1981; reviewed in Woodgett, 1994). Fotlowing peptide sequencing of

enzyme purified from skeletal muscle, a screen of a rat brain cDNA library revealed that

GSK-3 is encoded by two independent genes, GSK-3a and GSK-30, with molecular

weights of 51 and 47 kDa, respectively (Woodgett, 1990). Purified GSK-3a and GSK-3B

exhibit sïmilar biochemical and substrate properties (reviewed in Woodgett, 1991). The

two genes display 85% overaii sequence identity, which is even higher in the catalytic

domains (93%). ChromosomaI mapping identified the cytologicd location of human

GSK-3a as 19q13.2 whereas hurnan GSK-3B is maps to 3qI3.3 (Shaw et al., 1998). The

GSK-3B promoter contains several CMTboxes as well as positive and a negative

transcriptional response elements (Lau et al., I999b). Northern blot anaiysis has shown

that GSK-3a and GSK-30 are sornewhat variably expressed in different mamrnalian

tissues and there is poor correlation between the 1eveIs of mRNA and protein (Lau et al.,

1999a; Woodgett, 1990). This may represent a differentid mode of transcriptionai and

translational regdation for the two isoforms. Substrate Specific phosphorylation sequence Reference acetyl CoA carboxylase n.d. (Hughes et al., 1992) adenornatous polyposis coli protein FXVEXTPXCFSRXSSLSSLS (Groden et al., 199 1) ATPtitrate LLNASGSTSTPAPSRTASFSESR (Benjamin et al., 1994) kvin SANDSEQQS~'~ (Ikeda et al., 1998) SDADTLSL?" SLTDS'" bcatenin DSGLHSGA'ITTAPS (Yost er al., 1996) C/EBP- TPPPTPVPSP (Ross et al., 1999) cJun EEPQTVPEMPGETPPLSPIDMESQER (Boyle et al., 1991) C-Myb APVSCLGEHHHCTPSPPVDH (Humer et al., 1988) c-Myc DLWKKFELLPTPPLSPSRRSG (Pulverer er al., 1994) CREB KRREILSRRPSYR (Fiol et al., 1987) cyclin Dl EEVDLACTPTDVRDVDI (Diehl et al., 1998) efF-2B translation factor DSEELDSRAGSPQLDDIKVF (Welsh er al., 1993) G subunit of phosphatase 1 AIFKPGFSPQPSRRGSSESSEEVY (Fiol et al., 1988) glycogen synthase RPASVPPSPSLSRHSSPHQSEDEE (Fi01 et al., 1987) heat shock factor-1 KEEPPSPPQSP (He et al., 1998) inhibitor-2 GLMKlDEPSTPYHSMIGDDDDAYSD (Park et al., 1994) insulin receptor substrate 1 n.d, (Eldar-Finkelman et al., 1997) JunD (Nikolakaki et al., 1993) L-rnyc DIWKKFELVPSPPTSPPWGL (Saksela et al., 1992) NF-ATc n.d. (BeaIs et al., 1997) RI1 subunit of CAMP-dependent LREARSRASTPPAAPPS (Hemmings et al., protein kinase 1982) Tau TPPKSfSAAK (Hanger et a[., 1992) SPWSGDTSPR

Table 3: Protein substrates of GSK-3

GSK-3 substrates are listed ~iththe peptide sequences surrounding their phosphorylation sites. GSK-3-targeted serines and threonines are indicated in bold font. Sites of "priming" prephosphory lation required for GSK-3 phosp horylation (if known) are underlined. Some substrates have nurnerous GSK-3 phosphory!ation sites in multiple regions and could aii not be listed here. The notation "n.d." indicates that the phosphorylation sequence has not yet been characterized. 1.4.2 Regulation af GSK-3 activity

Besides glycogen synthase, a number of other GSK-3 substrates have been identified

(reviewed in Nusse, 1999; Woodgett et al., 1993). These include: translation initiation factor ~IF~BE,PKA, phosphatase subunit inhibitor-2, and ATP-citrate lyase (Table 3).

GSK-5 also phosphorylates severai transcription factors such as CREB, c-Jun, c-Myc, c-

Myb, as well as heat shock transcription factor HSF-l in vivo (Xavier et d.,2000) and in vitro (He et al.. l998a). Brain-associated prote& such as amyloid precursor protein, Tau and neurofilament protein are also targets of GSK-3 (Plyte et al., 1993). Phosphorylated

Tau has lower afinity for microtubules and hy perphosphorylation appears to promote association into paired heIical filaments, a pathoIogical feature of Mzheimer's disease

(Lovestone & Reynolds, 1997; Spittaels er al., 2000; Yamaguchi et al., 1996).

GSK-3 is a serine/threonine selective kinase tha: recognizes and phospho~latesa consensus sequence SXXXS(P) in certain proteins (Fiol et al., 1990). For some substrates, Like giycogen synthase, phosphorylation by GSK3 requires prior phosphorylation of a serine residue C-terminai to the target site. The specificity of this priming site has been probed with a synthetic peptide corresponding to a GSK-3 phosphoryIation site in eIF2Bc. Peptides phosphorylated at senne or threonine in the priming position were condusive to subsequent GSK-3 phosphorylation (Wiiliams et al.,

1999). Phosphotyrosine, however, was not a specificity determinant for GSK-3.Although prior phosphorylation of the consensus site seem to be a common mode by which GSK-

3 targets its substrates, this priming mechanism is not requisite for al1 substrates. For example, GSK-3 phosphorylation of p-catenin does not appear to require priming (Yost et al., 1996). This differential requirement for substrate interaction may provide a control mechanism by which the specificity of signal transduction is influenced by the contemporaneous activation of other signalLing pathways (which modulate the activity of the various priming kinases).

In studies in huit fly and fiog embryos, expression of kinase-deficient GSK-3 exhibits phenotypes which mimic WnWingless signalling and result in the stabilization of B-catenin/Armadillo (He et al., 1995; Peifer et al., 1994b; Pierce & Kimelrnan, 1996;

Siegfried er al., 1994; reviewed in Wodarz & Nusse, 1998). This is consistent with a mode1 in which GSK-3 is constitutiveiy active and negatively regulated by upstream signals. in Drosophila, the GSK-3 homologue is termed Shaggy or Zeste-white3 (herein denoted as 2~3~~~).A senes of epistasis anaiyses have indicated that 2~3~~~activity is inhibited by Wg protein, an effect mimicked by expression of downstrearn components of

Wg signalling, namely Frizzled-2 (Fz-2) and Dishevelled (Dsh; Rue1 et al., 1999). The

IeveI of 2~3'~kinase activity reguiates the stability of Annadillo protein and modulates the level of phosphorylation of D-Axin and Armadiilo. The mechanism by which activity of ntrjSegis regulated by Wg is uncIear. Work in our laboratory has show that Wg- dependent inactivation of is accompanied by serine phosphorylation although the identity of the relevant kinase acting on GSK-3 is unkno~vn.The most proximal upstream regulatory protein in this pathway is Dishevelled. This protein is dso phosphorylated by a

Dsh-associated kinase (DAK). The finding that casein kinase-II (CK-II)can bind Dsh suggest a possible role for CK-II (and DAK) in the transduction of the Wmgless signal

(Sakanaka et al., 2000; Willert et al., 1997). Dishevelled lives two functional iives. In addition to modulatirtg the Wingless pathway, it is dso a component in regulation of pIanar poI& which involves the DrosophiCa SAPK signalhg pathway. Interestingly, DAK drarnaticaliy enhances the bction of Dsh in the Wnt pathway while inhibiting in the establishment of planar polarity. Dsh thus appears to act as a teeter-totter in controiing flux through two distinct signaihg systems.

Under resting conditions, GSK-3i~w3'~~is highiy phosphorylated at a tyrosine residue (Hughes et al.: 1993; Wang et al., 1994). This constitutive phosphorylation is important in facilitating kinase activity. The Schizosaccharomyces pornbe homolog of

GSK-3, Skpl, is also phosphorylated on a homologous tyrosine residue to that in higher eukaryotes, and this phosphorylation is required for its efficient activity (PLyte et al.,

1996). In marnmalian cells, GSK-3a is constitutively phosphorylated at tyrosine-279,

GSK-38 at tyrosine-216 (Hughes et al., 1993)- Dephosphorylation of these residues is accompanied by kinase inactivation. Since mitogen-stimulated GSK-3 inhibition can be reversed by serine/threonhe-specific phosphatases, it is believed that the cause of such inhibition is serinefthreonine phosphorylation of GSK-3 rather tyrosine dephosphorylation (Cross et al., 1994). Serine 9 in GSK-3f3 (equivaient to serine-21 residue in GSK-3u) is the target residue for inhibitory phosphorylation upon insuiin stimuiation (Shaw et aL, 1997; Stambolic & Woodgett, 1994; Sutherland et al., 1993).

Several protein kinases target serine 9/21 inciuding PKB/Akt, pp90 Rsk and cyclic AMP- dependent protein kinase (Cross er al., 1995; Fang et al.. 2000; Li et al., 2000; Stambolic

& Woodgett, 1994).

Inactivation of GSK-3 by phosphorylation at serine 9/21 or via the Wat pathway invokes different celluiar consequences. The mechanism underlying this signai- dependent regdation is Iikely via the existence of distinct pools of, or complexes containing, GSK-3. For exsimpIe, in cells, cytoplasmic fi-catenin is associated with a scaffold protein called Axin which also binds GSK-3. The population of GSK-3 molecules bound to Axin is seiectively sensitized to Wnt signailing. However, a difFerent pool of GSK-3 appears to be responsibIe for transducing the effects of insuiin/PI3' kinase signalling. The fuiding that hcin/conductin-associated GSK-3 is insensitive to insuiin-

induced serine-9 phosphorylation but is instead is coupled to Wnt activation further

supports this notion (Ding et al., 2000). Since GSK-3 has multiple substrates, the

existence of distinct pools of responsive kinase molecules may allow specific responses

to certain agonists, leaving other targets unaffected. This idea is further strengthened by

the fmding that inhibition of GSK-3 by certain Wnt signalling components (such as

FRAT-IIGBP), leads to the inhi bition of GSK-3 induced phosphorylation of B-catenin

without effecting the phosp horylation of other substrates such as glycogen synthase

(Thomas et al., 1999).

In mammalian cells, protein kinase C (PKC)-like activity has been shown to be

required for the effect of Wg protein on GSK-3 activity (Cook er al., 1996; Goode et al.,

1992). PKC can phosphorylate GSK-3 in vim and resdts in its inactivation. Wg-induced

inactivation of GSK-3 is sensitive to both the PKC inhibitor Ro3 1-8220 and prolonged

pre-treatment of LOT112 fibroblasts with phorbol ester, suggesting a role for PKC in the

regulation of GSK-3. In addition it has been demonstrated that serum enhances the effect

of lithium on the accumulation of cytoplasmic p-catenin through the activation of PKC

(Chen et al., 2000). However, since Wnt-induced accumulation of cytopIasmic f3-catenin

is only partiaiiy inbibited by PKC inhibitors, PKC likely serves to enhance the effects of

Wnt signailing rather than being directly coupted to this pathway. 48

Species Full-len~ protein Catalytic dornain % identity imiiarity) ?6 identity imiiarity) GSK-3~ GSK-38 GSK-3a GSK-3 8 (human) (human) (human) (human)

Homo sapiens GSK-3a 100 83(89) LOO 9u97) Homo sapiens GSK-38 76(84) 100 9 l(97) 1O0 Raim norvegicus GSK-3a 93(93) 83(89) 99(99) 9 l(96) Rarrus norvegicu GSK-3 B 76(84) 92(92) 90(9'3 99(99) Mus muscutus GSK-38 76(84) 93(93) 9N97) loo(1oo) Danio rerio ZGSK-3c~ 80(86) 78(84) 9x96) 9 l(97) Danio rerio ZGSK-38 77(84) 92(93) 90(96) 99(99) Xenopus laevis XGS K-3 8 75(83) 90(92) 90(96) 9 L(97) Ciona inresrinalis GSK-3 80(87) 83(89) 90(96) 9 l(97) Paracentrotus lividus SUGSK-3 71(83) 79(88) 84(92) 87(94) Drosophila melanogarer av3sBg ïl(8 1) 79m 82(92) 85(94) Hydra vulgark GSK-3 66(80) 72(86) 78(9 1) 80(92) Caenorhabdirk elegans gs k-3 77(87) 77(87) 80(89) 80(89) Petunia hybrida shagf3Y 69(82) 67(80) 80(89) 8 l(89) Nicoiiana tabacum shagf3Y 62(76) 69(80) 7 I(84) 73(85) Arabidopsis rhaliana ASiüATK 1 62(76) 65(8 1) 70(85) 7 1@7) Oyasariva OSK 63(77) 65(80) 7 l(84) 7 I(85) ~tredicagosativa MSKI-3 62-65(76) 64-68(76- 70-71(83- 70-72(83- 79) 84) 85) Dictyosrelitrm discoideum gskA 58(72) 6 l(73) 69(85) 69(83) Schkosaccharomyces Skp 1 60(75) 65(78) 58(73) 67(82) pombe Saccharomyces cerevisiae MCKLMRKI 45-58(62- 53(68) 54-60(73- 54-57(72- 76) 77) 76) Plarmocrium fracipamm GSK-3 53(74) 54(72) 56(76) 58(76) Trifolkm repens GSK-3 7 l(84) 68(8 1) 7 l(84) 7 l(85) Ricinw cornmunis shaggy-like 74(88) 70(84) 74(88) 74(89) Cicer arietinum GSK-3 66(77) 7 l(82) 73(85) 74(85)

Table 4. CSK-3 is highly coaserved throughout evolution

GSK-3 homologues are represented as percentage identity and similarity (in parenthesis) over the fÙiI-length and catalytic region of hurnan GSK-3aand GSK-3p proteins. GSK-3 proteins share a remarkable degree of identity within the catalytic domain and in vertebrates the identity over this region is greater than of 90%. GSK-3 family members have also been identified in numerous invertebrates (nematode, fiuit fly, ascidians, ticks, sea urchin and malaria parasite), plants (Arabidopsis, Hydra, garden petunia, rice, clover, sunflower, alfalfa, chickpea and tobacco) and hgi(baker's yeast, fission yeast and sIime rnold), 1.4.3 GSK-3 homologues

GSK-3 homologues have been isolated from a number of organisrns (Table 4), and display diversity with respect to their physiological roles. In lower eukaryotes for example, the GSK-3 homoloueg in Schizusaccharomyces pombe, SKP1, plays a dein cytokinesis (Plyte et al., 1996). The slime moM Dic~osteliumdiscoidezcm harbours a

single GSK-3 related gene, tenned gsk-A, that piays a critical role in speciQing cell fate

in a CAMP-dependent process that controls the differentiation process leading to spore

and stalk formation (reviewed in hall, 1995). During normal development aggregated

progenitor cells differentiate into three ce11 types: prespore, prestalk-A, and prestalk-B. In

gsM* cells, the differentiation proceeds with a reduction of prespore ce11 population and

enhanced formation of prestaik-B ceIIs. One of the targets of CAMP activation of gsk-A

is a Dicryostelium STAT transcription factor. GSK-A phosphorylation at a serine residue

enhances nuclear export causing fiinctional inactivation of the STAT (Ginger er al.?

2000). The physiological significance of STAT regulation by GSK-A with respect to

cellular differentiation has yet to be addressed.

Sea urchin GSK-3 has been impIicated in the establishment of animal-vegetal

axis in early development. During embryogenesis, animal ceils give rise to ectodermal

tissues whereas vegetal cells cary a mesodemai and endodermal fate (reviewed in

Angerer & Angerer, 2000). Overexpression of wild-type GSK-3 results in the

animalization of the embryos by giving rïse to ectodermal phenotype (Emily-Fenouil et

al.: 1998). By cornparison, the inhibition of GSK-3 by treating embryos with Lithium or

by ectopic expression of kinase-dead GSK-3 or mutant fi-catenin, promotes vegetal ce11

fate. These results are cornparabLe to those fiom experiments using Xenopus oocytes, where transcripts encoding the known components of Wnt pathway are present as matemal products (reviewed in Itoh er al.3 1998; Moon et al., 1997; SokoI, 1999). After fertiIization, cortical rotation relocaiizes these moIecuies by the movement of outer cytoplasm with respect to the inner. This rotation b~gsthe dorsdizing components to the future dorsal side of the embryo. Thus, by the moruia stage of development the decision for specifjhg the dorsal and ventral symmetry has already been estabIished within the developing embryo. Misexpression of a dominant-negative form of GSK-3b in the ventral side of embryo results in the formation of secondary dorsal axis (He et al.,

1995). Consistent with this finding, ectopic expression of wild-type GSK-3 within the dorsal side ieads to ventralization. Taken together, the similar effects of ectopic expression of GSK-3 homologs from both Xerroptcs and sea urchin illustrates the conserved function of the Wnt pathway in regulating the specification of bilateral symmetry in vertebrates and invertebrates.

ln Xenopiis oocytes, another role of GSK-3 in the release of ceU cycle arrest has recently been uncovered (Fisher et ai., 1999). Prior to maturation the arrested oocytes contain constitutively active GSK-)I; progesterone-mediated inhibition of this activity appears to contnbute to meiotic maturation.

Caenorhabditis elegans gsk-3 displays a high degree of similarity to vertebrate

GSK-3 homolog and has been assigned rules in endoded induction and spindle orientation (Schlesinger et al., 1999; reviewed in Thorpe et ai., 2000). It seems iikely that gsk-3 serves as branch point for setectively regulating two Wntdependent processes during nematode deveIopment (reviewed in Thorpe et al., 2000). Drosophila melanogasrer represents the archetypa1organism for understanding of the WnüWingless pathway. The GSK-3 homoiogue 2~3~~~was 0riginaUy characterized as a segment-polarity gene and its role was impiied in ce11 fate determination (Bourouis et al., 1990; Siegfried et al., 1990). in Drosophila embryos, mutation of this gene gives rise to a phenotype where embryos Lack most of the ventral denticles normally present in the anterior region of each segment. Sequence comparison (90% identity in the kinase domain) as well as genetic studies demonstrate conservation of GSK-~IZW~~~~in mammals and insects (Ruel et al., 1993a; Ruel et al., 1993b; Siegfried et al,, 1992; reviewed in Woodgett, 1994). ZwPg is also essential for events such as mesoderm formation and cardiogenesis in Drosophila (Park et al., 1998). Using genetic and biochemical approaches, various components of Wnt/Wingless signalling have been ordered within the pathway. The embryonic phenotypes of armadillo//3-carenin and dishevelled mutants are very sirnilar to the dismption of wingless, whereas nvPghas a mutant phenotype very similar to that of embryos in which Wingless protein has been expressed in ail cells. These data Mersupport the notion that the functions of 2~3'~~ are antagonized by Wingless signalling (Peifer er ai., 1994a; Peifer et al., 1994b).

1.4.4 Mammalian GSK-3 and Wnt signaihg

Studies on the physiologicai functions of mammalian GSK-3 have taken important cues from the role of GSK-3 in simpier organisms, especiaily with respect to its regulation by the Wnt pathway. In mammals, however, the Wnt signaihg pathway is not ody important in early embryonic development, but is a target for tumorigenesis (reviewed in

Polakis, 2000). Inappropriate activation of Wnt signalling has been observed in various hurnan cancers, including hepatomas, colon carcinomas, melanomas, and uterine and ovarian cancers. The founding member of the Wnt family, Wnt-1, was originally identified as an oncogene comrnonly activated by the insertion of mouse marnmary tumor virus in virus-induced mammary adenocarcinomas (Nusse & Vannus, 1982). In addition to Wnts, the downstream signalling cornponents fi-catenin and FRAT Bequently

~arrangedin advanced -ceIl lymphomas, also known as BK-3 binding potein, GBP), both positive effectors of the pathway, have been identified as proto-oncogenes. The two negative regulators of the pathway, MC(adenomatous polyposis coli) protein and Auin, hction as tumor suppressors. Mutations of these Wnt signailhg pathway components have a similar consequence, namely accumulation of fi-catenin in the cytoplasm. increased cellular levels of P-catenin induce activation of downstrearn genes, some of which such as cyclin Dl (Tetsu & McCormick, 1999), WISP-1 (Xu et al., 2000) and c-

Myc (He er al.. 1998b): have been implicated in cancer.

The molecular cascade of events in mammaiian Wnt signaiiing is very similar to that in simpler organisms. In the absence of a Wnt signal, cytoplasrnic f3-catenin leveis are kept low via GSK-3 phosphorylation which targets P-catenin for ubiquith-mediated degradation by SlimbIf3TrCP (Jiang & Struhi, 1998; Maniatis, 1999; Figure 3). As in invertebrates, Axin also negatively regulates Wnt signalling and foms a complex with

GSK-3, p-catenin and APC (ikeda et al., 1998). Binding of Axin to GSK-3 promotes

GSK-3-dependent phosphorylation of both fi-catenin and Axin, the latter of which is necessary for Ain stability. GSK-3 also phosphorylates APC and this helps promote binding of APC to fi-catenin (Rubinfeld et al., 1996). This interaction is faditated by

Axin wsch associates with APC via a RGS kegulator of (i-protein simiallin) domain. Whether the interaction between APC and Axin is critical for Wnt signalling cemains to be established: although deletion of this domain has a dominant-negative effect on Wnt signalling in Xenopzcs, it is dispensable in Drosophila (Hamada et al., 1999; Itoh et al.,

1998).

ln the presence of a Wnt signai, the protein complex containhg GSK-3, Axin,

MCand B-catenin is disrupted, possibly by recruithg the GSK-3 inhibitor GBP/FRATl

(Yost et al., 1998; Jonkers er al., 19971, and GSK-3 kinase activity is dom-regulated.

GSK-3 phosphorylation of B-catenin is thus prtvented leading to accumulation of f3- catenin and subsequent interaction with HMG-box transcription factors of the LEF-l/Tcf

(Leukemia gnhancer lactoc-LE ceIl factor) family, and hence activation of specific target genes (reviewed in Eastman & Grosschedl, 1999). plasma membrane CKll

nuclear translocation ubiquitination gene transcription and degradation Figure 3. Role of GSK-3 in the Wnt signahg patbway In the absence of a Wnt signal, GSK-3 in the cytoplasm interacts with B-catenin, Axin and MC. GSK-3 phosphorylates these proteins, leading to the Slimbl$TrCP-mediated ubiquitination and proteolytic degradation of p-catenin. Upon binding of Wnt to a seven transmembrane domain receptor (Frizzled), Disheveiled is activated resulting in the downregulation of GSK-3 kinase activity. These, and other, signalling events and interactions stabilize fi-catenin, teading to the activation of LEF-11Tcf-mediated transcription (see text for Merdetails). Abbreviations: casein kinase-iI (Cm), protein kinase C-cc (PKCcc), Disheveiied @sh), Dishevelled-associated kinase (DAK), GSK-3- binding protein (GBP),adenornatous poryposis coii protein (APC). 1.4.5 GSK-3 phosphorylation of batenin and APC

GSK-3 can phosphorylate B-catenin on serine adthreonine at positions 33, 37, 41 and

45 (PoIakis, 1999; see Table 3). An N-tedly tnrncated form of 0-catenin, that Iacks the GSK-3 phosphorylation sites and thereby circurnvents the normal requirement for

Wnt signalling, is stabilized constitutively in vivo and is not impaired in its ability to bind

E-cadherin, u-catenin, and LEF-1/Tcf members. Mutations in P-catenin that impair or block phosphorylation of these sites have been identified in certain colon cancers

@articularly those in which APC is wild type). These P-catenin mutants have been usehi in delineating the physiological outcome of GSK-3 mediated P-catenin phosphorylation.

For instance, when an N-terminally tmcated human P-catenin mutant (AN87-Bcatenin) is placed under the control of the keratin KI 4 promoter and expression is driven in the basal Iayer of the mouse epidermis, these cells become capable of inducing foilicle morphogenesis normally occuning only in embryogenesis (Gat es al., 1998). In addition, these mice develop haïr fotKcle mors such as trichofollicuiomas and piiornatrichomas, which are often found in certain human familial polyposis syndromes (Chan et al., 1999).

An approach relying on Cre recombinase excision of 0-catenin exon 3, which contains

the GSK-3-phosphorylated residues, in the intestinal epithelium provided evidence for

intestinal adenornatous polyps in rnice at young ages (Harada et al., 1999). Interestingly,

the intestinal and coIonic tumors in these mice resembled those of the APC*"~ knockout

(Oshima et al., 1995).

Mammalian snidies have also suggested that a more complex rnechanism for the

regdation of p-catecin levels by GSK-3 involves MC(Rubinfeld et al., 1996). APC is

directly phosphoryiated by GSK-3 via Axin, which increases binding of APC to 0- catenin and its subsequent degradation. p-catenin contains a criticai sequence, made up of the N-terminal Armadillo repeats, that provides binding sites for the cytoplasmic

hgment of E-cadherin, the 15- and 20-amino acid repeats of APC, the N-terminal region of LEF-l/Tcf, and a central domain of Auin. E-cadherin, LEF-l/Tcf and APC compete

for binding to this region. Deletion mutagenesis indicates that ail these binding sites are

located in armadillo repeats 3-8 of P-catenin, which fonn a tightly packed superhelix.

Xenoprts extracts contain an activity that promotes the binding of P-catenin to

Axin at low concentrations (Salic et al., 2000). This activity is inhibited by expression of

RGS domain of Axin, suggesting that it is either APC or APC complexed to other

proteins. Also, aximîRGS (which cannot bind APC) does not bind P-catenin in extracts.

AIso, with purified components in vitro, APC accelerates the binding of B-catenin to

Auin, thereby recapitulating the effect of Wnt-stirnulated extracts on binding between the

two proteins. Taken together, these results identi@ APC as the activity stirnuiating the

axin-0-cdenin interaction. As a caveat, mutation of a Drosophila MChomolog does not

affect WntWg function (Hayashi et al., 1997), however additional APC-like molecules

exist in nies and there may be some functional redundancy between hem (McCartney et

al., 1999).

These general results are supported by experiments performed using SW480

colon carcinoma cells, which zontain a truncated nonfunctional APC, showing that

endogenous P-catenin is stabilized (Munemitsu et al., 1995). If APC is reintroduced in

these cells, fi-catenin is degraded. WhiIe endogenous Axin (and a related protein,

Conductin) is present in these cells, it is presumably incapable of regulating &catenin

phosphorylation in the absence of wild-type APC. In this way, overexpression of the Auin-related protein, Conductin, can induce j3-catenin degradation in SW480 cells, while a Conductin mutant which lacks the B-catenin binding domain canuot (Behrens et al.,

1998). More recentIy, SW48O ceils have been successfully utilized to Merexamine the functional interaction and stability of various B-catenin mutants (von Kries et al., 2000).

These experiments showed that P-catenin mutants that do not interact with Conductin

(and were therefore resistant to degradarion induced by ectopic expression of Conductin) were, however, efiecuvely degraded in the presence of wild-type APC. A double point- mutant of @-cateninthat bound to neither 15- nor 20-amino acid repeats of APC was still degraded by exogenous wiId type APC or a fragment of APC containhg only the 20- amino acid and SAMP (for Ser, Ala, Met and Pro) repeats, indicating that APC indeed does not need to bind directly to Pcatenin to induce bcatenin degradation. This is consistent with previous findings that APC and p-catenin can interact indirectly via

Conductin or Axin since they bind the SAMP repeats of APC (Behrens et al., 1998)-

Findly. only a mutant g-catenin incapable of binding either the 15- and 20-amino acid

repeats of APC or Conductin binding was tùlly stable in the presence of exogenous wild-

type MC.Thus, ail possibIe interactions between p-catenin and APC appear functionally

equivalent in the in vitro assays, namely the direct interaction via either the 15- or 20-

amino acid repeats, or the indirect interaction via ConductinlAwi. However, it was found

that APC with mutated SAhW repeats did not induce degradation of batenin in SW480

cells. The stability of B-catenin cm thus be regulated if a degradation complex with

Conductin and APC is fomed, that is, if B-catenin makes at least one interaction, with

either Conductin or APC. However, direct interaction between Conductinlaxin and APC

is required for this cornplex to be regulatory. This is supported by sequence data fiom human tumors, which showed that the majority of mutations in APC result in deletion of the SAMP repeats that provide an Axin interaction domain (Polakis, 1997). Many of these same mutants retain 15- and 20-amino acid repeat units that ailow fi-catenin binding. Thus, the tumocigenic potential of mutated APC conelares with the loss of binding to A.uin, rather than to B-catenin.

1.4.6 Regulation of GSK-3 by other signalling pathways

Glycogen synthase contains four GSK-3 targeted phosphorylation sites (reviewed in

Woodgett, 1994). Inhibition of GSK-3 reduces the level of phosphorylation of glycogen

synthase, which then becomes active for converthg glucose into glycogen. Stimulation of

glycogen synthesis by insulin also involves the dephosphorylation of serine residues in

glycogen synthase.

Several signal-dependent mechanisms have been proposed to lead to inhibition of

GSK-3 without involvement of the Wnt pathway (Figure 4). Although some studies

support the involvement of mitogen-activated protein kinase (MAPK)in GSK-3

regulation, other data do not. For example, in rat adipocytes EGF activates MAPK

without effecting GSK-3 activity (Moule et al., 1995). Similarly, the inhibitors of MAPK

have little effect on GSK-3 kinase activity. In vitro, the MAPKAPK-1 (MAPK-activated

protein kinase-1)-activated ribosomal S6 kinases p70 @70S6k) and p90 @90rsk) inhibit

GSK-3, however specific inhibitors of these protein kinases also fail to block signal

induced inhibition of GSK-3 in vivo (Cross et ai., 1995; Stambolic & Woodgett, 1994).

Many of the agonists that lead to suppression of GSK-3, including mitogens and

insulin, do so in a wortmannin-sensitive marner, implicating P13'K in the pathway. One of the most important physiologicd mediators of PI3'K signalhg is the protein-serine kinase PKB, also termed AKT, which exists as three isoforms in mammalian cells (a, and y). Mitogenic stimulation results in the phosphorylation of PKBu at two conserved sites: threonine 308 (T308) and serine 473 (S473; Alessi et al., 1996; reviewed in Coffer et ai., 1998). These phosphorylation events are dependent on P13' kinase activity and are required for full activation of PKB. In vitro, activated PKB phosphorylates serine 2 1 of

GSK-3u and serine 9 of GSK-8, both modifications causing inactivation. Antibodies that selectively bind the phosphorylated forms these sites on both GSK-3 subtypes have demonstrated a good correlation behveen P13'K activity and phosphorylation of serine

2119, dthough there are exceptions (see below).

Elevation of cyclic AMP within cells causes rapid activation of cyclic AMP- dependent protein kinase (PKA),and recently PKA has been show to directly bind and phosphorylate GSK-3 at serines 21 (GSK-3u) and 9 (GSK-3$) respectively (Fang et al.,

2000; Li et al., 2000). Another enzyme iniplicated in GSK-3 inactivation is integrin- linked kinase (ILK) which, like PKB, is also regdated by PU'K signalling. Human ILK phosphorylates and inhibits GSK-3 activity in vitro at an as yet undetermined site

(Delcornmenne et al., 1998).

Although GSK-3 can be inactivated by at least two distinct rnoiecular rnechanisms within cells, the relevant pathways appear to be independentiy insdated tiom each other to the point that effects on GSK-3 substrates are dependent upon the pathway by which the kinase is inactivated. This distinction is illustrated by genetic studies in Drosophila where overexpression of actived PKB faiis to generate a naked cuticle phenotype which is typicaI of activation of the WingIess pathway or disruption of 2~3'~~(Verdu et al., 1999). Furthemore, unlike injection of GSK-3, microinjection of activated PKB does not result in formation of a secondary axis in Xenopirs. These data suggest that distinct populations of GSK-3 eist witiiin cells that are somehow attentive to certain signals but not others. It is tempting to speculate that GSK-3 molecules sequestered into the AxinfB-catenin complex have adapted to a specific function that is entirely independent of the bulk of the GSK-3 moIecuIes within a cell. Proliferation

Apoptosi s Figure 1. Role of GSK-3 in phospbatidylinositol3-kinasesignalling

Upon activation P13' kinase is recruited to the plasma membrane and phosphorylates phosphoinositides at the 3'-position of their inositol ring. This, in tum, recruits PH-domain-containing proteins such as PKB and the PDKs. Once phosphorylsted by the PDG, PKB is activated and phosphorylates GSK-3 leading to its inhibition. See te= for details. Abbreviations: growth factor (GF), phosphatidyiiiositol 3 -kinase (PI3 'K), PI3'K-dependent protein kinase- 112 (PDK1/2), protein kinase B (PKB), phosphoinositide (Pi), phosphoinositides (J,j)P2 (PIPz), phosphoinositides (3,4,5)P3 (PIP3), integrin-linked kinase (LK). 1.47 Lithium as an inhibitor of GSK-3

The alkali metal lithium was frrst discovered in 1817, and over the ensuing period of

time, lithium has been utilized in various forrnuIations as a remedy for a multitude of

human maladies (Johnson, 1984). With the seminal work of Australian

physicidscientist, John Cade, and subsequent clinical studies by Mogens Schou in the

eady 1950s, Iithium was introduced as an effective therapy for manic-depressive illness

(bipolar affective disorder; Cade, 2000; Schou 1979).

More recently, lithium has also been shown to pecturb the development of diverse

organisms, including Xenoptrs, zebrafish, sea urchùis and Dicryosrelium. For instance, in

Dicyosrelitrm, lithium alters ce11 fate determination, blocking spore ce1L formation and

promoting stalk development (Maeda, 1970). In Xenopits embryos, lithium treatment

causes an expansion of dorsal mesoderm, leading to formation of a second dorsal axis

(Mao er al,, 1986). Treatment of sea urchin anima1 blastomeres with Lithium causes them

to dispiay a morphology resembling that of isolated vegetd bIastomeres (Horstadius,

1973). Importantly, these effects have since been shown to be similar to embryonic

patteming defects resdting from either disruption of the GSK-3 gene or overexpression

of inactive forms of GSK-3 that have a dominant-negative activity.

The developmentd and neuropsychiatrie effects of lithium have been attributed to

a variety of biochemical processes including modulation of G proteins and inhibition of

inositol monophosphatase (IMPase), an enzyme important for recyciing of rnyo-inositol

in the phosphatidylinositide (PI) pathway (Berridge et al., 1989). Exposure to lithium

reduces the ievel of myo-inositol in the rat brain (Ailison & Stewart, 1971) via

noncornpetitive inhi'bition of MFase (HalIcher & Sherman, 1980) and to a lesser extent, inositol polyphosphate- 1-phosphatase wP-1; Inhom & Majerus, 1987). However, an aitemative target of lithium was suggested by the close similarity between lithium action

(Kao & Elinson, 1989; Maeda, 1970) and the effect of either GSK-3 disruption in

Dicryostelirrm (Harwood et al., 1995) and ectopic expression of Wnt genes in Xenopus embryos (McMahon & Moon, 1989). While examining lithium-mediated effects on dorsaihentral patterning in Xenopus embyos, Klein and Melton (1 9%) observed that injection of a potent and selective IMPase inhibitor [L690,330; Atack et al., 1993) did not phenocopy lithium. The nature of the connection between lithium and the Wnt pathway was revealed by the finding that lithium inhibits the activity of purified GSK-3a and B

(Klein & Melton, 1996). This frnding was extended to GSK-3 function in intact cells

(Stambolic et al., 1996). These resdts directed attention towards GSK-3 as an important cellular target for lithium action,

Lithium is highiy selective for GSK-3. While one report suggested that lithium could activate PKB/AKT (Chdccka-Franaszek & Chuang, 1999), this result has not been confirmed. Lithium has thus been ernployed as a cool to study the role of GSK-3 various cellular processes. Although the pleiotropic nature of the ion on cells means that if a

Lithium effect is observed this may not be mediated by GSK-3, if lithium has no effect on a process, a role for GSK-3 in that process can usually be excluded.

Additional ATP competitors specific for GSK-3 (Ki 10-30 nM) have been recently reported to stimulate glycogen synthesis in liver ce11 Iines and induce the transcription of 0-catenin-regulated genes in epitheliai cells (Coghian et al., 2000). The availability of these inhibitors wiii greatiy assist dissection of the physiologicai processes

influenced by GSK-3. 1.4.8 Summary

As described, GSK-3 is an unusual enzyme that has surfaced in several unexpected areas of biology. True to fom, mutational inactivation of GSK-3g in mice revealed a surprising and unpredicted role in the regulation of NF-KB and in TNF sensitivity

(Chapters 3 and 4). This thesis describes my efforts to Merunderstandimg of TNF

function through two distinct pathways. These results demonsirate how much more is to

be learned about TNF biology and provide new insights and targets for possibk

therapeutic interventions- Mediation of TNF receptor-associated factor effector functions by

apoptosis signal-regulating kinase4 (ASK1)

A version of this chapter is published in Oncogene 1999 Oct 14;18(42):5814-20 2.1 Abstract

Tumor necrosis factor-cr (TNF), a major inflanmatory cytokine, generates a wide variety of cellular responses via key cytoplasmic adaptor molecules named TNF receptor-associated factors

(TRAFs). We report that TRAF2, TRAFS and TM6 associate with apoptosis signai- regulating kinase 1 (ASKI), and a catalytically-inactive ASKl mutant blocks stress-activated protein kinase (SAPK)/ Jun Ml-terminal kinase (M)activation by these TRAFs. A tnincated derivative of TRAFî, which inhibits SAPK activation by TNFI blocks TNT-induced ASKl activation. Furthemore, protection from TNF-induced ceH death confered by an ASKl mutant is dependent upon TRAFî. Hence, ASKI is a common mediator of W-regulated SAPK and apoptosis signallulg, and the TRAF2-ASKI connection completes the signalling cascade fiom

TNF to SAPWJNK activation.

2.2 Attributions

Plasnids for wild-type and kinasedead ASKI were provided by Zhengbin Yao. Mike Rothe provided the TRAFS cDNA and Wen-Chen Yeh kindly supplied the TRAF2 knock-out fibroblasts. 2.3 Introduction

Tumor necrosis factor-a (TNF)is a pieiotropic cytokine whose varied biological activities are sipaled by triggering the aggregation of receptor monomers (Tartagiia and Goeddel, 1992). These receptors, tenned TNFRl (p55) and TNFK @75), are expressed in most ceU types and are members of the growing TNF receptor superfarnily, which inciudes the Fas antigen and CD40

(Smith er al., 1994; Nagata, 1997; Gruss, 1996; Naismith and Sprang, 1998). Upon receptor aggregation, several cytoplasmic proteins, including TNFR1-associated death domain protein

(TRADD)(Hsu et al., 1995; Hsu et al., 1996b), receptor-interacting protein (RIP) (Hsu et al.,

1996a). and TNF receptor-associated factor (TRAF) faniily members (Rothe et al., 1995) are then recruited to the intraceliular domain of TNFRl where they then form an active signahg complex.

To date, six members of the TRAF farniiy have been found (Rothe et al., 1995; Cheng et al., 1995; Regnier et al., 1995; [shida et al.) 1996b; Cao et al., 1996; [shida et ai., 1996a; kch et al., 1998). AI1 contain a conserved C-temiinal TRAF domain that mediates homo- or hetero- oligomerization among the TRAF family and interactions with the cytoplasmic regions of the

TNF receptor superfamily (Rothe et a[., 1995). In addition to the TRAF domain, most TRAF proteins contain an N-terminal RiNG finger and several zinc hger structures that appear critical for theu effector fûnctions (Arch et al., 1998). In addition to the signalling events initiated by the

TM: receptor farnily members, T'Rus may also link other receptor-mediated simialling pathways. For example, TRAF6 is a component of interleukin-1 receptor @-IR) signaiüng compIex, in which it induces the activation of nuclear factord (NF-KB) and the stress-activated protein kinase (SAPK)/c-Jun N-temiinal kinase (M)farnily foliowing IL-1 binding (Cao et al.,

1996).

Although three of the known TRAF proteins (TRAFS, TRAFS and TRAF6) are able to mediate the activation of SAPK (Song et al., 1997), the rnolecular mechanisms coupling the

TNFR signaüiig complex to SAPK have not ken well characterized. However, considerable progress has been achieved in identifjkg the kinases immediately upstrearn to SAPK and the other mernbers of the MAPK family (Kiefer et al., 1997; Kyriakis et al., 1994). SAPK is phosphorylated and activated diectiy by the duai-specificity MAPK kinases (MAPKKs),

SEKllMKK4 (Sanchez et al., 1994; Lin et al., 1995) and MKK7 (Tournier et al., 1997; Holland et al., 1997; Yao et al., 1997; Moriguchi et al., 1997), which in turn are substrates of a group of

highly divergent MAPK kinase kinases (MAPKKKs) (Kiefer et al., 1997). A recently

characterized MAPKKK, apoptosis signal-regulating kinase 1 (ASK1; also cailed MAPKKKS),

can activate SEKl and MKK7 (SAPK pathway), and MKK.3 and MKK6 (p38 MAPK

pathway) by direct phosphorylation (Ichijo et al., 1997; Wang et al., 1996). ASKl is responsive

to TNF treatment in many ce11 types, and expression of dominant-negative ASKl inhibits TNF-

induced apoptosis in Jurkat and 293 cells (Icbijo et al., 1997). In addition, ASKl has ken shown

to be activated by reactive oxygen species (ROS) generation, and to be inhibited by the

antioxidant, thioredoxin (Saitoh et al., 1998; Gotoh and Cooper, 1998). It remains to be

determined how ASKl is regulated by TMF signalling,

In the present study, we show that ASKl associates with members of the TRAF fady,

and that these interactions reqyire the conserved Mi1-terminal zinc EüNG and TRAF-N motifs

typical to TRAF farnily members. Activation of ASKl by TNF cmbe blocked at the level of the TRAFs, while a catalyticaiiy inactive ASKl mutant inhibits SAPK activation by TRAFî-,

TRAFS and W6.interestingly, cataIyticaiIy inactive ASKl can only rescue cells fiom TNF- mediated apoptosis in the presence of endogenous TRAFî, as determined by comparing TRAF2- deficient and wild-type control fibroblasts. Together, these data demonstrate a role for the

TRAFî-ASK1 interaction in both SAPK activation and ceIl death signalhg

2.4 Materials and Methods

2.4.1 cDNA constructs

Hemaggiutinin (HA) epitope-tagged expression plasmids for ASKl/MAPKKKS, cataiytically

inactive ASKlX709E (Wang ez al., 1996), pS4-0-SAPK (Pulverer et ul., 1991), kinase-inactive

SEK1-KR (Yan et al., 1994), and FLAG epitope-tagged TRAM (Song et al., 1997) have ken previously described. TRAF2 and TRAF6 cDNAs were isolated by reverse transcription-PCR,

and the fuii-length coding regions were cloned into the eukaryotic expression vector pCR3.1

(Invitrogen), and in the process were hsed to an N-terminai FLAG epitope tag (SIGMA).

Expression vectors for TRAF2 deletion mutants were constructed by PCR amplification of

TRAFî coding sequences using oligonucleotide pairs as shown, and tigating the resulting

fragments into pCR3.1: for TRAF2(98-SOI), primers 69 (5'-CTA CCA TGG ACT ACA AGG

ACG ACG ATG ACA AGA GGG AGG TGG AGA GCC TG-3') and 54 (5'-ïTA GAG CCC

TGT CAG GTC-3'); for TRAE(272-Sol), primers 63 (5'-CTA CCA TGG ACT ACA AGG

ACG ACG ATG ACA AGT GCG AGA GCC TGG AGA AG-3') and 5% for TRAF2(1-271),

primers 48 (5'-CTA CCA TGG ACT ACA AGG ACG ACG ATG ACA AGA TGG CTG CAG CTA GCG TC-3') and 59 (5'-TTA CCT CTG CAG GAG CTC TGA-3'); for TRAFS(1-

353, prirners 48 and 64 (5'-TTA GAT GAA GAC CCC ATC GTA G-3'); for TRAF2(272-

358), primers 65Rev (5'-CTA CCA TGG ACT ACA AGG ACG ACG ATG ACA AGT CAG

AGC TCC TGC AGA GG-3') and 66Rev (5'-GAA AGT CTG AGA TCT TCC A-3'); and for

TRAF2(1-336), primers 48 and 88 (5'-TTA GTC GAG CAG CAT TAA GGT C-3'). GST-c-

Jun(5-89) fusion proteins were produced in the pLysS(BL21DE3) strain of Escherichia coli

using the pGEX expression system (Promega). Proteins were affinity purified on glutathione-

agarose beads and eluted with 10 mM reduced glutathione, 50 mM Tris pH 8. Eiuted proteins

were dialysed against a buffer containing 50 rnM Tris pH 8, 150 rnM NaCI, 5 mM EDTA, 0.1%

Triton X-LOO, 0.1% 2-mercaptoethanol, and 50% glyceroi. Restriction enzymes, and

polymerase were obtained From New England Biolabs or as indicated. Oligonucleotides were

purchsised frorn GIBCO.

2.42 Cell culture, transfection and treetment

293, COS-7 and mouse ernbryonic fibroblast ceUs were maintained in Dulbecco's modified

EagIe's media supplemented with 10% fetal caifserum. CeUs were plated at a density of 2 x 106

ceIid100-mm dish and transfected the next day using the caicium phosphate transfection

protocol. The amounts of DNA used for transfection were as indicated, and the empty vectors

were used to achieve equal amounts of DNA in each transfection. Where added, recombuiant

human RJF-a (Genetech) was used at a final concentratron of 100 ng/mI (30 min prior to lysis).

Ce11 iysis was performed in Gentle-Soft Iysis baer with inhibitors (ImM dithiothreito1, 100

pA4 sodium orhovanadate, Spg/mi Ieupeptin, 50 pM sodium fluonde, and 1 mM beIlzamidine) at SOC. Lysates were cleared by centdùgation at 13,000xg for LOmin, and then normaliztd for

total protein by Bradford assay before hunoprecipitation.

2-43 Immunoprecipitations and immunoblotting

Immunoprecipitations were performed using monoclond antibody agaht the HA tag generated

from cuIture supernatants of 12CA5 hybridoma (ATCC), monocIona1 antibody against the

FLAG epitope (Kodak), polyclonal rabbit antibody mked against murine TRAFî (UBI), or

poiyclod ASKl antibody (Santa Cruz Biotechnology). Epitope-tagged proteins were

imunoprecipitated by incubating the appropriate ceil lysates with 20 pl anti-HA or 0.2 pg anti-

FLAG antibodies for one hour and harvested with 10 p1 of protein G-Sepharose (Phmacia).

Endogenous TRAF2 was precipitated by incubation with t pg TRAF2 antibody and harvested

on protein A-Sepharose beads. The hune complexes were washed four tirnes in PBST

( 1SOrnM NaCL, 16 mM Na2HP04,4 mM NaH2P04, 0.1% Triton X-100, ZmM dithiothreitol,

100 pM sodium orthovanadate, 5pghd leupeptin, 50 pM sodium fluoride, and 1 rnM

benzamidine). For analysis of protein-protein interactions, immune complexes were resolved by

SDS-PAGE and immunobiotted with the indicated antibodies. A polyclonal rabbit antibody

nised to phosphorylated threonine 223 of human SEKLfiMK4 (New England Biolabs) was

used to detect p hosphorylated SEKl pro teins. 2.44 Kinase assay

To assay SAPK and ASKl activity, an equai volume of protein nonnaiized HA-SAPK or HA-

ASKl ce11 lysates, respectively, were incubated with 20p1 anti-HA antibodies for 1.5 hours and harvested with 10 pl of protein G-Sepharose (Pharrnacia). After 4 washes with PBST, the beads were hcubated with glutathione S-transferase (GST)-c-Jun(5-89) (SAPK substrate), GST-SEKI-

KR (ASKI substrate) or myelin basic protein (ASKI substrate), 40 pM ATP, 10 mM MgCl>

50 mM Tris-HCL pH=7.5, 1 dlEGTA, and 40 pCW [Y-~'P]-ATP at 30°C (Adler et al.,

1992). After 20 minutes, the reactions were terminated by addition of Laemmli sarnple buffer.

Sarnples were boiled, and phosphorylated proteins were resolved on sodium dodecyl sulfate-

12.3% polyacrylamide gels and visuaiized by PhosphoImager anaiysis (Molecular Dynarnics).

2.4.5 Embryonic fibroblast death assay

EF cells were grown in Dulbecco's modified Eagie's medium supplemented with 10% fetal caif senun. Cells were transfected at 50% confiuency in 6-weii dishes with 0.5 pg of pCMV-fi- galactosidase plus 2.5 pg of empty control vector or pCDNA3-ASK1-K709E. Mer 24 hr, recombinant mouse TNF-cc (Gibco BRL) was used at IO ndml, and cycloheximide (Sigma) was used at 250 ng/d for 24 hr, fbGaiactosidase cotransfection assays were performed as previously desmied (Kumar et ai., 1994), and cells were visualized by phase-contrast rnicroscopy, and the number of blue cens per weii deterrnined by counting. 2.5 Results

2.5.1 ASK1 and TRAF proteins associate in vivo

To begin to assess whether TM2, TRAFS and TRAF6 may regdate ASKl we determined whether TRAFs and ASKl could associate in a transfection-based, CO-immunoprecipitation assay. 293 and COS cells were transiently transfected with equal amounts of hemagglutinin (HA) epitope-tagged ASKl and untagged TRAF2, and their interaction was determined by co- imrnunoprecipitation followed by Western blotting. We found that ASKl associated with

TRAF2 as demonstrated by the CO-irnmunoprecipitationof HA-epitope-tagged ASKI with anti-

TRAF2 antibodies (Figure 5A1 left panel) and CO-immunoprecipitationof TRAF2 with anti-HA

(Figure SA! right panel). Western blots confirming equd expression of TRAF2 or ASKl are show. Shilarly, we also coexpressed HA-ASK1 with FLAG epitope-tagged TRAF5 or TRAF6 proteins and found association of ASKl with these TRAFs (Figure 33 and SC, respectively). A HA-ASK1 + + + + TRAF2 + + + + IP: HA T2 T2 T2 T2 HA HA HA 1234 5 678 WB: HA

B HA-ASK1 + + + + TRAF5 + + + + IP: HA F F F F HA HA HA 1234 5 678 WB: HA

C HA-ASK1 + + + + TRAF6 + + + + IP: HA F F F F HA HA HA 1234 5 678 WB: HA WB: FLAG

TRAFP lP ASK1 w- - anti-HA blot I .- -- anti-TRAF2 blot

ASKl .) anti-HA blot Figure 5. ASUassociates with TRAF2,TM5 and TRAF6 (A) 293 ceLis were transiently cotransfected with HA epitope-tagged ASKL and TUF2 constructs as indicated. ln lane 4, HA-ASK1 is present in the TRAF:! immunocomple.u, demonstrating an interaction between ASK and TRAï2. Shown is an anti-TRAF:! immunoblot of ceU Iysates to venfy the presence of TRAF2. Sùnilarily, in the next panel, lysates were immunoprecipitated with anti-HA antibody and immunobloned with anti-TRAF2 antibody to demonstrate that only CO-expressionof TRAF2 with ASKL resulted in detection of TRAFZ in the HA-ASK1 imrnunocomplex. The presence of HA-ASK1 was verified by immunoblotting. A sùnilar approach was taken to demonstrate the interaction between (B) HA-ASK1 and FLAG- TRAFS, and (C) HA-ASKL and FLAG-TRAF6. (D) Mouse embryonic fibroblasts generated hmwild-type (lanes L,4 and 5) and TRAF2-deficient (laues 2 and 3) mice were transfected with HA-ASKL (lanes 1, 3 and 5) or control vector (lanes 2 and 4). Mer 36 hr, lysates were prepared and incubated with the indicated antibody (anti-HA monoclonal antibody or anti- TRAF2 polyclonal antibody). HA-ASK1 CO-precipitating with endogenous W2was detected by immunoblot anaiysis using the anti-HA antibody, As a control for the presence of TRAF2, an equal amount of each imrnunoprecipitate was resolved by 12.5% SDS-PAGE and imunoblotted with anti-TRAFî, (bottom panel). To coh that these interactions were specific, ASKl and TRAF2 co- imrnunoprecipitations were repeated in embryonic fibroblasts (EFs) generated fiom TRAFî- deficient or wild-type rnice. Endogenously expressed TRAF2 was precipitated with anti-TRAF2 antibody and the immune complexes were subjected to immunoblotting with anti-HA antibody to detect the presence of transfected HA-ASKl. J4.A-ASKl was only found in TRAF2 irnmunoprecipitates from wild-type EFs (Figure 5D, lane j), and not in rraj2 '-extracts (lane 3).

The presence of TRAFS (Figure 5D. middle panel) and HA-ASKI (bottom panel) in the same lysates was shown. Taken together, these results suggest that ASKl tightly associates with

TRAF proteins.

2.5.2 Mapping of in vivo interaction domains between ASKl and TRAF2

Sequence digrnent oFTRAF family members reveals several highly conserved regions which are knotvn to mediate interactions with targets (Takeuchi et al., 1996). To defme the sites of interaction between ASKl and TRAFî,, we performed CO-immunoprecipibtionswith wild-type

ASKl and TRAF2 mutant proteins cotransfected into 293 ceils (Fig. 6A). Deletion of the entire

C-terminal TRAF domain (TRAF2(1-272)) resulted in a significant reduction of ASKI binding

(Figure 6B, lane 6), while deletion of only the TRAF-C subdomain (TRAF2(1-355)) had no adverse effect of TRAF2 association with ASKl (lanes 4 and 5). Hence, the TRAF-N domain, which is predicted to adopt a coiled-coi1 configuration, appears to be essential for TRAF2-ASK1 interaction (Iane 6). However, neither the whole TRAF domain (TRAF2(272-501)) nor the

TRAF-N region (TRA.F2(272-358)) alone are capable of CO-immmoprecipitatingASKl (ianes 8 and 9). Rather, the RING hger domain is &O required for effective ASKl binding, since in the absence of this region binding to ASKl was strongly impeded in this assay (lane 7). Taken together, these results indicate that the TRAF-N and RING hger domains both contribute to

ASKl binding by TRAF2.

2.5.3 TNF-dependent endogenous interaction between TRAFZ and ASKl

Given that the in vivo physical association between TRAF2 and ASKl might simply result from their overexpression with a constitutive promoter, we tested whether this interaction occured under more physiological conditions, As a control, our anti-ASK1 antibody effectively irnmunoprecipitated HA-ASKl (Fi,we 7A). We then did a tirne-course of TNF-stimulation (200 ng!ml) and harvested the cells. Immunoprecipitation of proteins from cellular extracts with an anti-TRAF2 antibody, followed by immun0blotting with antibody against ASK 1 demonstrated coprecipitation of these two proteins (Figure 7B). The maximum interaction was observed at 15-

30 minutes of TNF treatrnent and decreased thereafler. As a control, ASKl was not identified foliowving immunoprecipitation with an irreievant antibody, Therefore, we can conclude that endogenous TRAF:! and ASKl physicaily associate in a TNF-dependent manner. A RING Zn finger TRAF-N TRAF-C 1 TRAF2(1-501) I 1 1 98 271 355 50 1

1TRAF2(1-436) 1 TRAFP(1-355) TRAF2(1-271) -y--1 TRAF2(98-501) =-1 =-1 TRAF2(272-501) = TRAF2(272-358) Figure 6. Interaction of ASKl with TRAF2 mutants

(A) The domain structure of Ml-Iength TRAF2 is shown, with amino acid numbers iisted underneath, The various TRAFS mutants are indicated diagrammaticaiiy. (B) 293 celis were transientiy cotransfected with expression vectors for HA epitope-tagged ASKl and the indicated FLAG epitope-tagged wild-type and mutant TRAFZ proteins. Mer 36 hr, aliquots of ceil lysates were immunoprecipitated with anti-FLAG monoclonal antibody. Coprecipitating HA- ASKl was detected by ïmmunoblot andysis with anti-HA antibody. The relative expression levels of HA-ASKL and each of the FLAG-TRAF2 deletion mutants was measured by Western bIotting of whole-ce11 extracts Fom transfected cek with anti-HA or anti-FLAG antibody, respectively, A HA-ASKI ASKl IP anti-HA blot

TRAF2 IP anti-ASK1 blot

1mti-ASKl blot

anti-TRAF2 blot Figure 7. Physical association of endogenous TRAF2 and ASKL (A) Verification of ASKl polyclonal antibody. 293 cells were transfected with HA-ASK1 (lane 1) or contro1 vector (Iane 2). CeU lysates were immunoprecipitated with anti-ASK1 antibody, separated by SDS-PAGE and blotted with anti-HA antibody. (B) Immunoprecipitation-Western andysis of extracts from ceUs stimulated with 200 ng/rnl hTNF-a for the indicated times (in minutes). Equal arnounts of cellular lysates were immunoprecipitated with either anti-TRAFî antibody or a nonspecific control antibody, followed by immunoblotting andysis with anti- ASKl antibody. Cellular lysate (1120 the arnount imrnunoprecipitated) was analyzed directly to show the migration position of native ASKl (middle) and TRAF:! (bottom). 2.5.4 TRAF activation of the SAPK pathway is mediated by ASK1

The ability of ASKl to bind the three TRAF farnily rnembers known to activate SAPK and NF-

KB caised the possibility that ASK1 fùnctions as a component of TRAF-mediated signahg events. In particular, the obsewed interaction of ASKl with the SAX-kinases SEKl (Ichijo et ai., 1997; Wang et al., 1996) and MKK7 (Z. Yao, unpublished results) suggested that ASKl might be a dowmtream event in TRAF-mediated signai transduction. Consequently, we meamued the effect of a kinasedetiçient ASKl mutant, ASK1-K709E, on N2-,TRAFS- and TRAF6- rnediated SAPK and NF-KR activation. To examine whether ASKl mediates SAPK activation by

TRAFs. wc CO-transfectedCOS and 293 cells with mammaiian expression plasmids encoding

ASK1-K709E, HA-SAPK and TRAFs. Recombinant SAPK was imrnunoprecipitated fiom ceU

Iysates and assayed with GST-c-Junas substrate. Transfection of ceils with ASKI resulted in strong activation of SAPK (data not shown), conhnkg previous observations (ichijo er al.,

1997; Wang ef al., 1996), whereas transfection oicells with vector done or ASKl-K709E had no effect on SAPK activity (Figure 8A, lanes I and 2). Co-transfection of ASKI-K709E dong with

W7, and HA-SAPK into COS cells reduced SAPK activity 75% in the HA inununoprecipitates as compared with Wî, and HA-SAPK alone (lanes 3 and 4). Comparable results were obtained with WSand TRAF6 (lanes 5-8), akhough TRAFS-induced SAPK activation consistentiy was the least sensitive to the inhibition by ASK1-K709E. These resuits indicated that TRAF activation of the SAPK pathway rnight be mediated by direct effects on

ASKI. A TRAF2 + + TRAF5 + + TRAF6 HA-ASKI -K709E + + HA-SAPK + 4- + +

-1 -1 anti-HA blot

ASKl anti-HA blot Figure 8. ASKl is downstream of TRAFs (A) Constitutively inactive ASKl mutant inhibits TRAF-mediated SAPK activation. COS-7 cells were transiently cotransfected ulth HA-SAPK (5 pg; lanes 1-8), expression vectors for ASK1- K709E (5 pg; lanes 2,4,6,8), TRAFî (5 pg, lanes 4,5), TRAF5 (5 pg, lanes 5,6), TRAF6 (5 pg, lanes 7,8), or vector control to bring total DNA to 15 pg. FoUowing immunoprecipitation with ad-HA antibodies, SAPK activity was assessed using GST-c-Jun(5-89) as a substrate. TRAF2, TRAFS and TRAF6 induce activation of SAPK (lanes 3, 5, 7, respectively). However, in the presence of ASK1-K709E, SAPK levels are significantly reduced (ianes 4,6, 8). As a control for SAPK or ASK1-K709E protein expression, an equal arnount of each ce1 lysate was resolved by 12.5% or 6.5% SDS-PAGE, respectively, and immunoblotted with anti-HA antibody. Simiiar results were obtained in atleast three independent experiments. (B) Dominant-negative TUF2 can suppress ASKl activation by TNF-u as rneasured by ASKl phosphorylation of SEKl-KR. 293 cells were CO-transfectedwith HA-ASK1 (lanes 2-4), TRAF2(272-501), a dominant-negative TRAFî,expression constnict (lane 4), or control vector to bring total DNA for each transfection to 15 pg. Cells were stimulated with TNF-u as indicated. HA-tagged ASKl proteins were imrnunoprecipitated with anti-HA antibody and then incubated in kinase buffer with GST-SEK1- KR and ATP at 30°C. Proteins were resolved by SDS-PAGE, transfered to PVDF membrane by Western blotting, and immunobIotted with a polyclonal antibody raised to phosphorylated SEK1. In lane 1, no ASKl was transfected to assess background activity. ASKl activity increased signif~cantlyupon addition of TNF-u (lanes 2 and 3). However, upon TRAF2(272- 501) CO-transfection(lane 4), ASKl levels were reduced to levels lower than its basal activity. The relative expression levels of HA-ASK1 were rneasured by immunoblotting of whole-ce11 extracts fiom transfected cells with anti-HA antibody. (C) Inhibition of ASKl by TRAF2(272- 501) in MBP kinase assay. 293 ceUs were transfected as described in (B). Immunoprecipitated HA-ASK1 kinase activity was meaured by MBP as a direct substrate for ASKI. ASKl activity was then determined by Phosphoimage analysis. ASKl activity was elevated by RIF-a (lane 3), but was attenuated upon CO-transfectionof dominant-negative TRAF2 (lane 4). The uihibitory effect of ASKl-K709E on TNF- and TEL4.F-mediated NF-@ activation was uivestigated using a NF-KB-dependent reporter gene. In this experiment, overexpression of

ASK1-K709E had no effect on the activation of the KB-containing E-selech promoter tuciferase reporter gene by TNF, TRAF2, TELUS or TRAF6 (data not shown). This indicates that ASKl is specifically required for TELU-mediated SAPK activation.

2.5.5 TRAFZ regulates ASKl activation by TNF if ASKl acts downstream of TRAFs, a dominant negative TRAF:! mutant should block TNF stimulation of ASKi kinase activity. Therefore, we examineci whether TRAF2 could modi@ the

TNF-induction of kinase activity of transfected ASKI cDNA. lmmunoprecipitation of ASK 1 followed by kinase assays demonstrated that TNF consistently increased ASKl phosphorylation of SEKl and myelin basic protein (MBP) approxirnately Zfold in each case

(Figure 8B and 8C, lane 3). By cornparison, in the presence of TRAFZ(272-5011, a deletion mutant known to be defective in activating SAPK, ASKl kinase activity in response to TNF was significantly reduced (approximately 6-fold, lane 4). Western blot analysis of extracts used in

Figures 8B and 8C showed simiiar expression of ASK1. The fïndùig that a TRAF2 mutant can block ASKl activation by TNF is consistent with the positioning of ASKl downstrearn to

W2in TNF signabg.

Figure 9. Mecianism of ceIl death inhibition by ASK1

The indicated ceik ünes were transiently CO-transfectedwith 0.5 pg pCW-&al and 2.5 pg of ASKI -K709Eor control vector as indicated. After transfection (24 hr), 10 ng/d TNFa and 250 ngmi cyc!oheximide (final concentrations) were added for 24 hr. Subsequently, cells were fixeci and stained with X-Gd. Data (mean k SEM) is plotted as the number of blue celis per well (as a percentage of the unstimulated control) for at least three independent transfections. HIAPllHIAP2 anti-Myc blot

TRAF2 anti-FLAG blot .. . . . - - -. . .. , . - .. .

ASKl anti-HA blot Figure 10. ASKi does not inhibit the association of TRAF2 with HIAPl or W2 293 cells were transfected with the indicated pIasmids ztnd the lysates immunoprecipitated with anti-FLAG antibody. Western blot analysis of immunoprecipitated TRAF2 and coirnmunoprecipitated HlAPl and HIAP2 (iipper panel). Western blot anaiysis of total ceIl lysates are also shown. 2.5.6 Overexpression of dominant-negative ASKl inhibits TNF-mediated apoptosis only in the presence of TM2

To Merstudy the functional rok of the ASK1-TRAFS interaction, wild-type and TRAF2- deficien: mouse embryonic fibroblasts (EFs) were transfected with pairs of expression ptasmids encoding kgal and either ASKlX709E oc control vector at a 1:s' ratio, respectively (Yeh et al.,

1997). 24 hr later. cells were treated with murine TNF-CI (10 &ni) and cyclohe.uimide (0.25

@ni) for a subsequent j0 hr. Under &ese conditions, TRAF2-deficient cells begin to die as early as 6 hr after treatment and less than 5% of the mutant cells survive by 30 hr dertreatment.

Similar treatment has a less adverse effect upon wild-type EFs and approximately 20% survival is generally seen after 30 tir. Following treatment, cells were stained with X-Ga1 and morphologically normal. positive blue cells were visualized and counted. A quantitative representation of the data fiorn three independent transfectioos is show in Figure 9. in wiid- type, but not trap'*,fibroblasts a 3-fold increase in the viability of fLgalactosidase-positive ceUs was observed following expression of ASK1-K709E compared with the control vector

(Bonferroai :-test, p<0.05). The wiId-type EFs transfected with ASKIX709E were also signifïcantly protected fiom the accompanying morphologicai alterations of apoptosis (data not shown). In contmst, TNT-treated wild-type EFs expressing control vector displayed morphobgical aiterations rypical of adherent cens undergohg apoptosis, becomiug rounded, condensed, and detacfiing fiom the dish (data not show@- Taken together, this data suggests that

ASKI-K709E overexpression effectively blocks a TRAF2dependent, ïNF-induced apoptosis cascade. Hence, in addition to tbe ce1 suMvd pathways mediated by TRAF2, its interaction with ASKI cm also likely lead to programmed ceIl death.

2.6 Discussion

The RIF receptor-associated factor (TRAF) family of adaptor proteins mediate signais from ce11 surface receptors to various protein kinase cascades, including the stress-activated protein kinase

(SAPK) pathway (kch et ai., 1998). Since TNF-induced SAPK activation is defective in

TRAF2-deficient cells (Yeh et ai., 1997), TRAF2 appears to regulate SAPK activation at a very proximd step. The data presented here suggest that TRAF2 transduces signais for SAPK activation via the MAPKKK ASK1. During preparation of this manuscript, similar Lindigs were reported by Nishito et al. (Nishitoh et al., 1998). Our analysis of the association of TRAF2 with

ASKl is consistent with the current ~0ddof RJF signaiiing, wherein bifùrcation of the TNF- induced NF-KB and SAPK activation pathways occurs at the level of TRAF2 (Song et ai., 1997).

Comparable to the role of ASKl in SAPK activation, another MAPKKK homologue, NF-&- inducing kinase (NIK), has recently been identifïed as a TRAF'I-interacting protein and a common downstream mediator of NF-KB activation by TNF and TRAFs (Malinin et al., 1997;

Natoli et al., 1997). Overexpression of MK appears to be specific for NF-KB induction and has no effect on SAEK activity (Natoli et al., 1997; Song et al., 1997)- By cornparison, TRAF2,

TRAFS and TRAF6 dso activate ASKI ,which in tum activates the SAPK pathway, but not NF-KB. Collectivtly, these results further demonstrate that despite several cornmon mediators,

SAPK and NF-@ activation by TEiSF are two separate responses.

TNF is aiso known to upregulate a third MAPKKK, MAPEUERK kuiase kinase 1

[MEKKl), another activator of the SAPK pathway (Yan et al., 1994). It has been proposed that the carboxy-terminal cegulatory domain of germinal centre kinase (GCK), can bind both TRAFî and MEKKI, thereby constituting an ASKI-independent signalLing pathway fiom TRAF2 to

SAPK (Pombo et al., 1995; Yuasa et al., 1998). However, the physiologicd effect of this interaction is likely to difYer sornewhat fiom that of W2-mediated ASKl activation. For instance, whiie we show that in addition to impeding SAPK activation, a dominant-nzgative

ASKl mutant also blocks TNF-induced apoptosis in the presence of TRAF2 (Figure IO), no apoptosis function has been yet assigned to GCK. Also, through W2's interaction with the cytoplasmic death domain-containhg proteins TRADD and FUP (Hsu et al., 1995), ASK1- mediated apoptosis likely precedes caspase activation. When analyzed for cleavage by caspases during progammed ceti death, ASKI was unaffected (K.P.H and J.R.W, unpublished data), In contrast, MEKKl is activated by caspase-cleavage (Cardone et aLn 1997; Widmann el al., 1998) and is thus predicted to be involved in apoptosis-induced SAPK signaiiing downstream of caspase activation during the degradation phase of apoptosis (Frisch et al., 1996; Park et al.,

1996). Recent data has aiso implicated MEKKI as having a role in the activation of Id-kinase

(IKK), a component in the anti-apoptotic, NF-KB activation pathway (Karin and Delhase,

1998). Efforts to elucidate the mechanisms of MAPKKK regdation have been hampered by the fact that dl mammdian SAPK-activating MAPKKKs identified thus far are constitutively active upon overeqression. Correspondingly, while ASKI can associate with components of the

TNFRl cornplex, mere overexpression of ASKl results in its potent activation, overwhelrning any endogenous inhibitors present in Iimiting concentrations. Since ASKl interacts predominantly with the TRAF-N domah of TUF2 (Figure 7B), one such class of inhibitors couid be the iAPs (inhibitor of apoptosis proteins) which also bind the same TRAF2 domain

(Lacasse et al., 1998; Rohe et al., 1995). We can hypothesize that the pro-apoptotic ASKl and anti-apoptotic iAPs could compete for TRAFî binding. However, upon testing, no evidence was found for mutually exclusive bindiing of ASKl and HiAPlIHiAP2 to TRAF2 (Figure 10). In addition. ASKl has recently been show to interact with the Fas receptor-associated protein,

Da-(Chang et al., 1998). Hence, future work in evaluating the roIe of ASKl in TNF- or Fas- mediated apoptosis will be aided by the targeted disruption of ASKl in mice. Chapter 3

Requirement for glycogen synthase kinase-3B in ceii survival and NF-- activation

A version of this chapter is published in Nature 2000 Jul6;406(6791):86-90 3.1 Abstract

Glycogen synthase kinase3 (GSK-3) a and P are highly related protein-serine kinases which act as inhibitory components of Wnt signdling during ernbryoniç development and ce11 proliferation

in adult tissues (Welsh et al., 1996; Dale, 1998). Insight into the physiological hction of GSK-

3 has emerged from genetic analysis in Drosophila (Siegfkied et al., 1990; Rue1 et al., 1993),

Dicryostelizrm (Harwood et al., 1995) and yeast (Ehziss er al., 1994; Plyte et al., 1994). Here we

show that disruption of the murine GSK-3$ gene redts in embryonic lethality caused by severe

liver degeneration during mid-gestation, a phenotype consistent with excessive tumour necrosis

factor (TNF) toxicity observed in mice lacking genes involved in NF-KB activation.

Accordingly, GSK-3P deficient embryos were rescued by inhibition of RIF using an anti-W-cr

antibody. GSK-3B-deficient embryonic fibroblasts demonstrated hypersensitivity to TNF-

a accompanied by a reduction in NF-KB function. Lithium treatment, which inhibits GSK-3

(Klein et al., 1996; Stambolic et al., 1996), sensitized wild type fibroblasts to TNF and inhibited

NF-KB transactivation. The early steps leading to NF-KB activation (1-KB degradation and NF-

KB nuclear translocation) were unaffected by Ioss of GSK-3fi, indicating that regulation of NF-

KB by GSK-3P occurs at the level of the transcriptional complex. Together, these findings

Uidicate a novet role for GSK-3B in facilitating NF-KB function.

3.2 Attributions

The GSK-3p-deficient mouse was generated in the Woodgett laboratory by Juan Luo, with

technical assistance Eom Tak Mak's laboratory, including Drew Wakeham and Arda Shahhian. The knock-out embryo phenotype and patholugy were initiaIly described by Ming-Sound Tsao and Juan Luo. TUNEL staining was performed jointly by Jin Ou, Juan Luo and Ming-Sound

Tsao. Elizabeth Rubie cultured the embryonic fibroblasts and performed the EMSA in Figure

17A, B.

3.3 Materials and Methods

3.3.1 Reagents

Cytoplasmic and nuclear lysates were prepared as prcviously described (Yeh et al., 1997).

Irnmunoblotting was carried out with 1-@-a (polyclonal, New England Biolabs), p65 NF-KB

(polyclonal, Santa Cruz Biotechnology) and GSK-3 (rnouse monoclonal, Upstate Biotechnology) antibodies.

3.3.2 Apoptosis assays

Where indicated, TNF-cc-treated ceils (10 ng/mI, 24 hr) were collected, mixed with 4 pghl acridine orange (final concentration), and assessed by fluorescence microscopy. Survival of ce11 lines following TNF-a treatment was detennined by negative staining with trypan blue. Viability is expressed normalized to untreated controis. AU experiments were repeated at least three times, and the data (as a percentage) is show as the mean * standard error. In situ apoptosis was detected by terminal deoxynucleotidyl tramferase-mediated düTP nick end labelling (TUNEL) according to the manufacturer's instructions (Boehringer Mannheim), or by another fragmented

DNA end-IabeIing protocol (Wijsman et d,1993). For ~galactosidaseviabiiity assays, two days following transfec tion with pCMV-B-galactosidase plus either pCDNA.3-HA-GSK-3 b or control plasmid, ceiis were treated as indicated and analyzed.

3.3.3 Electrophoretic mobility shift assay

KB-binding activities of embryonic fibroblasts incubated with lithium or potassium (30 nM,4 hr) and murine RIF-ct (100 ng/ml, 30 min), as indicated, were compared by EMSA. Nuclear

Lysates were prepared and EMSAs were performed as described previously (Beyaert et al.,

1989). For oligonucleotide cornpetition assays, equivalent amounts of nuclear extract protein (3 pg) were preincubated for 5 min with a 200-fold excess of either NF-KB-specific oligonucleotide probe containing nvo tandem NF-KB-binding sites (5'-ATC AGG GAC TTT CCG CTG GGG

ACT TTC CG-3' and 5'-CGG AAA GTC CCC AGC GGA AAG TCC CTG AT-3') or mutant

NF-KB oligonucleotides (5'-GAT CAC TCA CTT TCC GCT TGC TCA CTT TCC AG-3' and 5'-

CTG GAA AGT GAG CAA GCG CAA AGT GAG TGA TC-3') pnor to the addition of radio- labeled NF-KB-specific probe. The EMSAs for Oct-1 were performed as descnbed foc NF-KB using the oligonucleotides 5'-TTC TAG TGA TTT GCA TTC GAC A-3' and 5'-TGT CGA ATG

CAA ATC ACT AGA A-3'.

3.3.4 Luciferase hsay

EF ceIls were trmfected with plasmids expressing endothelid leukocyte adhesion molecule

(ELAM) promoter-luciferase and fi-gaiactosidase. Transfected cells were incubated in the presence of 30 ng/ml TNF-a or IL-Ib for 6 hr. Luciferase assays were carried out using the

Promega assay kit and a Berthold luminometer. Activity was nomalized to b-galactosidase activity and is plotted as the mean SD of tripiicates from a representative experiment. To examine the effect of lithium treatment on NF-KB-mediated gene transcription, 293 epithelial ceIls were preincubated overnight with 30 tnM lithium or potassium prior to stimulation with either 10 or 20 ng/rnl TNF-a. Luciferase values were normalized to the unstimuiated potassium and lithium controls.

3.4 Results

In vertebrates, GSK-3 has been show to be crucial for the definition of embryonic axes (He er al., 1995; Dominguez er al.. 1995; Nasevicius et al., 1998). To investigate the role of GSK-3 in mammalian development, we disrupted the GSK-3p gene in murine 1295 embryonic stem (ES) cells using a targeting vector in which the exon encoding the ATP-binding loop was deleted

(Figure 11A). Chimeric mice derived from two independent heterozygous ES clones were back- crossed to C57BW6J mice, and heterozygous mice were crossed to generate homozygous mutant offspring. In addition to tail DNA genotyping, the nul1 mutation of GSK-3P was confmed by

Southem bIot (Figure 11B) and Western blot (Figure 1 IC) andyses of embryonic fibroblasts

(EFs) derived from mice on day 12.5 of genaticn (E12.5). Aithough GSK-~~-male and fernale mice were healthy and fertile, upon intercmssing they did not give cise to live GSK-~~progeny.

We therefore analyzed embryos fiom thed pregnancies of GSK-~P-'-intercrosses.

Before E12.5, the Mendelian ratio of nul1 embryos was normal and most ~~~-3fl~-embryos showed no morphological abnormality (no< show). However, betweeo El 3.5-14.5, GSK-~B~ embryos appeared pale and non-viab[e (Figure 12A). Histological examination of E13.5 GSK-

38" and GSK-J~~embryos demonstrated multifocal hemonhagic degeneration in the iivers of homozygous embryos (Figure 12%). In these areas, the hepatocytes showed pyknosis and karyorrhexis suggesting they were undergoing apoptosis. This was confhed by terniinal deoxytransferase-rnediated deoxyuridine triphosphate nic k-end labelling (ML)assay which nvealed numeiour positively-stahed nuclei (Figure 128). GSK-~/~'-embryos derived fiom the two original GSK-3B heterozygous ES ce11 clones were indistinguishable in phenotype (not shown). Fetal liver transkr experiments demonstrated that haematopoietic cells fiom GSK-~P- smbyos were capable of reconstituting haematopoiesis in Iethally irradiated mice (not shown).

No evidence for perturbation of Wnt signalling, cyclin D levets, p-catenin accumulation or giycogen metabolism was observed (not shown). These results suggested that hepatocyte

apoptosis is likely the major cause of lethality in the GX-36-embryos.

This phenotype is similar to that of iKK$ or ReiA-deficient mice which has been

demonstrated to result fiom increased sensitivity to TNF-cr (Li, Q. et al., 1999; Li, S. W. et al.,

1999; Beg et al., 1995). To inhibit endogenous TNF-a, we injected anti-murine TNF-a

monoclonal hamster antibody (or hamster imrnunoglobins as a control) into the peritoneum of

pregnant heterozygous femdes (sired by heterozygous males) on El05 or El 1.5. The femaies

were sacrificed and embryos collected on El45 From 3 females treated with anti-TNF-a

antibody injection, Il embryos (42% of total) were identified as GSK-3fl. As detennined by

gross morphology (not shown) and histoIogicaI analysis of the liver, substantial protection and

hepatocyte rescue of the anti-TNF-a- injected embryos was observed, as compared to control

embryos (Figure 12C). These results pointed to an unexpected role for GSK-3p in suppressing

RIF-mediated apoptosis. 8.0 kb Xbal 4 b

H HXH i E EH X Endogenous 1 III i I locus --. / 1. .: .-. . EZZZZaprobe 5, '-. p.+-+- . 1 Ne0 .. *..

X H H HXH H X Targeted l Ill Ne0 I I locus

b 6.4 kb Xbal Figure 11. Targeted disruption of the GSK-3b locus (A) A portion of the mouse GSK-3/l wild-type lacus (top) showing exons (solid box) and a 8.0 kb XbaI Fragment in the wild-type alleie. The mutated GSK-3p locus (battom) contains a 6-4 kb XbaI fragment. The positions of the probes used for Southern analysis are shown (hatched boxes). X, E and H represent XbaI, EcoRi and HindIII, respectively. @) Southern blot of EF ce11 DNA digested with XbaI and hybridized to the flanking probe. (C) GSK-3f3 protein is undetectable in GSK-~F-EF cells.

Figure 12. Phenotype of GSK-3F embryos (A) ~slC-3~'-ernbryos at E14.5 showed focal parenchymal haemorrhage. (B) Repiesentative histological sections of GSK-~B-" embryos stained with hematoxylin and eosin showed abnormality specifically localized to the liver with nuclear debris and apoptotic cells at varying stages of degeneration. increased TUNEL-positive cells are,-observed ~~,GsK-J/?*'mice as compared to controls. (C) Pregnant females from GSK-3/3 x GSK-3/3 intercrosses were injected interperitonealIy with either 250 pg anti-murine TNF-a hamster monoclonal antibody or control hamster IgG. Liver section analysis was performed on GSK-3$ nul1 genotyped embryos and apoptosis was detected by in siru end-labeling of fragrnented DNA. To probe the mechanism by which GSK-3$ affected TNF cytotoxicity, immortalized mouse embryonic fibroblast (EF) ce11 lines from two wild-type (+/+1 and +/+2), two GSK-3B heterozygous (+/-1 and +/-2) and two GSK-3$ homozygous mutant (-1-1 and -1-2) embryos were examined. OP note, expression of proteins important for TNF signalling, such as TNF-R1 and

IKKP, were normal in these Iines (not show). We next examined the sensitivity of the ce11 limes to TNF-induced ce11 death. Following TNF-GItreatment (10 ng/ml, 24 hr), nuclear rnorphology of GSK-3r'"and GSK-3B"- cells was visualized by acridine orange staining and fluorescence microscopy. Chromatin condensation characteristic of apoptosis was apparent only in the GSK-

38/- ce11 lines (Figure 13A). We then investigated the susceptibility of wild-type and GSK-3p- deficient cells to TNF in the absence or presence cycloheximide (CHX; Figure 13B). As measured by viability dye staining, wild-type and GSK-3B-heterozygous cells were relatively unaffected by exporure to TNFa. However, significantly fewer GSK-3@ cells retained viability

24 hr following maximum TNF (18% and 56%) or TNFKHX (2% and 28%, respectively)

treatment. As a reference, TNF receptor-associated factor 2 (TRAF2)homozygous mutant MEFs

were used in the sarne assay. TR.AFJ3- ce11 survival in OUI* hands was consistent with previous

reports (-20% at IO nghl TNF-d0.25 @ml CHX; Yeh et al., 1997).

Lithium ions inhibit GSK-3a and $ both in vitro and in intact cells (Stambolic et al.,

1996) and have been reported to exacerbate TNF cytotoxicity (Beyaert et al., 1989). Cells were

incubated with 20 mM lithium (or 20 mM potassium as control) during treatment with CHX and

an intermediate (10 nglmi) TNF-a concentration. Consistent with the hyper-sensitivity of the

GSK-3B nul1 cells to TNF-a, treatment of wild-type and heterozygous ce11 lines with the

combination of EthiumlTNF-dCHX resuited in a simiiar increase in TM?cytotoxïcity (Figure

13C). O IO 20 30 40 50 O 10 20 30 40 54 TNF* (flgtml) m~-a(Wml) plus CHX (0.25 uglml) Figure 13. GSK-3p-deficient ceüs show increased sensitivity to TNF-induced apoptosis (A) Chromatin condensation demonstrated by acridine orange staining. (B) Cel5,yiability foilowing 24 hr TNF-a (1 O, 25 or 50 nglml) treapent. Independently derived GS'-3/3 EF cefi clones are indicated by +/+1 and +/+2; GSK-3/3 - by +/-1 and +/-2; and GSK-3/3 - by 4-1, and - /-2. Survival after treatment with TNF-a (4, 10 or 50 nglmi) plus cycloheximide (0.25 pg/rni) is shown in the right panel. (C) Lithium potentiates TM: killing in wild-type ceils. Cells were treated ivith 0.25 pg/ml cycloheximide, 20 mM K', 20 mM ~i'and 10 nglml TNF-a for 24 hr as indicated. To demonstrate that the apoptotic sensitivity of GSK-3~'- fibroblasts was a direct consequence of GSK-3$ defciency, a rat GSK-38 gene was reintroduced in ~sK-30~-cells by transient transfection. GSK-J/~* fibroblasü expressing the GSK-38 transgene exhibited reduced sensitivity to TNF-CYtreatment compared with controls (Figure 14A). The specificity of apoptotic sensitivity upon disruption of GSK-3@ fuoction was assessed by treating ~X-30~-cells with a range of pro-apoptotic agents (anisomycin, sorbitol, daunorubicin or y-irradiation); no differences From wild type responses were observed (Figure 14B). Taken together, these data indicate that GSK-3P is required for protection fiom RIF-a cytotoxicity.

Several studies have demonstrated that activation of the transcription factor NF-KB is essential for suppression of TNF-induced apoptosis (Beg et al., 1996; Van Antwerp et al., 1996;

Wang et al., 1996; Liu et al., 1996). To determine whether TNF-induced NF-KB activation was affected in the GSK-~/~+'cells, ouclear extracts fiom TNFa treated EF ce11 Iines were analyzed by elecuophoretic mobility shift assay (EMSA) using an NF-KB-specific probe. The specificity of the shifted band was tested using a 200-fold excess of non-labelled wild-type or mutant KB

DNA (Figure 1SA). NF-KB activation in response to RIF-a treatment was reduced by greater than 50% in GSK-~Fcells, as cornpared to wild-type cells (Figure 1 SB). The effect of GSK-3P dimption was specific to NF-KB binding since DNA-binding of other transcription factors such as Oct-1 was unaEected (Figure 132). CHX O CHX tTNF-a

O 250 500 750 1000 Sorbitol (mM)

y-irradiation (Gy) Daunorubiin (nglmi) Figure 14. Apoptotic seoritivity of GSK-3f fibroblnsts

(A) Expression of exogenous GSK-3,B in GSK-3p-deficient cells protects against TNF-induced apoptosis. (B) Sensitivity of GSK-3p EF cells to various apoptotic stimuli. & +/+1 4+2 -1-1 -1-2 -1-- Paip TNF* ++ +f ++ ++ 8 S' Lithium ++++++++ TNF-a t + Potassium t + + + + + + +

* -. *-. --..a & 4P' . .* ..- -- .

C +/+1 4+2 -1-t 6-2 i--- TNFa +t ++ f+ + + Lithium +++CC+++ Potassium + + + + c + + + nuclear Figure 15. Inhibition/inactivation of GSK-3p reduces TNF-inducible NF-- activation (A) Specificity was evaiuated by oligonucleotide competition and elegrophoretic mobility shift assay (EMSA). (B) Activation ON-KBin GSK-3p and GSK-~Fcells treated with TNF-a. (C) GSK-30 does not interfere with DNA-binding of Oct-1. @) Effect of GSK-3p disruption on TNF-uand (E) IL-lfl signalhg. (F) GSK-3fl deficiency or (G) lithium treatment blocks NF-KB- mediated gene transcription. (H) Immunoblot aaalysis of cytoplasmic 1-KB-uand nuclear p65 NF-KB. Cells were stimulated with 10 ng/rnL TNF-u for 0,0.25,0.5, 1,2,4 or 8 hr (lanes 1-7). Non-specific (m.) bands are indicated. Since NF-KB-mediated transcription of the endothelial leukocyte adhesioa molecule-1

(ELAM-1 or E-selectin) gene is induced by TNF-a and interleukin-1fl (IL-If!), we examined

ELAM-luciferase induction by these agonists in EF cells. TNF or IL-lfi induction of ELAM- luciferase transcription was reduced by approximately 3-4-fold in GSK-38eeUs (Figure lSD,

E) and could be restored by expression of exogenous GSK-30 (Figure 15F). The diminished NF-

KB induction by IL-1 fi in GSK-~/~+'-cells indicates that the defect caused by lack of GSK-30 resides in NF-KB function, rather than a TNF-specific response. Lithium inhibited TNF-a induced NF-KB transactivation by approximately 70% in 293 epithelial cells (Figure 15G).

In resting cens, NF-KB complexes are retained in the cytoplasm by association with a fmiIy of inhibitory proteins termed 1-KBs (Baeuerle and Baltimoret 1988). Activation of NF-@ requires the phosphorylation of 1-KB which triggers its polyubiquitination and subsequent degradation by the 26s proteosorne. The liberated NF-KB rapidly translocates to the nucleus where it binds KB sites and regulates gene expression. To assess the mechanism by which NF-KB activity was inhibited in GSK-3f! nul1 cells, we studied the kinetics of 1-KB proteolysis and NF-

KB translocation following treatrnent of cells with TNF-or.Western blotting of cytoplasmic and nuclear extracts indicated no observable differences bem-een wild-type and GSK-38-deficient cdls (Figure 15H). In al1 ce11 lines tested, 1-KB-a was rapidly degraded following exposure to

TNF-a followed by a Iag phase and reaccumulation of the protein by 2 hr. p65 NF-KB translocation to the nucleus was also independent of GSK-38. These fïndings place the defect in

NF-KB activation in GSK-3fi'- cells downstream of 1-KB phosphorylation and nuclear translocation of NF-KB and impty that GSK-3p activity is required for nuclear activity of the complex. Since 1-KB re-synthesis (which is under the transcriptionai controI of NF-KB) is unchanged in ~X-36-cells, only a subset of NF-KB-inducible genes are dependent on GSK-38 function.

3.5 Discussion

Unexpectedly, mice lacking GSK-3fl suffer fiom defects simiIar to those of mice that are mutant for essential components of the NF& pathway. GSK-3f3-deficient rnice are morphologically normal up to approximately day 12 of embryonic deveiopment but die between days 13.5-14.5 due to massive Iiver degeneration and hepatocyte apoptosis (Figure 12). This death can be prevented by blocking the function of RSF-a. Embryonic fibroblast cells fiom GSK3$ mutant mice are highly sensitive to TNF-cc-mediated apoptosis but not to other apoptotic stimuli (Figure

13). In addition. GSK-3B-deficient fibroblasts show a significant reduction in NF-KB DNA binding and reporter gene response following treatment with TNF-a or IL-1. As has been previously described, partial inhibition of NF-KB DNA binding alone does not seem to be sufficient for dom-regdation of 1-&-a gene expression (Oliver et al., 1999).

The mechanism by which GSK-3p impacts NF-KB activity is not known but the kinetics of NF-& nuclear translocation and the haif-life of the regdatory protein 1-KB-a is unaffected in

GSK-3$ mutant cells (Figure 15). These data mie out an effect on the cascade of proteins that culminates in phosphorylation of 1-KB and its degradation. Presumably then, GSK-3b is required for a later step in the signalling pathway, such as a direct or indirect requïrement for phosphorylation of the NF-@ subunit, p65 or an as yet unidentified transcriptional CO-activator of p65. Whatever the molecular mech;tnism, it is clear that this effect is specific for GSK-3f3 since GSK-3a is unable to compensate. This is in stark contrast to the lack of effect of loss of GSK-30 on the Wnt pathway. GSK-3p mutant mice do not exhibit overt dysregulation of this parhway as might have been predicted indicating that the Wnt pathway does not discriminate between these two GSK-3 isoforms. Indeed, to date, NF-KB activation is the only known differentiator between these two enzymes. WhiIe the fmding that GSK-3P is required for NF-KB activation was surprising, there are other dues in the literanite. For instance, additional support for a connection between GSK-3 and NF-KB during vertebrate development has emerged fiom studies of Xenopzis laevis development. The mRNA of at lest one Xenopus member of the rel family, XrelA, is expressed in oocytes and early embryos. XrelA inpacts the dorsovenîral patterning process (Kao & Lockwood, 1996). XrelA induces a ventralizing effect early in embcyonic development and attenuates morphogenetic movements characteristic of dorsal mesodem. XrelA RNA was also show to reverse the strong dorsai ais-promoting effects of a dominant mutant of Xenopus GSK-3P. This complementation effect argues in favor of a p65 NF-

KB homologue being downstream of GSK-3p, at least in some systems.

Overall, the phenotype observed in p65- or NEMO-deficient mice is a more severe than that of GSK-3p- or IKKPmice (Beg et al., 1995; Rudolph et al., 2000; Li et al. 1999b). In particular. the onset of the liver apoptosis occurs a IittIe earlier in NEMO-deficient embryos and even though GSK-~B/*and I.@- embryos show some liver hemorrhaging at E13.0, the overall liver architecture is still intact. Moreover, MEFs lacking either GSK-30 or IKKB still show residual NF43 DNA-binding activity in response to stimulation with TNF, whereas in p65- or

NEMO-deficient cells it is completely abolished. This suggests that the residual NF-KB DNA- binding activity obsemed in GSK-~F-rnice might provide limited protection to liver cells against apoptosis induced by endogenous TNF. As show in this chapter and by Li et al.

(1 999b), this residual NF-KB activity is stilI sufficient to induce a low Ievel of expression of NF- &-dependent iarget genes, such as 1-@-a,in GSK-~B/'and K.-MEFs. In contrast, 1-KB-a resynthesis is undetectable in p65- or NEMO-deficient cells.

Treatment with lithium, a potent inhibitor of GSK-3, sensitized wild-type fibroblasts to

TNF-induced apoptosis at levels comparable with genetic disruption of GSK-3b (Figure 15).

These results are consonant with those of Beyaert et ai. (1989) who fust reported that lithium causes a dose-dependent enhancement of TNF cytotoxicity in human and murine ce11 lines

(Beyaert er al., 1989). Lithium also enhances the in vivo anti-tumor action of TNF. Similar results have been obtained when using interleukin-2 (IL-2), another NF-KBagonist (Wu & Cai,

1992). IL-? in combination with lithium demonstrated a stronger inhibitory effect on tumor

~rowththan alone, as detecmined by reduction of tumor size and prolongation of survivai in h IL-2 tumor-bearing rnice.

in sumrnary, these results demonstrate that GSK-3B function is required for the NF-KB- mediated anti-apoptotic response to TNF-a (see Figure 16). These data ais0 demonstrate distinct biological roles for GSK-3a and since the former is unable to compensate for loss of the latter. caspase activation inactive l-KB kinases Y apoptosis Figure 16. Model of NF-KB regulation by GSK-3B GSK-3P is required for NF-KB transactivation independent of 1-KB-a degradation and NF-KB nuclear translocation. See text for details. Abbreviations: nuclear factorxi3 (NF-KB), tumor necrosis factor (TNF), inhibitor of NF-KB (I-KB). Chapter 4

Further analysis of NF43 regulation by GSK-3f! and its inhibitors 4.1 Abstract

NF-KB is an essential transcription factor in the control of gene expression involved in immune and inflamrnatory responses. As described in Chapter 3, GSK-3b has been shown to prornote NF-KB mediated gene transcription, but the underlying signahg rnechanisms are still not well understood. Previously, it has been shown that TNF-cr induces p65 NF-KB phosphorylation on serine residues and that this phosphorylation increases NF-KB transcriptional activity when assayed using an exogenously supplied promoter. In this chapter, reagents for assessing the role of GSK-36 in p65 phosphoryiation are described. To Merevahate the effect of GSK-3Q on embcyonic development, TNFRl -"-GsK-~~and TNFR~GSK-SB"- mice have been generated.

Ongoing mouse breeding expenrnents and preliminary analysis of TNFR~~'-GSK-~~-

embryos are also presented.

4.2 Attributions

FR I" mice were obtained Çorn Tak W. Mak and pCDNA3-KA-p65RelA plasrnid was

provided by Hiroaki SAwai. The TNFRI'-GSK-~@- embryo phenotype was analyzed

with the assistance of Ming-Sound Tsao. 1.3 Materials and methods

4.3.1 Reagents

~WFRI-'mice were obtained fiom Dr. Tak W. Mak (Pfeffer et al., 1993). pCDNA3-HA-

p6jReiA was provided by Dr. Hiroaki Sakurai (Sakurai et al., 1999).

43.2 Assay of GSK-3 activity

The activity of immunoprecipitated GSK-3fl proteins was determined by phosphory lation

of GS1 peptide. GSI is a synthetic 25 arnino acid peptide,

YRRAAVPPSPSLSRHSSPHQSEDEE, derived from glycogen synthase. GS1 is

synthesized to contain a phospho-serine at position 20 (underlined) to mimic

prephosphorylation by casein kinaseII, thereby allowing efficient phosphoryiation of

three consensus sites by GSK-3 (bold). HA-GSK-3p immunoprecipitates were incubated

at 30°C with 30 pM of GSI peptide in the presence of [y-3'~]-~TPin 2.5 mM Tris pH

7.5, 1 mM DTT, 25 mM MgCl?. Quantification of GSK-3 kinase assays was performed

by resolving the phosphorylated peptides on tricine gels followed by Phosphorimage

analysis of resulting 31~-containingbands (Molecuiar Dynamics).

43.3 CeU culture

HEK293 celis were grown in Dulbecco's modified medium supplemented with 10% fetal

bovine semplus antibiotics. Stable ce11 lines that express HA-p6S were generated by

selection with high G418 (800 pg/mi) and maintained in 400 pg/d G418. 1.4 Results

1.4.1 Mechanistic analysis of NF-- regulation by GSK-3g

GSK-3 is acutely inactivated in response to insulin or growth factors (see section 1.4.6).

This inactivation of GSK-3 is mediated by increased phosphorylation of a serine residue near the N-terminus (Ser-9 of the B-isofonn; Ser-21 of the a-isoform; Sutherland et al.,

1993). To assess whether GSK-36 could be acutely reguiated by TNF signalling, GSK-3 activity was assessed as follows. As shown in Figure 17, IGF-1 decreases GSK-3B kinase activity, as monitored by phosphorylation of a GS1 peptide substrate or by phospho-

GSK-3(Ser9/21) blotting. However, RIF-a treatment did not affect GSK-3 activity or phospho~lation.

4.4.2 GSK-3B does not phosphorylate C-KB-~in vitro

Frorn the evidence that RJF-induced [-K~-cLdegradation is unaffected by GSK-3f3 status in fibroblasts, it would be predicted that GSK-3$ does not phosphorylate 1-KB, at least not on the N-terminal serine residues crucial for its targeted ubiquitination and degradation. However, it was necessary to confirm this since the surroundhg sequence of these serines in I-~B-cf/p/y (DSGLDS,DSGLGS and DSQCDS, respectively) closely resemble the GSK-3 phosphorylation motifs of bonafide subsmtes (Figure MA). In vitro kinase reactions using full-length GSK-3p and 1-~B-crGST fusion proteins purilied fiom bacteria did not show diect phosphorylation (Figure 188). However, fi-catenin would also not be phosphorylated under such conditions since it needs Axin as a scaffold. It is plausible that GSK-3B exerts its effects on NF-KB directly or indirectly via phosphorylation. Recent data indicates that NF-KB p65 is a highly phosphorylated protein and serine phosphorylation may modulate its transcriptional activity. LPS induces the phosphorylation of Ser276 by , increasing NF-KB transcriptional potential in T and B cells (Zhong et al., 1998). Stimulation by TNF results in phosphorylation of p65 at Serj29 by CKII (Wang & Baldwin, 1998; Wang et al., 2000) and Ser471 by an as yet unidentified kinase (Martin & Fresno, 2000). Purified recombinant IKKcrlB phosphorylates p65 at Ser536, although it is not known whether this event is physiologicaIly relevant (Sakurai et al., 1999). None of the sequences surrounding these serines fit the consensus site for GSK-3 phosphorylation (S-X-X-X- sP). Hence, to characterize potential changes in p65 phosphorylation in vivo, 1 have generated epitheliai 293 ce11 lines that stably express HA-p65 (Figure 18C). These ceIl

Iines will be instrumentai in [1'~]-~hosphatemetabolic labeling experiments to assess the phosphorylation status of p6j (see Chapter 5). 100 84.5 97.1 (% activity)

PhospheGSK-3 GSK-3a (Set 9/21) blot GSK-3f3 Figure 17. Effect of TNF-a and IGF-1 on GSK-3$ activity

(A) FolIowing transfection of HEK293 cells with pCDNA3-HA-GSK-30, ceUs were serum-starved and incubated with 50 nghl IGF-1 or TNF-cc for 10 minutes. GSK-3B kinase activity is show as a percentage of the unstimulated control. (B) Phosphorylation of serine-9/21 of GSK-3PIcc was detected by blotting with phospho-specific GSK-3 antibodies. I-KB-~protein levels are show as a control for TNF stimulation, Note IGF-1 does not induce 1-KB degradation. consensus D s G f X S 1KB-ff DSGLDS IKB-~~ DSGLGS IKB-y DSQCDS Cyclin Dl D s G v F S P-catenin DSGIHS

GST-GSK9P Phosphorimage GST-I-KB-a

Coomassie t GST-GSK-3B 4- GST-l-KB-cc

Cell lines Figure 18. GST-GSK-3$ does not phosphorylate EKB-a (A) Aiignment of potential GSK-3 phosphorylation sites on 143swith recognition sites on known targets. (B) Increasing amounts of bacteriaiiy purified GST-GSK-38 were incubated with GST-1-KB-aand kinase reaction b6rcontaining y-32~-~TP.GST-GSK- 3 autophosphorylation is indicated, and protein levels were monitored by Coomassie staining. (C) HEK393 ce11 lines stably expressing HA-p65 were generated. Non-specific bands are indicated by "n.~.". 4.4.3 Debromohymenialdisine

Hymeniddisine (HD) and related compounds such as debromohymenialdisine (DBH), were originally purified fiom the marine sponge Axinella verntcosa and have recently been identified in a screen for anti-idammatory compounds (Figure 19A). For example,

DBH exhibits anti-idammatory activity in a mode[ of adjuvant-induced arthritis in the rat (DiMartino et al., 1995).

HD was evaluated for its effects on the activation of NF-KB using electrophoretic rnobility shifi assay (EMSA) and luciferase reporters under the control of either the HIV-

LTR or IL-8 promoter. Shilar to lithium treatment or genetic disruption of GSK-30, DH

inhibited NF-KB DNA binding and NF-KB-regulated gene expression without affecting 1-

KB degradation (Breton & Chabot-FIetcber, 1997; Roshak er al., 1997). In particular, HD caused a concentration-dependent inhibition of Iuciferase production regardless of the

stimulus used (TNF-a, LPS or PMA), and HD-inhibited DNA-binding in Li937 ceiis was

rnost evident at 1.0 p.M HD, conmensurate with 50% inhibition of TNF-stimulated

DNA-binding. Although HD is ais0 known to inhibit PKC, the selective PKC inhibitor

RO 324432 has no effect on TNF-a-stimulated luciferase reporter activity or IL-8

production at low concentrations. However, as expected the RO 32-0432 only inhibited

Iuciferase production in response to PMA stimulation (with ICSo values of 0.2 mM),

reflecting the abïlity of this compound to inhibit PKC.

The striking similarities between HD effects and GSK-3p inhibition were noted

by Dr. Michel Roberge (University of British Cohmbia) who had been workiag with

DBH. DBH is a potent in vitro inhibitor of GSK-3p (Figure 19B) and aiso interferes with

in vivo phosphorylation of specific proteins by GSK-3p. In particuiar?phosphorylation of B-catenin (Figure 19C) and as recently demonstrated, Tau (Meijer et al., 2000), are comp1eteIy inhibited by DBH and HD in vivo. These effects occur at two-orders of magnitude lower concentrations of DBH than Lithium (Figure 19D). These data support the idea that the anti-infiammatory activity of HD compounds may be related to their abiiity to block the activity of the transcription factor NF-KB via inhibition of GSK-3b. ,@ GO HA-GSK-3B transfection

DBH concentration (nM)

Li+ DBH (nM) control 20mM 40 80 160

$-catenin densitometry

ns. n.s. Li+ DBH 30mM 160nM Figure 19. Debromohymeaiafdisine is an effective inhibitor of GSK-3$ (A) Structure of hymeniddisine (HD) family mernbers. For RI and R2 subgroups, Br/H is hymenialdisine,H/H is debrornohymenialdisine (DBH), and Br/Br is 3- bromohymenialdisine. DBH has powerful effects on the in vitro and in vivo activity of GSK-3p as shown by GS1 peptide kinase assays (B) and endogenous $-catenin stabilization (C). Cornparison of inhibition of GSK-j by lithium and DBH O). 4.4.4 Generation of GSK-J~~I'~mice

While the absence of p65 NF-KB/R~~Aresults in massive hepatocyte apoptosis and embryonic death benveen El4 and E15, mice that are deficient in both ReiA and 'IUFRl

survive embryonic development (Rosenfeld et al., 2000). At birth R~~A"'TNFRIJ* pups have normal liver rnorphology, but ultimately have [United survival because RelA provides critical postnatal hctions. These data suggest that RelA is not critical for Iiver development but may serve a vitai protective fiuiction against TNF and TNFRl apoptotic signalling during late stages of embryogenesis. Comparably, elimination of in vivo TNF signalling in GSK-3p- embryos. by maternai injection of antagonistic TNF-or antibodies, was shown to rescue the liver apoptosis occurring at E13.5 (Chapter 3). However, the effect of these antibodies on embryonic deveiopment is onty transient. Therefore, to further examine the contribution of signalling through TNFRl to the embryonic lethal phenotype of GSK-3B/' fetuses, mice deficient in both GSK-3p and WRl were generated by the scheme described in Figure 20.

Thed matings of WFRI *'-GSK-~T'*mice were conducted and pups at E13.5-

14.5 were coilected and analyzed (Figure 21)- Histological examination of TNFRI~-and

TNFRI"GSK-~T'embryos revealed that their livers were indistinguishable from those of wiId-type embryos and showed no evidence of apoptosis. One EI4.5 WFRI'"-GSK-

3@ embryo showed an incomptete rescue of the GSWB knofkout phewtrpe and rome liver apoptotic activity was present (Figure 2 1B). O TNFRI +/- GSK-3P+/- TNFR1 +I- GSK-3flt/- X X TNFRl+/-GSK-3fW TNFR1-/- Figure 20. GSK-3fbTNFR1 mouse breeding strategy Breeding undertaken between GSK-38"-and TNFRI'/- mouse strains. Expected probabilities of given genotypes are show in parentheses. TNFR lm/- TNFR 1-hGSK-3fP

B TNFR1 +4GSK-3/?A Figure 21. Phenotype of ~NFRI~-GSK-~~embryos. Representative histological sections of GWK- embryos at E13.5 (A)and E14.5 (B) stained with hematoxylin and eosin. 4.5 Discussion

4.5.1 Involvement of PKB in NF-KB regdation by GSK-3B

Phosphatidylinositol-3-OH kinase (P13'K) and PKB, which can inhibit GSK-3 in response to insulin or IGF-1, are also implicated in the activation of NF-KB by TNF, IL-

lP, PMA, pervanadate, and PDGF signailhg (Kane et al., 1999; reviewed in Scheid &

Woodgett, 2000). However, TNF-a treatment does not affect the activity of GSK-3P and converseiy IGF-1 has no effect on 1-KB-a levels (Figure 17).

Several papers indicated that IKK activity is involved in the ability of PKB to stimuIate NF-KB transcriptional activity, presumably through direct rnechanisms

involving enhanced 1-KB degradation and NF-KB nuclear translocation (Burow et al.,

2000; Ozes et al., 1999; Romashkova & Makarov, 1999). Meanwhile, other groups

indicate that constitutively-active forms of PI3'K or PKB stimulate NF-KB transcriptionai

activity predominantly through signalling pathways that target the tramactivation domain

of the p65 subunit and are separate or parallel to the cascade leading to 1-KB degradation

(Madrid et al., 2000; Sizemore et al., 1999). To Meradd to the controversy, stiii other

groups have failed to detect any direct involvement of PKB in the signalling pathway

though which TNF leads to NF-KB activation (Delhase, Li & Karin, 2000; Madge &

Pober, 2000). Taken together, it is unlikely that PKB activity alone is sufficient to

manifest a significant NF-KB effect, and that the PKB-NF43 comection may be ceU-

type and stimulus-specific (Jones et al., 2000). Importantly, Etom the perspective of this

thesis, these disparate observations point to deficiencies in our understmding of NF-KB

activation, and merinvestigation is clearly requüed to explore any possible

relationship between the opposing effects of PKB and GSK-3p on NF-KB. 45.2 Potential role of the SCF ubiquitin ligase complex

Ubiquitin-dependent proteolysis by the proteasome plays an essential role in a number of key biologicai processes, including ce11 cycle progression, transcription, and signai transduction (reviewed in Hershko, 1997; Peters, 1998). Frequently the target protein is first rnarked for degradation or processing by phosphorylation. The phosphorylated protein is then recognized and ubiquitinated in a process that requires three proteins: a ubiquitin-activation enzyme (E 1), a ubiquitin-conjugating enzyme (E2), and a ubiquitin ligase (E3). Ubiquitin is first attached to an El protein in an ATP-dependent reaction to form a high-energy thio-ester bond. The ubiquitin is then transferred fiom the E 1 to an E2 enzyme, which functions in conjunction with an E3 protein to link ubiquitin to lysine residues in the targeted protein. A specific Lysine residue in the conjugated ubiquitin can then attach to a second ubiquitin and reiteration of this process results in the assembly of a polyubiquitin chah The golyubiquitinated protein is recognized by the 26s proteasome and is subsequentIy degraded. übiquitination of both p-catenin and 1-KB is targeted by F- box/m/D40 repeat-containing proteins, such as Drosophiia Slimb hupernumerary Ms) and its mammalian homologue B-TrCP, that are components of a class of E3 , termed the ~kpll~linl-box(SCF) complex (Jiang & Struhi, 1998; Spencer, Jiang &

C hen, 1999; Wionet al., 1999; Yaron et al., 1998).

Recently, components of the ubiquitin proteasome pathway have been impiicated

as a common link between the Wnt and NF-KB signal transduction pathways.

Overexpression of a stabiiized f3-catenin mutant, in which the GSK-3-targeted serines

were altered to alanine, augments NF-KB activation by a constitutively-active IKKP (Spiegelman et al., 2000). The mechanism for this has been suggested to occur via

LEF-11Tcf-dependent upregulation of PTrCP levels and subsequently facilitated ubiquitination of phosphorylated 1-KB.

A comrnon structural feature between 1-KB and fi-catenin is that phosphorylation occurs on two closely located serines at positions 32/36 (1-&cc) and 33/37 (fi-catenin;

Figure 18). This suggests that Slimb/fl-TrCP recognizes a DSpGXXSp amino acid motif.

Since phosphorylation and degradation of batenin is mediated by a constitutively active

GSK-3B and 1-KB is inducibly phosphorytated by IKKs in response to various extracellular stimuli, what then distinguishes 1-KB and @-catenin?Perhaps it is only the kinases that are responsible for phosphoryiation of these substrates: phosphorylated

DSGXXS motifs on these substrates might be the signal for the comrnon ubiquitination pathway through SlimbIP-TrCP. In support of this possible explanation we have not been able to demonstrate 1-KB phosphorylation by GSK-3B (Figure 18).

4.5.3 Genetic rescue of the GSK-3B kaockout phenotype

Since the embryonic Iethality and liver apoptosis of animals lacking either p65 NF-KB or

IKKP cm be attributed solely to TNFRl signaling, animals lacking both p65 and TNFRl

(Rosenfeld et al., 2000), like those lacking both IKKP and Ml(Li et al., 1999b),

develop to term. As shown in Figure 2 1, rescue of El4.S GSK-3B-deficient mice fiom

premature lethality by elimination of TNFR1 Mersuggests that TNF-alevels are

critical for the embryonic phenotype observed in GSK-3@ embryos.

Preliminary data fiom newborn pups of WFN'/-GSK-~~'-x ~NFRI*'-GsK-~~~'-

breeding, which were monitored for survivai and examined for GSK-3fl status, shows that some GSK-3r- pups die at or just prior birth (presumably due to a deficiency of

TNFRI). This would suggest that double-knockout mice stiii do not survive gestation,

and that other GSK-30-dependent events, independent of TNFRI, exist during embryonic

development. Chapter 5

Conclusions and future direction This thesis addresses fundamentai questions regarding the signalling mechanism of TNF to hvo of the principal transcription factors activated by this inflammatory cytokine, NF-

KB and c-Jun. These transcription factors have been recognized as being downstream of

TNF sinct 1989 (Brenner et al., 1989; Conca, Kaplan & Krane, 1989; Israel et al., 1989).

Until recently Little progress has been made in elucidating the downstream kinases and adaptor proteins which are required for transducing signals from these molecules.

5.1 Apopiosis signal-regulating kinase-l

The resutts presented herein shed light on the mechanisms by which TNF stimulates the

SAPKs, thus contributing to activation of AP- 1. These findings, combined with previous results, suggest a rnodel with two parallel mecfianisms for TNF activation of SAPK. In this model, ASKl interacts in vivo with TRAF2 in a TNF-regulated fashion and thereby signals activation of both SAPKs and p38 MAPKs (Chapter 2)- The second mechanism

involves TRAFî,activation of MEKKl via GCWGCKR to activate only SAPKs (Shi &

Kehrl, 1997). it is possible that the TRAF2-,ASKL and TRAF2->GCKIGCKR-,

MEKKl pathways may be employed independently on a ceii- or sthuius-specific basis,

responding differentiaiiy to TNF signals of varying htensity or duration, It will be

important to Merdetermine how and why activation of the SAPKs by TNF involves

the cornbined effects of two apparently redundant mechaaisms.

While RIF-a can induce apoptosis through ASKl (Ichijo et al,, 1997), ASKl is

aIso involved in apoptosis triggered by chemotherapeutic agents such as paciitaxel

(~axol~'),an anti-neoplastic agent that specifcaiiy targets microtubules and arrests cel1.s

at the Gukl phase of the ce11 cycle (Jordan & Wilson, 1998). Whilt regulation of ceil death is essentiai for normal development and elhination of celis that have suffered serious damage, it is also an important defense against the emergence of cancer and other diseases (Thompson, 1995). Furthermore, as chemotherapy can also induce apoptosis in cancer cells both in vitro and in vivo (Searle et al., 1975), apoptosis plays a very important role in cancer and cancer therapy. Thus, discovering ways to modulate apoptosis may have therapeutic potentiai, and the importance of ASKl in these processes awaits furtfier evaluation,

In the six years since the initial identification of SAPKs, Iess than ten candidate substrates have been identified (only half of which have been confmed by more than one group). Genomic and proteomic-based approaches may therefore be needed to make significant strides in understanding events downstream of ASKl activation in normal and disease States. Such large-scale approaches to studying protein function should aid in indentiQing the TNF-dependent genes that are specifically regulated by ASKl or other

WKKKS.

5.2 Glycogen synthase kinase-3p

The generation of GSK-~B/-mice revealed a suprising and essential role in NF-KB activation (Chapter 3). Breeding experiments using GSK-3b heterozygous mice in a

TNFRI" background are ongoing (Chapter 4) and embryos wilI be Meranalyzed thtoughout gestation. Mice deficient in p65 NF-KB, EX$, NEMO, T2K or GSK-3b exhiiit an embryonic lethal phenotype, dying around day 13-14 of gestation due to severe tiver degeneration and apoptosis (Beg et al., 1995; Bonnard et al., 2000; Li et al., L999;

Rudolph et al., 2000). In contrast, murine knock-outs of genes encoding for proteins located more upstream in the NF-KB pathway such as TNFR1,TRAF2, W5and RF do not die of liver apoptosis (Kelliher et al., 1998; Nakano et al., 1999; PfeRer et al.,

1993; Yeh et al., 1997; see Chapter 1). Taken together, this suggests that components downstream in the NF-KB pathway have more impact (than those located upstream) in controlling hepatocyte survival during development.

The mechanism by which GSK-3P regulates NF-KB is still unclear. The means by which cells effectively sequester different populations of GSK-3 that are responsive to distinct pathways and that culminate in specific responses also remains poorIy understood. Analogous to the GSK-31Axin complex in Wnt signalling (Chapter l), we couId specuiate that a scaffold protein could be required for GSK-3B to efficientiy

phosphorylate NF-KB. GSK-3P's potential participation in the phosphorylation- dependent steps required for NF-KB activation also remains to be fully examined. In addition to targeting 1-KB for degradation, phosphorylation also plays other roies in

govecning the fùnction of NF-KB family members. For instance, serine phosphorylation

of NF43 pSO (which dimerizes with p65) is involved in stabilizing DNA binding (Li et

al., 1994). Proteolytic processing of the precursor pl05 molecule to the active p50

subunit is also accelerated by inducible phosphorylation occuring on multiple serines in

the 150 most C-terminal amino acids of plOs (MacKichan et al., 1996). Recent evidence

suggests that phosphorylation may be a general mechanism for regulating signal-Ïnduced

NF-KB nuclear translocation, as shown by signai-induced phosphorylation occurring on

Ser3 17, a residue completely conserved among ReVNF-KB members, of Drosophila

Dorsal (Drier, Huang & Steward, 1999). Finaily. inducible phosphoryIation of p65

increases the transcriptiona1 activity of NF-KB without affecthg its nuclear trandocation or DNA binding activity (Wang and Baldwin, 1998). Therefore, it is possible that GSK-

3$ may regulate the transactivation potentid of NF-KB through phosphorylation of NF-

KB proteins, either directiy or indirectiy. It is aIso conceivable that GSK-38 plays a role in inducing coactivators of the NF-KB-dependenttranscriptional complex (Zhong et al.,

1998). Hence, to characterize potentiai GSK-5g-dependent changes in p65 phosphorylation in vivo, epithelial cells stably expressing p65 (Chapter 4) will be treated with combinations of TNF-cc and lithium/debromohymenialdisine and used in phosphopeptide mapping experiments. Results hm these inhibitor studies can be validated using GSK-3fbdeficient fibroblast lines or site-directed mutagenesis of any p65 phosphorylation site(s).

Among the targets of GSK-3 regulation is a "Who's Who" list of transcription factors including $-catenin, NF-AT, CfEBP, cJun, c-Myc, NF-KB and CREB, al1 of which are negatively regulated in manimaLian celk As such, GSK-3 carries significant responsibility in regulating gene expression and it is therefore perhaps not too surprising that its own regulation is complex and redundant. Hence, in light of the role of GSK-3fl in the Wnt signalling pathtvay, it is remarkable that GSK-~~S/-embryos are normal for early embryonic patterning and cell-fate determination. In fact, defects in Wnt signailhg

(cyclin DI and B-catenin accumulation) or glycogen metabolism were not evident. in simpler organisms, such as Xenoprts, zebdtsh and sea urchins, overexpression of dominant-negative GSK-3 mutants (of either isoforrn) leads to improper axial formation

(Chapter 1). GSK-3a (or other reIated proteins) rnay functionalIy complement for GSK-

38 dwing murine early embryonic development- Conditional inactivation of GSK-3a by

Cre-loxP recombinase technology is already underway and will alIow for mer assessment of GSK-3's in vivo physiologie role (Hoeflich and Woodgett, unpublished data). A double-floxed GSK-3a and GSK-3f3 would allow examination of the consequences of total GSK-3 dektion throughout development and within adult tissues

(Rossant & MchIahon, 1999).

The ultimate merit of basic research in the so-called "life sciences" is largely gauged by its impact on medicine and human health. Although much research is focussed on understanding the action of disease-related/causing genes such as NF-KB and c-Jun, these gene products may not be amenable to pharmaceutical dmg intervention. It is therefore important to also concentrate efforts on the ceceptors, channels or enzymes (like

kinases, which make good drug targets) upstream of these disease genes. Significant success has already been achieved with the development of small molecule inhibitors to a

number of protein-tyrosine kinases (with several inhibitors in clinical trials) and simiiar

efforts for serinelthreonine protein kinases are under way. Like debromohymenialdisine,

most of these inhibitors target the ATP binding pocket (Chapter 4). The ability of

debromohymenialdisine to inhibit GSK-3b/NF-~B-driven idammatory gene products

lends support to the proposai that an inhibitor of GSK3B could provide an anti-

inflarnrnatory therapeutic agent. Severd infiammatory mouse models which are highly

dependent on NF-KB, and that couid be applicable for analysis with lithium or DBH,

have been described. For instance, a relatively straightforward mode1 for NF-KB-

dependent rheumatoid arthritis is as follows: DBAlIJ mice are immunized with lype II

coIIagen; 21 days later, coliagen is re-injectai and inhibitors are adrninistered three times

per week; the grade of arthntis is scored and the whole experirnent takes approxirnately

45 days (ümezawa et al.. 2000). 5.3 Evolutionary Perspective

Understanding the function of TM: signalling in vertebrates may benefit from studies conducted in Drosophila. In Drosophila embryos, formation of the dorsoventral axis is controlled by the NF-KB family member, Dorsal. A matemally encoded signal transduction pathway controls the formation of a ventrai-to-dorsal nuclear gradient of

Dorsal. Cytoplasmic retention of Dorsal is mediated by Cactus, an 1-KB protein. In response to upstrearn stimuli, Cactus is phosphorylated on regulatory serines that are

conserved with other 1-KBS. This phosphorylation causes Cactus degradation and

liberation of Dorsal, which then translocates to the nucleus (Belvin et al., 1995). Dorsal

stimulates transcription of Twist, which codes for a helix-loop-heh protein that initiates

rnesoderm differentiation (Jiang et al., 1991). In congruence with the rnodel predicted

tiom the murine studies described in Chapter 3 (see Figure 16): in the absence of the

Drosophila GSK-3 homohgue, Sgg/.w3, ernbryos exhibit a substantid reduction in Twist

mRNA expression (Dr. Amen Manoukian, personal communication). This opens a new

avenue for studying the physiologicai role of GSK-3 on NF-KB. Furthemore, genetic

screens could be conducted to identifi the components fûnctioning upstrearn and

downstream of 2~3'~~in the Drosophiia NF-KB pathway.

Two TRAF homologues are also present in the Drosaphiia genome and they

exhibit a variety of properties cornmoniy observed in the TRAF-family proteins of

mammais, including the abiiity to modulate NF-KB and SAPK activities (Liu et al.,

1999). Thus, in addition to mammaiian ceii culture experiments geared to biochemicaiiy

determine which downstream kinase couples TRAFs to the SAPK pathway under specific situations (see Chapter 21, it is exciting that we now have the tools in hand to assess genetically in Drosophila the role of MAPKKKs downstreatn of TRAFs. ASKl shows the highest amino acid identity (55%) to the putative cataiytic domain of the Drosophila

MAPKKK, PK93B. PK93B was fmt isolated fiom an eye-antemal imagina1 disc cDNA library and the gene was mapped to 9288-10 on the third chromosome by in situ hybridization to polytene chromosomes (Wasserman et al., 1996). Analysis of

Drosophila TRAFs and PK92B would greatiy facilitate our understanding of the normal physiological functions of these proteins and their roles in innate immune responses to pathogens. In tum, this could be helpfui in unraveling cornplex signdling pathways regulating SAPK activation downstream of TNF-like receptors. References Adams, R. H., Porras, A., Alonso, G., Jones, M., Vintersten, K., Panelli, S., Valladases, A., Perez, L., Klein, R, & Nebreda, A. R. (2000). Essentiai role of p38alpha MAP kinase in placental but not embryonic cardiovascular development. Mol Cell6, 109-16.

Adler, V., Polotskaya, A., Wagner, F. and Kraft, A.S. (1992). Affmity-purified cJun arnino-terminal protein kinase requires serinelthreonine phosphorylation for activity. J Biol Chem, 267, l7OOl-l'iOO5.

Alessi, D. R., hdjelkovic, M., Caudwell, B., Cron, P., Morrice, N., Cohen, P. & Hemmings, B. A. (1996). Mechanism of activation of protein kinase B by insulin and IGF-1. Embo J 15,6541-51.

Allen, M., Svensson, L., Roach, M.? Hambor, J., McNeish, J. & Gabel, C. A. (2000). Deficiency of the stress kinase p38alpha results in embryonic lethaiity: characterization of the kinase dependence of stress responses of enzyme-deficient embryonic stem cells. J Exp Med 191,859-70.

AIlison, J. H. & Stewart, M. A. (1971). Reduced brain inositot in lithium-treated rats. Nat New Biol 233,267-8.

Angerer, L. M. & Angerer, R. C. (2000). Animal-vegetal axis patternhg mechsinisms in the eady sea urchin ernbryo. Dev Bi01 218, 1-12.

Arch, R.H., Gedrich, R.W. and Thompson, C.B. (1998) Tumor necrosis factor receptor- associated factors (TRAFs)-a farnily of adapter proteins that regulates life and death. Genes Dev, 12,2821-2830.

Atack, J. EL, Cook, S. M., Watt, A. P., Fletcher, S. R. & Ragan, C. 1. (1993). tn vitro and in vivo inhibition of inositol monophosphatase by the bisphosphonate L-690,330. J lveurochern 60,652-8.

Baeuerle, P. A. & Baltimore, D. (1988). IKB: a specific inhibitor of the NF-KB transcription factor. Science, 242,540-546.

Baker, S. J. & Reddy, E. P. (1998). Modulation of life and death by the TM: receptor superfamily .Oncogene 17,326 1-70.

Barkett, M. & Gilmore, T, D, (1999). Control of apoptosis by ReWF-kappaB transcription factors. Oncogene 18,69 10-24.

Beais, C, R., Sheridan, C. M., Turck, C. W., Gardner, P. & Crabtree, G. R. (1997). Nuclear export of NF-ATc enhanced by glycogen synthase kinase-3. Science 275, 1930- 4.

Beg, A. A. & Baltimore, D. (1996). An essentiai role for NF-kappaB in preventing TNF- alpha-induced cell death. Science 274,782-4. Beg, A. A., Sha, W. C., Bronson, R. T., Ghosh, S. & Baltimore, D. (1995). Embryonic lethality and liver degeneration in mice lacking the RelA component of NF-kappa B. iVarure 376, 167-70.

Behrens, A., Jochum, W., Sibilia, M. & Wagner, E. F. (2000). Oncogenic transformation by ras and fos is mediated by cJun N-terminal phosphorylation. Oncogene 19,2657-63.

Behrens, A., Sibilia, M. & Wagner, E. F. (1999). Amino-terminal phosphorylation of c- Jun regulates stress-induced apoptosis and cellular proliferation. Nat Genet 21,326-9.

Behrens, J., Jerchow, B. A., Wurtele, M., Grimm, J., Asbrand, C., Wirtz, R., Kuhi, M., Wedlich, D. & Birchmeier, W. (1998)- Functional interaction of an axin homolog, conductin, with beta-catenin, APC, and GSK3beta. Science 280,596-9.

BeIvin, M.P., Jin, Y., Anderson, K.V. (1995). Cactus protein degradation mediates Drosophila dorsal-ventral signalling. Genes Dev. 9,783-793.

Benjamin, W. B., Pentyala, S. N., Woodgett, J. R., Hod, Y. & Marshak, D. (1994). ATP citrate-lyase and glycogen synthase kinase-3 beta in 3T3-L1 cells during differentiation into adipocytes. Biochem J300,477-82.

Beraud, C., Hemel, W. J. & Baeuerle, P. A. (1999). Involvement of regulatory and catalytic subunits of phosphoinositide 3- kinase in NF-@ activation. Proc Nat1 Acad Sci USA, 96,429-434.

Berberich, I., Shu, G., Siebelt, F., Woodgett, J.R., Kyriakis, J.M. and Clark, E.A. (1996) Cross-linking CD40 on B cells preferentially induces stress-activated protein kinases rather than mitogen-activated protein kinases. EMBO Journal, 15,92-101.

Berridge, M. J., Downes, C. P. & Hanley, M. R. (1989). Neural and developmental actions of lithium: a unifying hypothesis. Cell59,411-9.

Beyaert, R., Vanhaesebroeck, B., Suffys, P., Van Roy, F. & Fiers, W. (1989). Lithium chloride potentiates tumor necrosis factor-mediated cytotoxkity in vitro and in vivo. Proc Nat1 Acad Sci USA86,9494-8.

Bishop, A. L. & Hall, A. (2000). Rho GTPases and their effector proteins. Biochem J348 Pt 2,241-55.

Bonnard, M., Mirtsos, C., Suzuki, S., Graham, K., Huang, J., Ng, M., Itie, A., Wakeham, A., Shahinian, A., Hemel, W. J., Elia, A. J., Shiliingiaw, W., Mak, T. W., Cao, 2. & Yeh, W. C. (2000). Deficiency of T2K leads to apoptotic liver degeneration and impaired NF- kappaB-dependent gene transcription. Embo J 19,4976-85. Borsch-Haubold, A. G., Pasquet, S. & Watson, S. P. (1998). Direct inhibition of cyclooxygenase-l and -2 by the hase inhibitors SB 203580 and PD 98059. SB 203580 also inhibits thromboxane synthase. JBiol Chem 273,28766-72.

Bourouis, M., Moore, P., Ruel, L., Grau, Y., Heitzler, P. & Simpson, P. (1990). An early embryonic product of the gene shaggy encodes a serine/threonine protein kinase related to the CDC28/cdc2+ subfarnily. Embo J9,2877-84.

Boyle, W. J., Smeal, T., Defize, L. H., Angel, P., Woodgett, J. R., Karin, M. & Hunter, T. (1991). Activation of protein kinase C decreases phosphorylation of c-Jun at sites that negatively regulate its DNA-binding activity. Ce11 64, 573-84.

Brenner, D. A., O'Hara, M., Angel, P., Chojkier, M. & Karin, M. (1989). Prolonged activation of jun and colIagenase genes by tumour necrosis factor-alpha. Nature 337, 661-3.

Breton, J. J. & Chabot-Fletcher, M. C. (1997). The natural product hymenialdisine inhibits interleukin-8 production in U937 cells by inhibition of nuclear factor-kappaB. J Pharmacol lGXp Ther 282,459-66.

Burbelo, P. D., Drechsel, D. & Hall, A. (1995). A conserved binding motif defines numerous candidate target proteins for both Cdc42 and Rac GTPases. J Bi01 Chem 270, 2907 14.

Burow, M. E., Weldon, C, B., Melnik, L. I., Duong, B. N., Coilins-Burow, B. M.! Beckrnan, B. S. & McLachlan, J. A. (2000). PI3-K/AKT regdation of NF-kappaJ3 signaling events in suppression of ï'NF-induced apoptosis. Biochem Biophys Res Commun 271,342-5.

Cade, J. F. (2000). Lithium saits in the treatment of psychotic excitement. 1949. Bull World Health Organ 78,j1 8-20.

Cadigan, K. M. & Nusse, R. (1997). Wnt signaling: a common theme in animai development. Genes Dev 11,3286-305.

Canagarajah, B. J., Khokhlatchev, A., Cobb, M. H. & Goldsmith, E. J. (1997). Activation mechanism of the MAP kinase ERK2 by dual phosphorylation. CelZ 90,859-69.

Cao, Z., Xiong, J., Takeuchi, M., Kurama, T. & Goeddel, D. V. (1996). TRAF6 is a signal transducer for interleukin-1. Nature 383,443-6.

Cardone, M.H., Saivesen, G.S., Widmann, C., Johnson, G. and Frisch, S.M. (1997) The re,@ation of anoikis: MEKK-1 activation requires cleavage by caspases. Cell, 90,3 15- 323. Chalecka-Franaszek, E. & Chuang, D. M. (1999). Lithium activates the serine/threonine kinase Akt-I and suppresses glutamate-induced inhibition of Akt-1 activity in neurons. Proc Nari Acad Sci USA 96,8745-50.

Chan, E.D., Winston, B.W.,Jarpe, M.B., Wynes, M.W. and Riches, D.W. (1997) Preferential activation of the p46 isoform of JNK/SAPK in mouse macrophages by TNF aipha. Proc Nat1 Acad Sci USA, 94,13 169-13 174.

Chan, E. F., Gat, U., McNiff, J. M. & Fuchs, E. (1999). A cornmon human skin turnour is caused by activating mutations in beta-catenin. Nat Genet 21,4 10-3.

Chang, H.Y., Nishitoh, H., Yang, X., Ichijo, H. and Baltimore, D. (1998) Activation of apoptosis signal- regulating kinase 1 (ASKI) by the adapter protein dm.Science, 281, 1860-1863,

Chen, R. H., Ding, W. V. & McComick, F. (2000). Wnt signaling to beta-catenin involves two interactive components. Glycogen synthase kinase-3 beta inhibition and activation of protein kinase C. JBiol Chem 275, 17894-9.

Cheng, G., Cieary, A.M., Ye, Z.S., Hong, DA., Ledeman, S. and Baltimore, D. (1995) Involvement of CRAF1, a relative of TRAF, in CD40 signaling. Science, 267, 1494- 1498.

Chow, C. W., Rincon, M., Cavanagh, J., Dickens, M. & Davis, R. 1. (1997). Nuciear accumulation of NFAT4 opposed by the JNK signai transduction pathway. Science 278, 1638-41.

Coffer, P. J., Jin, J. & Woodgett, J. R. (1998). Protein kinase B (c-Ab): a multifunctionai mediator of phosphatidyIinositol3-kinase activation. Biochem J 335, 1-1 3.

Coghlan, M. P., Culbert, A. A., Cross, D. A., Corcoran, S. L., Yates, J. W., Pearce, N. J., Rausch, O. L., Murphy, G. J.: Carter, P. S., Roxbee Cox, L., Mills, D., Brown, M. J., Haigh, D., Ward, R. W., Smith, D. G., Murray, K. J., Reith, A. D. & Holder, J. C. (2000). SelecUve smaii molecule inhibitors of glycogen synthase kinase-3 modulate glycogen metabolism and gene transcription. Chem Bioi 7,793-803.

Cohen. P. (1997). The search for pbysiological substrates of MAP and SAP kinases in manundian celis. Trends Celt Bi01 7,353-61.

Conca, W., Kaplan, P. B. & Krane, S. M. (1989). Increases in Ievels of procollagenase mRNA in human fibroblasts induced by interleukin-i, tumor necrosis factor-alpha, or senun folIow c-jun expression and are dependent on new protein çynthesis. Tram Assoc Am Physicians 102, 195-203, Cook, D., Fry, M. J., Hughes, K., Sumathipala, R., Woodgett, J. R. & Dale, T. C. (1996). Wingless inactivates glycogen synthase kinase-3 via an intracellular signalling pathway which involves a protein kinase C. Embo J 15,4526-36.

Coso, O. A., Chiariello, M., Yu, J. C., Teramoto, H., Crespo, P., Xu, N., Miki, T. & Gutkind, J. S. (1995). The small GTP-binding proteins Racl and Cdc42 regdate the activity of the JNWSAPK signalhg pathway. Cell81, 113746.

Courchesne, W. E., Kunisawa, R. & Thonier, J. (1989). A putative protein kinase overcomes pheromone-induced anest of ceIl cycling in S. cerevisiae. CeIl 58,1107-19.

Cross, D. A., Alessi, D. R., Cohen, P., Andjelkovich, M. & Hemmings, B. A. (1995). Inhibition of glycogen synthase kinase-3 by insulin mediated by protein kinase B. Nature 378,785-9.

Cross, D. A., Alessi, D. R,, Vandenheede, J. R., McDowell, H. E., Bundal, H. S. & Cohen, P. (1 994). The inhibition of glycogen synthase kinase-3 by insulin or insulin-like growth factor 1 in the rat skeletal muscle ce11 line L6 is blocked by wortmannin, but not by rapamycin: evidence that wortmannin blocks activation of the mitogen-activated protein kinase pathway in L6 cells between Ras and Raf. Biochem J303,21-6.

Cuenda, A., Rouse, J., Doza, Y. N., Meier, R.. Cohen, P., Gallagher, T. F., Young, P. R. & Lee, J. C. (1995). SB 203580 is a specific inhibitor of a MAP kinase homologue which is stimulated by cellular stresses and interleukin-1. FEBS Len 364,229-33.

Dale, T. C. (1998). Signa1 transduction by the Wnt farnily of ligands. Biochem J, 329, 209-223.

Delcommenne, M., Tan, C,, Gray, V., Rue, L., Woodgett, J. & Dedhar, S. (1998). Phosphoinositide-3-OH kinase-dependent regulation of glycogen synthase kinase 3 and protein kinase B/AKT by the integrin-linked kinase. Proc Nat1 Acad Sci USA 95, 1121 1- 6.

Delhase, M., Li, N. & Karin, M. (2000). Kinase regulation in infiammatory response. Nature 406,367-8.

Decijard, B., Hibi, M., Wu, I.H., Barrett, T., Su, B., Deng, T., Karin, M. and Davis, R.J. (1994) JNKl: a protein kinase stimulated by UV light and Ha-Ras that binds and phosphorylates the c-Jun activation domain. Cell, 76, 1025-1037.

Derijard, B., Rahgeaud, J., Barrett, T., Wu, 1. H., Han, J., Ulevitch, R. J. & Davis, R I. (1995). Independent human MAP-kinase signai transduction pathways defined b;. MEK and MKK isoforms. Science 267,682-5, Dickens, M., Rogers, J. S., Cavanagh, J., Raitano, A., Xia, Z., Halpern, J. R., Greenberg, M. E., Sawyers, C. L. & Davis, R. J. (1997). A cytopiasmic inhibitor of the JMC signai transduction pathway. Science 277,693-6.

DiDonato, J.A., Hayakawa, M., Rothwarf, D.M., Zandi, E., Karin, M. (1997). A cytokine-responsive IkappaB kinase that activates the transcription factor NF-kappa& Nariire 388,548-554.

Diehl, J. A., Cheng, M., Roussel, M. F. & Sherr, C. J. (1998). Glycogen synthase kinase- 3beta regulates cyclin Dl proteolysis and subcellular localization. Genes Dev 12, 3499- 511-

Diehl, J. A., Tong, W., Sun, G. & Hannink, M. (1995). Tumor necrosis factor-alpha- dependent activation of a RelA homodimer in astrocytes. Increased phosphorylation of RelA and MAD-3 precede activation of RelA. J Bi01 Chem 270,2703-7.

DiMartino, M., Wolff, C., Patil, A. & Nambi, P. (1995). Effects of a protein kinase C inhibitor (PKCI) on the development of adjuvant-induced arthritis (AA) in rats. Infiamm Res 44 Suppl2, S123-4.

Ding, V. W., Chen, R. H. & McConnick, F. (2000). Differential regulation of glycogen synthase kinase 3beta by insulin and wnt signaling. J Bi01 Chem 275,32475-8 1.

Doi, T. S., Marino, M. W., Takahashi, T., Yoshida, T., Sakakura, T., Old, L. J. & Obata, Y. (1999). Absence of tumor necrosis factor rescues RelA-deficient mice fiom ernbryonic lettiality. Proc Nd1 Acad Sci USA 96,2994-9.

Dominguez, I., Itoh, K. & Sokol, S. Y. (1995). Role of glycogen synthase kinase 3 beta as a negativc regulator of dorsoventraI avis formation in Xenopus embryos. Proc Nad Acad Sei USA. 92,8498-8502.

Dong, C., Yang, D. D., Wysk, M., Whitrnarsh, A. J., Davis, R. J. & Flavell, R. A. (1998). Defective T ce11 differentiation in the absence of Jnkl. Science 282,2092-5.

Drîer, E. A., Huang, L. H. & Steward, R. (1999). Nuclear import of the Drosophila Re1 protein Dorsai is regulated by phosphoryIation. Genes Dev 13,556-68.

Dungy, L. J., Siddiqi, T-A. & Khan, S. (1991). C-jun and jun-B oncogene expression during placental development. Am J Obstec Gynecol165,1853-6.

Eastman, Q. & Grosschedl, R (1999). Regulation of LEF-l/TCF transcription factors by Wnt and other signais. Curr Opin Cell Bi01 11,23340,

Eldar-Fhkelman, H. & Krebs, E. G. (1997). Phosphorylation of insuiin receptor subsh-ate 1 by glycogen synthase kinase 3 impairs insulin action. Proc Nat1 Acad Sci USA 94, 9660-4. Elion, E. A., Grisafi, P. L. & Fink, G. R. (1990). FUS3 encodes a cdcZ+/CDC28-reiated hase required for the transition fiom mitosis into conjugation. Cell60,649-64.

Embi, N., RyIatt, D. B. & Cohen, P. (1980). Glycogen synthase kinase3 fiom cabbit skeletal muscle. Separation from cyclic-AMP-dependent protein kinase and . Etcr J Biochem 107,s 19-37.

Ernily-Fenouil, F., Ghiglione, C., Lhomond, G., Lepage, T. & Gache, C. (1998). GSK3betdshaggy mediates patternhg along the animal-vegetal axis of the sea urchin ernbryo. Developmenr 125,2489-98.

Engel, M. E., McDonnell, M. A., Law, B. K. & Moses, H. L. (1999). Interdependent SMAD and JNK signaling in tcansfocming growth factor-beta-mediated transc&%ion. J Bioi Chem 274,374l3-20.

Eyers, P. A., Craxton, M., Momce, N., Cohen, P. & Goedert, M. (1998). Conversion of SB 203580-insensitive MAP kinase family rnembers to dmg-sensitive foms by a singk amino-acid substitution. Chem Bi01 5,321-8.

Fang, X., Yu, S. X., Lu, Y., Bast, R. C., Jr., Woodgett, J. R. & Mills, G. B. (2000). Phosphorytation and inactivation of glycogen synthase kinase 3 by protein kinase A Pn Process Citation]. Proc Nat1 Acad Sci USA 97, 11960-5.

Fanger, G. R,, Johnson, N. L. & Johnson, G. L. (1997). MEK kinases are regulated by EGF and selectively interact with RacICdc42. Emba J 16,496 1-72.

Fm,G. H., 3rd, Ferkey, D. M., Yost, C., Pierce, S. B., Weaver, C. & Kimelman, D. (2000). Interaction among GSK-3, GBP, min, and APC in Xenopus axis specification. J Ceil Bi01 148,69 1-702.

Fiol, C.J., Haseman, J. H., Wang, Y. H., Roach, P. J., Roeske, R. W., Kowalcnik, M. & DePaoli-Roach, A. A. (1988). Phosphoserine as a recognition determinant for glycogen synthase kinase-3: phosphorylation of a synthetic peptide based on the G-component of protein phosphatase-1. Arch Biochem Biophys 267,797-802,

Fiol, C. J., Mahrenholz, A- M., Wang, Y., Roeske, R. W. & Roach, P. J. (1987). Formation of protein kinase recognition sites by covalent modification of the substrate. Molecdar mechanism for the synergistic action of casein kinase II and giycogen synthase kinase 3. J Biol Chem 262, 14042-8.

Fiol, C. J., Wang, A., Roeske, R. W. & Roach, P. J. (1990). Ordered mdtisite protein phosphorylation. Analysis of glycogen synthase kinase 3 action using mode1 peptide substrates. J Biol Chem 265,6061-5. Fisher, D. L., Morin, N. & Doree, M. (1999). A novel role for glycogen synthase kinase-3 in Xenopus development: maintenance of oocyte ceil cycle arrest by a beta-catenin- independent mechanism. Development 126,567-76.

Frisch, S.M., Vuori, K., Kelaita, D. and Sicks, S. (1996) A roIe for Jun-N-terminal kinase in anoikis; suppression by bcl-2 and crmA. JCell Biol, 135, 1377-1382.

Fuchs, S. Y., Adler, V., Pincus, M. R. & Ronai, S. (1998). MEKKl/JNK signalhg stabilizes and activates p53. Proc Nat1 AcadSci USA 95,10541-6.

Ganiatsas, S., Kwee, L., Fujiwara, Y., Perkins, A., Ikeda, T., Labow, M. A. & Zon, L. 1. (1998). SEKl deficiency reveals mitogen-activated protein kinase cascade crossregdation and leads to abnormai hepatogenesis, Proc Nat1 Acad Sci USA 95,688 1- 6.

Gat, U., DasGupta, R., Degenstein, L. & Fuchs, E. (1998). De Novo hair folIicie morphogenesis and hair tumors in mice expressing a truncated beta-catenin in skin. Ce12 95,605-14.

Ginger, R. S., Dalton, E. C., Ryves, W. J., Fukuzawa, M., Williams, 1. G. & Harwood, A. J. (2000). Glycogen synthase kinase-3 enhances nuclear export of a dictyostelium STAT protein. Embo J 19,5483-9 1.

Glise, B., Bourbon, H. & Noselli, S. (1995). hemipterous encodes a novel Drosophila MAP kinase kinase, required for epithelial ce11 sheet movement. Ce11 83,45 1-6 1.

Goode, N., Hughes, K., Woodgett, J. R. & Parker, P. J. (1992). DEerential regdation of glycogen synthase kinase-3 beta by protein kinase C isotypes. J Biol Chem 267, 16878- 82.

Gotoh, Y. and Cooper, J.A. (1998) Reactive oxygen species- and dimerization-induced activation of apoptosis signal-regulating kinase 1 in tumor necrosis factor-alpha signal transduction. J Biol Chem, 273, 17477-17482.

Groden, J., Thliveris, A., Samowitz, W., Carlson, M., Gelbert, L., Albertsen, W., Joslyn, G., Stevens, J., Spirio, L., Robertson, M. & et al. (1991). Identification and characterization of the familial adenomatous polyposis COLgene- CelZ 66,589-600.

Gms, H.J. (1996) Molecuiar, structural, and biological characteristics of the tumor necrosis factor ligand superfamiiy. Int JClin Lab Res, 26, 143-159.

Gum, R. J., McLaughiin, M. M., Kumar, S., Wang, Z., Bower, M. J., Lee, S. C., Adams, J. L., Livi, G. P., Goldsmith, E. J. & Young, P. R. (1998). Acquisition of sensitivhy of stress-activated protein kinases to the p38 inhibitor, SB 203580, by aiteration of one or more amino acids within the ATP binding pocket. JBiol Chem 273,15605-10. Gupta, S., Barrett, T., Whitmarsh, A. J., Cavanagh, J., Sluss, H. K., Derijard, B. & Davis, R. J. (1996). Selective interaction of JNK protein kinase isoforms with transcription factors. Embo J 15,2760-70.

Gupta, S., Campbell, D., Derijard, B. & Davis, R. J. (1995). Transcription factor ATF2 regulation by the JNK signai transduction pathway. Science 267,389-93.

Hall-Jackson, C. A., Goedert, M., Hedge, P. & Cohen, P. (1999). Effect of SB 203580 on the activity of c-Raf in vitro and in vivo. Oncogene 18,2047-54.

Hallcher, L. M. & Sherman, W- R. (1980). The ef3ects of lithium ion and other agents on the activity of myo-inositol-1-phosphatase fiom bovine brain. J Bi01 Chem 255, 10896- 901.

Hamada, F., Tomoyasu, Y., Takatsu, Y., Nakamura, M., Nagai, S., Suzuki, A., Fujita, F., Shibuya, H., Toyoshima, K., Ueno, N. & Akiyama, T. (1999). Negative regulation of Wingless signaling by D-axin, a Drosophila homolog of min. Science 283, 1739-42.

Han, J., Jiang, Y., Li, Z., Kravchenko, V. V. & Ulevitch, R. J. (1997). Activation of the transcription factor MEF2C by the MAP kinase p38 in inflammation. Nature 386,296-9.

Han. J.. Lee, J. D.. Bibbs. L. & Ulevitch, R. J. (1994). A MAP kinase targeted by endotoxin and hyperosmolarity in mammalian cells. Science 265, 808-1 1.

Han, 2. S., Enslen, H., Hu, X., Meng, X., Wu, 1. H., Barrett, T., Davis, R. J. & Ip, Y. T. (1998). A conserved p38 mitogen-activated protein hase pathway regulates Drosophila immunity gene expression. Mol Ce11 Bi01 18,3527-39.

Hanger, D. P., Hughes, K., Woodgett, J. R., Brion, J. P. & Anderton, B. H. (1992). Glycogen synthase kinase-3 induces Alzheimer's disease-like phosphorylation of tau: generation of paired helicai filament epitopes and neuronaf localisation of the kinase. Netrrosci Lett 147,58-62.

Harada, N., Tamai, Y., Ishikawa, T., Sauer, B., Takaku, K., Osba, M. & Taketo, M. M. (1999). Intestinai polyposis in mice with a dominant stable mutation of the beta-catenin gene. Embo J 18,593 1-42.

Harwood, A. J., Plyte, S. E., Woodgett, J., Strutt, H. & Kay, R. R. (1995). Glycogen synthase kinase 3 regdates cell fate in Dictyostefim. Ce11 80, 139-48.

Hatai, T., Matsuzawa, A., inoshita, S., Mochida, Y., Kuroda, T., Sakamaki, K., Kuida, K., Yonehara, S., Ichijo, H. & Takeda, KI (2000). Execution of apoptosis signal- regulating kinase 1 (ASKI)-induced apoptosis by the mitochondria-dependent caspase activation. J Biol Chem 275,26576-8 1. Hayashi, S., Rubinfeld, B., Souza, B., Polakis, P., Wieschaus, E. & Levine, A. J. (1997). A Drosophiia homolog of the turnor suppressor gene adenornatous polyposis coli down- reguIates beta-catenin but its zygotic expression is not essential for the regulation of Armadillo. Proc Nat1 Acad Sci USA 94,242-7.

He, B., Meng, Y. H. & Mivechi, N. F. (1998a). Glycogen synthase kinase 3beta and extracelIular signal-regulated kinase inactivate heat shock transcription factor 1 by facilitating the disappearance of transcriptionally active granules afte-r heat shock. il Cell Biol18,6624-33.

He, T. C., Sparks, A. B., Rago, C., Hermeking, H., Zawel, L., da Costa, L. T., Morin, P. J., Vogelstein, B. & Kinzler, K. W. (1998b). Identification of c-MYC as a target of the APC pathway. Science 281, 1509-12.

He, X., Saint-Jeannet, J. P., Woodgett, J. R., Varmus, H. E. & Dawid, 1. B. (1995). GIycogen synthase kinase-3 and dorsoventral patternhg in Xenopus embryos. Nature 374617-23.

Hemrnings, B. A., Aitken, A., Cohen, P., Rymond, M. & Hofmann, F. (1982). Phosphorylation of the type-II regulatory subunit of cyclic-AMP-dependent protein kinase by glycogen synthase kinase 3 and glycogen synthase kinase 5. Eur J Biochem 127,473-8 1.

Hernmings, B. A., Yellowlees, D., Kernohan, J. C. & Cohen, P. (1981). Purification of glycogen synthase kinase 3 from rabbit skeletal muscle. Copurification with the activating factor (FA) of the (Mg-ATP) dependent protein phosphatase. Eur J Biochem 119,443-51.

Hershko, A. (1997). Roles of ubiquitin-mediated proteolysis in ce11 cycle control. Curr Opin Cell Bi01 9,788-99,

Herskowitz, 1. (1995). MAP kinase pathways in yeast: for mating and more. Cell80, 187- 97.

Hibi, M., Lin, A., Smeal, T., Minden, A. & Karin, M. (1993). Identification of an oncoprotein- and UV-responsive protein kinase that binds and potentiates the cJun activation domain. Genes Dev 7,2 135-48.

HiIberg, F., Aguzzi, A., Howells, N. & Wagner, E. F. (1993). c-jun is essential for normal mouse development and hepatogenesis. Nature 365,179-8 1.

Hoeflich, K. P., Luo, J., Rubie, E. A., Tsao, M. S., Ji, O. & Woodgett, J. R. (2000). Requirement for glycogen synthase kinase3beta in ce11 survival and NF-kappaB activation. Nature 406,86-90. HoeBich, K. P., Yeh, W. C., Yao, Z., Mak, T. W. & Woodgeît, J. R (1999). Mediation of TNF receptor-associated factor effector fùnctions by apoptosis signai-regulaiing kinase-1 (ASK1). Oncogene 18,5814-20.

Holland, P.M., Suzanne, M., Campbell, J.S., NoseUi, S. and Cooper, J.A. (1997) MKK7 is A stress-activated mitogen-activated protein kinase kinase functionally related to hemipterous. J Bi01 Chem, 272,24994-24998.

Horstadius, S. (1973). Two metabolic systems with different reactions to temperature in sea urchin Iarvae. kpCell Res 78,251-5,

Hsu, H., Huang, J., Shu, H.B., Baichwal, V. and Goeddel, D.V. (1996a) TNF-dependent recruitment of the protein kinase IUP to the TNF receptor- 1 signaling complex. Immuniiy, 4,387-396.

Hsu, H., Shu, H.B., Pan, M.G. and Goeddel, D.V. (1996b) TRADD-TRAF2 and TRADD-FADD interactions define two distinct TNF receptor 1 signal transduction pathways, Cell, 84,299-308,

Hsu, H., Xiong, J. & Goeddel, D. V. (1995). The TNF receptor 1-associated protein TRADD signals ce11 death and NF-kappa B activation. Cell81,495-504.

Hu, H.M., O'Rourke, K., Boguski, M.S., Dixit, V.M. (1994). A novel EUNG finger protein interacts with the cytoplasmic domain of CD4O. J Biol Chem 269,30069-30072.

Hu, Y., Baud, V., Delhase, M., Zhang, P., Deerinck, T., Ellisman, M., Johnson, R., Karin, M. (1999). Abnormal morphogenesis but intact IKK activation in mice lacking the IKKalpha subunit of ikappaB kinase. Science 284,3 16-320.

Hughes, K., Nikolakaki, E., Plyte, S. E., Totty, N. F. & Woodgett, J. R. (1993). Modulation of the glycogen synthase kinase-3 farniiy by tyrosine phosphorylation. Embo J 12,803-8.

Hughes, K., Ramakrishna, S., Benjamin, W. B. & Woodgett, J. R (1992). Identification of multifunctional ATP-citrate lyase kinase as the alpha-isofonn of glycogen synthase kinase-3. Biochem J288,309-14.

Hunter, T., Angel, P., Boyle, W. J., Chiu, R., Freed, E., Gould, K. L., Isacke, C. M., Karin, M., Lindberg, R. A. & van der Geer, P. (1988). Targets for signal-transducing protein kinases. Coid S'ring Harb S'p Quant Bi01 53 Pt 1,I31-42.

Ichijo, H., Nishida, E., Irie, K., ten Dijke, P., Saitoh, M., Moriguchi, T., Takagi, M., Matsumoto, K., Miyazono, K. & Gotoh, Y. (1997). induction of apoptosis by ASK1, a mammaiian MAPKKK that activates SAPWJNK and p38 sigoaling pathways. Science 275,904. Ihle, J. N. (2000). The challenges of translating knockout phenotypes into gene function. Ce11 102,13 14.

Ikeda, S., Kishida, S., Yamamoto, H., Murai, H., Koyama, S. & Kikuchi, A. (1 998). Auin, a negative regulator of the Wnt signaling pathway, forms a complex with GSK- 3beta and beta-catenin and promotes GSK-3beta-dependent phosphorylation of beta- catenin. Embo J 17, 137 1-84,

Inhoni, R. C. & Majems, P. W. (1987). Inositol polyphosphate 1-phosphatase fiom calf brain. Purification and inhibition by Li+, Ca2+, and Mn2+. J Bi01 Chem 262, 15946-52.

Insall, R, (1995). Glycogen synthase kinase and Dictyostelium development: old pathways pointing in new directions? Trends Genet 11,37-9.

Iordanov, M., Bender, K., Ade, T., Schmid, W., Sachsenmaier, C., Engel, K., Gaestel, M., Rahmsdorf, H. J. & Herrlich, P. (1997). CREB is activated by WC through a p3 8/HOG-1 -dependent protein kinase. Embo J 16, 1009-22.

Ishida, T., Mizushirna, S., huma, S., Kobayashi, N., Tojo, T., Sunik*, K., Aizawa, S., Watanabe, T., Mosialos, G., Kieff, E., Yamamoto, T. and Inoue, J. (1996a). Identification of TRAF6,a novel tumor necrosis factor receptor- associated factor protein that mediates signaling from an arnino- terminal dornain of the CD40 cytoplasmic region. J Bi01 Chem, 271,28745-28748.

Ishida, T.K., Tojo, T., Aoki, T., Kobayashi, N., Ohishi, T., Watanabe, T., Yamamoto, T. and Inoue, J. (1996b) TRAFS, a novel tumor necrosis factor receptor-associated factor farnily protein, mediates CD40 signaiing. Froc Nat1 Acad Sci US A, 93,9437-9442.

IsraeI, N., Hazan, U., Alcarni, J., Munier, A., Arenzana-Seisdedos, F., Bachelecie, F., Israel, A. & Virelizier, J. L. (1989). Tumor necrosis factor stimdates transcription of HIV-1 in human T lymphocytes, independently and synergistically with mitogens. J Immzrnol113,3956-60.

Itoh, K., Krupnik, V. E. & Sokol, S. Y. (1998). Avis detedation in Xenopus involves biochemical interactions of axin, glycogen synthase kinase 3 and beta-catenin. Curr Bi01 4591-4.

Jiang, J., Kosman, D., Ip, Y.T., Levine, M. (1991). The dorsal morphogen gradient regulates the mesoderm determinant twist in early DrosophiIa embryos. Genes Dev. 5, 1881-1891.

Jiang, J. & Struhl, G. (1998). Reguiation of the Hedgehog and Wingless signalling pathways by the F-boxlWD40-repeat protein Slimb. Nature 391,493-6.

Johnson, G. (1984). Lithium. Med JAust 141,595-601. Johnson, R. S., van Lingen, B., Papaioannou, V. E, & Spiegelman, B. M. (1993). A nul1 mutation at the c-jun locus causes embryonic IethaIity and retarded ce11 growth in cuIture. Genes Dev 7,1309-17.

Jones, R. G., Parsons, M., Bonnard, M., Chan, V. S., Yeh, W. C., Woodgett, J. R. & Ohashi, P. S. (2000). Protein kinase B regulates T lymphocyte survival, nuclear factor kappaB activation, and Bcl-X(L) levels in vivo. J Exp ~Med191, 1721-34.

Jonkers, J., Korswagen, H. C., Acton, D., Breuer, M. & Bem, A. (1997). Activation of a novel proto-oncogene, Fratl, contributes to progression of mouse T-ceil lymphomas. Embo J 16,441-50.

Jordan, M. A. & Wilson, L. (1998). Microtubules and actin filaments: dynamic targets for cancer chemotherapy, Curr Opin Cell Biol10, 123-30.

Kallunki, T., Deng, T., Hibi, M. & Karin, M. (1996). c-Jun can recruit JNK to phosphorylate dimerizrition partners via specific docking interactions. CelZ 87,929-39-

Kane, L, P., Shapiro, V. S., Stokoe, D. & Weiss, A, (1999). Induction of NF-KB by the AktPKB kinase. Curr Biol, 9,601-604.

Kao, K. R. & Elinson, R. P. (1989). Dorsalization of mesoderm induction by lithium. Dev Biol 132,8 1-90.

Kao, K. R. & Lockwood, A. (1996). Negative regulation of dorsal patternhg in early embryos by overexpression of XrelA, a Xenopus homologue of NF-kappa B. Mech Dev 58, 129-39.

Karin, M. and Delhase, M. (1998) JNK or IKK, AP-1 or NF-kappaB, which are the targets for MEK hase 1 action? Proc Nat1 Acad Sci USA, 95,9067-9069.

Kelliher, M. A., Grimm, S.. Ishida, Y., Kuo, F., Stanger, B. S. & Leder, P. (1998). The death dornain kinase RP mediates the TNF-induced NF-kappa signal. Immunity 8,297- 303.

Keyse, S. M. (2000). Protein phosphatases and the regulation of mitogen-activated protein kinase signalling. Curr Opin Ce11 Bi01 12, 186-92.

Kharbanda, S., Ren, R., Pandey, P., Shafman, T.D., Feller, S.M., Weichselbaum, R.R. and Kufe, D.W. (1995) Activation ofthe c-Ab1 tyrosine kinase in the stress response to DNA- darnaging agents. Narure, 376,785-788.

Kiefer, F., Tibbies, LA., Lassam, N., Zanke, B., Iscove, N. and Woodgett, I.R. (1997) Novel cornponents of mammalian stress-activated protein kinase cascades. Biochem Soc Trans, 25,49 1-498. Klein, P. S. & Melton, D. A. (1996). A molecular mechanism for the effect of lithium on deveiopment. Proc Natl Acad Sci USA 93,8455-9.

Knight, R.J. and Buxton, D.B. (1996) Stimulation of c-Jun kinase and mitogen-activated protein kinase by ischemia and reperfusion in the perfused rat heart. Biochem Biophys Res Commun,218,83-88,

Kuan, C. Y., Yang, D. D., Samanta Roy, D. R., Davis, R. J., Rakic, P. & Flavell, R. A. (1999). The Jnkl and Jnk2 protein kinases are required for regiond specific apoptosis during early brain development. Neuron 22,667-76.

Kumar, S., Kinoshita, M., Noda, M., Copeiand, N.G. and Jenkins, N.A. (1994) Induction of apoptosis by the mouse Nedd2 gene, which encodes a protein similar to the product of the Caenorhabditis elegans ce11 death gene ced-3 and the mammalian IL-1 beta- converthg enzyme. Genes Dev, 8,16 13-1626.

Kyriakis, J. M. (1999). Signaiing by the germinal center kinase family of protein kinases. 3 Bioi Chem 274,5259-62.

Kyriakis, J. M., Baneqee, P., Nikolakaki, E., Dai, T., Rubie, E. A., Ahmad, M. F., Avmch, J. & Woodgett, J. R. (1994). The stress-activated protein kinase subfamily of c- Jun kinases. Nature 369, 156-60.

LaCasse, E.C., Baird, S., Korneluk, R.G. and MacKenzie, A.E. (1998) The inhibitors of apoptosis (IAPs) and their emerging role in cancer. Oncogene, 17,3247-3259.

Lali, F. V., Hunt, A. E., Turner, S. J. & Foxwell, B. M. (2000). The pyridinyl imidazole inhibitor SB703580 blocks phosphoinositide-dependent protein kinase activity, protein kinase B phosphorylation, and retinoblastoma hyperphosphorylation in interleukin-2- stimulated T cells independently of p38 rnitogen-activated protein kinase. J Biol Chem 27597395402.

Latinis, K.M. and Koretzky, G.A. (1996) Fas ligation induces apoptosis and Jun kinase activation independently of CD45 and Lck in human T ceiis. Biood, 87,871-875.

tau, K. F., C. C., Anderton, B. H. & Shaw, P. C. (1999a). Expression analysis of glycogen synthase kinase-3 in human tissues. J Pepr Res 54,85-9 1.

Lau, K. F., Miller, C. C., Anderton, B. H. & Shaw, P. C. (1999b). Molecular ctoning and characterization of the human glycogen synthase kinase3beta promoter. Genomics 60, 121-8.

Lee, F. S., Hagler, J., Chen, 2. J. & Maniatis, T. (1997). Activation of the IkappaB aipha kinase complex by MEKK1, a kinase of the JNK pathway. Ceil88,213-22. Li' C. C., Dai, R. M., Chen, E. & Longo, D. L. (1994). Phosphorylation of NF-KB1-pSO is involved in NF-kappa B activation and stable DNA binding. J Bi01 Chem 269,30089- 92.

Li, L., Yuan, H., Weaver, C. D., Mao, J., Fm,G. H., 3rd, Sussman, D. J., Jonkers, J., Kimelman, D. & Wu, D, (1999a). Axin and Fratl interact with dvl and GSK, bridging Dvl to GSK in Wnt-mediated regdation of LEF-1. Embo J 18,4233-40.

Li, M., Wang, X., Meintzer, M. K., Laessig, T., Birnbaum, M. J. & Heidenreich, K. A. (2000). Cyclic AMP promotes neuronal survival by phosphorylation of glycogen synîhase kinase 3beta. Mol Ce11 Bi01 20,9356-63.

Li, Q., Van Antwerp, D., Mercurio, F., Lee, K-F. & Venna, 1. M. (1999b). Severe liver degeneration in mice lacking the Ikappd kinase 2 gene. Science 284,321-5.

Li, Z. W., Chu, W., Hu, Y., Dehase, M., Deerinck, T., Ellisman, M., Johnson, R., Karin, M. (1999). The IKK$ subunit of kB kinase (IKK) is essential for nuclear factor KB activation and prevention of apoptosis. J EY~Med, 189,1839-1845.

Lin, A., Minden, A., Martinetto, H., Claret, F. X., Lange-Carter, C., Mercurio, F., Johnson, G. L. & Karin, M. (1995). identification of a dual specificity kinase that activates the Jun kinases and p38-Mpk2. Science 268,286-90.

Liu, H., Nishitoh, H., Ichijo, H. & Kyriakis, J. M. (2000). Activation of apoptosis signal- regulating kinase 1 (ASKl) by turnor necrosis factor receptor-associated factor 2 requires pnor dissociation of the ASKl inhibitor thioredoxin. hfol Cell Biol20,2198-208.

Liu, H., Su, Y.C., Becker, E., Treisman, J., Skolnik, E.Y. (1999). A Drosophila TNF- receptor-associated factor (TRAF) binds the se20 kinase Misshapen and activates Jun kinase. Curr Biol. 9, 10 1-104.

Liu, Z. G., Hsu, H., Goeddel, D. V. & Karin, M. (199Q Dissection of TNF receptor 1 effector functions: JNK activation is not linked to apoptosis while NF-KB activation prevents ce11 death. CeIl, 87,565-576.

Lovestone, S. & Reynolds, C. H. (1997). The phosphorylation of tau: a critical stage in neurodevelopment and neurodegenerative processes. Neuroscience 78,309-24.

MacKichan, M. L., Logeat, F. & Israel, A. (1996). Phosphoryiation of pl05 PEST sequence via a redox-insensitive pathway up-regdates processing of p5O NF-kappaB. J Biol Chem 271,608491.

Madge, L. A. & Pober, J. S. (2030). A phosphatidylinositol 3-kinaseIAkt pathway, activated by nimor necrosis factor or interleukin-1, inhibits apoptosis but does not activate NFkappaB in human endothehi celis. J Biol Chem 275,15458-65. Madhani, H. D. & Fink, G. R. (1998). The riddle of MAP kinase signaling specificity. Trent& Genet 14,15 1-5.

Madrid, t. V., Wang, C. Y., Guttridge, D. C., Schottelius, A. J., Baldwin, A. S., Jr. & Mayo, M. W. (2000). Akt suppresses apoptosis by stimulating the transactivation potential of the ReWp65 subunit of NF-kappaB. Mol Ce11 Bi01 20,1626-38. hIaeda, Y. (1970). Influence of ionic conditions on ce11 differentiation and rnorphogenesis of the ceilular sbemolds. Dev Growth D@er 12,217-27.

Malinin, N. L., Boldin, M. P., Kovalenko, A, V. & Wallach, D. (1997). MAP3K-related kinase involved in NF-kappaB induction by TNF,CD95 and iL-1. Nature 385,5404

Maniatis, T. (1999). A ubiquitin ligase complex essential for the NF-kappaB, Wnt/Wingless, and Hedgehog signaling pathways. Genes Dev 13,505-10.

Mao, K. R., Masui, Y. & Elinson, R. P. (1986). Lithium induced respecification of pattern in Xenopzis laevis embryos. Nature 322,371-3.

Martin, A. G. & Fresno, M. (2000). Turnor necrosis factor-alpha activation of NF-kappa B requires the phosphorylation of Ser-471 in the transactivation domain of c-Rel. J Bi01 Chem 275,24383-91.

Martin, J.H., Mohit, A.A. and Miller, C.A. (1996) Developmental expression in the mouse nervous system of the p493F12 SAP kinase. Brain Res Mol Brain Res, 35,4757.

McCartney, B. M., Dierick, H. A., Kirkpatrick, C., MoIine, M. M., Baas, A., Peifer, M. & Bejsovec, A. (1999). Drosophila APC2 is a cytoskeletally-associated protein that regdates wingless signaling in the embryonic epidermis. JCell Biol 146, 1303-18.

McMahon, A. P. & Moon, R. T. (1989). Ectopic expression of the proto-oncogene int-l in Xenopus embryos leads to duplication of the embryonic axis. Ce11 58, 1075-84.

Meijer, L., Thunnissen, A. M., White, A. W., Garnier, M., Nikolic, M., Tsai, L. H., Walter, J., Cleverley, K. E., Salinas, P. C., Wu, Y. Z., Biernat, J., Mandelkow, E. M., Kim, S. H. & Pettit, G. R. (2000). inhibition of cyclin-dependent kinases, GSK-3beta and CK1 by hymeniddisine, a marine sponge constituent. Chem BioI7,S 1-63.

Moon, R. T., Brown, I. D. & Torres, M. (1997). WNTs modulate ce11 fate and behavior during vertebcate development. Trends Genet 13, 157-62.

Moriguchi, T., Kawasaki, H., Matsuda, S., Gotoh, Y. & Nishida, E. (1995). Evidence for multiple activators for stress-activated protein kinase/c-Jun amino-terminal kinases. Existence of novel activators. J Bi01 Chem 270, 12969-72. Moriguchi, T., Toyoshima, F., Masuyama, N., Hanafusa, H., Gotoh, Y. & Nishida, E. (1997). A novel SAPWJNK kinase, MKK7, stimulated by TNFalpha and cellular stresses. Embo J 16,7045-53.

MosiaIos, G., Birkenbach, M., Yalamanchili, R., VanArsdale, T., Ware, C., Kieff, E- (1995). The Epstein-Barr Wus transforming protein LMPl engages signaling proteins for the tmor necrosis factor receptor family. Cell80,389-399.

Mode, S. K., Edgell, N. J., Welsh, G. I., Diggle, T. A., Foulstone, E. J.: Heesom, K. J., Proud, C. G. & Denton, R. M. (1995). Multiple signalling pathways involved in the stimulation of fatty acid and glycogen synthesis by insulin in rat epididymal fat cells. Biochern J 311,595-60 1.

Mudgett, J. S., Ding, J., Guh-Siesel, L., Chartrain, N. A., Yang, L., Gopal, S. & Shen, M. M. (2000). Essential role for p38alpha mitogen-activated protein kinase in placental angiogenesis. Proc Natl Acad Sci USA 97, 10454-9.

Munemitsu, S., Albert, I., Souza, B., Rubideid, B. & Polakis, P. (1995). Regulation of intracellular beta-catenin levels by the adenomatous polyposis coli (APC) turnor- suppressor protein. Proc Natl Acad Sci USA 92,3046-50.

Nagata. K.. Puls. A.. Futter, C.. Aspenstrorn. P., Schaefe. E., Nakata, T., Hirokawa, N. & Hall, A. (1 998). The MAP kinase kinase kinase MLK2 CO-localizeswith activated JNK dong microtubules and associates with kinesin superfarnily motor KE3. Embo J 17, 149- 58.

Nagata, S. (1997) Apoptosis by death factor. Cell, 88,355-365.

Naismith, J.H. and Sprang, S.R. (1998) Modularity in the TNF-receptor family. Trends Biochem Sci, 23,74-79.

Nakano, H., Sakon, S., Koseki, H., Takemori, T., Tada, K., Matsumoto, M., Munechika, E., Sakai, T., Shirasawa, T., Akiba, W., Kobata, T., Santee, S. M., Ware, C. F., Rennert, P. D., Taniguchi, M., Yagita, H. & Okumuca, K. (1999). Targeted disruption of TrafS gene causes defects in CD40- and CD27-mediated lymphocyte activation. Proc Nd Acad Sci USA 96,9803-8.

Nasevicius, A., Hyatt, T., Eh, W., Guttman, J., Walsh, E., Sumanas, S., Wang, Y., Ekker, S. C. (1998). Evidence for a frizzled-mediated wnt pathway required for zebrafïsh dorsal rnesoderm formation. Developmenr, 125,4283-4292.

Natoli, G., Costanzo, A., Moretti, F., Fulco, M., Balsano, C. and Levrero, M. (1997) Tumor necrosis factor (TNF) receptor 1 signaling downstream of TNF receptor- associated factor 2. Nuclear factor kappaB (NFkappaB)-inducing kinase requirement for activation of activating protein 1 and NFkappaB but not of cJun N-terminal kinase/stress-activated protein kinase. J Bi01 Chem, 272,26079-26082. Nebreda, A. R. & Porras, A. (2000). p38 MMkinases: beyond the stress response. Trends Biochem Sci 25,257-60.

Nikolakaki, E., Coffer, P. J., Hemelsoet, R., Woodgett, J. R. & Defize, L. H. (1993). Glycogen synthase kinase 3 phosphoryIates Jun family members in vitro and negatively regulates their transactivating potential in intact cells. Oncogene 8,833-40.

Ninomiya-Tsuji, J., Kishimoto, K., Hiyama, A., houe, J., Cao, 2. & Matsumoto, K. (1999). The kinase TAKl can activate the NIK-1 kappaB as well as the MAP kinase cascade in the IL-1 signalling pathway. iVatiire 398,252-6.

Nishina, H., Bachmann, M., Oliveirados-Santos, A. J., Kozieradzki, I., Fischer, K. D., Odermatt, B., Wakeham, A., Shahinian, A., Takimoto, W., Bernstein, A., Mak, T. W., Woodgett, J. R., Ohashi, P. S. & Penninger, J. M. (1997a). Impaired CD28-mediated interleukin 2 production and proliferation in stress kinase SAPKIERKl kinase (SEK1)Imitogen-activated protein kinase kinase 4 (MKK4)-deficient T lymphocytes. J Exp Med 186,941-53.

Nishina, H., Fischer, K. D., Radvanyi, L., Shahinian, A., Hakem, R., Rubie, E. A., Bernstein, A., Mak, T. W., Woodgett, J. R. & Penninger, J. M. (1997b). Stress-signalling kinase Sekl protects thymocytes from apoptosis mediated by CD95 and CD3. Nature 385,3 50-3.

Nishina, H., Vaz, C., Billia, P., Nghiem, M., Sasaki. T., De la Pompa, J. L., Furlonger, K., Paige, C., Hui, C., Fischer, K. D., Kishimoto, H., Iwatsubo, T., Katada, T., Woodgett, J. R. & Penninger, J. M. (1999). Defective liver formation and Iiver ce11 apoptosis in mice lacking the stress signaling kinase SEKl/MKK4. Development 126,505-16.

Nishitoh, H., Saitoh, M., Mochida, Y., Takeda, K., Nakano, H., Rothe, M., Miyazono, K. & ichijo, H. (1998). ASKl is essential for JNK/SAPK activation by TRAF2. Mol Ce11 2, 389-95.

Noguchi, K., Kitanaka, C., Yamana, H., Kokubu, A., Mochizuki, T. & Kuchino, Y. (1999). Regdation of c-Myc through ptiosphorylation at Ser-62 and Ser-71 by c-Jun N- terminal kinase. J Biol Chem 274,32580-7.

Nusse, R. (1999). WNT targets. Repression and activation. Trends Genet 15, 1-3.

Nusse, R. & Varmus, H. E. (1982). Many tumors induced by the mouse mammary tumor Wus contain a provins integrated in the same region of the host genome. Ce11 31, 99- 109.

Oiiver, F. J., Menissier-de Murcia, J., Nacci, C., Decker, P., Andriantsitohaina, R., Mulier, S., de la Rubia, G., Stoclet, J. C., de hcia, G. (1999). Resistance to endotoxic shock as a consequence of defective NF-KB activation in poIy (ADP-ribose) polymerase- 1 deficient mice. EMBO J, 18,4446-4454.

Oshima, M., Oshima, H., Kitagawa, K., Kobayashi, M., Itakura, C. & Taketo, M. (1995). Loss of Apc heterozygosity and abnormal tissue building in nascent intestinal polyps in rnice carrying a tnincated Apc gene. Proc Natl Acad Sci USA 92,4482-6.

Ozes, O. N., Mayo, L. D., Gustin, J. A., Pfeffer, S. R., Pfeffer, L. M., Donner, D. B. (1999). NF-KB activation by tumour necrosis factor requires the Akt serine- threonine kinase Nature, 40 1,82-85.

Park, 1. K., Roach, P., Bondor, J., Fox, S. P. & DePaoli-Roach, A. A. (1994). Molecular mechanism of the synergistic phosphorylation of phosphatase inhibitor-2. Cloning, expression, and site-directed mutagenesis of inhibitor-2. J Biol Chem 269,944-54.

Park, D.S.,Stefanis, L., Yan, C.Y.I., Farinelli, S.E. and Greene, L.A. (1996) Ordering the ce11 death pathway. Differential effects of BCL2, an interIeukn-1-converting enzyme family protease inhibitor, and other survival agents on JNK activation in serumlnerve growth factor-deprived PC12 cells. J Biol Chem, 271,21898-21905.

Park, M., Venkatesh, T. V. & Bodmer, R. (1998). Dual role for the zeste-white3lshaggy- encoded kinase in mesoderm and hem development of Drosophila. Dev Genet 22,201- 11.

Park, Y.C., Burkitt* V., Villa, A.R., Tong, L., Wu, K. (1999). Structural basis for self- association and receptor recognition of human TRAF'L. Nature 398,533-538.

Peifer, M., Pai, L. M. & Casey, M. (1994a). Phosphorylation of the Drosophila adherens junction protein Armadillo: roIes for wingless signai and zeste-white 3 kinase. Dev Bi01 166,543-56.

Peifer, M., Sweeton, D., Casey, M. & Wieschaus, E. (1994b). wingless signal and Zeste- white 3 kinase trigger opposing changes in the intraceltular distribution of Armadillo. Development 120,369-80.

Perkins, N. D. (2000). The ReM-kappaü family: triend and foe. Trends Biochem Sci 25,434-40.

Peters, J. M. (1998). SCF and APC: the Yin and Yang of celi cycle regulated proteolysis. Curr Opin Ce11 Bi01 10,759-68.

Pfeffer, K., Matsuyama, T., Kundig, T, M., Wakeham, A., Kishihara, K., Shahhian, A., Wiegmann, K., Ohashi, P. S., Kronke, M. & Mak, T. W. (1993). Mice deficient for the 55 kd tumor necrosis factor receptor are resistant to endotoxic shock, yet succumb to L. monocytogenes infection. Cell73,457-67. Pierce, S. B. & Kimelman, D. (1996). Overexpression of Xgsk-3 disrupts anterior ectodermal patterning in Xenopus. Dev Biol175,256-64.

Plyte, S. E., Feoktistova, A., Burke, J. D., Woodgett, J. R. & Gould, K. L, (1996). Schizosaccharomyces pombe skpl+ encodes a protein kinase related to mammalian glycogen synthase kinase 3 and complements a cdcl4 cytokinesis mutant. Mol Ce11 Biol 16, 179-91.

Plyte, S. E., Hughes, K., Nikolakaki, E., Pulverer, B. J. & Woodgett, J. R. (1992). Glycogen synthase kinase-3: functions in oncogenesis and development. Biochim Biophys Acra 1114, 147-62.

Polakis, P. (1997). The adenornatous polyposis coli (APC) tumor suppressor. Biochim Biophys Acra 1332, F127-47.

Polakis, P. (1999). The oncogenic activation of beta-catenin. Curr Opin Genet Dev 9, 15- 31.

Polakis, P. (2000). Wnt signaling and cancer, Genes Dev 14, 1837-51.

Pombo, C.M., Kehrl, J.H., Sanchez, I., Katz, P., Avruch, J., Zon, L.I., Woodgett, J.R.? Force. T. and Kyriakis, J.M. (1995) Activation of the SAPK pathway by the human STEZO homologue germinai centre kinase. Natlire, 377,750-754.

Pulverer, B. J., Fisher, C., Vousden, K., Littlewood, T., Evan, G. & Woodgett, J. R. (1994). Site-specific modulation of c-Myc cotransformation by residues phosphorylated in vivo. Oncogene 9,59-70.

Pulverer, B.J., Kyriakis, J.M., Avruch, J., Nikolakaki, E. and Woodgett, J.R. (1991) Phosphorylation of c-jun mediated by MAP kinases. Nature, 353,670-674.

Puziss, J. W., Hardy, T. A., Johnson, R. B., Roach, P. I. & Hieter, P. (1994). MDSI, a dosage suppressor of an rnckl mutant, encodes a putative yeast homolog of glycogen synthase kinase 3. iMol Cell Biol, 14,83 1-839.

Regnier, C.H., Song, H.Y., Gao, X., Goeddel, D.V., Cao, Z., Rothe, M. (1997). Identification and characterization of an IkappaB kinase. Ce11 90,373-383.

Regnier, C.H., Tomasetto, C., Moog-Lutz, C., Chenard, M.P., Wendling, C., Basset, P. and Rio, M.C. (1995) Presence of a new conserved domain in CART1, a novel member of the tumor necrosis factor receptor-associated protein famiiy, which is expressed in breast carcinoma J Biol Chem, 270,25715-25721.

Riesgo-Escovar, J. R., Jenni, M., Fritz, A. & Hafen, E. (1996). The Drosophila Jun-N- terminai kinase is required for ce11 morphogenesis but not for DJun-dependent ce11 fate specification in the eye. Genes Dev 10,2759-68. Rincon, M., Whitmarsh, A., Yang, D. D., Weiss, L., Derijard, B., Jayaraj, P., Davis, R. J. & FlaveIl, R. A. (1998). The JNK pathway regulates the In vivo deletion of immature CD4(+)CD8(+) thymocytes. J Erp Med 188, 18 17-30.

Romashkova, J. A. & Makarov, S. S. (1999). NF-KB is a target of AKT in anti-apoptotic PDGF signalling. Natzire, 401, 86-90.

Rosenfeld, M. E., Prichard, L., Shiojiri, N, & Fausto, N. (2000). Prevention of hepatic apoptosis and embryonic lethality in RelA/TNFR-1 double knockout mice. Am J Parhol 156,997-1007.

Roshak, A., Jackson, 1. R., Chabot-Fletcher, M. & Marshall, L. A. (1 997). Inhibition of NFkappaB-mediated interleukin-lbeta-stimulated prostaglandin E2 formation by the marine natural product hymenialdisine. J Pharmacol & Ther 283,955-61.

Ross, S. E., Erickson, R. L,, Hemati, N. & MacDougaid, O. A. (1999). Glycogen synthase kinase 3 is an insulin-regulated CEBPalpha kinase. Mol Cell Bi01 19,8433-41.

Rossant, J. & McMahon, A. (1999). "CreW-atingmouse mutants-a meeting review on conditionai mouse genetics. Genes Dev 13,1425.

Rothe, M., Pan, M.G., Henzel, W.J.,Ayres, T.M. and Goeddel, D.V. (1995) The TNFR2- TRAF signahg complex contains two novel proteins related to baculoviral inhibitor of apoptosis proteins. Cell, 83, 1243-1252.

Rothe, M., Wong, S. C., Henzel, W. J. & Goeddel, D. V. (1994). A novel family of putative signal tramducers associated with the cytoplasmic domain of the 75 kDa tumor necrosis factor receptor. Cell78,681-92.

Rubinfeld, B., Albert, I., Porfiri, E., Fiol, C., Munemitsu, S. & Polakis, P. (1996). Binding of GSK3beta to the APC-beta-catenin complex and regulation of complex assembly. Science 272,1023-6.

Rudolph, D., Yeh, W. C., Wakeham, A., Rudolph, B., Nallainathan, D., Potter, J., Elia, A. J. & Mak, T. W. (2000). Severe liver degeneration and lack of NF-kappaB activation in NEMO/iKKgamma-deficient mice. Genes Dev 14,854-62.

Ruel, L., Bourouis, M., HeitzIer, P., Pantesco, V. & Simpson, P. (1993a). Drosophila shaggy kinase and rat glycogen synthase kinase-3 have conserved activities and act downstream of Notch. Nature 362,557-60.

Ruel, L., Pantesco, V., Lutz, Y., Simpson, P. & Bourouis, M, (1993b). Fuuctionai significance of a family of protein kinases encoded at the shaggy locus in DrosophiIa. Embo J 12,1657-69. Ruel, L., Starnbolic, V., Mi, A., Manoukian, A. S. & Woodgett, J. R (1999). Regulation of the protein kinase activity of Shaggy(Zeste-white3) by components of the wingless pathway in Drosophila cells and embryos. J Bi01 Chem 274,S 1790-6.

Sabapathy, K., Hu, Y., Kaiiunki, T., Schreiber, M., David, J. P., Jochum, W., Wagner, E. F. & Karin, M. (1999a). JNK2 is required for efficient T-ce11 activation and apoptosis but not for normal lymphocyte development. Curr Bi01 9, 116-25.

Sabapathy, K., Jochurn, W., Hochedlinger, K., Chang, L., Karin, M. & Wagner, E. F. (1999b). Defective neural tube rnorphogenesis and altered apoptosis in the absence of both JNKl and JNK2. Mech Dev 89, 115-24.

Saitoh, M., Nishitoh, H., Fujii, M., Takeda, K., Tobiume, K., Sawada, Y., Kawabata, M., Miyazono, K. & khijo, H. (1998). Marnmalian thioredoxin is a direct inhibitor of apoptosis signal-regulating kinase (ASK) 1. Embo J 17,2596-606.

Sakanaka, C., Sun, T. Q. & Wiltiams, L. T. (2000). New steps in the Wntheta-catenin signal transduction pathway. Recenr Prog Horm Res 55,225-36.

Saksela, K., Makela, T. P., Hughes, K., Woodgett, J. R. & Alitalo, K. (1992). Activation of protein kinase C increases phosphorylation of the L-myc tram-activator domain at a GSK-3 target site. Oncogene 7,347-53.

Sakurai, H., Chiba, H., Miyoshi, H., Sugita, T. & Tonumi, W. (1999). IkappaB kinases phosphorylate NF-kappaB p65 subunit on serine 536 in the tramactivation domain. JBiol Chem 274,30353-6.

Salic, A., Lee, E., Mayer, L. & Kirschnec, M. W. (2000). Conaol of beta-catenin stability: reconstitution of the cytoplasmic steps of the wnt pathway in Xenopus egg exûacts. hl01 Cell 5, 523-32.

Sanchez, L, Hughes, R. T., Mayer, 0. J., Yee, K., Woodgett, I. R., Amch, I., Kyriakis, J. M. & Zon, L. 1. (1994). Role of SAPWERK kinase-1 in the stress-activated pathway regulating transcription factor c-Jun. Nartrre 372,794-8.

Schaeffer, H. J., Catling, A+D., Eblen, S. T., Collier, L. S., Krauss, A. & Weber, M. J. (1998). MP1: a MEK binding partner that enhances enqmatic activation of the MAP kinase cascade. Science 281,1668-71,

Schaeffer, H. J. & Weber, M. J. (1999). Mitogen-activated protein kinases: specitTc messages fiom ubiquitous messengers. Mol Ce11 Biol 19,24354.

Scheid, M. P. & Woodgett, J. R. (2000). Protein kinases: six degrees of separation? Curr Biol 10, R191-4. Schlesinger, A., Shelton, C. A., Maloof, J. N., Meneghim, M. & Bowerman, B. (1999). Wnt pathway components orient a mitotic spindle in the eady Caenorhabditis elegans embryo without requiring gene transcription in the responding cell. Genes Dev 13,2028- 38.

Schorpp-Kistner, M., Wang, Z. Q., Angel, P. & Wagner, E. F. (1999). JunB is essential for mammalian placentation. Embo J 18,934-48.

Schou, M. (1979). Lithium in the treatment of other psychiatric and nonpsychiatric disorders. Arch Gen Psychiatry 36,856-9.

Searle, J., Lawson, T. A., Abbott, P. J., Hamon, B. & Kerr, J. F. (1975). An electron- microscope study of the mode of ce11 death induced by cancer-chemotherapeutic agents in populations of proliferating normal and neoplastic cells. JPathol116, 129-38.

Shaw, M., Cohen, P. & Alessi, D. R. (1997). Further evidence that the inhibition of glycogen synthase kinase3beta by tGF-1 is mediated by PDKlIPKB-induced phosphorylation of Ser-9 and not by dephosphorylation of Tyr-216. FEBS Lett 416,307- II.

Shaw, P. C., Davies, A. F., Lau, K. F., Garcia-Barcelo, M., Waye, M. M., Lovestone, S., Miller, C. C. & Anderton, B. H. (1998). Isolation and chrornosomal mapping of hurnan glycogen synthase kinase-3 alpha and -3 beta encoding genes. Genome 41,720-7.

Shi, C. S. & Kehri, J. H. (1997). Activation of stress-activated protein kinaselc-Jun N- terminal kinase, but not NF-kappaB, by the tumor necrosis factor (TNF) receptor 1 through a TNF receptor-associated factor 2- and gennuial center kinase reiated-dependent pathway. JBiol Chem 272,32102-7.

Siegfried, E., Chou, T. B. & Perrimon, N. (1992). wingless signaling acts through zeste- white 3, the Drosophila homolog of glycogen synthase kinase-3, to reguiate engtailed and establish ce11 fate. Ce11 71, 1167-79.

Siegfried, E., Perkins, L. A., Capaci, T. M. & Pecrimon, N. (1990). Putative protein kinase product of the Drosophila segment-polarity gene zeste-white3. Nature 345,825-9.

Siegfïied, E., Wilder, E. L. & Perrimon, N. (1994). Components of wingless signahg in Drosophila. Nature 367,7640.

Sizemore, N., Leung, S. & Stark, G. R. (1999). Activation of phosphatidylinositol 3- kinase in response to interleukin-1 Ieads to phosphorylation and activation of the NF-KB p651RelA subunit. Mol Ce11 Bi02 19,47984805.

Sluss, H.K.,Barrett, T., Derijard, B. and Davis, RJ. (1994) Signal transduction by turnor necrosis factor mediated by JNK protein kinases. Molecular & Cellular Biology, 14, 8376-8384. SIuss, H. K., Han, Z., Barrett, T., Davis, R. J. & Ip, Y. T. (1996). A JNK signal transduction pathway that mediates morphogenesis and an immune response in Drosophila. Genes Dev 10,2745-58.

Smith, C.A., Farrah, T. and Goodwin, R.G. (1994) The TNF receptor superfamily of cellular and viral proteins: activation, costimulation, and death. Cell, 76,959-962.

Sokol, S. Y. (1999). Wnt signaling and dorso-ventral axis specification in vertebrates. Curr Opin Genet Dev 9,405-10.

Song, H. Y., Regnier, C. H., Kirschning, C. J., Goeddel, D. V. & Rothe, M. (1997). Turnor necrosis factor (TNF)-mediated kinase cascades: bifiucation of nuclear factor- kappaB and c-jun N-terminal kinase (JNWSAPK) pathways at TNF receptor-associated factor 2. Proc Nat1 Acad Sei iJ S A 94,9792-6.

Spencer, E., Jiang, J. & Chen, Z. J. (1999). Signal-induced ubiquitination of 1kappaBalp ha by the F-box protein Slimbheta-TrCP. Genes Dev 13,284-94.

SpiegelmanoV. S., Slaga, T. J., Pagano, M., Minamoto, T,, Ronai, Z. & Fuchs, S. Y. (2000). Wntfbeta-catenin signaling induces the expression and activity of betaTrCP ubiquitin ligase receptor. Mol Ce11 5,877-82.

Spittaels, K., Van Den Haute, C., Van Dorpe, J., Geerts, H., Mercken, M., Bruynseels, K., Lasrado, R., Vandezande, K., Laenen, I.? Boon, T., Van Lint, J., Vandenheede, J., Moechars, D., Loos, R. & Van Leuven, F. (2000). Glycogen synthase kinase-3beta phosphorylates protein tau and rescues the axonopathy in the central nervous system of hurnan four-repeat tau transgenic mice. J Bi01 Chem 275,4134041349-

Stambolic, V., Ruel, L. & Woodgett, J. R. (1996). Lithium inhibits glycogen synthase kinase-3 activity and mimics wingless signding in intact cells. Curr Biol6, 1664-8.

Stambolic, V. & Woodgett, J. R. (1994). Mitogen inactivation of glycogen synthase kinase-3 beta in intact cells via serine 9 phosphoryiation. Biochem J303,701-4.

Stanger, B.Z.,Leder, P., Lee, T.H., Kim, E., Seed, B. (1995). NP: a novel protein containhg a death domain that interacts with FadAPO-1 (CD99 in yeast and causes ce11 death. CeIl 81,513-523.

Stokoe, D., Campbell, D. G,,Nakielny, S., Hidaka, H., Leevers, S. J., Marshail, C. & Cohen, P. (1992). MAPKAP kinase-2; a novel protein kinase activated by mitogen- activated protein kinase. Embo J 11,3985-94.

Su, B., Jacinto, E., Hibi, M., Kdunki, T., Karin, M. and Ben-Neriah, Y. (1994) JNK is involved in signal integration during costirnulation of T lymphocytes. Cell, 77,727-736. Sutherland, C., Leighton, 1. A. & Cohen, P. (1993). Inactivation of glycogen synthase kinase-3 beta by phosphorylation: new kinase comections in insulin and growth-factor signalling. Biochem J 296, 15-9.

Swat, W., Fujikawa, K., Ganiatsas, S., Yang, D., Xavier, R. J., Harris, N. L., Davidson, L., Ferrini, R., Davis, R. J., Labow, M. A., Flaveil, R. A., Zon, L. 1. & Alt, F. W. (1998). SEKl/MKK4 is required for maintenance of a normal peripheral lymphoid cornpartment but not for lymphocyte development. Immunity 8,625-34.

Takeuchi, M., Rothe, M. and Goeddei, D.V. (1996) Anatomy of TRAF2. Distinct domains for nuclear factor-kappaB activation and association with tumor necrosis factor signaling proteins. J Biol Chem, 271,1993519942,

Tamura, K., Sudo, T., Senftleben, U., Dadak, A. M., Johnson, R. & Karin, M. (2000). Requirement for p38alpha in erythropoietin expression: a role for stress kinases in erythropoiesis. Ce11 102,22 1-3 1.

Tartaglia, L.A. and Goeddel, D.V. (1992) Two TNF receptors. Immunoi Toduy, 13, 15 1- 153.

Tetsu, O. & McConnick, F. (1999). Beta-catenin regulates expression of cyclin Dl in coIon carcinoma cells. Nature 398,422-6.

Thomas, G. M., Frame, S., Goedert, M., Nathke, I., Polakis, P. & Cohen, P. (1999). A GSK3-binding peptide from FRATl selectively inhibits the GSK3-catalysed phosphorylation of axin and beta-catenin. FEBS Lett 458,247-51.

Thompson, C. B. (1995). Apoptosis in the pathogenesis and treatment of disease. Science 267, 1456-62.

Thorpe, C. J., Schlesinger, A. & Bowerman, B. (2000). Wnt signalling in Caenorhabditis eIegans: regulating repressors and poIarizing the cytoskeleton. Trends Ce11 Biol10, 10-7.

Tibbtes, L. A. & Woodgett, J. R. (1999). The stress-activated protein kinase pathways. Ce11 Mol Life Sci 55, 1230-54.

Tong, L., Pav, S., White, D. M., Rogers, S., Crane, K. M., Cywin, C, L., Brown, M. L, & Pargellis, C. A. (1997). A highly specific inhibitor of human p38 MAP kinase binds in the ATP pocket. iVat Smct Bi01 4,3 11-6.

Tournier, C., Hess, P., Yang, D. D., Xu, J., Turner, T. K., Nimnual, A., Bar-Sagi, D., Jones, S. N., Flaveil, R. A. & Davis, R. J. (2000). Requirement of JNK for stress-induced activation of the cytochrome c-mediated death pathway. Science 288,8704 Tournier, C., Whitmarsh, A, J., Cavanagh, J., Barrett, T. & Davis, R. J. (1997). Mitogen- activated protein kinase kinase 7 is an &vator of the cJun NHZterminai kinase. boc Nat1 Acad Sci USA 94,7337-42.

Tournier, C., Whitmarsh, A, J., Cavanagh, J., Barrett, T. & Davis, R. J. (1999). The MKK7 gene encodes a group of c-Jun NHZterminal kinase kinases. Mol Cell Biol 19, 1569-8 1.

Umezawa, K., Ariga, A. & Matsumoto, N. (2000). Naturally occurring and synthetic inhibitors of NF-KB functions. Anti-Cancer Dmg Design 15,239-244.

Van Antwerp, D. J., Martin, S. J., Kafri, T., Green, D. R. & Verma, 1. M. (1996). Suppression of TNF-cl-induced apoptosis by M.'-KB. Science 274,787-789.

Verdu, J., Buratovich, M. A., Wilder, E. L. & Birnbaum, M. J. (1999). Cell-autonomous regulation of ce11 and organ growth in Drosophila by Akt/PKB. Nat Cell Biol 1,500-6. von Kries, J. P., Winbeck, G., Asbrand, C., Schwarz-Romond, T., Sochnikova, N., Dell'Oro, A., Behrens, J. & Birchmeier, W. (2000). Hot spots in beta-catenin for interactions with LEF-1,conductin and APC. Nat Struct Bioi 7,800-7.

Wang, C. Y., Mayo, M. W. & Baldwin, A. S., Jr. (1996). TNF-and cancer therapy- induced apoptosis: potentiation by inhibition of NF-KB. Science 274,784-787.

Wang, D. & Baldwin, A. S., Jr. (1998). Activation of nuclear factor-kappaB-dependent transcription by turnor necrosis factor-alpha is mediated through phosphory Iation of RelNp65 on se~e529. J Biol Chem 273,2941 1-6.

Wang, D., Westerheide, S. D., Aanson, J. L. & Baldwin, A. S., Jr. (2000). Tumor necrosis factor alpha -induced phosphorylation of RelAfp65 on sec529 is controlled by casein kinase II. J Bi01 Chem 275,32592-7.

Wang, Q. M., Fiol, C. J., DePaoli-Roach, A. A. & Roach, P. J. (1994). Glycogen synthase kinase-3 beta is a dual specificity kinase differentialty regulated by tyrosine and serine/threonine phosphorylation. J Bi01 Chem 269, 14566-74.

Wang, X.S., Diener, K., Jannuzzi, D., Troiiinger, D., Tan, T.H., Lichenstein, H., Zukowski, M. and Yao, Z. (1996) MoiecuIar cloning and characterization of a novel protein kinase with a catalytic domain homologous to mitogen-activated protein kinase kinase kinase. J Biol Chem, 271,3 1607-3 16 11.

Wang, X. S., Diener, K., Manthey, C. L., Wang, S., Rosennveig, B., Bray, J., Delaney, J., Cole, C. N., Chan-Hui, P. Y., Mantlo, N., Lichstein, H. S., Zukowski, M. & Yao, Z. (1997)- MoIecular cloning and characterization of a novel p38 mitogen-activated protein kinase. J Biol Chem 272,23668-74. Wang, X. 2. & Ron, D. (1996). Stress-induced phosphorylation and activation of the transcription factor CHOP (GADD153) by p38 MAP Kinase. Science 272, 1347-9.

Wassarman, D.A., Solomon, N.M., Rubin, G.M. (1996). Pk92B: a Drosophila melanogaster protein kinase that belongs to the MEKK family. Gene 169,283484,

Wehrli, M., Dougan, S. T., Caldwell, K., O'Keefe, L., Schwartz, S., Vaizel-Ohayon, D., Schejter, E., Tomlinson, A. & DiBardo, S. (2000). anow encodes an LDL-receptor- related protein essential for Wingless signalling. Nature 407,327-30.

Welsh, G. 1. & Proud, C. G. (1993). Glycogen synthase kinase-3 is rapidIy inactivated in response to insulin and phosphorylates eukaryotic initiation factor eIF-2B. Biochem J 294,625-9.

Welsh, G.I., Wilson, C. & Proud, C.G. (1996). GSK-3: a SHAGGY Erog story. Trends Cell Biol. 6,274-279.

White, R, A., Hughes, R, T., Adkison, L. R., Bruns, G. & Zon, L. 1. (1996). The gene encoding protein kinase SEKl maps to mouse chromosome 11 and human chromosome 17. Genomics 34,430-2.

Whitmarsh, A.J.. Cavanagh. J.. Tournier, C., Yasuda, 1. and Davis, R.J. (1998) A mammalian scaffold complex that selectively mediates MAP kinase activation. Science, 281, 1671-1674.

Whitmarsh, A. J., Shore, P., Sharrocks, A. D. & Davis, R. J. (1995). Integration of MAP kinase signal transduction pathways at the semresponse element. Science 269,403-7.

Widmann, C., Gerwins, P., Johnson, N.L., Jarpe, M.B. and Johnson, G.L. (1998) MEK kinase 1, a substrate for DEVD-directed caspases, is involved in genotoxin-induced apoptosis. Mol Ce11 Biol, 18,2416-3429.

Widmam, C., Gibson, S., Jarpe, M. B. & Johnson, G. L. (1999). Mitogen-activated protein kinase: conservation of a three-kinase module fiom yeast to human. Physiol Rev 79, 143-80.

Wijsman J. H., Jonker R. R., Keijzer R., van de Velde C. J., Cornelisse C. J., van Dierendonck I. H* (1993). A new rnethod to detect apoptosis in paraffin sections: in situ end-labeling of hgmented DNA. J Histochem Cytochem 41,7-f 2.

Willert, K, Brink, M., Wodarz, A., Varmus, H-& Nusse, R. (1997). associates with and phosphorylates dishevelled. Embo Jl6,3089-96.

Williams, D. D., Marin, O., Pinna, L. A. & Proud, C. G. (1999). Phosphorylated seryl and threonyl, but not tyrosyl, residues are efficient specificity determinants for GSK-3beta and Shaggy. FEBS Lert 448,86-90. Wilson, K. P., McCaffiey, P. G., Hsiao, K., Pazhanisamy, S., GaIuilo, V., Bernis, G. W., Fitzgibbon, M. J., Caron, P. R., Murcko, M. A. & Su, M. S. (1997). The stmctural bais for the specificity ofpyridinylimidazoIe inhibitors of p38 MAP kinase. Chem Bi01 4,423- 31.

Winston, J. T., Strack, P., Beer-Rornero, P., Chu, C. Y., Elledge, S. J. & Harper, J. W. (1999). The SCFbeta-TRCP-ubiquitin ligase cornplex associates specifically with phosphorylated destruction motifs in IkappaEMpha and beta-catenin and stimulates IkappaBalpha ubiquitination in vitro. Genes Dev 13,270-83.

Wodarz, A. & Nusse, R. (1998). Mechanisrns of Wnt signaling in developrnent. Annu Rev CeIl Dev Bi01 14,59438,

Woodgett, J. R. (1990). Molecular clonhg and expression of glycogen synthase kinase- jlfactor A. Embo J 9,243 1-8.

Woodgett, J. R. (1991). cDNA cloning and properties of glycogen synthase kinase-3. ~tfethodsE~ymol200,564-77.

Woodgett, J. R, (1994). Regulation and functions of the glycogen synthase kinase-3 subfarnily. Semin Cancer Bi01 5,269-75.

Woodgett, J. R., Plyte, S. E., Puiverer, B. J., Mitchell, J. A. & Hughes, K. (1993). Roles of glycogen synthase kinase-3 in signal transduction. Biochem Soc Tram 21,905-7.

Woronicz, J.D., Gao, X., Cao, Z., Rohe, M., Goeddel, D.V. (1997). IkappaB kinase-beta: NF-kappaB activation and complex formation with IkappaB kinase-alpha and NIK. Science 278,866-869.

Wu, Y. & Cai, D. (1992). Study of the effect of lithium on lymphokine-activated killer ce11 activity and its antitumor growth. Proc Soc Erp Bi01 Med 201,284-8.

Wysk, M., Yang, D. D., Lu, H. T., FIaveii, R. A. & Davis, R. J. (1999). Requirement of mitogen-activated protein kinase kinase 3 (MKK3) for tumor necrosis factor-induced cytokine expression. Proc Narl Acad Sci USA 96,3763-8.

Xavier, 1. J., Mercier, P. A., McLoughlin, C. M., Ali, A., Woodgett, J. R. & Ovsenek, N. (2000). Glycogen synthase kinase 3beta negatively regulates both DNA-binding and transcriptional activities of heat shock factor 1. JBiol Chem 275,29147-52.

Xia, Y., Makris, C., Su, B., Li, E., Yang, J., Nemerow, G. R. & Karin, M. (2000). MEK kinase 1 is critically required for c-Jun N-terminal hase activation by proinfiammatoty stimuli and growth factor-induced cell migration. Proc Nat1 Acad Sci USA 97,5243-8. Xu, L., Corcoran, R. B., Welsh, J. W., Pennica, D. & Levine, A. J. (2000). WISP-1 is a Wnt-1- and beta-catenin-responsive oncogene. Genes Dev 14,585-95.

Yamaguchi, H., Ishiguro, K., Uchida, T., Takashima, A., Lecnere, C. A. & Imahori, K. (1996). Preferential tabeiing of Alzheimer neurofibrillary tangles with antisera for tau protein kinase (TPK) Vglycogen synthase kinase-3 beta and cych-dependent kinase 5, a component of TPK II. Acta Neuropathol92,232-41.

Yamaoka, S., Courtois, G., Bessia, C., Whiteside, S.T., Weil, R., Agou, F., Kkk, HE, Kay, R.J., Israel, A. (1998). CompIementation cloning of NEMO, a component of the IkappaB kinase complex essential for NF-kappaB activation. Ce11 93, 1231-1240.

Yan, M., Dai, T., Deak. J-C., Kyriakis, J.M., Zon, L.I., Woodgett, J.R. and Templeton, D.J. (1994) Activation of stress-activated protein kinase by MEKKl phosphorylation of its activator SEKI . Nature, 372,798-800.

Yang, D., Tournier, C., Wysk, M., Lu, H. T., Xu, J., Davis, R. J. & Flavell, R. A. (1997a). Targeted disruption of the MKK4 gene causes embryonic death, inhibition of c- Jun NH2-terminal kinase activation, and de fects in AP- 1 transcriptional activity . Froc Nat1 Acad Sci Cl S A 94,3004-9.

Yang, D. D., Conze, D., Whitmarsh, A. J., Barrett, T., Davis, R. J., Rincon, M. & Flavell, R. A. (1998). Differentiation of CD4+ T cells to Th1 cells requires MAP kinase JNK2. Immunity 9,575-85.

Yang, D. D., Kuan, C. Y., Whitmarsh, A. J., Rincon, M., Zheng, T. S., Davis, R. J., Rakic, P. & Flriveil, R. A. (1997b). Absence of excitotoxicity-induced apoptosis in the hippocampus of rnice lacking the Jnk3 gene. Nature 389,865-70.

Yang, J., Boerm, M., McCarty, M., Bucana, C., Fidler, 1. J., Zhuang, Y. & Su, B. (2000). Mekk3 is essentid for early embryonic cardiovascular development. Nat Genet 24,309- 13.

Yao, Z., Diener, K., Wang, X. S., Zukowski, M., Matsumoto, G., Zhou, G., Mo, R., Sasaki, T., Nishina, H., Hui, C. C., Tan, T. H., Woodgett, J. P. & Penninger, J. M. (1997). Activation of stress-activated protein kinaseslc-Jun N-terminal protein kinases (SAPKsIJNKs) by a novel mitogen-activated protein kinase kinase, J Biol Chem 272, 32378-83.

Yaron, A., Hatzubai, A., Davis, M., Lavon, I., Amit, S., Manning, A. M., Andersen, J. S., Mann, M., Mercurio, F. & Ben-Neriah, Y. (1998). Identification of the receptor component of the IkappaBaIpha-ubiquitin ligase, Nature 396,5904

Yasuda, J., Whitmarsh, A. J., Cavanagh, J., Sharma, M. & Davis, R. J. (1999). The JIP group of mitogen-activated protein kinase scaEo1d proteins. Moï Cell Bioll9,7245-54. Yeh, W.C., Shahinian, A., Speiser, D., Krauus, J., Billia, F., Wakeham, A., de la Pompa, J.L., Femck, D., Hum, B., Iscove, N., Ohashi, P., Rothe, M., Goeddei, D.V. and Mak, T. W. (1997) Early lethality, functional NF-kappaB activation, and inçreased sensitivity to RIF-induced celi death in TRAF2deficient mice. Immirniv, 7,715-725.

Yost, C., Farr, G. H., 3rd, Pierce, S. B., Ferkey, D. M., Chen, M. M. & Kirneiman, D. (1998). GBP, an inhibitor of GSK-3, is implicated in Xenopus development and oncogenesis. Cell93, 103 1-41.

Yost, C., Torres, M., Miller, J. R., Huang, E., Kimelman, D. & Moon, R. T. (1996). The ais-inducing activity, stability, and subcellular distribution of beta-catenin is regulated in Xenopus embryos by glycogen synthase kinase 3. Genes Dev 10,1443-54-

Yuasa, T., Ohno, S., Kehl, J. H, & Kyriakis, J. M. (1998). Tumor necrosis factor signaling to stress-activated protein kinase (SAPK)/Jun NH2-terminal kinase (JNK) and p38. Germinal center kinase couples TRAF2 to mitogen-activated protein kinase/ERK kinase kinase 1 and SAEK white receptor interacting protein associates with a mitogen- activated protein kinase kinase kinase upstream of MKK6 and p38. J Bi01 Chem 273, 22681-92.

Yujiri, T., Ware, M., Widmann, C., Oyer, R,, Russell, D., Chan, E., Zaitsu, Y., Clarke, P., Tyler, K., Oka, Y., Fanger, G. R., Henson, P. & Johnson, G, L. (2000). MEK kinase 1 gene disruption alters ce11 migration and c-Jun NH2-terminal kinase regdation but does not cause a measurable defect in NF-kappa B activation. Proc Nail Acad Sci USA 97, 7272-7.

Zanke, B. W., Rubie, E. A., Winnett, E., Chan, J., Randall, S., Parsons, M., Boudreau, K., McInnis, M., Yan, M., Templeton, D, J- & Woodgett, J. R. (1996). Mammaiian mitogen- activated protein kinase pathways are regulated through formation of specüic kinase- activator complexes. J Biol Chem 271,29876-8 1.

Zhong, H., Voll, R E. & Ghosh, S. (1998). Phosphorylation of NF-kappa B p65 by PKA stimulates transcriptional activity by prornoting a navel bivalent interaction with the coactivator CBP/p300. Mol CeIll, 66 1-71. Appendix 1

Activation of endogenous stress-activated protein kinase (SAPK)

isoforms by osmotic stress and anisomycin Al.1 Abstract

The cellular response to treatment with proinflammatory cytokines or exposure to various environmental stresses is mediated, in part, by the stress-activated protein kinases (SAPKs), also commonly known as c-Jun NH2-terminai kinases (JNKs). In marnmalian cells, the stress- activated protein kinase group consists of 10 members that are derived by alternative spticing of three genes. Because of the extensive sequence homology between these distinct isoforms, biochemicd identification of endogenous proteins using antibodies is very difficult, if at ail possibIe. Here we report the partial purification of the 46 kDa and 54 kDa SAPK isofonns by

Iiquid chromatography. We demonstrate that osrnolar stress and the protein synthesis inhibitor, anisomycin. differentiaily activate these SAPK isoforms in U937 cells, Notably, the rnajority of the anisomycin- and sorbitol-dependent SAPK activity was found in p46, and not the p54,

SAPK-containhg fractions. This preferentiai coupling of the hyperosrnolarity and anisomycin signal transduction pathways to certain SAPK isoforms has important implications for the specificity of signal transduction cascades.

A12 Introduction

The response of ceiis to the extracellular environment is mediated by a variety of signa1 transduction pathways, including the stress-activated protein kinase (SAPK)/c-Jun NH2-terminai kinase (JNK) cascade (reviewed in (Kiefer et al., 1997; Widmann et al., 1999). SAPK is preferentialiy activated by cellular stresses, such as UV irradiation, thed shock, ischemia reperhsion, alkylating agents, hyperosmolarity, and treatment with translational inhibitors, such as anisomycin (Kharbanda et ol., 1995; Knight and Buxton, 1996; Kyriakis er al., 1994). Very strong SAPK activation is aIso observed after co-stirnulatory activation of T cells, treatment with sphingomyelinase, or exposure to certain inflammatory cytokines (tumour necrosis factor-ol, interleukin- 1 and Fas-ligand; Kyriakis er d.,1994; Latinis and Koretzky, 1996; Su et al., 1994).

For the most part, SAPK is poorfy induced by mitogenic stimuli and growth factors, however in certain cell types the kinases are activated by phorbor esters, epidermal growth factor and CD40

(Berberich er al., 1996; Kyriakis et al., 1994).

Mammalian cells encode at least three pardogous SAPK genes (SAPK a, $ and y; or

JNK2, JNK3 and JNKL , respectivety) which are 88-98% related to each other in the catalytic domain (Derijard ef a(.. 1994; Gupta et al., 1996; Kyriakis et ul., 1994). Through differential splicing, these kinases yield as many as I0 distinct isoforms. In particular, each gene is expressed as two morecdar mass species due to the alternate insertion of five nucleotides (CACAG) within codon 377 just prior to the C-terminal region. This results in the generation of 54 or 46 kDa proteins with identical catalytic domains. Though it is possibte that these various SAPK isofonns have completeIy redundant fictions in the cell, findings to date indicate that there are differences in their regulation and in vivo properties. The first key evidence was the observation that SAPKy/JNKl, but not SAPKdJNK2, functionally complements a hyperosmotically sensitive Sacchmomyces cerevisiae strain defective in HOG1 expression (Sluss et a[., 19943. In addition, recent studies utilizing gene targeting to generate mice deficient in the particular SAPK paratogues have shown that individual isoforms can play an important roIe in cell proIiferation, apoptosis and differentiation. For instance, in ~apkyl~nkl"mice, CD4 helper T (TH) ceUs hyperproliferated, exhibited decreased activation-induced death, and preferentially differentiated to TH2 cells @ong et al., 1998). in the absence of SapkcdJnk2, murine T celis only revealed an impainnent in TH1 dxerentiation (Yang et al., 1998). Furthemore, dismption of the gene encoding SapkP/Jnkj, which is selectively expressed in the nervous system, caused the mice to be resistant to the excitotoxic glutamate-receptor agonist kainic acid, thereby showing a reduction in seinire activity and hippocampai neuron apotosis (Yang et al., 1997).

The existence of multiple SAPK isoforms provides the potential for the generation of stimuIus-specific and ce11 type-specific responses to activation of the SAPK signalhg pathway.

The identification of which SAPK isofoms are activated by which signals, and their mechanism of activation, represents one step that is required for understanding the physiological role of this kinase farnily in mammalian cells. Here we investigated whether various well-established SAPK agonists differentially stimulate the 46 kDa and 54 kDa SAPK splice forms. in this appendix, 1 describe the partial purification of p46 and p54 SAPK by anion-exchange Liquid chromatography, and their respective activation by hyperosmolar stress and the protein synthesis inhiiitor, anisomycin. A13 Materials and Methods

A1.3.1 cDNA constructs

HemaggluGnin (KA) epitope-tagged expression plasmids for p46-S APKy and p54-SAPKf3 have been previously described (Kyriakis et al., 1994). GST-c-Jun(5-89) hsion proteins (Ader et al.,

1992; Kyriakis er al., 1994) were produced in the pLysS(BL2 lDE3) sîrain of Escherichia coli using the pGEX expression system (Promega).

Al.3.2 Cell culture and treatment

Human rnonocytic leukemia U937 cells were maintained in Dulbecco's modified Eagie's media supplemented with 10% fetal calf sem. For rneasurement of SAPK activity, cells were stimulated with 400 rnM D-Sorbitol (in water; SIGMA)or 10 ughi anisomycin (in 50% ethanol;

SIGMA).

A13.3 Partial purification of SAPK isoforms

Approlcimately 4 x 10' U937 ceiis were lysed in 25 mM Tris-HCl pH=7.5, 1 mM EDTA, 0.2 mM EGTA, lmM dithiothreitol, 100pM sodium orthovanadate, 5pgiml leupeptin, 50 pM sodium fluoride, and 1 mM benzamidine. After clarification at 10,000 x g for 30 min, the supernatant was loaded ont0 an anion exchange column (Mono Q, Pharmacia Biotech) and equiiiirated with 25 mM Tris-HCI pH=7.6, containing 1 mM EDTA, 0.2 mM EGTA, 1mM dithiothreitol and 50 mM sodium chloride, Protein was eluted in equilibration buffier at a flow rate of 1 dmin for 40 ml with a 0.05-0.25 M hear gradient of sodium chloride. One-millilitre fractions were collected. To anaiyze hctions for SAPK isoforms, fractions were resolved by

SDSPAGE, transferred to PVDF membrane (NEN Life Science Products), and immunoblotted

~vithpolyclonal antibodies raised to either p54 SAPKP (Kyriakis et ai, 1994) or phosphorylated threonine 183 and tyrosine 185 of human SAPK (New England Biolabs).

A1.3.1 Assay of SAPK activity

SAPK activity was measured by a solid phase assay using glutathione S-transferase (0-c-lun

(5-89) coupled to glutathione-agarose beads (Adler et al., 1992; Kyriakis et ai., 1994). Equal arnounts of each fraction (20 pl) were mked with 20 pl of a 1:I slurry of GSTc-Jun-agarose

(approximately 10 pg). Kinase assays were performed at 30°C in 40 pM ATP, 10 mM MgCu,

50 mM Tris-HCl pH=7.5, 1 mM EGTA, and 40 @ihL [y-j2~]~TP.Mer 30 minutes, the beads were washed twice in cold PBST (150 mM NaCl, 16 mM Na2HP04, 4 mM NaH2P04,

0.1 % Triton X- 100, IrnM dithiothreitol, I O0 pM sodium orthovanadate, 5pg/ml leupeptin, 50 pM sodium fluoride, and 1 mM benzamidine), and the reactions were terminated by addition of

Laemmli sample buffer. Samples were boiled, and phosphorylated proteins were resolved on

SDS, 12.5% polyacrylamide gels and visualized by Phosphorhager analysis (ushg Molecdar

Dynamics ImageQuant software). A13.5 Immunoprecipitations and immunublotting

Cell lysis was performed in Gentle Soft buffer (10 mM NaCl, 20 mM Pipes, 0.5% NP-40, 5 mM

EDTA, I mM dithiothreitol) with inhibitors (t00pM sodium orthovanadate, 5 jtgiml leupeptin,

50 FM sodium fluoride, and 1 mM benzamidine) at 4OC. Lysates were cleared by centrifugation at 13,000xg for lOmin, and then norrnalized for total protein by Bradford assay.

Irnmunoprecipitations were performed using monocional antibody against the HA tag generated kom culture supernatants of 12CA5 hybridoma (ATCC). *Epitope-tagged proteins were immunoprecipitated by incubating the appropriate ceii Lysates with 20 pl ami-HA for one hour and harvested with 10 pl of protein G-Sepharose (Pharmacia). The immune complexes were washed four times in PBST, resolved by 10% SDS-PAGE and immunoblotted with anti-SAPK antibody.

A1.J Results

A1.4.1 Endogenous teveis of p46 and p54 SAPK isoforms

We sought first to determine whether the p46 and p54 SAPK isoforms were present in equal amounts in U937 tek Whole ceii lysates (Figure 22, lanes 3 and 4), together with HA-tagged recombinant p54 SMKg and p46 SAPKy (lanes 1 and 2, respectively) were separated by

SDSPAGE, transferred to PVDF membrane, and immunoblotted with SAPK antisera cross- reacting with al1 known forms of the kinase (Kyriakis et al., 1994), Elizabeth Rubie and J-kW., unpublished observations). As show in Figure 22, the p54 SAPK isoform was more highly expressed than p46 in U937 cek. The positions of p54 (upper band) and p46 SAPK (Iower band) are indicated by arrowheads. Using scanning densitometry ta compare the p46 and p54 signals in the whole celi lysate, we estimated that of the total U937 cellular SAPK, approximately 65% is 54 kDa and 35% is 46 kDa.

A1.4.2 Examining the kinetics and amplitude of SAPK activity in U937 cells

To optimize the conditions for SAPK activation in U937 cells, we analysed the activity of these proteins following sorbitol or anisomycin addition (Figure 22). Exposure to 400 mM D-sorbitol induced maximal activity after 30 min, as measured by GST-c-Jun(5-89) phosphorylation, By comparison, we observed maximal SAPK activation after 45 min of 10 Wml anisomycin treatment. induction of SAPK activity by sorbitol was somewhat weaker when compared with the response elicited by anisomycin. Consistently, cJun phosphorylation was detectable 5 min after stimulation, and SAPK activity was reduced after 60 min of exposure to either of the agonists. a~nisorn~dn Sorbitol

0 5 10 15 20 30 45 60 Time (min) Figure 22. Detection of endogenous SAPK protein and activity in U937 ceiis (A) U937 cells were harvested, and equal amounts of proteins were separated by SDS-PAGE and subjected to imrnunoblotting with anti-SAPK antibody (lanes 3 and 4). As a control, 293 ceils were transiently transfected with 5 pg of pCDNA3-HA-p54-SAPKfl (lane 1) or pCDNA3-HA- p46-SAPKy (lane 2), and epitope-tagged SAPKs were immunoprecipitated fiom these cell extracts using anti-HA antibody. The two major bands that were recognized by anti-SAPK antibody is indicated by arrows. (B) Activation of SAPK induced by anisomycin or sorbitol in U937 cells. U937 cells were stimulated with the 400mM D-sorbitol or 10 ughi anisomycin at 37°C for the indicated times. The cells were lysed and assayed for SAPK activity. The results are show as the GST-c-Jun phosphorylation fold-increase (over nonnalized control). A1.1.3 Partial pufication of p46 and p54 SAPK isoforms by Liquid chromatography identification of the SAPK isoforms using antibodies is complicated by their extensive sequence homology. An alternative approach would be to separate the two molecular weight SAPK species by liquid chromatography. Hence, U937 ce11 Iysates were chromatographed on a Mono Q anion-exchange column using a 0.05-0.25 M linear gradient of sodium chloride. Collected fiactions were subjected to SDS-PAGE and immunoblot analysis using anti-SAPK antibody (Figure 23).

In both control and sorbitol-stimulated lysates (using the optimized conditions fiom Figure 22), p46 SAPK was eiuted fmt at a salt concentration of approximately 90 mM NaCl (top and rniddle panels). By cornparison, p54 SAPK was eluted around 165 mM NaCl and appeared to be hlly separated Erom p46 SAPK. kowheads indicated the positions of p46 and p54 SAPK.

Since SAPK is activated in response to extracellular stimuli by dual phosphorylation on threonine and tyrosine, the phosphorylation state of endogenous SAPK could be determined by immunoblotting with an antibody that specifically recognizes the threonine 183- and tyrosine

185-phosphorylated Forms of SAPK, but not SAPK that is unphosphorylated at this site.

Hence, the sarne sorbitol-treated Fractions were probed with an antibody against dual- phosphorylated SAPKs. SAPK was shown to be prorninently phosphorylated in the appropriate fractions (lower panel). As a control experiment, Thr 183iTyr 18 5-phosp hory Iated

SAPK was hardly detectabIe in unstimdated ceiis (data not shown). Thus, according to the data show in Figure 23, the Mono Q separation p46 and p54 SAPKs was sufficient to examine the activation of these isofom. Unstimulated --___------c p54 (SAPK blot) - - - p46 .------Sorbitol - -- - p54 (SAPK blot) --- - - cp46 Sorbitol - + p54 (phospho -m.--. SAPK blot) - p46

L 1 1 1 1 1 1 1 5 10 15 20 25 30 35 40 Fraction

Fraction

Fraction Figure 23. Separation of p46 SAPK and pS4 SAPK by anion exchange liquid chromatography (A) Lysates from unstimuiated and stimuiated U937 ceUs were separately chromatographed on Mono Q to resolve the daerent SAPK isoforms. (9 Western blot analysis of the Mono Q fractions using anti-SAPK antibody to detect unstimuiated p46/p54 SAPK or (ii) stimulated p46fp54 SAPK. (iii) Anti-phospho-SAPK (threoninel83ftyrosine185) Westerns were done to defme the SAPK isofonns more clearly. Similar results were obtained in at least five different experiments. (B) Sorbitol-stimulated (30 min, 400 mM fW) or control U937 ceU lysates were Fractionated by ion-exchange chromatography (MonoQ column) at a gradient of 50 mM to 250 mM NaCl (flow rate of I ml/min). Aliquots of each fiaction (20 pl) were subjected to kinase assay with 10 pg of bacterially expressed GST-c-Jun(5-89) and quantitated by PhosphorIrnager analysis rneasured in arbitrary units. Values shown are the fold-increase in activity over unstimulated, control fractions. The results are representative of the separate experhents. (Cl Activation of p46 SAPK and p54 SAPK by anisomycin. Extracts were prepared fiom 4 x 10 U937 cells incubated in the presence or absence of 10 ug/d anisomycin at 37°C for 35 min, and subjected to anion-exchange chromatography. Each fiaction was analyzed for anisomycin- dependent c-Jun-phosphorylating activity by an in vitro reconstitution assay. Quantitation was done by Phosphorlmager analysis. Similar results were obtained in three different experiments. -41.4.4 Differential activation of endogenous SAPKs in response to sorbitol or anisomycin

Although anti-SAPK (Thr183-P/Tyr18S-P) antibody is a specific indicator for activity of the

SAPK isoforms, to better quantitate the extent of sorbitol-activated p46 versus p54 SAPK, proteins isolated at various fractions der exposure of U937 cells to sorbitol (400 mM, 30 min) were subjected to an immunecornplex kinase assay by using glutathione S-transferase-c-Juri(5-89) as a substrate (Figure 23). Phosphorylated proteins were visualized and quantitated by

Phosphorimager analysis. In the absence of sorbitol, only very weak SAPK activity was observed in the Mono Q fiactions (data not shown). The addition of sorbitol to the ceiis induced the appearance of three peaks. One of peaks observed in the Mono Q fractions was located at the same position as p46 (fractions 12-15), and another peak at p54 (fractions 27-28). A peak of low level of SAPK activity (fraction 21) was associated with the shoulder of the largest peak of activity and may comprise another sorbitol-inducible kinase capable of c-Jun phosphorylation which is unrelated to the SAPK faniily. Notably, the addition of sorbitol produced a 4-fold higher peak in the p46 elution fractions, as compared to the p54 peak. These results are representative of three separate experiments.

Similady, to identi& any anisomycin-dependent SAPK isoform activation in U937 cells, we fiactionated lysates fiom U937 ceiis optimally stimuiated with anisomycin (10 pg/mi, 45 min) and assayed each fraction for in vitro GST-c-Jun(5-89) phosphorylation. Comparable resdts for SAPK activation were observed when anisomycin was added instead of sorbitol. The ciifference in the relative height of each peak in response to these agonists reflects the different specific activity of sorbitol and anisomycin on the different SAPK isoforms present in each peak.

FolIowing nomahmion of the kinase activity dues to the relative amount of cellular p46 and p54, the activity of the p54 fractions was approximately 2.5 fold than the activity of hctions containing p46. Hence, p46 SAPK displays substantially higher kinase activity than p54 SAPK in response to sorbitol- and anisomycin-stimulation of U937 cells.

A1.5 Discussion

The stress-activated protein kinase group consists of 10 members that are derived by alternative splicing of three genes. These SAPK isoforms differ in their tissue distribution and in their interaction with substrate proteins. In particular, SAPKa and SAPKy are widely expressed in several tissues, whereas SAPKB is more selectively expressed in brain, testis, and heart (Gupta et al., 1996; Kyriakis er al., 1994; Martin et al., 1996). The expression of multiple SAPK isoforms in mammalian tissues suggests that these kinases rnay differ in their physiological function, and may be coupled to different upstrearn signaling pathways. However, the specific contribution of these isoforms in SAEK signaling remains unknown.

Since specifk SAPK isoforms rnay be activated in response to different stimuli, we have compared the regdation of endogenous p46 and p54 SAPK in response to two environmental stress signais. U937 cells were exposed to sorbitol (osmotic shock) or anisomycin (an inhibitor of protein synthesis), ceU lysates were hctionated, and the SAPK activity was detected by measurement of protein kinase activity in an immune complex kinase assay using GST-c-Jun(5-

89) as the substrate. The evidence presented here suggests that the 46 kDa and 54 kDa SAPK isoforms do not reqond equally to stimulation by these two diierent agonists k U937 cells. in both cases, p46 was consistently activated to severai fold higher than p54 (Figure 23). These resuit. are consistent with other reports indicating that SAPKs do not have identical biochemical properties. For instance, in mouse macrophages s-iuiated with TM;,46 Da SAPK is activated whereas the 54 Daisoform is not (Chan et al., 1997).

SAPK family members are directly activated by two MAPK kinases, SEKl/MKK4

(Sanchez et al.. 1994) and MKK7 (Hoiland et al., 1997; Moriguchi er al., 1997; Tournier et al.,

1997; Yao et al., 1997). The conûibution of these two activators in the regdation of SAPK isoforms in vivo remains tu be determined, Interestingly though, in sek14 ernbryonic stem cebs, sorbitol-induced osmolarity changes activate SAPKs to wild-type sekl'" levels, while the eEect of anisornycin was abolished in sek14' cells (Nishiia et al., 1997). Specific stimuli can therefore cause the differential activation of MAPK kinases. It follows that MAPK kinases could lead to the activation of different groups of SAPK isoforms. Altematively, p46 and pS4 SAPK isoforms could aiso be differentiaily sensitive to phosphatases. in our lab, we are currently exminhg the activation of specific SAPK isoforms in embryonic fibroblasts denved fiom sek14-, sekl*'- and sekl"* mice to address these possibilities.

The differential activation of SAPKs provides a mechanism for the genetation of stimulus-specific responses of cells to their environment. It is currently not clear whether the selective activation of p46 over p54 SAPK by sorbitol and anisomycin is a cornmon event in

SAPK activation or whether it is atypical. There may also be situations where simultaneous activation of p46 and p54 SAPK act synergisticalIy to eIicit a response that is distinct fkom the response to the activation of either isoform alone. Furthemore, the prospect of recentiy identifkd scaffolding proteins pIaying a rote in reguIating the SAPK isoform family aeeds to be investigated (Whitmarsh er al., f 998). However, detailed comparative studies of the SAPK isoforms have not yet been completed and Merstudies are required to provide answers to

these underlying questions.

A1.6 References

Adler, V., Polotskaya, A., Wagner, F., and Kraft, A. S. (1992). Afiïnity-pded c-Jun amino- terminal protein kinase requires serinelthreonine phosphorylation for activity. J Biol Chem 267, 1700 1-1 7005.

Berberich, I., Shu, G., Siebelt, F., Woodgett, 1. R., Kyriakis, J. M., and Clark, E. A. (1996). Cross-linking CD40 on B cells preferentially induces stress-activated protein kinases derthm mitogen-activated protein hases. ElMBO J 15,92-10 1.

Chan, E, D., Wiston, B. W., larpe, M. B., Wynes, M. W., and Riches, D. W. (1997). Preferential activation of the p46 isofom of JWSAPK in mouse macrophages by RIF alpha. Proc Nat1 Acad Sci CI S ri 94,13 169-13 174.

Derijard. B., Hibi, M., Wu, 1. H., Barrett, T., Su, B., Deng, T., Karin, M,,and Davis, R. J. (1994). JNK1: a protein kinase stimuiated by UV light and Ha-Ras that binds and phosphorylates the cJun activation domain. Ceil 76, 1025-1037.

Dong, C., Yang, D. D., Wysk, M., Whitmarsh, A. J., Davis, R. J., and Flaveii, R A. (1998). Defective T ceii differentiation in the absence of Jnkl. Science 282,2092-2095.

Gupta, S., Barrett, T., Whitrnarsh, A, J., Cavanagh, J., Sluss, H. K., Derijard, B., and Davis, R. J. (1996). Selective interaction of JNK protein kinase isoforms with transcription factors. Embo J 15,2760-2770.

Holland, P. M., Suzanne, M., Campbell, 1. S., Noselii, S., and Cooper, J. A. (1997). MKK7 is A stress-activated mitogen-activated protein kinase kinase functionaüy reIated to hemipterous. J Bi02 Chem 272,24994-24998.

Kharbanda, S., Ren, R, Paadey, P., Shafinan, T. D., Feiler, S. M., Weichselbaum, R R,and Kufe, D. W. (1995). Activation of the c-Ab1 tyrosine kinase in the stress response to DNA- damaging agents. Nature 376,785-788. Kiefer, F., Tibbles, L. A,, Lassa.,N., Zanke, B., Iscove, N., and Woodgett, J. R. (1997). Novel components of mammalian stress-activated protein kinase cascades. Biochem Soc Tram 25,491- 498.

Knight, R J., and Buxton? D. B. (1996). Stimulation of c-Jun kinase and mitogen-activated protein kinase by ischemia and reperfusion in the perfused rat heart. Biochem Biophys Res Commun 218,83-88.

Kyriakis, J. M., Banerjee, P., Nikolakaki, E., Dai, T., Rubie, E. A., Ahmad, M. F., Avnich, J., and Woodgett, J. R (1994). The stress-activated protein kinase subfamily of c-Jun kinases. Naitrre 369, 156-160.

Latinis, K. M., and Koretzky, G. A. (1996). Fas ligation induces apoptosis and Jun kinase activation independently of CD45 and Lck in human T cells. Blood 87,871-875.

Martin, I. H., Mohit, A, A., and Miller, C. A. (1996). Developmental expression in the mouse nervous system of the p493F12 SAP kinase. Brain Res Mol Brain Res 35,47-57.

Moriguchi, T., Toyoshima, F., Masuyama, N., Hanafusa, H., Gotoh, Y., and Nishida, E. (1997). A novel SAPEUJNK kinase, MEX7, stimulated by TNFalpha and cellular stresses. Embo J 16, 7045-7043.

Nishina, H., Fischer, K. D., Radvanyi, L., Shahinian, A., Hakem, R., Rubie, E. A., Bernstein. A., Mak, T. W., Woodgett, J. R., and Penninger, J. M. (1997). Stress-signalling kinase Sekl protects thymocytes fiom apoptosis mediated by CD95 and CD3. Nature 385,350-353.

Sanchez, I., Hughes, R. T., Mayer, B. J., Yee, K., Woodgett, J. R., Avruch, J., Kyriakis, J. M., and Zon, L. 1. (1994). Role of SAPWER.kinase-1 in the stress-activated pathway regulating transcription factor c-Jun. Nature 372,794-798.

Sluss, H. K., Barrett, T., Derijard, B., and Davis, R. J. (1 994). Signal transduction by turnor necrosis factor mediated by JNK protein kinases. Molecular & Cellular Biology 14,8376-8384.

Su, B., Jacinto, E., Hibi, M., Kallunki, T., Karin, M., and Ben-Neriah, Y. (1994). JNK is involved in signal integration during costimulation of T lymphocytes. Ce11 77,727-736.

Tournier, C., Whitmarsh, A. J., Cavanagh, J., Barrett, T., and Davis, R. J. (1997). Mitogen- activated protein kinase kinase 7 is an activator of the c-Jun NH2-terminai kinase, Proc Nat1 Acad Sci USA 94,7337-7342.

Whitmarsh, A. J., Cavanagh, J., Tournier, C., Yasuda, J., and Davis, R J. (1998). A mammalian scafKold complex that selectively mediates MM kinase activation. Science 281,1671-1674. Widmann, C., Gibson, S., Jarpe, M. B., and Johnson, G. L. (1999). Mitogen-activated protein kinase: conservation of a three-kinase module fiom yeast to human, Physiol Rev 79,143-180.

Yang, D. D., Conze, D., Whitmarsh, A. J., Barrett, T., Davis, R. J., Rincon, M., and Flavell, R A. (1998). Differentiation of CD4+ T ceils to Th1 cells requires MAP kinase JNK2. Immunity 9, 575-585.

Yang, D. D., Kuan, C. Y., Whitmarsh, A. J., Rincon, M., Zheng, T. S., Davis, R. J., Rakic, P., and Flavell, R. A. (1997). Absence ofexcitotoxicity-induced apoptosis in the hippocarnpus of mice lacking the Jnk3 gene. Nature 389,865-870.

Yao, Z., Diener, K., Wang, X. S., Zukowski, M., Matsumoto, G., Zhou, G., Mo, R, Sasaki, T., Nishina, H., Hui, C. C., Tan, T. H., Woodgett, J. P., and Penninger, J. M. (1997). Activation of stress-activated protein kinasesk-Jun N-terminal protein kinases (SAPKsIJNKs) by a novel mitogen-activated protein kinase kinase. JBiol Chem 272,32378-32383. Assessrnent of postnatal survival of TNFRI~GSK-3pmice

20 1 Newbom progeny were coiiected and genotyped from 'IT~~FRI?'-GsK--~/~+'- x WFR~ +'-

GsK-~~'-mouse matings. In Figure 24, pups L-9 were stillborn, and their genotype was deterrnined by Western blotting for GSK-3a and GSK-38 protein, as well as PCR reactions for the presence of the neo and wild-type alleles of WFRI and GSK-3/3. No live WFRI'-GSK-~#T"* mice were detected, While disruption of TNFRl is capable of rescuing the hepatocyte apoptosis observed at E13.5-14.5, the cause for this later embryonic lethality warrants further investigation. anti-GSK-3 blot

+ TNFRI wt PCR -TNFRI neo

+GSK-3p neo

+GSK-30 ~t

Figure 24. Post-natal survival of TM;RIJJGSK-38/-mice