MJClin Pathol: Mol Pathol 1995;48:M292-M301

Leaders

The mammalian mitogen activated proteinkinase network

P Lenormand

The families ofMAP kinases cascades (mitogenactivated protein kinases) form major signallingsystems by which cells transduce extracellularsignals into cellular responses. Studying theregulation of these MAP kinase signalling path-ways is of utmost importance in the under-standing of the triggering of cell proliferationversus cell differentiation and the responses ofthe cell to stress or osmotic factors.

This review will describe in detail thearchetypal MAP kinase cascade (Raf-MEK-ERK); it will present the more recently dis-covered MAP kinase cascades (SAPK and p38)and provide some indications on the regulationof this network of signalling pathways. Thescope of the review is not to provide an ex-haustive listing of all aspects of research in thisfield; it aims rather to provide a basis for readersunfamiliar with the subject to scrutinise specificpoints of interest. Special attention will begiven to introduce clearly the nomenclature ofmammalian MAP kinase cascades, since it isoften confusing with up to five names beingused for the same mammalian enzyme. Readersrequiring more detailed information on theMAP kinase cascades are referred to otherrecent reviews. 1-5

Centre de Biochimie-CNRS UMR 134,Universite de Nice,Parc Valrose,06108 Nice Cedex 2,FranceP Lenormand

Correspondence to:Dr Philippe Lenormand.

Accepted for publication8 June 1995

Importance of protein phosphorylationin cell signallingPhosphorylation is a reversible means of co-

valently modifying proteins. The phosphategroup being added is highly hydroxylated andnegatively charged; thus it is sterically largeand able to alter markedly the tridimensionalconformation ofproteins. Phosphorylation maybe permanent and necessary for correct foldingof many proteins. Reversible phosphorylationconsists of the addition of a phosphate fromATP by a specific protein kinase, and sub-sequent removal of this phosphate through a

protein phosphatase. Phosphorylation of manyenzymes affects their catalytic properties. Theimportance of protein phosphorylation in reg-ulating cellular activities is illustrated by. thelarge number of protein kinase genes that are

present in eukaryotic genomes. The latest es-

timate is that humans have as many as 2000conventional protein kinase genes.3 It is be-coming increasingly obvious that protein phos-phatases play a specific role in regulatingcellular activities and some aspects of theseregulating activities will be dealt with in thefinal section of this review.

Protein kinases have broadly been classifiedinto two main categories according to the aminoacid that they phosphorylate: the tyrosine kin-ases and the serine/threonine kinases. Recentlya new class has been discovered, the dual speci-ficity protein kinases that phosphorylate bothtyrosine and serine/threonine residues. Ad-ditionally, in bacteria6 and yeasts,7 histidinekinases have been identified which functionas osmosensors. In yeast the histidine kinaseosmosensor triggers activation of one of thethree yeast MAP kinase cascades.7 Each newlycloned kinase can be classified in one of thesecategories by sequence homology, and thus itsspecificity rapidly revealed.

The MAP kinase cascade moduleMAP kinases (MAPK) are serine/threoninekinases which are activated by a dual phos-phorylation on threonine and tyrosine residues8in a sequence of three amino acids: TXY (T=threonine, Y = tyrosine).9 The central aminoacid (X) of this triad defines each family ofMAP kinase as is indicated in fig 1. Theirspecific upstream activators, called MAP kinasekinases (MAPKK),10 constitute a family of dualspecific threonine/tyrosine kinases, which inturn are activated byMAP kinase kinase kinases(MAPKKK) which are serine/threonine kin-ases.11 A schematic diagram at the left of fig 1illustrates the MAP kinase cascade module.There are several families of MAP kinase cas-cades, and members of each family are con-served among various species such asmammals, frogs, flies, worms, or yeasts (re-viewed in 1). This high degree of conservationduring evolution points towards a fundamentalrole for these enzymes in cell physiology.

Theoretical enzymological analyses indicatethat an amplification factor of 10 can readilybe achieved at each stage of these kinase cas-cades.3 A three protein kinase cascade, such asthe MAP kinase cascade, could generate atheoretical amplification of 1000-fold. Thisprovides an easy means to augment a specificsignal received at the cell surface by a relativelysmall number of receptors. The actual am-plification achieved by a MAPK pathway invivo is not yet known, but experimental evid-ence shows that binding of less than 5% ofRaf- 1 molecules to Ras is sufficient to inducemaximum activation of the archetypal MAPkinase ERK (cascade illustrated in fig 1).12 As

M292

on July 23, 2020 by guest. Protected by copyright.

http://mp.bm

j.com/

Clin M

ol Pathol: first published as 10.1136/m

p.48.6.M292 on 1 D

ecember 1995. D

ownloaded from

Mammalian MAP kinase network

Nomenclature of the MAP kinase network

ERK cascade

MAPKKKSerine/Threonine

kinase

MAPKKTyrosine + Threonine

kinase

MAPK

Serine/Threoninekinase

SAPK cascade

/ Alternativenames

MEKK

SEK

SAPK

MAPKKK

MKK4

JUNK

Sequence ofdual phosphorylation TEY TPY TGY

at the MAPK level

Figure 1 Diagram indicating the nomenclature of the mammalian MAP kinase network. On the left are represented the three levels of the MAP kinasecascade module. Each box represents a cascade and provides alternative names for each step. At the bottom is indicated the amino acid sequence of thephosphorylation sites of the MAPKs (T=threonine; Y=tyrosine; E=glutamic acid; G=glycine; P=proline).

will be described later, activation of the ERKcascade is an absolute requirement for growthfactor mediated cell division. Therefore it maybe expected that any deregulation of the sig-nalling component leading to ERK activationwould have a profound effect on cell pro-liferation because of signal amplification by thecascade. For analytical purposes, amplificationof the signal at the end of the cascade mayprove to be a useful tool for detecting aberrantcellular pathologies initiated by undetectableabnormal regulation of an upstream com-

ponent.

NomenclatureWhile searching for cytoplasmic substrates oftyrosine kinase receptors, Ray and Sturgillwere the first to purify two proteins of 42 kDaand 44 kDa phosphorylated on tyrosine andthreonine residues when 3T3-L1 adipocyteswere stimulated by insulin.813 The preferentialsubstrate of these proteins was the brain Mi-crotubule Associated Protein 2, and thus theywere named MAP kinase (mapk).8 13 At thesame time, other laboratories working withthese p42maPk and p44mapk gave the followingnames to the same enzymes: MBP kinase (My-elin Basic Protein) kinase'415 and RSK kinase(90 kDa Ribosomal S6 Kinase-kinase).'6As numerous agonists were able to stim-

ulate MAP kinase activity, Sturgill and Weberproposed that the same acronym MAP kinasecould stand for Mitogen Activated Protein ki-nase. Cobb and colleagues'718 then identifiedthe cDNAs corresponding to several isoforms

of the MAP kinase proteins and named themERK, for Extracellularly Regulated Kinases(ERK1 for the p44mapk, ERK2 for the p42maPk).

Accordingly, the same protein may be calledMAP kinase, ERKI, p44maPk, MBP kinase, andRSK kinase. Similar confusion occurred laterwith the direct upstream activator of ERKwhich was named MAP kinase kinase(MAPKK),'° or MEK for "MAP or ERK Ki-nase" in an attempt to reconcile the two ac-

ronyms MAP and ERK.'9Recently, two other families of enzymes pre-

senting some homologies to ERKs have beencloned and termed SAPK for Stress ActivatedProtein Kinases2021 and p38.22 Their activationis also achieved by dual specificity threonine/tyrosine kinases presenting homologies to theERK activator.2324 Thus a reasonable con-

sensus exists at the time of writing for calling"MAP kinase cascade" a protein kinase cascadeof three enzymes activated sequentially as fol-lows: a serine/threonine kinase (MAPKKK)which activates a dual specificity tyrosine-serine/threonine kinase (MAPKK), whichactivates a serine/threonine kinase (MAPK)(fig 1).MAPK (interchangeably MAP kinase) is

thus the generic name of the third member ofeach cascade, whereas the name ERK is re-served for the third enzyme of one of the MAPcascades, the Raf-MEK-ERK cascade. Figure2 gives a synopsis of the actual nomenclatureused for the three mammalian MAP kinasecascades and provides alternative names thatcan be encountered in published reports. Thenames may correspond to isoforms or enzymes

0

E

cu

InCu

C._

0~

cLli

p38 cascade

Alternativenames

Raf

MEK MAPKK

ERK MAPKp44maPk

MBP kinaseRSK kinase

AlternativeI names

RKKK

MKK3 RKK

p38 CSPB kinase

K RK

M293

on July 23, 2020 by guest. Protected by copyright.

http://mp.bm

j.com/

Clin M

ol Pathol: first published as 10.1136/m

p.48.6.M292 on 1 D

ecember 1995. D

ownloaded from

Lenormand

Tyrosine kinase receptor

a)

-0ECD-o

~0Cn

cDC.,.,_

a-

MAPKKK

MAPKK

MAPK

Example of nuclear substrate*

G-Protein coupled receptor Receptor

1<1~ DAG

1/

s222

s228

183

185

Elk-1

ERKcascade

Outcome of activation* "Proliferation-Differentiation"

SAPK p38cascade cascade

"Stress response" "Osmotic response"

Figure 2 Schematic representation of the identified and putative signalling pathways leading to activation of the three MAP kinase cascades. Theblack stars represent cloned members of the MAP kinase cascades. Known phosphorylation sites are indicated on the left of the stars. Arrows representdirect activation or protein-protein interactions; see text for details.* Indication of the cellular responses resulting from the activation of these cascades should be interpreted with care, as crosstalk and multiple activationoccur, and some responses may be cell specific.

from different species, and thus a precise equi-valence between the names of each isoenzymeis still lacking.

Activation of the Raf-MEK-ERKcascade, the archetypal MAP kinasecascadeACTIVATION BY TYROSINE KINASE RECEPTORS(TKR)Following ligand binding, TKRs dimerise andtransphosphorylation occurs on several tyrosineresidues of the receptor. Src homology 2 (SH2)domain containing proteins are recruited tothese phosphorylated tyrosine residues (re-viewed in 25). Binding specificity for each SH2domain is provided by the four amino acidsupstream of the tyrosine.26 Among the numer-ous SH2 proteins recruited, the adaptator pro-tein Grb2 leads to activation of the Raf-MEK-ERK pathway, at least in the case of the EGF(epidermal growth factor) and PDGF (plateletderived growth factor) pathways.25 Grb2 hasno intrinsic catalytic activity; it is composed ofone SH2 and two SH3 domains which promoteprotein-protein interactions. In this case, therole of Grb2 is solely to recruit proteins boundto its SH3 domain (Src homology 3), such asthe Ras guanine-nucleotide exchange factor,Sos, to the phosphorylated receptor bound toits SH2 domain. The complex Grb2-Sos is

brought to the membrane where Sos interactionwith Ras-GDP catalyses the dissociation ofGDP from Ras allowing GTP loading andactivation of Ras. Then by direct interactionwith Ras-GTP, Raf-1 is brought to themembrane22730 where it is likely to be activatedby another mechanism which may involvephosphorylation. At present this specific issueis not resolved and two groups have shownrecently that the targeting of Raf to the mem-brane was sufficient to induce Raf activationfully.3' 32 Activated Raf-1 phosphorylatesMEK33 on two serine residues S222 and S228.34In turn MEK phosphorylates ERK on a tyros-ine (Y183) and a threonine residue (T185) inthe sequence TEY (E = glutamic acid).An alternative pathway to transmit Ras ac-

tivation to MEK is thought to involve MEKkinase (MEKK),35 which phosphorylates MEKon the same sites as Raf. Furthermore, it ap-pears that Ras may have a third effector, P13kinase,"6 and thus activated Ras could exert amultiplicity of effects. Understanding all thecellular signalling pathways induced by Ras isa great challenge for future research; mattersare complicated by the existence of three iso-forms of Ras in each cell: H-Ras-1, K-Ras-2,and N-Ras (reviewed in 37). What is the ra-tionale for having three isoforms in one cell?How many direct partners interact with Ras?The contribution of multiple Ras functions to

Ras

RKKK

I*

T

_ y

ATF-2

SEK

TaSAPKr P

Y

c-Jun

M294

on July 23, 2020 by guest. Protected by copyright.

http://mp.bm

j.com/

Clin M

ol Pathol: first published as 10.1136/m

p.48.6.M292 on 1 D

ecember 1995. D

ownloaded from

Mammalian MAP kinase network

mammalian cell transformation has recentlybeen demonstrated.38ERK1 and ERK2 are widely expressed in

most cell types, and in virtually all cells thesetwo isoforms are coordinately regulated anddisplay the same substrate specificity. Hence itis not understood why cells need both iso-forms.39 ERKI and ERK2 have been shownto phosphorylate numerous substrates on theconsensus site PXS/TP (reviewed in 1), but invitro many kinases show an apparently broadersubstrate specificity; it is thus difficult toidentify unambiguously substrates of ERK invivo. Putative in vivo substrates ofERK includesurface receptors, cytosolic protein, and nuc-lear proteins.'One ofthe best candidates for nuclear targets

of ERK is the transcription factor Elk-i (alsonamed p62TCF).4"2 ERK phosphorylates Elk-1 on identical sites in vitro and in vivo. Elk-iplays an essential role in increasing the rate oftranscription of many immediate early genessuch as c-Fos (reviewed in 43). Elk-I interactswith SRF (serum response factor) and forms aternary complex with an SRF dimer.44 Thephosphorylation of ELK on the sites phos-phorylated by ERK has been shown to increasethe transcription in vitro of genes containingan SRE (Serum Response Element, bindingsite for SRF).4

In quiescent cells, ERK is cytosolic4647 andthus unable to be in contact with nuclear pro-teins such as Elk-i. Upon mitogenic stimu-lation, ERK translocates to the nucleus,4647 andcan thus phosphorylate several nuclear targets.It is noteworthy that MEK47 and Raf (48 andLenormand P, personal communication) areexclusively located in the cytoplasm and Rasis membrane bound following specific post-translational modifications at the C terminus(reviewed in 49); thus ERK is the first memberof this MAP kinase cascade able to translocateand transmit extracellular signals to the nuc-leus.Another means of transmitting signals from

the Ras-MEK-ERK cascade to the nucleus isperformed by ERK activation of the cytosolickinase p90`RK (p90 ribosomal S6 kinase), whichin turn can translocate to the nucleus uponstimulation.46 The mechanism ofnuclear trans-location of ERK is not understood as ERKdoes not possess nuclear localisation sequences(NLS), and we have shown in this laboratorythat the nuclear translocation of ERK occurseven when the two phosphorylation sites aremutated and the kinase is no longer ac-tivatable.47 As will be discussed later, nucleartranslocation ofERK is thought to be a criticalstep in triggering cell proliferation or differ-entiation; thus understanding the mechanismof this translocation will prove to be very im-portant.

ACTIVATION BY NON-TYROSINE-KINASERECEPTORSMany hormones and neurotransmitters actingthrough G protein coupled receptors also ac-tivate the Raf-MEK-ERK pathway by a mech-anism that involves Ras activation. It has been

demonstrated that the thrombin, muscarinicacetylcholine M2, and c2 adrenergic receptorsactivate Ras (conversion from Ras-GDP toRas-GTP).50 The link by which the G proteincoupled receptors activate Ras is not at presentwell established, and may involve proteins ofthe shc, src family, and (or) PTPiD (Syp). Itwas shown recently that coexpression of the fand y subunits of heterotrimeric G proteinswas able to activate Ras."Many G protein coupled receptors activate

protein kinase C (PLC) and ERK, but it is notclear yet how PKC activation is transmitted tothe Raf-MEK-ERK pathway, nor which iso-forms of PKC are implicated.' One report in-dicates that PKC can activate Raf-1 directly byphosphorylation52; alternatively PKC is knownreadily to phosphorylate members of the 14-3-3 protein family (reviewed in 53) which havebeen shown to be bound to Raf. There is somecontroversy as to whether or not the 14-3-3protein is indeed able to activate Raf.5"57 Thisshows that deciphering the molecular eventsleading to Ras and (or) Raf activation will bean area of intense research.

Roles of the Raf-MEK-ERK cascadeCells express many types of tyrosine kinasereceptors at their surface. Interestingly onlysome of them can induce cell proliferation(or cell differentiation in a different cell con-text). We shall see that the duration of the Raf-MEK-ERK cascade activation provides clues tounderstanding the final outcome of the cellularresponse elicited by various tyrosine kinase re-ceptors.

CELL PROLIFERATIONThe importance of the Raf-MEK-ERK cascadein conducting the mitogenic signal is deducedfrom an array of data. (1) Transfection ofoncogenic forms of Ras or Raf induce theproliferation of many cell lines.38 (2) Injectionof oncogenic Ras into quiescent fibroblastsis sufficient by itself to stimulate DNAsynthesis.5859 (3) Expression of constitutivelyactivated forms of MEK, generated by sitedirected mutagenesis, is sufficient to inducemitogenesis and transformation in fibro-blasts.6"62 Constitutively active MEKI mutantswere generated by replacing the two serinesthat are the sites of activation by Raf by neg-atively active amino acids that mimic the neg-ative charge brought by the phosphorylation(S222D and S228D, or S222E and S228E;D=Aspartic acid, E=glutamic acid). (4) Thedominant negative form of ERK can blockmitogenic signalling.63 These dominant neg-ative forms are generated by mutation of thephosphorylation sites (Tl83 or YI85) or muta-tion of the ATP binding site. The dominantnegative forms ofERK are only effective whenhighly expressed in a cell, and the mechanismof their action is not understood; they may actby making stable complexes with MEK.63

In most cell types, many agonists stimulateERK (reviewed in '), but there are several

M295

on July 23, 2020 by guest. Protected by copyright.

http://mp.bm

j.com/

Clin M

ol Pathol: first published as 10.1136/m

p.48.6.M292 on 1 D

ecember 1995. D

ownloaded from

Lenormand

noticeable differences in the pattern of ERKactivation according to the mitogenic potentialof these agonists. For example in CCL39fibroblasts it has been shown that non-mito-genic stimuli activate ERK in a transient man-ner (maximum of activation between 10 and30 minutes), whereas mitogenic stimuli inducea second phase of activation that persists forhours.64 In addition, only mitogenic stimuli areable to induce nuclear translocation of ERK.47Thus mitogenic stimuli activate ERK for a longduration and induce its nuclear translocation,which may then induce phosphorylation oftranscription factors. In this way several stimulican activate ERK, but the quantitative differ-ence in ERK activation that they induce (dur-able versus transitory) is translated into aqualitative difference (phosphorylation of tran-scription factors or not). Establishing the fullcascade of events linking the prolonged ERKactivation by mitogenic stimuli to cell pro-liferation is an immense challenge for futureresearch in this field.

DIFFERENTIATIONThe response ofPC 12 cells to receptor tyrosinekinase activation has been extensively used asan experimental system to study transductionand how activation of some receptors triggersdifferentiation, while the activation of othersleads to proliferation. Treatment ofPC 12 cellswith nerve growth factor (NGF) triggers neuriteoutgrowth and stops cell division,6566 whereastreatment with EGF leads to cell proliferation.67ERK activation is sustained for several hoursfollowing NGF stimulation, but it is transitorywith EGF stimulation.68-70 The association ofprolonged ERK activation with NGF stimu-lation of PC12 cells has led to the idea that itis sustained activation ofthis pathway that leadsto differentiation. Confirmation of this theorycame from three sets of experiments. (1) inPC 12 cells, stimulation ofthe endogenous EGFreceptor induces transient activation of ERKand does not lead to differentiation, but over-expression of the EGF receptor leads to EGFdependent differentiation and sustained ac-tivation of ERKs.7' (2) Expression of con-stitutively active MEK mutants in PC12 cellsinduces neurite outgrowth in the absence ofdifferentiating factors.60 (3) Dominant negativemutants of ERK block ligand activation ofendogenous ERK and stop differentiation ofPC 12 cells.60 These experiments show thatERK activation is necessary and prolongedactivation is sufficient to trigger PC 12 celldifferentiation.7' There is a correlation betweensustained activation of ERK induced by NGFand differentiation. Thus, as for ERK in-volvement in fibroblast proliferation, these ex-periments lead us to propose in PC12 thatthe sustained ERK activation induces markedchanges in gene expression which induce celldifferentiation. In conclusion, one can proposethat prolonged ERK activation triggers a cellfate that was preprogrammed in the cell, leadingto differentiation in PC 12 cells and cell pro-liferation in fibroblasts.

The newly discovered MAP kinasefamiliesSTRESS ACTIVATED MAP KINASES: SAPKIn 1991, Kyriakis et al72 discovered a 54 kDaprotein kinase that was phosphorylated on ser-ine, threonine, and tyrosine residues and ac-tivated in liver extracts from cycloheximidetreated rats. Subsequent protein purification,peptide sequencing, and cDNA cloning re-vealed a family of closely related stress activatedprotein kinases (SAPKs) activated by stressessuch as heat shock and by inflammatory cyto-kines such as tumour necrosis factor x.20 In-dependently, other laboratories identified twokinases that phosphorylate the amino terminusof c-un (on serines 63 and 73) and potentiatec-Jun transcriptional transactivating activity.7374These enzymes were thus named Jrun N ter-minal kinase: JNK1 and JNK2.2' JNKI isclosely related to SAPKy, andJNK2 to SAPKct.These enzymes constitute a new family ofMAPkinases since their activation requires dualphosphorylation on a tyrosine and a threonineresidue at the sequence TPY (P = proline), in-stead of the sequence TEY required for ERK.The race to identify the SAPK activator led to

the nearly simultaneous discovery of SEK12324and MKK475 which are highly homologous.Two initiation codons can potentially be usedto start translation of the SAPK activator, andit appears that the proteins engineered for invitro studies by these two groups differ: SEK1starts at the first potential initiation codon andis 34 amino acids larger than MKK4. TheSAPK activation cascade was further definedby discovering that MEKK is able to activatein vivo and in vitro both SEKW4 and MKK4.75Initially MEKK was thought to contribute inparallel with Raf-1 to the activation of MEK,and thus was named MEK kinase (MEKK).3However, MEKK is able to activate both MEKand SEK when highly expressed in a cell line,but upon mild (normal?) activation it solelyphosphorylates and activates SEK.2375 In fact,it is more likely that MEKK activates the stresspathway of SAPKs (thus SEK) than the mito-genic pathway of ERKs (thus MEK) as ex-pression of MEKK in fibroblasts arrests cellgrowth (24, and Brunet A, personal com-munication).

Matters are complicated by the knowledgethat both Raf and MEKK can be activated byRas35 76; the action of cofactors in Ras activationmust intervene to lead specifically to the ac-tivation of one of these two pathways. In ad-dition to Raf and MEKK, it has been shownrecently that Ras was able to activate PI3kinase.36 Understanding precisely the eventsleading to Ras and Raf activation and all thepathways initiated following Ras activation isa great challenge for the future, and for cancerresearch in particular, as the presence of onco-genic forms of Ras has been found in at least10% of all human cancers.37

OSMOREGULATED MAP KINASES: p38Three closely related proteins, p38, RK, andCSBP kinase, were cloned independently in1994 and belong to a new MAP kinase family,

M296

on July 23, 2020 by guest. Protected by copyright.

http://mp.bm

j.com/

Clin M

ol Pathol: first published as 10.1136/m

p.48.6.M292 on 1 D

ecember 1995. D

ownloaded from

Mammalian MAP kinase network

as they are activated upon dual phosphorylationof both a tyrosine and a threonine residue inthe sequence TGY (G = glycine).

p38p38 is a mammalian protein kinase phos-phorylated on tyrosine in response to lipo-polysaccharide (LPS) that was cloned frompeptide sequencing after its purification22 andwas shown to be closely related to the yeastprotein HOG1 (S cerevisiae). HOGI lies in asignalling pathway that restores the osmoticgradient across the yeast cell membrane inresponse to increased external osmolarity.77

RKRK was cloned after purification of the enzymethat phosphorylated MAPKAP 2 (MAP kinaseactivated protein 2) upon heat shock or arsenitestimulation, in the absence ofERK activation.78This protein was named RK after its capabilityto reactivate MAPKAP 2 which is then able tophosphorylate HSP27 (heat shock protein of27 kDa). HSP27 is induced in response tostress79 to protect cells against thermal in-juries.80

CSBP KINASECSBP kinase was identified as the target of anew drug, CSAID (cytokine suppressive anti-inflamatory drug), that inhibits inflammationcaused by the action of cytokines such asinterleukin- 1 and tumour necrosis factor(TNF).8' CSBP kinase was purified accordingto its ability to bind to radiolabelled and radio-photoaffinity labelled CSAID. It was thuscalled CSAID binding protein kinase, CSBPkinase. CSAID blocks the biosynthesis of cyto-kines through inhibition of the CSBP kinase;it has the extraordinary interest of being thefirst drug able to block specifically one of thethree known MAP kinase cascades.

The three kinases p38, RK, and CSBP kinaserepresent the same protein and will be referredto in this review as p38. One putative nuclearsubstrate ofp38 is the transcription factor ATF-2 (75, and Brunet A, personal communication).One activator ofp38 has been partially purified:RKK (RK kinase)78; two others, MKK3 andMKK4, were cloned by polymerase chainreaction (PCR) using degenerated oligo-nucleotides corresponding to PBS2, the knownactivator ofHOG1 which is the yeast equivalentof p38.75 MKK3 has been shown to activatep38 specifically, whereas MKK4 (= SEK) canalso phosphorylate SAPK, as was described inthe previous section.24 No upstream activatorof MKK3 has yet been cloned but it seemsthat Ras and Raf are not implicated in theactivation of this cascade.78 Interestingly if onemakes a parallel with the yeast cascade (PBS2and HOG177 which are highly homologous toMKK3 and p38 respectively), one can envisionthat an osmosensor capable of triggering theactivation of this kinase cascade could be ahistidine kinase, a type of kinase unknown inmammalian cells at present. The yeast osmo-sensor is called SLN17; its mutational deletion

leads to lethality, which can be suppressed byfour complementation groups. One suppressorof the SLN1 mutation is HOG1 and anotheris its activator PBS2. These data clearly indicatethat the yeast osmosensor SLN1 activates akinase cascade similar to the mammalianMKK3-p38 cascade. The two other com-plementation groups led to the discovery of acytoplasmic protein homologous to the reg-ulatory domain of the osmosensor SLN1, andof phosphatases.7 It remain to be seen whethermammalian proteins homologous to each yeastprotein of this osmoregulatory pathway will befound.

Understanding the regulation of themammalian MAP kinase networkStudying the regulation of such a complexkinase network is fraught with difficulties. Thissection will provide some reflections about lim-itations in interpreting current results, and in-dicate research trends in this field.

THE DIFFICULTIES IN IDENTIFICATION OF INVIVO SUBSTRATEIn vitro the catalytic properties of a kinase areoften modified, usually with a reduction insubstrate specificity. To identify substrates un-ambiguously in vivo, one can only correlate invitro and in vivo phosphorylation at the samesites, but these sites may be phosphorylated invivo by several kinases. Phosphorylation mayalso occur on sites that do not affect the func-tion of a particular protein as we can assess it.For example, c-Jun was thought to be an invivo substrate ofERK because ERK was shownto phosphorylate the full c-Jun protein readily invitro.82 The significance of this ERK dependentphosphorylation of c-Jun is not understoodbecause C terminal phosphorylation does notaffect c-Jun transcriptional transactivation. Itis now acknowledged that SAPKs (JNKs) arethe cellular activators of c-Jun as they spe-cifically phosphorylate both the serines 63 and73 implicated in activation of c-Jun tran-scriptional transactivation.83The discovery of the p38 specific inhibitor,

CSAID,8' will allow the specific blocking ofthis kinase cascade and should permit the iden-tification of specific roles of the cascade incell physiology. Specific inhibitors of the otherMAP kinases may be found, and this wouldallow us to assess the precise contribution ofeach cascade in cellular responses to mitogenicand non-mitogenic stimuli or to stress factors.One incentive for finding specific inhibitorsof ERK is the possibility of blocking manyupstream pathological disorders of this path-way, disorders that often lead to aberrant cellproliferation. For example, it is known thatmutations of Ras are implicated in about 70%of all pancreatic cancers37; as blocking ERKactivation was shown to stop mitogenicity infibroblasts,63 it would be of interest to testwhether this is also true for some types ofcancer cells, such as those with constitutivelyactive Ras.

M297

on July 23, 2020 by guest. Protected by copyright.

http://mp.bm

j.com/

Clin M

ol Pathol: first published as 10.1136/m

p.48.6.M292 on 1 D

ecember 1995. D

ownloaded from

Lenormand

OVEREXPRESSION OF THE KINASE STUDIED MAYCREATE ARTEFACTUAL EFFECTSMEKK was named from its ability to activateMEK both in vitro and in vivo when over-

expressed in fibroblasts.35 Recentwork indicatesthat transfection ofMEKK induces both MEKand SEK activation, whereas low expressionofMEKK upon conditional induction activatesSEK specifically, and not MEK.232475 The use

of inducible promoters to drive expression ofthe kinase of interest should be more widelyapplied. Unfortunately, all available inducibleexpression vectors present drawbacks such as

leaky expression without induction or weakinduction, or even secondary effects by theinducer. An alternative strategy is to make fu-sion proteins with the catalytic domain of theprotein of interest and the oestradiol receptorbinding domain (ER). In some cases, such aswith Raf, the oestradiol binding domain is ableto mask the catalytic domain of the kinase,and upon addition of oestradiol the catalyticdomain is revealed.8485 Cell lines expressingthis Raf-ER have been established, and it hasbe shown that addition of oestradiol leads toMEK activation within two minutes and ERKactivation within five minutes, the maximumof activation being reached by 30 minutes(85 and Lenormand P, unpublished data).

CELL TYPE SPECIFIC DIFFERENCESOne stimulus may provoke opposite effects intwo cell types, which may be the result ofmarked differences in the differentiation pro-

gramme engaged. I shall illustrate this com-

plexity by two examples, the response of MAPkinase pathways to osmotic shock, and therole of PKA activation on the Raf-MEK-ERKpathway.

Variability in osmotic shock responsesIn some fibroblasts, osmotic shock has no sig-nificant effects on ERK activation.22 On thecontrary, cells of renal origin such as MDCKcells (Madin-Darby canine kidney epithelialcells) respond to hyperosmolarity by activationof Raf, MEK, and ERK.86 Clearly fibroblastsare not exposed to the same variations of os-

motic pressure as are renal cells; none the lessno molecular mechanism can yet explain suchdifferences.

In the cell lines tested so far (mostly fibro-blasts), hyperosmolarity modulates both SAPKand p38.202287 This can be explained by thefact that p38 and SAPK can be activated bya SEK/MKK475; in addition they could beactivated by distinct MAPKKs that are bothosmoregulated.

Variability in response to PKA activationERK activation is required for fibroblast cellproliferation,6' and it has been appreciated forsome time that a rise in cAMP, which activatesPKA, inhibits cell proliferation.88 In contrast,in dog thyrocytes (cells of epithelial origin)PKA stimulates growth in the absence ofERK

stimulation.89 The molecular mechanisms ex-plaining these differences are lacking; howeverprogress was obtained by showing in NIH 3T3fibroblasts that PKA activation inhibits theERK pathway at the level of Rafl.90-94 Thiseffect may result from direct inhibitory phos-phorylation of Rafl by PKA.95 However, it hasrecently been shown in CCL39 fibroblasts thatPKA inhibition ofERK, although rapid, is verytransient (McKenzie F, personal com-munication). Finally, in PC12 cells, PKA stim-ulates ERK activation by NGF instead ofblocking it.""

Interestingly, in PC 12 cells PKA inhibits Rafactivated by NGF.9899 Thus cAMP activationof ERK would have to be accomplished by Rafindependent activation of a distinct MAPKKKsuch as MEKK, which is activated by a Rasdependent pathway and has been shown toactivate MEK in vitro or when overexpressed.35The different effects of PKA activation on

the ERK pathway in fibroblasts, PC12 cells,and thyrocytes provide another illustration offundamental cell type specific differences. Un-derstanding these differences at the molecularlevel is a great challenge.

SPECIFICITY IN MAP KINASE ACTIVATIONIn some cell lines the activation ofMAP kinasepathways can be kept separated; for examplein PC1 2 cells, NGF activates ERK but notp38, and arsenite activates p38 but not ERKs.78

This isolation of one MAP kinase pathwayfrom the others may be due to the specificitiesofthe kinases involved; it may also be helped byarranging the cascades of kinases in complexes.The only example of such complex in MAPkinase signalling is provided by the mating-type pathway in the yeast S cerevisiae. Thekinases of this pathway (homologous to MEKand ERK) seem to be held together by ascaffolding protein, STE5.1`010' Kinases fromthe two other yeast MAP kinase pathways (onefor osmosensing and one dedicated to cell wallbiosynthesis) cannot bind to STE5.1'° STE5may bring kinases into close proximity to ensurespecificity and speed of activation. Severallaboratories are searching for mammalian hom-ologues of STE5, but there is no report ofsuccess at present.

CROSS TALK BETWEEN MAP KINASE PATHWAYSNegative feedback is a widely used mechanismfor signal attenuation and desensitisation,though it is not clear yet whether negativefeedback operates in MAP kinase signallingpathways. Activated ERK was shown to phos-phorylate SOS,'02 Raf,'03 and MEK.'0"'06 Itseems that these feedback phosphorylationshave little or no effect on MEK and Raf ac-tivities, but this lack of effect could be due toour incapacity to measure fine regulation within vitro studies. Negative feedback by ERK wasthought to occur by phosphorylation of theEGF receptor at threonine 669, which doesnot seem to affect the kinase activity but couldincrease receptor intemalisation.'07 '08

M298

on July 23, 2020 by guest. Protected by copyright.

http://mp.bm

j.com/

Clin M

ol Pathol: first published as 10.1136/m

p.48.6.M292 on 1 D

ecember 1995. D

ownloaded from

Mammalian MAP kinase network

Two groups reported that ERK phos-phorylates SEK,2475 but the significance of thiscrosstalk among MAP kinase pathways is stillnot understood.

ROLE OF THE PHOSPHATASESIt is increasingly obvious that protein phos-phatases are involved in the regulation of signaltransduction (reviewed in 3109110). ERK ac-

tivation is biphasic, and some phosphatasesare likely to be responsible for both phases ofinactivation. The phosphatase responsible forinactivating the first peak of activation is un-

known, but for the second phase of activationMKP- 1 (MAP kinase phosphatase- 1, alsocalled CL100) is a good candidate.1' It belongsto a new family of dual specificity phosphatasesable to dephosphorylate ERK on both thethreonine and the tyrosine residue. MKP-1 isan early gene which is not expressed in arrestedcells: upon stimulation MKP-1 accumulatesafter 30 minutes and is nuclear."2 It could beresponsible for inactivating ERK in the nucleus;none the less its precise role still needs to bewell defined since it is now apparent that a

family ofMKP- 1 -like proteins exists which maybe specific inactivators of each MAP kinase.However, ERK kinase inactivation is often toorapid to be explained by the MKP-1 familyof proteins."3 Recent data suggest that rapidinactivation of the MAP cascade at the level ofERK is mediated by PP2A (protein phos-phatase 2A) which removes phosphothreonineand an as yet unidentified tyrosine phosphatasedistinct from MKP-1 .

Concluding remarksIn recent years, signalling by MAP kinase hasblossomed into three distinct kinase cascades.The crossover points between individual mem-bers of these cascades, and hence the specificityof agonist activation of each cascade, remainto be defined. One of the greatest challengesis to identify the specific in vivo substrates ofeach of the three end members of the ERK,SAPK, and p38 cascades. For pathologists,fundamental research on this pathway will pro-vide tools that may permit the correlation ofthe abberant status of an MAP kinase pathwaywith pathophysiological conditions. For ex-

ample it is at present impossible to evaluateRas activation in one cell, and it may remainimpossible because of the small percentage ofGTP-Ras sufficient to induce full activation ofERK. None the less signal magnification mayrender the indirect assessment ofRas activationpossible from the development of new toolssuch as monoclonal antibodies that specificallyrecognise phosphorylated ERK (Nishida E,personal communication). With these anti-bodies, it is possible to evaluate the subcellularlocalisation of activated ERK. Will it be possibleto distinguish between aberrantly proliferatingcells and normal ones by detecting nuclearaccumulation of activated ERK?

I am most grateful to Drs A Brunet, N Rivard, J-M Brondello,and F McKenzie critical reading of the manuscript; specialthanks to Drs J Lavoie and G L'Allemain for providing valuableinformation and DrJ Pouyssegur for excellent scientific support.

1 L'Allemain G. Deciphering the MAP kinase pathway. ProgGrowth Factor Res 1994;5:291-334.

2 Marshall CJ. Specificity of receptor tyrosine kinase sig-naling: transient versus sustained extracellular signal-reg-ulated kinase activation. Cell 1995;80: 179-85.

3 Hunter T. Protein kinases and phosphatases: the yin andyang of protein phosphorylation and signaling. Cell 1995;80:225-36.

4 Cooper JA. Straight and narrow or tortuous and in-tersecting? Cu-rr Biol 1994;4: 1118-21.

5 Cano E, Mahadevan LC. Paralled signal processing amongmammalian MAPKS. Trends Biochenm Sci 1995;20:117-22.

6 Bourret RB, Hess JF, Borkovich KA, Pakula AA, SimonMI. Protein phosphorylation in chemotaxis and two-component regulatorv systems of bacteria. J Biol Chem1989;264:7085-8.

7 Maeda T, Wurgler-Murphy SM, Saito H. A two-com-ponent system that regulates an osmosensing MAP kinasecascade in yeast. Nature 1994;369:242-5.

8 Rav LB, Sturgill TW. Rapid stimulation by insulin of aserine/threonine kinase in 3T3-LI adipocytes that phos-phorylates microtubule-associated protein 2 in vitro. ProcNatl Acad Sci USA 1987;84:1502-6.

9 Payne DM, Rossomando AJ, Martino P, Erickson AK, HerJH, Shabanowtiz J. et al. Identification of the regulatoryphosphorylation sites in pp42/mitogen-activated proteinkinase (MAP kinases). EMBOJ7 1991;10:885-92.

10 Matsuda S, Kosako H, Takenaka K, Moriyama K, SakaiH, Akiyama T, et al. Xenopus MAP kinase activator:identification and function as a key intermediate in thephosphorylation cascade. EAfBO J 1992;11:973-82.

12 Lange-Carter CA, Pleiman CM, Gardner AM, BlumerKJ, Johnson GL. MEK kinase and Raf define a divergencein the MAP kinase regulatory network. Science 1993;260:318-9.

12 Hallberg B, Rayter SI, Downward J. Interaction of rasand raf in intact mammalian cells upon extracellularstimulation. J Biol Chen, 1993;269:3913-6.

13 Ray LB, Sturgill TW. Insulin-stimulated microtubule-as-sociated protein kinase is phosphorylated on tyrosine andthreonine in vivo. Proc Natl Acad Sci USA 1988;85:3753-7.

14 Pelech SL, Tombes RM, Meijer L, Krebs EG. Activationof myelin basic protein kinases during echinoderm oocytematuration and egg fertilization. Dev Biol 1988;130:28-36.

15 Ahn N, Krebs E. Evidence for an epidermal growth factor-stimulated protein kinase cascade in swiss 3T3 cells. JBiol Chen, 1990;265:11495-501.

16 Chung J, Pelech S, Blenis J. Mitogen-activated SwissMouse 3T3 RSK kinases I and II are related to pp44 mapkfrom sea star oocytes and participate in the regulation ofpp9O RSK activity. Proc Natl Acad Sci USA 1991;88:4981-5.

17 Boulton TG, Yancopoulos GD, Gregory JS, Slaughter C,Moomaw C, Hsu J, et al. An insulin-stimulated proteinkinase similar to yeast kinases involved in cell cycle con-trol. Scienzce 1990;249:64-7.

18 Boulton TG, Nye SH, Robbins DJ, Ip NY, RadziejewskaE, Morgenbesser SD, et al. ERKs: a family of protein-serine/threonine kinases that are activated and tyrosinephosphorylated in response to insulin and NGF. Cell1991;65:663-75.

19 Crews CM, Alessandrini A, Erikson RL. The primarvstructure of MEL, a protein kinase that phosphorylatesthe ERK gene product. Science 1992;258:478-80.

20 Kyriakis JM, Banerjee P, Nikolakaki E, Dai T, Rubie EA,Ahmad MF, et al. The stress-activated protein kinasesubfamily of c-J7un kinases. Natuire 1994;369: 156-60.

21 Derijard B, Hibi M, Wu I-H, Barrett T, Su B, Deng T, etal. JNKI: a protein kinase stimulated bv UV light andHa-Ras that binds and phosphorylates and the c-junactivation domain. Cell 1994;76:1025-37.

22 Han J, Lee J-D, Bibbs L, Ulevitch RJ. A MAP kinasetargeted by endotoxin and hyperosmolarity in mammaliancells. Science 1994;265:808-11.

23 Sanchez I, Hughes RT, Mayer BJ, Yee K, Woodgett JR,Avruch J, et al. Role of SAPK/ERK kinase- 1 in the stress-activated pathway regulating transcription factor c-junl.Nature 1994;372:794-8.

24 Yan M, Dai T, Deak JC, Kyriakis JM, Zon LI, WoodgettJR, et al. Activation of stress-activated protein kinase bvMEKK1 phosphorylation of its activator SEKI. Natlure1994;372:798-800.

25 Schlessinger J. SH2/SH3 signalling proteins. Curr OpinGenet Dev 1994;4:25-30.

26 Songyang Z, Shoelson SE, McGlade J, Olivier P, PawsonT, Bustelo XR, et al. Specific motifs recognized by theSH2 domains of CSk, 3BP2, fps/fes, GRB-2, HCP, SHC,Syk, and Vav. Mol Cell Biol 1994,14:2777-85.

27 Koide H, Satoh T, Nakafuku M, Kaziro Y. Gtp-dependentassociation of raf-1 with Ha-Ras: identification of Ras asa target downstream of Ras in mammalian cells. Proc VailAcad Sci USA 1993;90:8683-6.

28 Moodie SA, Willumsen BM, Weber MJ, Wolfman A.Complexes ofRas-GTP with Raf- I and mitogen-activatedprotein kinase kinase. Scietnce 1993;260: 1658-61.

29 Van Aelst L, Barr M, Marcus S, Polverino A, Wigler M.Complex formation between Ras and Raf and otherkinases. Proc Natl Acad Sci USA 1993;90:6213-7.

30 Vojtek AB, Hollenberg SM, Cooper JA. Mammalian Rasinteracts directly with the serine/threonine kinase Raf.Cell 1993;74:505-14.

M299

on July 23, 2020 by guest. Protected by copyright.

http://mp.bm

j.com/

Clin M

ol Pathol: first published as 10.1136/m

p.48.6.M292 on 1 D

ecember 1995. D

ownloaded from

Lenormand

31 Stokoe D, Macdonald SG, Cadwallader K, Symons M,Hancock JF. Activation of Raf as a result of recruitmentto the plasma membrane. Science 1994;264: 1463-7.

32 Leevers SJ, Paterson HF, Marshall CJ. Requirement forRas in Raf activation is overcome by targeting Raf to theplasma membrane. Nature 1994;369:411-4.

33 Kyriakis JM, App H, Zhang X, Banerjee P, Brautigan DL,Rapp UR, et al. Raf- 1 activates MAP kinase-kinase. Nature1992;358:417-21.

34 Alessi D, Saito Y, Campbell DG, Cohen P, SithanandamP, Rapp U, et al. Identification of the sites in MAP kinasekinase-1 phosphorylated by p74 raf-1. EMBO_7 1994;13:1610-9.

35 Lange-Carter CA, Johnson GL. Ras-dependent growthfactor regulation of MEK kinase in PC12 cells. Science1994;265: 1458-61.

36 Rodriguez-Viciana P, Wame PH, Dhand R, Van-haesebroeck B, Gout I, Fry MJ, et al. Phos-phatidylinositol-3-OH kinase as a direct target of Ras.Nature 1994;370:527-32.

37 Barbacid M. Ras genes. Annu Rev Biochem 1987;56:779-827.

38 White MA, Nicolette C, Minden A, Polverino A, Van AelstL, Karin M, et al. Multiple ras functions can contributeto mammalian cell transformation. Cell 1995;80:533-41.

39 Davis RJ. The motigen-activated protein kinase signaltransduction pathway. Biol Chem 1993;268: 14553-6.

40 Marais R, Wynne J. The SRF accessory protein Elk-1contains a growth factor-regulated transcriptional ac-tivation domain. Cell 1993;73:381-93.

41 Hill CS, Marais R, John S, Wynne J, Dalton S, TreismanR. Functional analysis of a growth factor-responsive tran-scription factor complex. Cell 1993;73:395-406.

42 Hipskind RA, Baccarini M, Nordheim A. Transient ac-tivation of Raf-1, MEK and ERK2 coincides kineticallywith ternary complex factor phosphorylation and im-mediate-early gene promoter activity in vivo. Mol BellBiol 1994;14:6219-31.

43 Treishman R. The serum response element. Trends BiochemSci 1991;17:423-7.

44 Hipskind RA, Rao VN, Mueller CG, Reddy ES, NordheimA. Ets-related protein Elk-I is homologous to the c-fosregulatory factor p62TCF. Nature 1991;354:531-4.

45 Rao VN, Reddy SP. Elk-I proteins interact with MAPkinases. Oncogene 1994;9:1855-60.

46 Chen TH, Samecki C, Blenis J. Nuclear localization andregulation of ERK- and RSK-encoded protein kinases.Mol Cell Biol 1992;12:915-27.

47 Lenormand P, Sardet C, Pages G, L'Allemain G, BrunetA, Pouyssegur J. Growth factors induce nuclear trans-location of MAP kinases (p42mapk and p44mapk) butnot of their activator MAP kinase kinase (p45 mapkk) infibroblasts. _7 Cell Biol 1993;122: 1079-88.

48 Traverse S, Cohen P, Paterson H, Marshall C, Rapp U,Grand RJA. Specific association of activated MAP kinasekinase kinase (Raf) with the plasma membranes of ras-transformed retinal cells. Oncogene 1992;8:3175-81.

49 Grand RJA, Owen D. The biochemistry of ras p21. Biochem199 1;279:609-31.

50 Johnson GL, Vaillancourt RR. Sequential protein kinasereactions controlling cell growth and differentiation. CurrOpin Cell Biol 1994;6:230-8.

51 Crespo P, Xu N, Simonds WF, Gutkind JS. Ras-dependentactivation of MAP kinase pathway mediated by G-proteinbeta-gamma subunits. Nature 1994;369:418-20.

52 Kolch W, Heidecker G, Kochs G, Hummel R, Vahidi H,Mischak H, et al. Protein kinase C alpha activates Raf-1by direct phosphorylation. Nature 1993;364:249-52.

53 Aitken A. 14-3-3 proteins on the map. Trends Biochem Sci1995;20:95-7.

54 Freed E, Symons M, Macdonald SG, McCormick F,Ruggieri R. Binding of 14-3-3 proteins to the proteinkinase Raf and effects on its activation. Science 1994;265:1713-6.

55 Irie K, Gotoh Y, Yashar BM, Errede B, Nishida E, Mat-sumoto K. Stimulatory effects of yeast and mammalian14-3-3 proteins on the Raf protein kinase. Science 1994;265:1716-9.

56 Fu H, Xia K, Pallaq DC, Cui C, Conroy K, NarsimhanRP. et al. Interaction of the protein kinase Raf-1 with 14-3-3 proteins. Science 1994;266: 126-9.

57 Reuther GW, Fu H, Cripe LD, Collier RJ, PendergastAM. Association of the protein kinases c-Bcr and Bcr-Abl with proteins of the 14-3-3 family. Science 1994;266:129-33.

58 Feramisco JR, Gross M, Kamata T, Rosenberg M, SweetRW. Microinjection of the oncogene form of the humanH-ras (T-24) protein results in rapid proliferation ofquiescent cells. Cell 1984;38:109-17.

59 Morris JDH, Price B, Lloyd AC, Self AJ, Marshall CJ,Hall A. Scrape loading of Swiss 3T3 cells with Rasprotein rapidly activates protein kinase-C in the absenceof phosphoinositide hydrolysis. Oncogene 1989;4:27-31.

60 Cowley S, Paterson H, Kemp P, Marshall CJ. Activationof MAP kinase kinase is both necessary and sufficient for

PC 12 differentiation and for transformation ofNIH 3T3cells. Cell 1994;77:841-52.

61 Mansour SJ, Matten WT, Hermann AS, Candia JM, RongS, Fukasawa K, et al. Transformation of mammalian cellsby constitutively active MAP kinase kinase. Science 1994;265:966-70.

62 Brunet A, Pages G, Pouyssegur J. Constitutively activemutants of MAP kinase kinase (MEKI) induce growthfactor-relaxation and oncogenicity when expressed infibroblasts. Oncogene 1994;9:3379-87.

63 Pages G. Lenormand P, L'Allemain G, Chambard JC,Meloche S, Pouyssegur J. Mitogen-activated protein kin-ases p42maPk and p44"apk are required for fibroblast pro-liferation. Proc Natl Acad Sci USA 1993;90:8319-23.

64 Meloche S, Seuwen K, Pages G, Pouyssegur J. Biphasicand synergistic activation of p44"P'a (ERKI) by growthfactors: correlation between late phase activation andmitogenicity. Mol Endocrinol 1992;6:845-54.

65 Greene LA, Tischler AS. Establishment of a noradrenergicclonal line of rat adrenal pheochromocytoma cells whichrespond to nerve growth factor. Proc Natl Acad Sci USA1976;73:242-8.

66 Tischler AS, Dichter MA, Biales B, DeLellis RA, WolfeH. Neural properties of cultured human endocrine tumorcells of proposed neural crest origin. Science 1976;192:902-4.

67 Huff K, End D, Guroff G. Nerve growth factor-inducedalterations in the response of PC 12 phaechromocytomacells to epidermal growth factor. _7 Cell Biol 1981;88:189-98.

68 Heasley LE, Johnon GL. The beta PDGF receptor inducesneuronal differentiation of PC 12 cells. Mol Biol Cell 1992;3:545-53.

69 Traverse S, Gomez N, Paterson H, Marshall C, CohenP. Sustained activation of the mitogen-activated protein(MAP) kinase cascade may be required for differentiationof PC12 cells. Biochem 1992;288:351-5.

70 Nguyen TT, Scimeca J-C, Fillous C, Peraldi P, CarpentierJ-L, Van Obberghen E. Co-regulation of the mitogen-activated protein kinase, extracellular signal-regulated ki-nase and the 90 kDa ribosomal S6 kinase in PC12 cells._7 Biol Chem 1993;268:9803-10.

71 Traverse S, Seedorf K, Paterson H, Marshall C, Cohen P,Ullrich A. EGF triggers neuronal differentiation ofPC 12cells that overexpress the EGF receptor. Cuw Biol 1994;4:694-701.

72 Kyriakis JM, Brautigan DL, Ingebritsen TS, Avruch J.pp54 microtubule-associated protein-2 kinase requiresboth tyrosine and serine/threonine phosphorylation foractivity. Biol Chem 199 1;266:10043-6.

73 Adler V, Polotskaya A, Wagner F, Kraft AS. Affinitv-purified c-Jun amino-terminal protein kinase requiresserine/threonine phosphorylation for activity. 7 Biol Chem1992;267: 17001-5.

74 Hibi M, Lin A, Smeal T, Minden A, Karin M. Iden-tification of an oncoprotein- and UV-responsive proteinkinase that binds and potentiates the c-Jun activationdomain. Genes Dev 1993;7:2135-48.

75 Derijard B, Raingeaud J, Barrett T, Wu I-H, Han J,Ulevitch RJ, et al. Independent human MAP kinase signaltransduction pathways defined by MEK and MKK iso-forms. Science 1995;267:682-5.

77 Hall A. A biochemical function for Ras-at last. Science1994;264:1413-4.

77 Brewster JL, de Valoir Y, Dwyer ND, Winter E, GustinMC. An osmo-sensing signal transduction pathway in

yeast. Science 1993;259:1760-3.78 Rouse J, Cohen P, Trigon S, Morange M, Alonso-Li-

mazares A, Zamanillo D, et al. A novel kinase cascadetriggered by stress and heat shock that stimulates MAP-KAP kinase-2 and phosphorylation of the small heatshock proteins. Cell 1994;78: 1027-37.

79 Landry J, Chretien P, Lambert H, Hickey E, Weber LA.Heat shock resistance conferred by expression of thehuman hsp 27 gene in rodent cells. _7 Cell Biol 1989;109:7-15.

80 Lavoie JN, Gingras BG, Tanguay RM, Landry J. Inductionof Chinese hamster HSP27 gene expression in mousecells confers resistance to heat shock. HSP27 stabilizationof the microfilament organization. _7 Biol Chem 1993;268:3420-9.

81 Lee JC, Laydon JT, McDonnell PC, Gallagher TF, KumarS, Green D, et al. A protein kinase involved in theregulation of inflammatory cytokine biosynthesis. Nature1994;372:739-46.

82 Pulverer BJ, Kyriakis JM, Avruch J, Nikolakaki E, WoodgettJR. Phosphorylation of c-jun mediated by MAP kinases.Nature 1991;353:670-4.

83 Minden A, Lin A, Smeal T, Derijard B, Cobb M, DavisR, et al. c-jun N-terminal phosphorylation correlates withactivation ofthe JNK subgroup but not the ERK subgroupof mitogen-activated protein kinases. Mol Cell Biol 1994;14:6683-8.

84 Samuels ML, McMahon M. Inhibition of platelet-derivedgrowth factor- and epithelial growth factor-mediated mito-genesis and signaling in 3T3 cells expressing ARaf-1:ER,an estradiol-regulated form of Raf- 1. Moll Cell Biol 1994;14:7855-66.

85 Samuels ML, Weber MJ, Bishop JM, McMahon M. Con-ditional transformation of cells and rapid activation ofthe mitogen-activated protein kinase by an estradiol-dependent human Raf-1 protein kinase. Mol Cell Biol1993;13:6241-52.

86 Terada Y, Tomita K, Homma MK, Nonoguchi H, YangT, Yamada T, et al. Sequential activation of Raf- kinase,mitogen-activated protein (MAP) kinase kinase, MAPkinase, and S6 kinase by hyperosmolarity in renal cells._7 Biol Chem 1994;269:31296-301.

87 Galcheva-Gargova Z, Derijard B, Wu I-H, Davis RJ. Anosmosensing signal transduction pathway in mammaliancells. Science 1994;265:806-8.

88 Pastan I, Johnson GS, Anderson WB. Role of cyclic nuc-leotides in growth control. Ann Rev Biochem 1975;44:491-522.

M300

on July 23, 2020 by guest. Protected by copyright.

http://mp.bm

j.com/

Clin M

ol Pathol: first published as 10.1136/m

p.48.6.M292 on 1 D

ecember 1995. D

ownloaded from

Mammalian MAP kinase network

89 Lamy F, Wilkin F, Baptis M, Posada J, Roger PP, DumontJE. Phosphorylation of mitogen-activated protein kinasesis involved in the epidermal growth factor and phorbolester, but not in the thyrotropin/cAMP, thyroid mitogenicpathway. J Biol Chem 1993;268:8398-401.

90 Burgering BMT, Pronk GJ, Vanweeren PC, Chardin P,Bos JL. cAMP antagonizes p21 (ras)-directed activationof extracellular signal-regulated kinase-2 and phos-phorylation of mSos nucleotide exchange factor. EMBOJ 1993;12:4211-20.

91 Cook SJ, McCormick F. Inhibition by cAMP of Ras-dependent activation of raf. Science 1993;262:1069-72.

92 Graves LM, Bornfeldt KE, Raines EW, Potts BC, Mac-donald SG, Ross R, et al. Protein kinase-A antagonizesplatelet-derived growth factor-induced signaling by mi-togen-activated protein kinase in human arterial smoothmuscle cells. Proc Natl Acad Sci USA 1993;90:10300-4.

93 Sevetson BR, Kong XM, Lawrence JC. Increasing cAMPattenuates activation of mitogen-activated protein kinase.Proc Natl Acad Sci USA 1993;90:10305-9.

94 Wu J, Dent P, Jelinek T, Wolfman A, Weber MJ, SturgillTW. Inhibition of the EGF-activated MAP kinasesignaling pathway by adenosine 3',5'-monophosphate.Science 1993;262: 1065-9.

95 Hafner S, Adler HS, Mischak H, Janosh P, Heidecker G,Wolfman A, et al. Mechanism of inhibition of Raf-I byprotein kinase A. Mol Cell Biol 1994;14:6696-703.

96 Frodin M, Peraldi P, Van Obberghen E. Cyclic AMPactivates the mitogen-activated protein kinase cascade inPC 12 cells. J Biol Chem 1994;269:6207-14.

97 Vaillancourt RR, Gardner AM, Johnson GL. B-Raf de-pendent regulation of the MEK-l/mitogen-activated pro-tein kinase pathway in PC 12 cells and regulation by cyclicAMP. Mol Cell Biol 1994;14:6522-30.

98 Stephens RM, Sithanandam G, Copeland TD, KaplanDR, Rapp UR, Morrison DK. 95-kilodalton B-Raf serine/threonine kinase: identification of the protein and itsmajor autophosphorylation site. Mol Cell Biol 1992;12:3733-42.

99 Jaiswal RK, Mookied SA, Wolfman A, Landreth GE. Themitogen-activated protein kinase cascade is activated byB-Raf in response to nerve growth factor through in-teraction with p21 Ras. Mol Cell Biol 1994;14:6944-53.

100 Choi KY, Satterberg B, Lyons DM, Elion EA. Ste5 tethersmultiple protein kinases in the MAP kinase cascade re-quired for mating in S. cerevisiae. Cell 1994;78:499-512.

101 Marcus S, Polverino A, Barr M, Wigler M. Complexesbetween STE5 and components of the pheromone-re-sponsive mitogen-activated protein kinase module. ProcNatl Acad Sci USA 1994;91:7762-6.

102 Cherniack AD, Klarklund JK, Czech MP. Phosphorylationof the ras nucleotide exchange factor son of sevenless bymitogen-activated protein kinase. J Biol Chem 1994;269:4717-20.

103 Lee RM, Cobb MH, Blackshear PJ. Evidence that ex-tracellular signal-regulated kinases are the insulin-ac-tivated Raf-1 kinase kinases. J Biol Chem 1992;267:1088-92.

104 Matsuda S, Gotoh Y, Nishida E. Phosphorylation of Xen-opus mitogen-activated protein (MAP) kinase kinase byMAP kinase kinase and MAP kinase. 7 Biol Chem 1993;268:3277-81.

105 BrunetA, Pag&s G, PouyssegurJ. Growth factor-stimulatedMAP kinase induces rapid retrophosphorylation and in-hibition ofMAP kinase kinase (MEK1). FEBS Lett 1994;346:299-303.

106 Gardner AM, Vaillencourt RR, Lange-Carter CA, JohnsonGL. MEK-1 phosphorylation by MEK kinase, Raf andmitogen-activated protein kinase: analysis of phos-phopeptides and regulation of activity. Mol Biol Cell 1994;5:193-201.

107 Takishima K, Griswold-Prenner I, Ingebritsen T, RosnerMR. Epidermal growth factor (EGF) receptor T669 pep-tide kinase from 3T3-L1 cells is an EGF-stimulated MAPkinase. Proc Natl Acad Sci USA 1991;88:2520-4.

108 Northwood IC, Gonzalez FA, Wartmann M, Raden DL,Davis RJ. Isolation and characterisation of two growthfactor-stimulated protein kinases that phosphorylate theepidermal growth factor receptor at threonine 669. J BiolChem 1991;266:15266-76.

109 Clarke PR. Switching off MAP kinases. Curr Biol 1994;4:647-50.

110 Sun H, Tonks NK. The coordinated action of proteintyrosine phosphatases and kinases in cell signalling. TrendsBiochem Sci 1994;19:480-5.

111 Sun H, Charles CH, Lau LF, Tonks NK. MKP-1(3CH134), an immediate early gene product, is a dualspecificity phosphatase that dephosphorylates MAP ki-nase in vivo. Cell 1992;75:487-93.

112 Brondello J-M, McKenzie FM, Sun H, Tonks NK, Pouys-segur J. Constitutive MAP kinase phosphatase (MKP1)expression blocks GI specific gene transcription and S-phase entry in fibroblasts. Oncogene 1995;10:1895-904.

113 Wu J, Lau LF, Sturgill TW. Rapid deactivation of MAPkinase in PC12 cells occurs independently of inductionof phosphatase MKP-1. FEBS Lett 1994;353:9-12.

114 Alessi D, Gomez N, Moorhead G, Lewis T, Keyse SM,Cohen P. Inactivation of p42 MAP kinase by proteinphosphatase 2A and a protein tyrosine phosphatase, butnot CL100, in various cell lines. CurrBiol 1995;5:283-95.

M301

on July 23, 2020 by guest. Protected by copyright.

http://mp.bm

j.com/

Clin M

ol Pathol: first published as 10.1136/m

p.48.6.M292 on 1 D

ecember 1995. D

ownloaded from