Sending Signals From the Synapse to theNucleus: Possible Roles for CaMK, Ras/ERK,and SAPK Pathways in the Regulation ofSynaptic Plasticity and Neuronal GrowthJessica Curtis and Steven Finkbeiner*Division of Neuroscience, Department of Neurology, Harvard Medical School, Children’s Hospital,Boston, Massachusetts

The ability to learn and form memories depends onspecific patterns of synaptic activity and is in parttranscription dependent. However, the signal transduc-tion pathways that connect signals generated at syn-apses with transcriptional responses in the nucleus arenot well understood. In the present report, we discussthree signal transduction pathways: the Ca21/calmodu-lin-dependent kinase (CaMK) pathway, the Ras/ERKpathway, and the SAPK pathways that might functionto couple synaptic activity to long-term adaptiveresponses, in part through the regulation of new geneexpression. Evidence suggests that these pathwaysbecome activated in response to stimuli that regulatesynaptic function such as the influx of extracellularCa21 and certain neurotrophin growth factors such asbrain-derived neurotrophic factor. Once activated,the CaMK, Ras/ERK, and SAPK pathways lead to thephosphorylation and activation of transcription fac-tors in the nucleus such as the cyclic AMP responseelement binding protein (CREB). Genes regulated byCREB or other transcription factor targets of theCaMK, Ras/ERK, and SAPK pathways could mediateimportant adaptive responses to changes in synapticactivity such as changes in synaptic strength and theregulation of neuronal survival and death. J. Neuro-sci. Res. 58:88–95, 1999.r 1999 Wiley-Liss, Inc.

Key words: calcium; neurotrophins; CaMKs; Ras/ERK;JNK/p38

INTRODUCTIONFor the nervous system to learn and adapt, it must

transform brief synaptic impulses into specific intracellu-lar signals that culminate in lasting changes in itsstructure and function. The intracellular signaling path-ways that mediate synaptic plasticity are not well under-stood. However, results from a cellular model of learningand memory, known as long-term potentiation (LTP),have suggested that the influx of extracellular Ca21 into

the postsynaptic neuron is a critical signal that initiatessome forms of synaptic plasticity (Cain, 1997; Malenka etal., 1997). Two major pathways by which Ca21 flows intothe postsynaptic neuron are through theN-methly-D-aspartate (NMDA) subtype of glutamate receptor and theL-type voltage-activated calcium channel (Nicoll andMalenka, 1995; Christie et al., 1997). More recently,certain growth factors, known as neurotrophins, havebeen shown to regulate synaptic function potently (Kangand Schuman, 1995). Neurotrophins signal by bindingand activating receptors that contain intrinsic tyrosinekinase activity, whereas calcium influx directly activatesmultiple kinase cascades. These calcium- and neuro-trophin-activated intracellular signal transduction path-ways probably mediate neuronal adaptation in part bymodifying existing proteins and by activating transcrip-tion factors that regulate new gene expression (Stantonand Sarvey, 1984).

A common downstream target of some of thesignaling pathways activated by calcium influx and byneurotrophins is the cyclic AMP response element bind-ing protein (CREB) family of transcription factors (Shenget al., 1990, 1991; Ginty et al., 1994; Finkbeiner et al.,1997). The best known family member, CREB, must bephosphorylated on serine 133 for CREB to activatetranscription (Gonzales and Montminy, 1989). CREB-dependent gene transcription has been shown to beimportant for certain forms of synaptic plasticity, from

Contract grant sponsor: NIDS; Contract grant number: KO8NSO1817;Contraact grant sponsor: Mental Retardation Research Center; Con-tract grant number: NIHP30-HD18655.

Jessica Curtis and Steven Finkbeiner’s current address: The GladstoneInstitute for Neurological Disease and the Departments of Physiologyand Neurology, University of California, San Francisco, CA 94143.

*Correspondence to: Steven Finkbeiner, Department of Physiology andNeurology, University of California, San Francisco, CA 94143.E-mail: sfinkbeiner@gladstone.ucsf.edu

Received 11 May 1999; Accepted 20 May 1999

Journal of Neuroscience Research 58:88–95 (1999)

r 1999 Wiley-Liss, Inc.

Drosphilato mammals (Bourtchuladze et al., 1994; Yin etal., 1994). Thus, an understanding of the pathways bywhich synaptic signals lead to CREB phosphorylationcould show highly conserved mechanisms of synapticplasticity.

CREB can be phosphorylated on serine 133 bymultiple kinases including protein kinase C (PKC),protein kinase A (PKA), cGMP-dependent protein kinase,calcium/calmodulin-dependent kinase I(CamKI), CamKII,CamKIV, pp90 Rsk1–3, and p38 mitogen-activated pro-tein kinase-activated protein kinase-2 (MAPKAP ki-nase-2; Gonzales and Montminy, 1989; Sheng et al.,1991; Sun et al., 1996; Xing et al., 1996; Pende et al.,1998; Tang et al., 1998; Tan et al., 1998; Xing et al.,1998). In this review, we describe the signal transductionpathways that have been reported to lead to CREBphosphorylation in response to neuronal activity gener-ally and in response to neurotrophin stimulation andcalcium influx more specifically. We focus on the CaMKpathways, the Ras/ERK pathway, and SAPK pathways,which may couple synaptic signals such as Ca21 influxand the activation of neurotrophin receptors to thephosphorylation of CREB within the nucleus.

CA21/CALMODULIN-DEPENDENT KINASEPATHWAYS

Several laboratories have shown that one group ofkinases that couples Ca21 influx and synaptic activity toCREB phosphorylation comprises the Ca21/calmodulin-dependent protein kinases, CaMK I, CaMK II, and CaMKIV (Sheng et al., 1991; Deisseroth et al., 1998). All threeenzymes bind to Ca21/calmodulin, and each contains acatalytic domain and an autoregulatory domain; however,CaMK II differs from CaMK I and CaMK IV in manyrespects (Hanson and Schulman, 1992; Enslen et al.,1994; Braun and Schulman, 1995). CaMK II is a holoen-zyme made up of 8–12 homologous peptides calleda orb subunits, and its structure by electron microscopyresembles a flower with petals (Kanaseki et al., 1991;Colbran and Soderline, 1992). When Ca21/calmodulinbinds a subunit of CaMK II, calmodulin displaces theautoregulatory domain of the subunit, enabling it tobecome phosphorylated by the catalytic domain of aneighboring subunit (Tokumitsu et al., 1997). Oncephosphorylated, the subunit can remain active for aprolonged period, even in the absence of a sustainedelevation in Ca21 (Miller et al., 1988). By contrast,CaMK I and CaMK IV exist as monomeric proteins (Parkand Soderling, 1995). CaMK I and CaMK IV bindCa21/calmodulin, but these enzymes are most effectivelyactivated through phosphorylation by Ca21/calmodulin-dependent kinase kinase (CaMKK; Sugita et al., 1994;

Park and Soderling, 1995; Anderson et al., 1998). Thesubcellular distribution of CaMK I, CaMK II, and CaMKIV may also differ (Picciotto et al., 1995). CaMK I andCaMK II are reportedly enriched in neuronal processesand synapses, whereas CaMK IV is relatively enriched inthe nucleus (Picciotto et al., 1995; Finkbeiner et al., 1997;Deisseroth et al., 1998). Subcellular localization is impor-tant because any kinase that phosphorylates CREB inresponse to synaptic activation must reach the nucleuswhere CREB is localized. Although all three kinases arecapable of phosphorylating serine 133 on CREB, transfec-tion experiments in heterologous cells have suggestedthat CaMK I and CaMK IV are more effective activatorsof CREB-dependent transcription than CaMK II (Mat-tews et al., 1994; Enslen et al., 1995; Sun et al., 1996).One possible reason for the difference is that CaMK IIcan also phosphorylate CREB on a neighboring serineresidue (ser-142) that has been shown to inhibit CREBfunction (Mattews et al., 1994; Sun et al., 1996).

The role of CaMK II in CREB-mediated genetranscription may be unclear, but its role in synapticplasticity is better understood. Because the expression ofCaMK II is extremely high in the postsynaptic densityand it is a target of Ca21 influx as a result of NMDAreceptor activation, it is thought to play an important rolein LTP (Kennedy et al., 1983). Evidence supporting thishypothesis came from mice deficient for the gene thatencodes the enzymea-CaMKII (Silva et al., 1992).Although these mice developed normally, they showeddefects in LTP and long-term depression (LTD). Further-more, overexpression of a constitutively active form ofCaMK II in neurons resulted in enhanced synaptictransmission and prevented additional potentiation by anLTP-inducing protocol (Pettit et al., 1994). CaMK II issuited to transduce synaptic activity into graded biochemi-cal responses because it is activated in response to strongsynaptic stimulation and is capable of decoding thefrequency of Ca21 influx transients into distinct amountsof kinase activity (Dosemeci and Albers, 1996; DeKoninckand Schulman, 1998). Together, these observations pro-vide evidence suggesting that CaMK II activity maycorrelate with synaptic activation and CaMK II may acton synaptic targets to regulate plasticity.

Although there is good evidence that CaMK II actsat the synapse, CaMK IV may be more important forlate-phase CREB-dependent synaptic plasticity. Linden etal. (1998) showed that dominant-interfering forms ofCREB or CaMK IV selectively inhibit the late-phase ofLTD within cultured Purkinje neurons in response toglutamate stimulation and Purkinje neuron depolariza-tion. By contrast, blockade of PKA had no effect.Deisseroth et al. (1998) showed that a rapid, activity-dependent nuclear translocation of calmodulin is impor-

Signaling From Synapse to Nucleus 89

tant for CREB phosphorylation and that camodulin andCamKK translocate to the nucleus and activate CaMK IV.In addition, CaMKK has been implicated in mediatingCa21-dependent cell survival because it can directlyphosphorylate and activate the serine-threonine kinase,Akt (Yano et al., 1998). Akt promotes the phosphoryla-tion and inhibition of the pro-apoptotic Bcl familymember, BAD; the phosphorylation and inhibition ofcasapse 9; and the phosphorylation and inhibition of theforkhead family of transcription factors that regulate theexpression of death genes (Burgering and Coffer, 1995;Datta et al., 1997; Cardone et al., 1998; Brunet et al.,1999).

THE RAS/ERK PATHWAYIn addition to the CaMKs, the Ras/ERK pathway

can regulate CREB-mediated gene transcription. Bothcalcium influx and the activation of TrkB, the receptortyrosine kinase through which brain-derived neurotrophicfactor (BDNF) signals, can regulate the Ras/ERK path-way. Activation of this pathway requires the conversionof the small guanine nucleotide binding protein Ras froma GDP-bound state into its GTP-bound state. Ras-GTPthen leads to the phosphorylation and activation of theserine/threonine kinase Raf, recruiting Raf from thecytoplasm to the plasma membrane. Raf phosphorylatesthe mitogen-activated protein kinase (MEK), allowingMEK to phosphorylate the extracellular signal-relatedkinases 1 and 2 (ERKs 1 and 2). Erk phosphorylates pp90ribosomal S6 kinases 1–3 (Rsks 1–3), which translocateto the nucleus and phosphorylate transcription factorssuch as CREB (reviewed by Heumann 1994; Campbell etal., 1998).

Several studies have demonstrated that the Ras/Erkpathway is important for cellular functions such assurvival, plasticity, and long-term memory (Orban et al.,1999). For example, the Ras/ERK pathway is activatedduring the induction of LTP and pharmacological inhibi-tion of the Ras/ERK pathway greatly attenuates LTPinduction (English and Sweatt, 1997). Bailey et al. (1997)showed that the Ras/Erk pathway plays a role in theactivity-dependent regulation of the structure and func-tion of synapses in the sea slug,Aplysia. Evidently, theAplysiahomologue of the neural cell adhesion molecule(Ap-CAM) forms homotypic structures that span thesynaptic cleft and must be disrupted for the synapse toundergo long-lasting stuctural and functional changes.ERK activation disrupts Ap-CAM connections by phos-phorylating the cytoplasmic tail of Ap-CAM, therebypromoting its ubiquitination and degradation. Anotherpiece of evidence, which suggests that this pathway maybe required for lasting changes of active synapses, comesfrom mice in which thegene for guanine nucleotide

exchange factor 1(Ras-GRF 1) has been disrupted (Bram-billa et al., 1997). These animals had severely impairedLTP in the amygdala and a corresponding deficit inlong-term memory for aversive events. Ras-GRF is a Rasexchange factor that can activate Ras in response to Ca21

elevations (Farnsworth et al., 1995).In a variety of cell types, the Ras/Erk pathway is a

major convergence point for a wide range of signals withpleiotropic effects on cellular function, in particular onsurvival. Previous research has demonstrated that Ras isnecessary for neurotrophin-induced neuronal survivaland that the introduction of Ras into embryonic neuronsmimics growth factor treatment (Borasio et al., 1989).Ras may also indirectly mediate Ca21-induced survivalsignals. Extracellular application of anti-BDNF antibod-ies blocked depolarization-induced Ca21-dependent sur-vival of cortical neurons (Ghosh et al., 1994). This findingsuggests that depolarization promotes survival by induc-ing BDNF synthesis and secretion, TrkB receptor activa-tion, and possibly through Ras–dependent survival mecha-nisms. Ras may promote survival by multiple mechanisms.Some studies have suggested that Ras can activate theprosurvival kinase Akt (Khwaja et al., 1997). The Ras/ERK cascade member MEK has also been shown tophosphorylate directly and inhibit the pro-apoptotic mol-ecule BAD (Downward, 1998; Scheid and Duronio,1998).

JNK/P38 PATHWAYS

In addition to the CaMK and Ras/ERK pathways,several studies have implicated members of a subgroup ofMAPKs, c-Jun N-terminal kinases (JNKs) and p38, in theprocesses of CREB phosphorylation, synaptic plasticity,and neuronal survival. For example, p38 activation hasbeen shown to trigger CREB (ser-133) phosphorylationvia MAPKAPK-2 and JNK is known to phosphorylateand activate the CREB family member ATF-2 (Living-stone et al., 1995; Xing et al., 1998). JNKs and p38 aredownstream members of MAPK cascades that are orga-nized similarly to the Ras/ERK pathway (reviewed byHerdegen and Leah, 1998). For example, both p38 andJNK1 can be phosphorylated and activated by stress-activated protein kinase 1 or ERK kinase 1 (SEX1),which in turn can be activated by mitogen and extracellu-lar regulated kinase kinase (MEKK1). The small GTP-binding protein Ras can activate MEKK1. However, thereare many other specific signal transduction cascades thatculminate in the activation of JNKs and p38, and thereader is referred to several excellent reviews for addi-tional examples (Waskiewicz and Cooper, 1995; Cohen,1997). The upstream mechanisms by which growthfactors and Ca21 influx regulate the stress-response

90 Curtis and Finkbeiner

pathways remain poorly understood (Ko et al., 1998).However, Enslen et al. (1996) recently elucidated onepathway by which Ca21 influx activates the JNKs. In thepheochromocytoma cell line PC12, Ca21 influx inducesthe activation of p38 and JNK-1 by a CaMKK- andCaMK IV-dependent mechanism (Fig. 1).

JNKs and p38 are best known for their roles inregulating cell death. In PC12 cells, apoptosis induced byneurotrophin withdrawal was blocked by interfering withJNK or p38 function. Conversely, overexpression ofupstream activators of JNKs or p38 promoted apoptosisin PC12 cells and sympathetic neurons (Xia et al., 1995;Eilers et al., 1998). Additional support for a specific rolefor JNKs in the regulation of neuronal survival comesfrom mice in which the gene for the neuron-specific JNKisoform, JNK-3, has been disrupted. Compared withneurons from wild-type mice, neurons from mice thatlack JNK-3 were resistant to apoptosis induced by theneurotoxin kainic acid (Yang et al., 1997). Le-Niculescuet al. (1999) discovered a mechanism by which JNKactivation regulates neuronal survival. They found thatthe transcription of the gene that encodes Fas ligand isregulated by the JNK substrate c-Jun. Thus, JNK activa-tion induced increases in Fas ligand expression. Fasligand promotes apoptosis by binding the cell surface Fasreceptor, thereby activating the highly conserved pro-apoptotic caspase cascades.

However, several studies have also shown that therelationship between JNK activation and apoptosis iscomplex and likely depends on additional factors (Herde-gen et al., 1997). In one study, apoptosis of sympatheticneurons induced by neurotrophin withdrawal led to asustained activation of JNK, but suppression of JNKactivity was not sufficient to rescue all of the neurons(Virdee et al., 1997). In another study, neuronal insultssuch as ischemia or axotomy induced sustained JNKactivation as measured by c-Jun phosphorylation indepen-dently of whether the neurons subsequently survived ordegenerated (Herdegen and Leah, 1998). Indeed, onestudy found that mice that are restrained or are exploringa novel environment show a 3–15-fold increase in JNKactivity within neurons from certain brain regions (Xu etal., 1997). The fact that JNK activity can be induced by atask that does not lead to neurodegeneration suggests thatJNK may play a role in mediating normal physiologicfunctions in response to neuronal activity. The fact thatJNK3 2/2 mice are resistant to seizures has even raisedthe intriguing possibility that the JNK pathway may playa role in synaptic transmission or some forms of synapticplasticity. The small GTPases, Rac and cdc42, areupstream activators of the JNK signaling cascade and arealso able to promote dendritic growth and remodeling(Minden et al., 1995; Threadgill et al., 1997).

SIGNAL SPECIFICITYFrom the limited discussion above, it is evident that

synaptically generated signals are capable of triggeringthe activation of multiple distinct kinase cascades. In thisreview, we have focused on the CaMK, Ras/ERK, andSAPK signal transduction pathways because of theirpotential roles in coupling synaptic signals to CREBphosphorylation. However, synaptic activity stimulatesother cascades such as the NF-kB and nitric oxidepathways that also may be important for regulatingplasticity, death, or survival (Ohki et al., 1995; Arancio etal., 1996; Beg and Baltimore, 1996). In some instances,the CaMK, Ras/ERK, SAPK, NF-kB, and nitric oxidesignaling pathways appear to promote plasticity andneuronal survival, but under different circumstances activa-

Fig. 1. Possible signal transduction pathways by which cal-cium influx and neurotrophin withdrawal could activate genetranscription via phosphorylation of c-Jun, ATF-2, and CREB.Ca21, calcium; CaMKK, CaMK II and CaMK IV, calcium/calmodulin protein kinase kinase, II, and IV; JNK, c-JunN-terminal kinase; p38, p38 kinase; SEK 1, stress-activatedprotein kinase 1; MEK, mitogen-activated protein kinase; Rsk,pp90 ribosomal S6 kinase; MAPKAPK-2, mitogen-activatedprotein kinase-activated protein kinase-2.

Signaling From Synapse to Nucleus 91

tion of some of the same pathways appears to promotedeath (Trump and Berezensky, 1992). The observationthat these same signal transduction pathways may be ableto mediate distinct biological responses raises the puz-zling question of how synaptically generated signalsencode information to generate specific long-lastingbiological responses.

One widely held idea is that spatial or temporalfeatures of the signal that is generated within a neuron inresponse to synaptic activity determine the biologicalresponse. In vitro, synaptic stimulation protocols thatstrengthen synaptic transmission often generate high-frequency Ca21 transients that accumulate to producerelatively high elevations of intracellular Ca21. By con-trast, protocols that weaken synaptic connections oftenevoke low-frequency Ca21 transients thatmoderately el-evate intracellular Ca21. High-amplitude, high-frequencyCa21 transients very effectively activate CaMK II, akinase associated with synaptic potentiation. Calcineurin(CaN), a phosphatase implicated in LTD, is relativelywell activated by more moderate increases in Ca21

(Mulkey et al., 1994; Dosemeci and Albers, 1996). Thus,patterns of postsynaptic Ca21 transients could directqualitative and quantitative changes in synaptic strengthin part through the control of the balance of CaMK II andCaN activity. Temporal features of Ca21 transients havealso been shown to determine temporal patterns of CREBphosphorylation and the repertoire of Ca21-activatedgenes that are transcribed (Bito et al., 1996; Dolmetsch etal., 1998; Liu and Graybiel, 1998). For the Ras/ERKpathway inAplysia, moderate synaptic stimulation acti-vated ERK molecules close to the stimulated synapse andincreased transiently synaptic strength (Martin et al.,1997). By contrast, sustained stimulation of the Ras/ERKpathway led to the activation of ERK molecules at thesynapse and a translocation of activated ERK into thenucleus. Translocation of ERK into the nucleus correlatedwith long-lasting synaptic potentiation and the formationof new synaptic contacts. Distinct temporal profiles ofSAPK pathway activation have also been suggested tocontribute to specific adaptive responses (Herdegen andLeah, 1998).

Another important determinant of specificity insignal transduction is the spatial organization of the signalitself and the kinase cascades that the signal can activate.Deisseroth et al. (1996) exploited the biophysical differ-ences between two intracellular Ca21 chelators to showthat Ca21 ions flowing in at the synapse lead to CREBphosphorylation in the nucleus by coupling to signaltransduction cascades at sites very near (,1 µm) theplasma membrane. The group subsequently found thatCREB phosphorylation in the nucleus was highly corre-lated with the translocation of calmodulin from neuronalprocesses to the nucleus. Calmodulin couples Ca21 ions

to effector molecules such as CaMK II, CaMKK, andCaMK IV (Deisseroth et al., 1998). It is unknown how thesubcellular localization of calmodulin is controlled withinneurons, but work inDrosophilahas suggested that someproteins function as scaffolds to localize effector mol-ecules close to the source of Ca21 signals (Scott andZuker, 1998). Tsunoda et al. (1997) showed that thescaffolding proteinInaD is critical for signal transductionin the Drosophilaphotoreceptor.InaD contains multipleprotein–protein interaction domains, called PDZ domains(postsynaptic density protein, disc large, and Zo-1), thatenable it to colocalize important components of the signaltransduction pathway such as PKC and phospholipase Cg(PLCg). If these PDZ domains are mutated, PKC andPLCg no longer remain colocalized and the photorecep-tor fails to transduce a signal, even though each of thecomponents of the signaling pathway is expressed atnormal levels. Scaffolding proteins are also very impor-tant for transmitting specific signals at the membrane tothe Ras/MAPK cascade (Zanke et al., 1996). In addition,enzyme–substrate interactions at each level of Ras/MAPK cascades are critical for determining whethercross-talk occurs between Ras/MAPK cascades once anupstream member of a pathway has become activated.

Although the pathway that couples Ca21 to CREBmay begin at points very near the site of Ca21 entry, Ca21

signals within the nucleus may also be important forCa21-dependent gene transcription. Hardingham et al.(1997) showed that chelators that are restricted to thenucleus block Ca21-dependent transcription, suggestingthat Ca21 signals within the nucleus may regulate addi-tional components of the basal transcriptional machinery.A spatial requirement for simultaneous or nearly simulta-neous Ca21 signals at the synapse and within the nucleuscould form the basis for achieving specific adaptiveresponses by limiting the transcription of certain genes tostimuli that meet these criteria (Finkbeiner and Green-berg, 1997).

DISCUSSIONNeurons have the remarkable ability to undergo

changes in the strengths and patterns of their connectionswith other neurons in response to changes in synapticstimulation. Several recent studies have suggested thatsynaptic stimulation can initiate changes in synapticstrength through the covalent modification of existingproteins but that long-lasting changes in synaptic struc-ture and function depend in part on gene transcription andnew protein synthesis. This review has focused on themechanisms that link synaptic signals such as the influxof extracellular Ca21 or neurotrophin stimulation to thephosphorylation and activation of a particular transcrip-tion factor within the nucleus, i.e., CREB. Three signal-

92 Curtis and Finkbeiner

ing pathways have been shown to be activated withsynaptic stimulation and to culminate in CREB phosphor-ylation including members of the CaMK (CaMK II andCaMK IV), the Ras/ERK, and the SAPK pathways (e.g.,p38, JNK). Considerable evidence has linked CaMK andRas/ERK pathways to synaptic plasticity and neuronalsurvival, whereas activation of the SAPK pathways ismore commonly associated with neuronal death. Never-theless, the relationship between the activation of thesepathways and a particular biological response is complex,and there is considerable cross-talk between these cas-cades (Fig. 2). Thus, an important unanswered question ishow different synaptic signals can activate the samesignal transduction cascades but elicit distinct biologicalresponses. Although our understanding of signaling speci-ficity in the nervous system remains rudimentary, itappears that temporal and spatial features of the synapticsignal itself may be critical for determining which signaltransduction pathways are activated and whether the

activation of specific pathways leads to changes in geneexpression. In addition, the enzyme–substrate interac-tions of downstream pathway components and the overallspatial organization of the pathways likely govern thelevel of cross-talk between pathways.

ACKNOWLEDGMENTSSteven Finkbeiner was supported by NIDS grant

KO8NSO1817.

REFERENCES

Anderson KA, Means RL, Huang QH, Kemp BE, Goldstein EG,Selbert MA, Edelman AM, Fremeau RT, Means AR. 1998.Components of a calmodulin-dependent protein kinase cascade.Molecular cloning, functional characterization and cellularlocalization of Ca21/calmodulin-dependent protein kinase ki-nase beta. J Biol Chem 273:31880–31889.

Arancio O, Lev-Ram V, Tsien RY, Kandel ER, Hawkins RD. 1996.Nitric oxide acts as a second messenger during long-termpotentiation in cultured hippocampal neurons. J Physiol [Paris]90:321–322.

Bailey CH, Kaang B-K, Chen M, Martin KC, Lim C-S, Casadio A,Kandel ER. 1997. MAP kinase translocates into the nucleus ofthe presynaptic cell and is required for the long-term facilitationin Aplysia. Neuron 18:899–912.

Beg AA, Baltimore D. 1996. An essential role for NF-kB in preventingTNF-a induced cell death. Science 274:782–784.

Bito H, Diesseroth K, Tsien RW. 1996. CREB phosphorylation anddephosphorylation: a Ca(21)- and stimulus duration-dependentswitch for hippocampal gene expression. Cell 87:1203–1214.

Borasio GD, John J, Wittinghofer A, Barde Y, Sendtner M, HeumannR. 1989. Ras p21 protein promotes survival and fiber growth ofcultured embryonic neurons. Neuron 2:1087–1096.

Bourtchuladze R, Frenguelli B, Blendy J, Cioffi D, Schultz G, Silva AJ.1994. Deficient long-term memory in mice with a targetedmutation of the cAMP-responsive element-binding protein. Cell79:59–68.

Brambilla R, Gnesutta N, Minichiello L, White G, Roylance AJ,Herron CE, Ramsey M, Wolfer DP, Cestari V, Rossi-Arnaud C,Grant SGN, Chapman PF, Lipp L-H, Sturani E, Klein R. 1997. Arole for the Ras signalling pathway in synaptic transmission andlong-term memory. Nature 390:281–286.

Braun AP, Schulman H. 1995. The multifunctional calcium/calmodulin-dependent kinase: from form to function. Annu Rev Physiol57:417–445.

Brunet A, Bonni A, Zigmond MJ, Lin MZ, Juo P, Hu LS, Anderson MJ,Arden KC, Blenis J, Greenberg ME. 1999. Akt promotes cellsurvival by phosphorylating and inhibiting a forkhead transcrip-tion factor. Cell 96:857–868.

Burgering BMT, Coffer PJ. 1995.Protein kinase B (c-akt) in phosphatidyl-inositol-3-OH kinase signal transduction. Nature 376:599–602.

Cain DP. 1997. LTP, NMDA, genes and learning. Curr Opin Neurobiol7:235–242.

Campbell SL, Khosravi-Far R, Rossman KL, Clark GJ, Der CJ. 1998.Increasing complexity of Ras signaling. Oncogene 17:1395–1413.

Cardone MH, Roy N, Stennicke HR, Salvesen GS, Franke TF,Stanbridge E, Frisch S, Reed JC. 1998. Regulation of cell deathprotease caspase-9 byphosphorylation. Science 282:1318–1321.

Christie BR, Schexnayder LK, Johnson D. 1997. Contribution ofvoltage-gated Ca21 channels to homosynaptic long-term depres-sion in the CA1 region in vitro. J Neurophysiol 77:1651–1655.

Fig. 2. Biological responses. Calcium influx and neurotrophinstimulation and/or withdrawal can lead to the activation of thesignal transduction pathwaysthat could initiate death or survivalresponses. Ca21, calcium; PYK2, protein tyrosine kinase 2; Racand Cdc42, small GTP-binding proteins; CaMKK and CaMKIV, calcium/calmodulin protein kinase kinase and IV.

Signaling From Synapse to Nucleus 93

Cohen P. 1997. The search for physiological substrates of MAP andSAP kinases in mammalian cells. Trends Cell Biol 7:353–360.

Colbran RJ, Soderling TR. 1992. Calcium/calmodulin-dependent pro-tein kinase II. Curr Topics Cell Regul 31:181–221.

De Koninck P, Schulman P. 1998. Sensitivity of CaM kinase II to thefrequency of Ca21 oscillations. Science 279:227–230.

Deisseroth K, Bito H, Tsien RW. 1996. Signaling from synapse tonucleus: postsynaptic CREB phosphorylation during multipleforms of hippocampal synaptic plasticity. Neuron 16:89–101.

Deisseroth K, Heist EK, Tsien RW. 1998. Translocation of calmodulinto the nucleus supports CREB phosphorylation in hippocampalneurons. Nature 392:198–202.

Dolmetsch RE, Xu K, Lewis RS. 1998. Calcium oscillations increasethe efficiency and specificity of gene expression. Nature 392:933–936.

Dosemeci A, Albers RW. 1996. A mechanism for synaptic frequencydetection through autophosphorylation of CaM kinase II. Bio-phys J 70:2493–2501.

Downward J. 1998. Ras signaling and apoptosis. Curr Opin Genet Dev8:49–54.

Edelman AM, Mitchelhill KI, Selbert MA, Anderson KA, Hook SS,Stapleton D, Goldstein EG, Means AR, Kemp BE. 1996.Multiple Ca21/calmodulin-dependent protein kinase kinasesfrom rat brain. Purification, regulation by Ca21-calmodulin, andpartial amino acid sequence. J Biol Chem 271:10806–10810.

Eilers A, Whitfield J, Babij C, Rubin LL, Ham J. 1998. Role of the Junkinase pathway in the regulation of c-Jun expression andapoptosis in sympathetic neurons. J Neurosci 18:1713–1724.

English JD, Sweatt JD. 1997. A requirement for the mitogen-activatedprotein kinase cascade in hippocampal long term potentiation. JBiol Chem 272:19103–19106.

Enslen H, Sun P, Brichey D, Soderling SH, Klamo E, Soderling TR.1994. Characterization of Ca21/calmodulin-dependent proteinkinase IV. Role in transcriptional regulation. J Biol Chem269:15520–15527.

Enslen H, Tokumitsu H, Stork P, Davis R, Soderling TR. 1996.Regulation of mitogen-activated protein kinases by a calcium/calmodulin-dependent protein kinase cascade. Proc Natl AcadSci USA 93:10803–10808.

Farnsworth CL, Freshney NW, Rosen LB, Ghosh A, Greenberg ME,Feig LA. 1995. Calcium activation of Ras mediated by neuronalexchange factor Ras-GRF. Nature 376:524–527.

Finkbeiner S, Greenberg ME. 1997. Spatial features of calcium-regulated gene expression. Bioessays 19:657–660.

Finkbeiner S, Tavazoie SF, Maloratsky A, Jacobs KM, Harris KM,Greenberg ME. 1997. CREB: a major mediator of neuronalneurotrophin responses. Neuron 19:1031–1047.

Ghosh A, Carnahan J, Greenberg ME. 1994. Requirement for BDNF inactivity-dependent survival of cortical neurons. Science 263:1618–1623.

Ginty DD, Bonni A, Greenberg ME. 1994. Nerve growth factoractivates a Ras-dependent protein kinase that stimulates c-fostranscription via phosphorylation of CREB. Cell 77:713–725.

Gonzales G, Montminy M. 1989. Cyclic AMP stimulates somatostatingene transcription by phosphorylation of CREB at serine 133.Cell 59:675–680.

Hanson PI, Schulman H. 1992. Neuronal Ca21/calmodulin-dependentprotein synthesis. Annu Rev Biochem 61:559–601.

Hardingham GE, Chawla S, Johnson CM, Bading H. 1997. Distinctfunctions of nuclear and cytoplasmic calcium in the control ofgene expression. Nature 385:260–265.

Herdegen T, Leah JD. 1998. Inducible and constitutive transcriptionfactors in the mammalian nervous system: control of geneexpression by Jun, Fos and Krox, and CREB/ATF proteins.Brain Res Brain Res Rev 28:370–390.

Herdegen T, Skene P, Ba¨hr M. 1997. The c-jun transcription factor—bipotential meditor of neuronal death, survival and regenera-tion. Trends Neurosci 20:227–231.

Heumann R. 1994. Neurotrophin signalling. Curr Opin Neurobiol4:668–679.

Kanaseki T, Ikeuchi Y, Sugiura H, Yamauchi T. 1991. Structuralfeatures of Ca21/calmodulin-dependent protein kinase II re-vealed by electron microscopy. J Cell Biol 115:1049–1060.

Kang H, Schuman EM. 1995. Long-lasting neurotrophin-inducedenhancement of synaptic transmission in the adult hippocam-pus. Science 267:1658–1662.

Kennedy MB, Bennet MK, Erondu NE. 1983. Biochemical andImmunochemical evidence that the ‘‘major postsynaptic densityprotein’’ is a subunit of a calcium-dependent protein kinase.Proc Natl Acad Sci USA 80:7357–7361.

Khwaja A, Rodriguez-Viciana P, Wennesrom S, Warne PH, DownwardJ. 1997. Matrix adhesion and Ras transformation both activate aphosphoinositide 3-OH kinase and protein kinase B/Akt cellularsurvival pathway. EMBO J 16:2783–2793.

Ko H, Park K, Kim H, Han P, Kim Y, Gwag B, Choi E. 1998.Ca21-mediated activation of c-Jun N-terminal kinase and nuclearfactor kB by NMDA in cortical cell cultures. J Neurochem71:1390–1395.

Le-Niculescu H, Bonfoco E, Kasuya Y, Claret FX, Green DR, Karin M.1999. Withdrawal of survival factors in activation of the JNKpathway in neuronal cells leading to Fas ligand induction andcell death. Mol Cell Biol 19:751–763.

Linden DJ, Ahn S, Ginty DD. 1998. A late phase of cerebellarlong-term depression (LTD) requires activation of CREB andCaMK IV. Soc Neurosci Abstr 24:9.8

Liu FC, Graybiel AM. 1998. Dopamine and calcium signal interactionsin the developing striatum: control of kinetics of CREBphosphoryaltion. Adv Pharmacol 42:682–686.

Livingstone C, Patel G, Jones N. 1995. ATF-2 contains a phosphoryla-tion-dependent transcriptional activation domain. EMBO J14:1785–1797.

Malenka RC, Kauer JA, Zucker RS, Nicoll RA. 1997. Post-synapticcalcium is suffcient for potentiation of synaptic transmission.Science 242:81–84.

Martin KC, Michael D, Rose JC, Barad M, Casadio A, Zhu H, KandelER. 1997. MAP kinase translocates into the presynaptic cell andis required for long-term facilitation inAplysia. Neuron 18:899–912.

Mattews RP, Guthrie CR, Wades LM, Zhoa Z, Means AR, McknightGS. 1994. Calcium/calmodulin-dependent protein kinase type IIand IV differentially regulate CREB-dependent gene expres-sion. Mol Cell Biol 140:6107–6116.

Miller SG, Patton BL, Kennedy MB. 1988. Sequences of autophos-phorylation sites in neuronal type II CaM kinase that controlCa21-independent activity. Neuron 1:593–604.

Minden A, Lin A, Claret FX, Abo A, Karin M. 1995. Selectiveactivation of the JNK signaling cascade and c-Jun transcrip-tional activity by the small GTPases Rac and Cdc42Hs. Cell81:1147–1157.

Mulkey R, Endo S, Shenolikar S, Malenka RC. 1994. Involvement of acalcineurin/inhibitor-1 phosphatase cascade in hippocampallong-term depression. Nature 369:486–389.

Nicoll RA, Malenka RC. 1995. Contrasting properties of two forms oflong-term potentiation in the hippocampus. Nature 377:115–118.

Ohki K, Yoshida K, Hagiwara M, Harada T, Takamura M, Ohasi T,Matsuda H, Imaki J. 1995. Nitric oxide induces c-fos geneexpression via cyclic AMP response element binding protein(CREB) phosphorylation in rat retinal pigment epithelium.Brain Res 696:140–144.

94 Curtis and Finkbeiner

Orban PC, Chapman PF, Brambilla. 1999. Is the Ras-MAPK signallingpathway necessary for long-term memory formation? TrendsNeurosci 22:38–44.

Otani S, Connor JA. 1996. Rapid dendritic Ca21 influx is associatedwith induction of homosynaptic long-term depression in adultrat hippocampus. Eur J Pharmacol 318:R5–R6.

Park IK, Soderling TR. 1995. Activation of Ca21/calmodulin-dependent protein kinase (CaM-kinase) IV by CaM-kinasekinase in Jurkat T lymphocytes. J Biol Chem 270:19320–19324.

Pende M, Fisher TL, Simpson PB, Russel JT, Blenis J, Gallo V. 1998.Neurotransmitter and growth factor-induced cAMP responseelement binding phosphorylation in glial cell progenitors: roleof calcium ions, protein kinase C, and mitogen-activated proteinkinase/ribosomal S6 kinase pathway. J Neurosci 17:1291–1301.

Pettit DL, Perlman R, Malinow R. 1994. Potentiated transmission andprevention of further LTP by increased CaMKII activity inpostsynaptic hippocampal slice neurons. Science 266:1881–1885.

Piccitto MR, Zoli M, Bertuzzi G, Nairn AC. 1995. Immunochemicallocalization of calcium/calmodulin-dependent protein kinase I.Synapse 20:75–84.

Scheid MP, Duronio V. 1998. Dissociation of cytokine-induced phos-phorylation of BAD and activation of PKB/akt: involvement ofMEK upstream of BAD phosphorylation. Proc Natl Acad SciUSA 95:7439–7444.

Scott K, Zuker CS. 1998. Assembly of theDrosophilaphototransduc-tion cascade into signalling complex shapes elementary re-sponses. Nature 395:805–808.

Sheng M, McFadden G, Greenberg ME. 1990. Membrane depolariza-tion and calcium induce c-fos transcription via phosphorylationof transcription CREB. Neuron 4:571–582.

Sheng M, Thompson MA, Greenberg ME. 1991. CREB: a Ca21-regulated transcription factor phosphorylated by calmodulin-dependent kinases. Science 252:1427–1430.

Silva AJ, Stevens CF, Tonegawa S, Wang Y. 1992. Deficient hippocam-pal long-term potentiation ina-calcium-calmodulin kinase IImutant mice. Science 257:201–206.

Stanton PK, Sarvey JM. 1984. Blockade of long-term potentiation inrat hippocampal CA1 region by inhibitors of protein synthesis. JNeurosci 4:3080–3088.

Sugita R, Mochizuki H, Ito T, Yokokura H, Kobayashi R, Hidaka H.1994. Ca21/calmodulin-dependent protein kinase kinase cas-cade. Biochem Biophys Res Commun 203:694–701.

Sun P, Lou L, Mauer RA. 1996. Regulation of activating transcriptionfactor-1 and the cAMP response element-binding protein byCa21/calmodulin-dependent protein kinase type I, II, and IV. JBiol Chem 271:3066–3073.

Tan Y, Rouse J, Zhang A, Cariati S, Cohen P, Comb MJ. 1998. FGF andstress regulate CREB and ATF-1 via a pathway involving p38MAP kinase and MAPKAP kinase-2. EMBO J 15:4629–4642.

Tang K, Wu H, Mahata SK, Mahata M, Gill BM, Parmer RJ, O’ConnerDT. 1998. Stimulus coupling to transcription versus secretion inpheochromocytoma cells: convergent and divergent signal trans-duction pathways and crucial roles for route of cytosolic

calcium entry and protein kinase C. J Clin Invest 100:1180–1192.

Threadgill R, Bobb K, Ghosh A. 1997. Regulation of dendritic growthand remodeling by Rho, Rac and Cdc42. Neuron 19:625–634.

Tokumitsu H, Wayman GA, Muramatsu M, Soderling TR. 1997.Calcium calmodulin-dependent protein kinase kinase: identifica-tion of regulatory domains. Biochemistry 36:12823–12727.

Trump BF, Berezensky IK. 1992. The role of cytosolic Ca21 in cellinjury, necrosis and apoptosis. Curr Opin Cell Biol 4:227–232.

Tsunoda S, Sierralta J, Sun Y, Suzuki E, Becker A, Socolich M, ZukerCS. 1997. A multivalent PDZ-domain protein assembles signal-ling complexes in a G-protein-coupled cascade. Nature 388:243–249.

Virdee K, Bannister AJ, Hunt SP, Tolkovsky AM. 1997. Comparisonbetween the timing of JNK activation, c-Jun phosphorylation,and onset of death commitment in sympathetic neurons. JNeurochem 69:550–561.

Waskiewicz AJ, Cooper JA. 1995. Mitogen and stress responsepathways: MAP kinase cascades and phosphatase regulation inmammals and yeast. Curr Opin Cell Biol 7:798–805.

Xia Z, Dickens M, Raingeaud J, Davis R, Greenberg ME. 1995.Opposing effects of ERK and JNK-p38 kinases on apoptosis.Science 270:1326–1331.

Xing J, Ginty D, Greenberg ME. 1996. Coupling of the Ras-MAPKpathway to gene activation by RSK2, a growth factor CREBkinase. Science 273:959–963.

Xing J, Kornhauser JM, Xia Z, Thiele EA, Greenberg ME. 1998. Nervegrowth factor activates extracellular signal-regulated kinase andp38 mitogen-activated protein kinase pathways to stimulateCREB serine 133 phosphorylation. Mol Cell Biol 18:1946–1955.

Xu X, Raber J, Yang D, Su B, Mucke L. 1997. Dynamic regulation ofc-Jun N-terminal kinase activity in mouse brain by environmen-tal stimuli. Proc Natl Acad Sci USA 94:12655–12660.

Yang D, Kuan C, Whitmarsh A, Rincoon M, Zheng T, Davis R, Rakic P,Flavell R. 1997. Absence of excitoxicity-induced apoptosis inthe hippocampus of mice lacking theJnk3 gene. Nature389:865–870.

Yang E, Zha J, Jockel J, Boise LH, Waksman G, Korsmeyer SJ. 1995.Bad, a heterodimeric partner for Bcl-XL and Bcl-2, displacesBax and promotes cell death. Cell 80:285–291.

Yano S, Tokumitsu H, Soderling TR. 1998. Calcium promotes cellsurvival through CaM-K kinase activation of the protein-kinase-B pathway. Nature 396:584–587.

Yin JC, Wallach JS, Del Vecchio M, Wilder EL, Zhou H, Quinn WG,Tully T. 1994. Induction of a dominant negative CREB trans-gene specifically blocks long-term memory inDrosophila. Cell79:49–58.

Zanke BW, Rubie EA, Winnett E, Chan J, Randall S, Parsons M,Boudreau K, McInnis M, Yan M, Templeton DJ, Woodgett JR.1996. Mammalian mitogen-activated protein kinase pathwaysare regulated through formation of specific kinase activator-complexes. J Biol Chem 271:29879–29881.

Signaling From Synapse to Nucleus 95