Oncogene (1999) 18, 2055 ± 2068
ã 1999 Stockton Press All rights reserved 0950 ± 9232/99 $12.00
http://www.stockton-press.co.uk/onc
Nerve growth factor induces survival and di€erentiation through two
distinct signaling cascades in PC12 cells
LJ Klesse1, KA Meyers1, CJ Marshall2 and LF Parada*,1
1
2
Center for Developmental Biology, University of Texas, Southwestern Medical Center, Dallas, Texas 75235-9133, USA; and
Chester Beatty Laboratories, 237 Fulham Road, London SW3 6JB, UK
Nerve growth factor induces di€erentiation and survival
of rat PC12 pheochromocytoma cells. The activation of
the erk cascade has been implicated in transducing the
multitude of signals induced by NGF. In order to
explore the role of this signaling cascade in NGF
mediated survival, di€erentiation and proliferation, we
generated recombinant adenoviruses which express the
intermediates of the erk cascade in their wild type,
dominant negative and constitutively activated forms.
We show that di€erentiation of PC12 cells requires
activity of the ras/erk pathway, whereas inhibition of
this pathway had no e€ect on survival or proliferation.
Constitutively active forms of ras, raf and mek induced
PC12 cell di€erentiation, while dominant interfering
forms inhibited di€erentiation. Survival of PC12 cells in
serum-free medium did not require activity of the ras/
erk pathway. Instead, PI3 Kinase signaling was
necessary for PC12 cell survival. Interestingly, constitutively activated versions of raf and mek were able to
promote survival, but again this was dependent on
activation of PI3 Kinase. Therefore, at least two
distinct signaling pathways are required in PC12 cells
for mediation of NGF functions.
Keywords: PC12 cells; nerve growth factor; map kinase;
phosphatidylinositol 3-kinase
Introduction
The Neurotrophin family of soluble peptide factors is
required for the correct development and differentiation of the embryonic vertebrate nervous system, as
reviewed by (Kaplan and Miller, 1997; Segal and
Greenberg, 1996; Snider and Silos-Santiago, 1996). The
neurotrophins, which include Nerve Growth Factor,
Brain Derived Neurotrophic Factor, Neurotrophin-3
and Neurotrophin 4/5, bind to the Trk family of
receptor tyrosine kinases (RTKs). This binding
activates a variety of intracellular signaling molecules
reviewed by (Kaplan and Miller, 1997; Segal and
Greenberg, 1996; Snider and Silos-Santiago, 1996). As
a result of neurotrophin induced activation of signaling
intermediates, sensory and sympathetic neurons differentiate appropriately and evade programmed cell
death. Thus, neurotrophin induced signaling is
*Correspondence: LF Parada
Received 22 May 1998; revised 27 October 1998; accepted 27 October
1998
necessary for neuron survival and di€erentiation and
the identi®cation of the speci®c molecules involved in
vivo has attracted considerable attention.
Due to the relative diculty of studying signaling in
neurons, neurotrophin signaling has been primarily
studied using PC12 cells as a model system. PC12 cells,
a rat pheochromocytoma tumor cell line, respond to
NGF exposure by di€erentiating to resemble sympathetic neurons (Greene and Tischler, 1976). Upon
NGF exposure, PC12 cells cease division, extend
neurites, become electrically excitable and express
neuronal markers (Greene and Tischler, 1976; Tischler
and Greene, 1975). NGF has been shown to bind to its
high anity receptor, Trk A, on the surface of PC12
cells (Kaplan et al., 1991a,b) resulting in TrkA
dimerization and transphosphorylation (Jing et al.,
1992). The phosphorylated Trk then binds to a variety
of signaling e€ectors, several of which have been
characterized. Two well characterized e€ectors are the
SH2 containing proteins, PLCg1 and SHC (Dikic et al.,
1995; Obermeier et al., 1993, 1994). PLCg, after
binding to Trk, catalyses the formation of diacylglycerol and inositol triphosphate from PI 4,5-biphosphate. These second messengers stimulate protein
kinase C and increase intracellular calcium, which
may play a role in cytoskeletal rearrangements. SHC,
on the other hand, acts as an adapter protein by ®rst
associating with Grb2, another adapter protein
containing SH2 and SH3 domains and then with
SOS, a guanine nucleotide exchange protein (Nakamura et al., 1996). SOS then leads to the activation of
ras and consequently its signaling cascade. Another
signaling intermediate which forms a complex with
activated Trk is phosphatidylinositiol 3-kinase (PI3-K),
although it may not interact directly with the Trk
receptor (Raoni and Bradshaw, 1992; Solto€ et al.,
1992). PI3-K is composed of two subunits, an SH2
containing 85 kDa regulatory domain and a 110 kDa
catalytic domain. Activation of PI3-K catalyzes the
formation of lipid second messengers which then lead
to the activation of the serine threonine kinase akt,
also known as protein kinase B or Rac protein kinase
(Burgering and Co€er, 1995; Datta et al., 1996; Franke
et al., 1995, 1997; Klippel et al., 1996, 1997). Akt has
been reported to play a role in survival in a number of
di€erent cell types, including cerebellar neurons
(Dudek et al., 1997; Hemmings, 1997; Kennedy et al.,
1997; Kulik et al., 1997; Marte and Downward, 1997).
Another e€ector of Trk whose function is much less
clear is the protein SNT. SNT has been reported to
function in a novel, ras independent pathway to induce
PC12 cell di€erentiation (Rabin et al., 1993). Recent
work, however, has indicated that SNT may be
upstream of ras in the SHC-Grb2-SOS complex
NGF signals through distinct signaling cascades
LJ Klesse et al
2056
(Kouhara et al., 1997; Ong et al., 1996; Wang et al.,
1996).
Several reports have implicated members of the ras/
erk signaling cascade in a variety of PC12 cell
responses to NGF. A role for the pathway, for
instance, has been reported in NGF induced cell
di€erentiation. Ras activation appears to be critical
for the di€erentiation NGF induces in PC12 cells (BarSagi and Feramisco, 1985; Cowley et al., 1994; Kremer
et al., 1991; Szeberenyi et al., 1990, 1992; Wood et al.,
1993). Blocking ras signaling either by expression of a
dominant interfering form or introduction of function
blocking antibodies inhibits NGF induced differentiation of PC12 cells (Hagag et al., 1986; Kremer et al.,
1991; Szeberenyi et al., 1990). Conversely, a constitutively active form of ras protein injected into PC12 cells
induces NGF independent di€erentiation (Bar-Sagi and
Feramisco, 1985; Noda et al., 1985). The best
characterized e€ector pathway activated by ras-GTP
is the erk cascade, which includes raf, mek and erk
kinases. Ras activity has been demonstrated to be
necessary for erk phosphorylation induced by NGF
(Thomas et al., 1992). The intermediates in this
pathway are highly conserved; they can be found in
species ranging from yeast to C. elegans to mammals.
In this cascade, activated ras binds to the serinethreonine kinase raf bringing it to the cell membrane.
Activated raf then phosphorylates mek, a threonine
and tyrosine kinase whose substrates include the erks
(Lange-Carter and Johnson, 1994; Robbins et al., 1992,
1993). The activities of both raf and mek have also
been shown to be critical for PC12 cell di€erentiation
(Cowley et al., 1994; Wood et al., 1993). Expression of
a constitutively active form of mek, for instance, can
mimic NGF induced PC12 di€erentiation, while a
dominant interfering form will inhibit this NGF action
(Cowley et al., 1994). Activated erk serine threonine
kinases translocate to the nucleus where they can act
on a number of transcription factors.
Although the ras/erk signaling pathway appears to
be extremely important in mediating NGF induced
a€ects in PC12 cells, there is evidence for ras/erk
independent pathways. While constitutively active
forms of ras have been reported to be sucient for
cell survival (Rukenstein et al., 1991), other pathways
have also been implicated. PI3-K, for instance, has
been demonstrated as necessary for NGF induced
survival of PC12 cells. In one such study, inhibitors to
PI3-K were able to block NGF induced survival (Yao
Table 1
Signaling mutation
and Cooper, 1995). Therefore, it is still unclear if the
ras/erk cascade is the mediator of the anti-apoptotic
activity of NGF.
In order to explore the role of the ras/erk signaling
cascade in the variety of NGF mediated e€ects:
survival, di€erentiation, and proliferation; we generated recombinant adenovirus which expressed each of
these signaling intermediates in wild type, dominant
interfering, and constitutively active forms (Bruder et
al., 1992; Cowley et al., 1994; Feig and Cooper, 1988;
Heidecker et al., 1990; Robbins et al., 1993; Tabin et
al., 1982; Zhang et al., 1993). Recombinant adenovirus
permitted us to modulate the levels of expression,
allowing for distinction between endogenous roles of
the signaling molecules and their biochemical capabilities. Our results indicated that the ras/erk cascade is
necessary and sucient for di€erentiation, but not for
survival of PC12 cells.
Results
Table 1 shows the origin and nature of the ras/erk
signaling intermediate mutations which were used to
generate the recombinant adenovirus. We chose the
adenovirus system because of its broad spectrum of
infectivity, its low in vitro toxicity, its non-replicative
nature, and its ability to produce high titer stocks. We
routinely obtain clonaly puri®ed stocks with plaque
assay titers ranging from 108 ± 109 (see Materials and
methods). We ®rst tested the ability of the adenoviruses
to infect and express recombinant genes in PC12 cells.
Figure 1 indicates that a majority of the PC12 cells
infected with a b-galactosidase recombinant adenovirus
(Herz and Gerard, 1993) expressed the lacZ marker at
multiplicities of infection (MOI) as low as 20. We were
not able to achieve 100% infectivity, likely due to
continued cell division after infection. All further
experiments were performed at multiplicities of
infection (MOI) greater than 20 in order to achieve
at least 60% infection of cells. Dose response
experiments at MOI's ranging from 1 ± 500 exhibited
signs of cellular toxicity only at MOI above 150 (data
not shown). We thus observed no adverse a€ects on
PC12 cells in our experimental conditions.
We tested the expression levels of each recombinant
gene via Western and/or Northern analysis in infected
PC12 cells. Typical expression patterns are shown in
Figure 2. These Western blots show the correlation
Signaling mutations utilized for generation of recombinant adenovirus
Mutation
Mechanism
Species
Reference
Arg 12 to Val
Ser 17 to Asn
Does not hydrolyze GTP
Decreased anity for GTP
Human
Human
(Tabin et al., 1982)
(Feig and Cooper, 1988)
CAraf 1
DNraf1
Xraf1
Deletes aa 26-302
Truncated at aa 257
Truncated at aa 257;
Cys 168 to Ser
Deleted N terminus
Deleted kinase domain
Deleted kinase domain;
Decreased ras binding
Human
Human
Human
(Heidecker et al., 1990)
(Bruder et al., 1992)
(Zhang et al., 1993)
CAmek1
DNmek1
Ser 217/221 to Glu
Ser 217 deleted
Activated phosphor. sites
Phosphor. site deleted
Rabbit
Rabbit
(Cowley et al., 1994)
(Cowley et al., 1994)
DNerk2
Lys 52 to Arg
Mutations in ATP binding site,
no catalytic activity
Human
(Robbins et al., 1993)
CAras (Ha-ras)
DNras (Ha-ras)
NGF signals through distinct signaling cascades
LJ Klesse et al
2057
Figure 1 Expression of a b-galactosidase expressing recombinant adenovirus in PC12 cells. PC12 cells were infected with multiples
of infection (MOI) as indicated. Seventy-two hours post-infection the cells were ®xed with 0.5% glutaraldehyde and stained
overnight. Multiples of infection are: (a) No Virus; (b) MOI of 10; (c) MOI of 20; and (d) MOI of 50
between expression levels and increasing MOI. In
addition, recombinant protein levels increased between
24 and 72 h (data not shown). Therefore, the use of
adenoviral vectors enabled us to control and modulate
the levels of introduced protein in large populations of
PC12 cells.
Di€erentiation
In order to con®rm that the activated ras pathway
could mimic NGF in PC12 cells (Bar-Sagi and
Feramisco, 1985; Cowley et al., 1994), we employed
adenoviruses expressing the constitutively active forms
of the ras pathway intermediates (CA adenoviruses).
When PC12 cells are infected with these CA forms of
ras, raf and mek, neurite outgrowth is induced in the
absence of NGF stimulation (Figure 3). To control
against adenoviral gene expression mediated e€ects, a
recombinant adenovirus harboring an inactivated form
of raf (Xraf) was used in mock infections. Xraf infected
PC12 cells were indistinguishable from un-infected
controls (Figure 3a). Consistent with the biological
e€ects of our CA adenoviruses, biochemical downstream activation of erks was observed (Figure 3b).
Therefore, these results con®rm the capacity of the ras/
erk pathway to mimic NGF function in PC12 cells.
Since activation of the ras/erk cascade could induce
di€erentiation, we next tested whether endogenous
activity of the cascade was required. Since gene
ablation is not feasible in this system, the recombinant
adenovirus approach allowed us to systematically
interfere with the activity of each intermediate of the
ras pathway. Figure 4 shows that each of the dominant
negative intermediate expressing adenoviruses (DN
adenovirus) signi®cantly inhibits NGF induced neurite
outgrowth. Again, Xraf and uninfected cells exhibited
comparable percentages of neurite outgrowth
(49.2+4.3% and 56.9+5.5%; Figure 4a). Figure 4b
shows examples of the DN adenoviruses activity in
these neurite outgrowth assays. DN ras inhibited
neurite outgrowth most e€ectively, with only
9.7+1.4% of PC12 cells exhibiting neurites. DN raf
and DN mek adenoviruses inhibited neurite outgrowth
at comparable levels (23+1% and 22.1+1.8%
NGF signals through distinct signaling cascades
LJ Klesse et al
2058
in the presence of NGF, Figure 5b). Since rapid erk
phosphorylation has been associated with proliferation
of PC12 cells and sustained phosphorylation with
di€erentiation (Qui and Green, 1992), we next tested
the e€ects of the DN adenoviruses, particularly DN
ras, on baseline PC12 cell proliferation. Cells were
pulse labeled with BrdU for 48 h. The percentage of
cells exhibiting BrdU incorporation is shown in Figure
6. Although capable of inhibiting erk activity in
response to NGF, expression of the DN ras/erk
pathway molecules had no adverse e€ect on baseline
PC12 cell proliferation.
Survival
Figure 2 Western analysis of recombinant adenoviruses in PC12
cells. Cells were infected with indicated MOI as described in
Materials and methods. (a) Expression of recombinant
adenoviruses expressing wild type, DN and CA ras. (Probed
with conditioned media from rat hybridoma Y13-259 and goat
anti-rat HRP secondary antibody (Boehringer Mannheim)). (b)
Expression of recombinant adenoviruses expressing wild type,
DN and CA mek. (Probed with anti-Mek1 antibody (Vector
Laboratory) and goat anti-mouse HRP secondary antibody
(Santa Cruz)). (c) Expression of recombinant adenovirus
expressing inactive and DN raf (Probed with anti-raf1 from
Transduction Laboratories and goat anti-mouse HRP secondary
antibody (Santa Cruz))
respectively). DN erk2 inhibited neurite outgrowth as
well, with only 28+4.1% of PC12 cells exhibiting
neurites. Therefore, blocking the activity of any of the
signaling intermediates in the ras/erk pathway signi®cantly inhibits NGF induced neurite outgrowth of
PC12 cells, supporting the model that activation of this
pathway is required for neurite outgrowth and
di€erentiation.
Stimulation of PC12 cells with NGF has been
reported to induce rapid and sustained phosphorylation of the erk kinases (Qui and Green, 1992). We
therefore next tested how our battery of DN
adenoviruses a€ected NGF induced erk phosphorylation. PC12 cells were infected with each of the
recombinant adenoviruses and 72 h later stimulated
with NGF (50 ng/ml) for 1 h. Figure 5a shows that
expression of DN ras, DN raf, or DN mek resulted in
signi®cant reduction of erk activation as measured by
phosphorylation. The DN erk2 protein is subject to
phosphorylation and can be distinguished by its gel
migration at an intermediate molecular mass compared
to the endogenous erks (Figure 5a, lane 6). The
inactive form of raf (Xraf) had no e€ect on erk
phosphorylation, again controlling against possible
e€ects caused by the adenovirus vector system.
Proliferation
Although the DN adenoviruses inhibited the sustained
phosphorylation of erks, only DN ras inhibited rapid
erk phosphorylation upon NGF stimulation (5 ± 10 min
NGF has also been shown to mediate survival of
PC12 cells when cultured in the absence of serum, an
e€ect that mimics an important in vivo role of
neurotrophins on neurons (Greene, 1978; Rukenstein
et al., 1991). PC12 cells undergo programmed cell
death, apoptosis, when cultured in serum free medium
without NGF. We employed the CA adenoviruses to
test whether these molecules could mimic NGF
activity in this assay. Figure 7 illustrates the results
of these experiments in which expression of CA forms
of ras, raf and mek mediate PC12 cell survival in the
absence of serum or NGF. One of the hallmarks of
apoptosis is fragmentation of chromosomal DNA into
electrophoretically assayable ladders. The CA adenoviruses could also inhibit DNA fragmentation (data
not shown). To examine whether the activities of the
ras/erk intermediates were implicated in the endogenous pathway for survival, as they are for differentiation, we tested the activity of the DN adenoviruses in
serum free medium. In contrast to the ability of the
CA adenovirus to induce survival, the DN adenoviruses did not inhibit the NGF induced survival
(Figure 8). Consistent with our previous experiments
which examined di€erentiation, the DN adenoviruses
did inhibit di€erentiation of the PC12 cells (Figure
3b). Thus, in serum free media, PC12 cells expressing
the DN molecules survive, but fail to extend neurites
when stimulated with NGF (Figure 8). Therefore, our
data is most consistent with the interpretation that the
ras/erk cascade is not required endogenously for
survival.
Since the ras/erk pathway did not appear to be
necessary for PC12 cell survival, we next examined
alternative signaling pathways that could mediate this
signal. In conjunction with the recombinant adenovirues, we utilized speci®c pharmacological agents to
perturb signaling. Since PI3-K has previously been
implicated in survival signaling (Yao and Cooper,
1995), we tested the speci®c inhibitor of PI3-K,
LY294002, in our survival assay. LY294002 is capable
of inhibiting the majority of PI3-K activity at
concentrations as low as 10 mM (Vlahos et al., 1994).
PI3-K inhibition caused apoptosis of PC12 cells in
serum free media both in the presence of NGF (Figure
9a) and in the presence of the CA adenoviruses (Figure
9c ± e). Apoptosis of cells in the presence of NGF and
LY294002 followed a slightly slower time course than
cells treated with LY294002 in the absence of
neurotrophin. Inhibition of PI3-K induced DNA
fragmentation in PC12 cells regardless of prior
treatment with CA adenoviruses or NGF (Figure 9f,
NGF signals through distinct signaling cascades
LJ Klesse et al
2059
Figure 3 Constitutively active ras, raf and mek induce neurite outgrowth in the absence of NGF. Cells were infected at MOI of 20
and were photographed 6 days post-infection. (a) Xraf; (b) CA ras; (c) CA raf; (d). CA mek. (e) Expression of CA signaling
intermediates are sucient for erk phosphorylation. Top panel probed with Phospho-erk speci®c antibody. Bottom panel probed
with erk antibody as a loading control. Mock+NGF lane: PC12 cells stimulated with NGF at 100 ng/ml for 30 min
3 ± 5). DMSO, the carrier used for LY294002, had no
e€ect which di€ered from untreated cells (data not
shown). Therefore, neither NGF stimulation nor
expression of the CA forms of ras, raf or mek
prevented cell death when PI3-K activity was inhibited.
Our data indicates that the PI3-K activity is
necessary for PC12 cell survival, but CA forms of
ras, raf and mek, although not required, are capable of
supporting cell survival. In order to clarify the role of
the erk cascade in PC12 cell survival, we utilized both a
pharmacological inhibitor and the DN erk2 adenovirus
to perturb signaling of the ras/erk cascade. Inhibition
of mek activity with 50 mM PD98059 inhibited erk
phosphorylation (data not shown) but did not inhibit
PC12 cell survival induced by NGF in serum free
media. However, consistent with the DN adenovirus
results, inhibition of the ras/erk cascade did inhibit
neurite outgrowth (Figure 10a). Although NGF did
not require activity of the erk cascade to promote
survival, we tested whether the CA adenoviruses
required this cascade for their survival promoting
activity. Inhibition of mek with PD98059 blocked the
ability of CA raf and CA mek to lead to erk
phosphorylation (data not shown) and to promote
PC12 cell survival (Figure 10d and e). Survival induced
by the CA ras adenovirus was only partially inhibited
(Figure 10c). Cells coinfected with the DN erk2
adenovirus and the CA adenoviruses exhibited similar
results (data not shown). Consistent with our
observations, DNA fragmentation was observed in
PC12 cells which express CA ras, CA raf and CA mek
in the presence of PD98059, but not the cells in the
presence of NGF (Figure 10f lanes 1 ± 5). Our data
indicates that PC12 cells require activity of PI3-K, but
not the ras/erk cascade, to avoid apoptosis. The CA
adenoviruses, however, are capable of supporting
survival only when neither the ras/erk or PI3-K
cascade is inhibited. Therefore, the ability of the CA
adenoviruses to promote PC12 cell survival does not
appear to mimic NGF in this survival assay.
Ras and PI3-K have been shown to interact in a rasGTP dependent manner (Rodriquez-Viciana et al.,
1994, 1996). Since CA ras could promote cell survival
even in the absence of mek activity, we tested its ability
to activate PI3-K. Akt is a serine-threonine kinase
which functions as an e€ector of PI3-K. The PI3-K
inhibitor LY294002 blocked both Akt activity and
PC12 cell survival in a dose dependent manner (Figure
NGF signals through distinct signaling cascades
LJ Klesse et al
2060
11a). We tested the activation of this pathway in the
presence of the recombinant adenoviruses. CA ras
infected PC12 cells exhibited akt kinase activity at
levels comparable with NGF stimulation (Figure 11b).
CA raf and mek, however, did not increase akt activity
levels above base line. We also found that the DN
adenoviruses did not inhibit akt kinase activity induced
by NGF stimulation (Figure 11b). Therefore, although
ras is capable of activating the PI3-K pathway, NGF
induced activation of PI3-K does not require ras
Figure 4 Dominant negative forms of the erk cascade signi®cantly inhibit neurite outgrowth in PC12 cells stimulated with
NGF. (a) Quantitiation of the inhibition of neurites by the DN adenoviruses. Cells were infected with a multiplicity of infection
of 50. Cells exhibiting neurites (three cell bodies or greater in length) and cells without neurites were counted. All recombinant
adenovirus shown signi®cantly inhibition of neurites except the Xraf (P=0.39). Signi®cance indicated is calculated in comparison
to Xraf infected cells: DN ras ± P=0.0022; DN raf ± P=0.0022; DN mek - P=0.0022; DN erk 2 - P=0.0052. N=6. (b) PC12
cells were infected at MOI of 50 as described in Materials and methods. (a and b) PC12 cells infected with Xraf expressing
adenovirus in the presence (a) and absence (b) of NGF. PC12 cells infected with recombinant adenoviruses expressing (c) DN
ras; (d) DN raf; (e) DN mek; (f) DN erk2 in the presence of NGF. Cells were photographed 6 days after addition of NGF at
50 ng/ml
NGF signals through distinct signaling cascades
LJ Klesse et al
ras/erk cascade has previously been implicated in
survival, di€erentiation and proliferation of these
cells, we demonstrate that this cascade is necessary
and sucient only for di€erentiation. Inhibition of the
ras/erk cascade had no e€ect on PC12 cell survival
induced by NGF. Instead, the PI3-K pathway
appeared to be the primary cascade for NGF
mediated survival.
By utilizing recombinant adenovirus, we were able
to expand on previous work in the PC12 cell system by
systematically exploring the roles of signaling intermediates induced by NGF stimulation. The adenovirus
system provides a reproducible method for expressing
signaling molecules at de®ned levels in large cell
populations. This powerful expression system also
allows for the distinction between endogenous functions of the molecules under study and e€ects which
were not endogenous, but were instead biochemically
possible. For instance, while CA ras is capable of
activating PI3-K, we ®nd that the endogenous
activation of PI3-K by NGF did not require ras
function.
Figure 5 (a and b) Dominant negative signaling intermediates
inhibit prolonged, but not rapid erk phosphorylation. Western
analysis of PC12 cells infected with recombinant adenovirus at
MOI of 20 as described in Material and methods. Seventy-two
hours post-infection, the cells were stimulated for (a) 1 h or (b)
10 min with 50 ng/mL NGF and then lysed. Whole cell lysates
were probed with (top panel) anti-phosphorylated erk antibody
(Promega) and (bottom panel) erk antibody (Santa Cruz) as a
loading control
Di€erentiation
The results of this work support a model whereby
NGF e€ects are mediated by at least two distinct
signaling pathways. The ras/erk cascade was both
required and sucient for the di€erentiation of PC12
cells. CA adenoviruses were capable of inducing
di€erentiation of PC12 cells in the absence of NGF,
while the DN adenoviruses could inhibit NGF induced
di€erentiation. Contrary to previous reports (Jackson
et al., 1996), inhibition of the PI3-Kinase pathway with
LY294002 did not inhibit neurite outgrowth induced
by NGF or the CA adenoviruses (data not shown).
The observation that the DN adenoviruses and the
mek inhibitor PD98059 blocked neurite outgrowth, but
not survival, again indicates that two distinct signaling
pathways mediate these functions of NGF.
Survival
Figure 6 Dominant negative signaling intermediates have no
signi®cant a€ect on PC12 cell proliferation. Percentages of cells
with BrdU incorporation are shown. P values indicated are in
comparison to Xraf. No Virus (P=0.106); DN ras (P=0.089);
DN raf (P=0.14); DN mek (P=0.845); DN erk2 (P=0.841)
activity. Overall, it appears that intermediates of the
ras/erk signaling cascade are necessary and sucient
for the di€erentiation of PC12 cells, while cell survival
is mediated predominantly by the activity of ras
independent PI3-K.
Discussion
In this study, we analyse the signaling pathways which
mediate the e€ects of NGF on PC12 cells. While the
Survival of PC12 cells induced by NGF required the
activity of PI3-K. LY294002, a speci®c inhibitor of
PI3-K, inhibited the survival of PC12 cells induced
either by NGF or the CA adenoviruses. Our data with
the CA recombinant adenoviruses indicate that the
PI3-K and ras signaling pathways are biochemically
capable of interacting. CA ras can activate the PI3-K
cascade, as measured by akt kinase activity, although
this pathway's activation via NGF does not require ras
activity. This result di€ers from that reported by
Rodriquez-Viciana et al. (1994) who demonstrate a
reduction in phosphatidylinositol (3,4) bisphosphate
and phosphatidylinositol (3,4,5) triphosphate levels in
the presence of DN ras. This reduction, however, is not
complete and may not be sucient to inhibit activation
of the downstream intermediate, in this case, akt. The
CA forms of raf and mek were also capable of
supporting cell survival in serum free medium,
consistent with previous reports. For instance, Xia et
al. (1995) demonstrates the ability of a constitutively
active form of mek to promote PC12 cells survival,
while Yan and Greene (1998) shows that Nacetylcysteine promotes survival by activation of the
2061
NGF signals through distinct signaling cascades
LJ Klesse et al
2062
ras/erk cascade. By utilizing the dominant negative
forms of these molecules, we extend upon these
®ndings and show that the activity of the erk cascade
intermediates are not necessary for NGF induced PC12
cell survival. Therefore, although biochemically capable
of supporting PC12 survival, the erk cascade is not
biologically required.
Akt, the serine threonine kinase downstream of PI3K, was activated upon NGF stimulation. Akt
activation has been implicated in the survival of
several cell types, including cerebellar neurons via
IGF-1 stimulation (Dudek et al., 1997). Akt has also
been reported to directly interact with and phosphorylate Bad, a member of the Bcl-2 family of proteins
Figure 7 Constitutively active ras, raf and mek, as well as wild type ras, promote survival of PC12 cells in the absence of serum
and NGF. PC12 cells were infected at multiplicities of infection of 100. Pictures were taken 48 h after addition of serum free media.
(a) Xraf with 100 ng/mL NGF; other wells are without NGF: (b) Xraf; (c) wtras; (d) CA ras; (e) CA raf; (f) CA mek. (g)
Quantitation of the ability of the CA adenoviruses to promote cell survival in the absence of serum and NGF. Cells were infected as
above at MOI of 100 and the data presented represents three experiments. ***Indicates P values of 50.0001 as compared to X raf NGF
NGF signals through distinct signaling cascades
LJ Klesse et al
(Datta et al., 1997; del et al., 1997). The Bcl-2 family is
composed of a variety of homologous proteins which
either promote or protect against cell death (Korsmeyer, 1995; Korsmeyer et al., 1993; Merry and
Korsmeyer, 1997; Rinkenberger and Korsmeyer,
1997). Members of the Bcl-2 family which promote
cell survival, such as Bcl-2 itself, do so by forming
homodimers. Cell death promoting members, like Bad
(Yang et al., 1995), inhibit homodimer formation by
binding to the cell survival promoting family members
(Zha et al., 1997). Yet, when Bad is phosphorylated, it
is no longer capable of binding to the anti-apoptotic
family members Bcl-2 and BclxL (Zha et al., 1996).
Therefore, NGF stimulation of the PI3-K cascade,
results in activation of akt that may in turn lead to the
phosphorylation of Bad and promotion of cell survival.
Figure 8 Dominant negative signaling intermediates do not inhibit NGF induced survival of PC12 cells in the absence of serum.
PC12 cells were infected at MOI of 100 and pictures were taken 48 h after addition of serum free media. All wells in the presence of
100 ng/mL NGF except B. (a and b) Xraf infected; (c) DN ras infected; (d) DN raf infected; (e) DN mek infected; (f) DN erk2
infected. (g) Quantitiation of the inability of the DN adenoviruses to block NGF induced cell survival in the absence of serum. Cells
were infected as above. P values were calculated in comparison to Xraf + NGF and are: Xraf - NGF (P=50.0001); DNras
(P=0.63); DNraf (P=0.85); DNmek (P=0.313); DNerk2 (P=0.07). n=3
2063
NGF signals through distinct signaling cascades
LJ Klesse et al
2064
Previous work has indicated that activation of a cell
death promoting cascade, in conjunction with inhibition of a cell survival cascade, is required for PC12 cell
apoptosis. Work by Xia et al. (1995) implicates the
JNK cascade as apoptosis promoting in PC12 cells.
The presented experiments identify activity of the PI 3
Kinase cascade as necessary for inhibition of apoptosis.
Therefore, the relative balance of activity in the JNK
cascade and the PI 3 Kinase cascade may regulate
PC12 cell apoptosis.
Proliferation
Di€erentiation of PC12 cells, as opposed to proliferation, has been postulated to be a function of the
kinetics of erk kinase activation (Qui and Green, 1992).
We found that the DN forms of the ras/erk signaling
pathway would inhibit di€erentiation of PC12 cells,
but had no apparent e€ect on their baseline proliferation. Each of the DN signaling intermediates blocked
long term phosphorylation, while only DN ras
appeared to inhibit short term phosphorylation. The
ability of DN ras to inhibit short term phosphorylation
was, however, not complete. Based on our results,
neurite outgrowth correlated with sustained erk
activation both in experiments with the DN intermediates and the CA intermediates. We were unable to
clearly determine if proliferation required the short
term activation. These results are consistent with
previous reports which show that the same dominant
Figure 9 Constitutively active forms of ras, raf and mek are not sucient to promote survival of PC12 cells in the presence of
LY294002. Cells were infected at MOI of 100 and pictures taken 48 h after addition of serum free media and 10 mM LY294002. (a)
Xraf with 100 ng/mL NGF; (b) Xraf; (c) CA ras; (d) CA raf; (e) CA mek; (f) DNA fragmentation (1) Xraf+NGF; (2) Xraf; (3) CA
ras; (4) CA raf; (5) CA mek
NGF signals through distinct signaling cascades
LJ Klesse et al
negative ras mutation used here, when stably
transfected into PC12 cells, did not inhibit proliferation, but did inhibit di€erentiation (Szeberenyi et al.,
1990).
SNT
SNT has been postulated to act independently of ras in
PC12 cell di€erentiation (Rabin et al., 1993). Although
our experiments did not directly address the function
of SNT, our results would appear to rule out
additional, ras independent di€erentiation pathways.
CA forms of ras, raf and mek were all capable of
mimicking NGF's di€erentiating e€ects on PC12 cells,
while blocking the pathway with the DN adenoviruses
inhibited NGF actions. Recent reports indicate that
SNT, instead of being ras independent, lies upstream of
ras signaling in the complex of SHC-Grb and SOS
(Kouhara et al., 1997; Ong et al., 1996; Wang et al.,
1996). Our results are consistent with this placement of
SNT in the signaling cascade required for differentiation.
The PC12 cell system has been used as a model
system for neutrotrophin signaling in primary neurons.
Both sensory and sympathetic neurons undergo
apoptosis when deprived of NGF support in vitro
(Edwards and Tolkovsky, 1994; Vogel et al., 1995).
Previous studies have implicated ras in the transduction of neurotrophin signals in sensory neurons.
Neurons derived from mice homozygous for a
mutation in NF1, a ras-GAP expressed in neural crest
derived cells, are independent of neurotrophic support
(Brannan et al., 1994; Vogel et al., 1995), perhaps due
to the loss of negative regulation of ras. On the other
hand, perturbing ras activity with function blocking
Fab fragments inhibits neurotrophin induced survival
Figure 10 Mek inhibitor PD98059 inhibits the ability of constitutively active raf and mek to promote survival of PC12 cell in
serum free media. Cells were infected with MOI of 100 and were photographed 48 h after the addition of serum free media and
25 mM PD98059. (a) Xraf+100 ng/ml NGF; (b) Xraf - NGF; (c) CA ras; (d) CA raf; (e) CA mek. (f) DNA fragmentation of PC12
cells in the presence of PD98059 (1) Xraf+NGF; (2) Xraf; (3) CA ras; (4) CA raf; (5) CA mek
2065
NGF signals through distinct signaling cascades
LJ Klesse et al
2066
returned to 37%. Two hundred and ninety-three cells were
grown in Cellgro DMEM with 10% heat inactivated fetal
bovine serum, 1% penicillin/strepavitin, and 1% Lglutamine. Cells were maintained post transfection by
replacement of 3 ml of media biweekly for 3 ± 4 weeks
until the entire plate had lysed. Lysates were collected,
underwent two rounds of freeze/thawing and spun to
remove cell debris. A plate of 293 cells was then reinfected
with 100 ml of the lysate. Thirty-six hours after addition of
the potential adenoviral lysate, DNA was collected from
the cells using Hirt bu€er (0.6% SDS, 10 mM EDTA,
100 mg/ml proteinase K). Southern analysis was performed
to con®rm the presence of the insert of interest. Two
hundred and ninety-three cells were then infected with
serial dilutions of the original lysate and overlayed with
agar. Ten days later, isolated plaques of virus were
collected and used to reinfect 293 cells to obtain clonal
stocks of the recombinant adenovirus. Southern blot
analysis was used to recon®rm the presence of the insert.
Once clonal stocks are con®rmed, small amounts of these
lysates were used to reinfect several 293 plates of cells to
obtain 30 ± 40 mL of recombinant adenovirus lysate.
Infection protocol
Figure 11 (a) Inhibition of akt kinase activity and PC12 cell
survival at increasing doses of LY294002 in serum free media.
Akt kinase activity was measured after stimulation of the cells in
serum free media with 100 ng/ml NGF for 5 min. PC12 cell
survival is presented as a percentage of the viable cells in no
LY294002 controls. ***Indicates P values of 50.0005. n=3. (b)
Akt kinase activity is induced in PC12 cells in serum free media
by NGF and constitutively active ras. Cells were infected as
previously at MOI of 100. Where indicated, cells were stimulated
with 100 ng/mL NGF for 5 min. PC12 cells in the absence of
serum or NGF were comparable to non-speci®c background
levels and were subtracted from each sample. Samples are
depicted as a percentage of the activity of PC12 cells in serum
free media stimulated with 100 ng/ml NGF. n=4
(Borasio et al., 1993; Nobes et al., 1996). Therefore, ras
and potentially the erk kinase cascade, may mediate
neurotrophin signals in primary neurons. The recombinant adenovirus utilized in this paper will be excellent
reagents to test the role of this cascade in primary
neurons.
Materials and methods
Generation of recombinant adenovirus
cDNAs for each of the signaling intermediates were cloned
into the vector pAC-CMV (Gomez-Foix et al., 1992). Five
mg of the pAC were then cotransfected with 5 mg pJM17
and 10 mg ssDNA into 293 cells using a calcium phosphate
technique. The DNA was added to 500 mL of 10 mM
HEPES bu€er saline, pH 7.12. Thirty-two ml of 2 M
CaCl2 was added and a precipitate formed at room
temperature for 20 ± 30 min. Precipitates were added
dropwise to the cells, incubated at 378C/5% CO2 for 6 h.
Cells were shocked with 15% glycerol in DMEM for 2 min
at room temperature. Media was then replaced and the cell
PC12 cells were maintained in growth media which
consisted of Gibco DMEM (high glucose) supplemented
with 10% heat inactivated horse serum, 5% heat
inactivated fetal bovine serum, 1% penn/strep, 1% Na
pyruvate. Cells were infected with viral lysates in a minimal
volume of media for 1 h at 378C. Growth media was then
added to increase the volume. Twenty-four hours after
infection, the media containing the adenovirus was
removed and replaced with fresh growth media.
Western analysis
The PC12 cells were plated 24 h prior to infection at 16106
cells per dish. Cells were infected with recombinant
adenovirus using the infection protocol described above.
For expression analysis, 72 h after the addition of the
recombinant adenovirus, the media was removed, the cells
were washed once with PBS and then lysed with NP40 lysis
bu€er (150 mM NaCl, 50 mM Tris pH 8, 1% NP-40)
containing protease inhibitors (aprotinin, pepstatin,
PMSF and leupeptin). For erk phosphorylation analysis,
media containing the recombinant adenovirus was removed
24 h post-infection and replaced with fresh media. Seventytwo hours after initial infection, the cells were stimulated
with 50 ng/mL NGF (Regeneron) for the speci®ed amount
of time. Cells were then washed once with PBS and lysed
with NP40 lysis bu€er with protease inhibitors. Whole cell
lysates were separated on a 12% polyacrylamide gel. All
antibodies used were diluted in PBS containing 0.1%
Tween-20 and include anti-P-Erk (Promega diluted
1 : 1000), anti-ras (Y259 condition media diluted 1 : 2),
anti-raf diluted 1 : 2000 (Transduction Laboratories), antimek1 diluted 1 : 2000 (Vector Laboratories), anti-erk
diluted 1 : 5000 (Santa Cruz), goat anti-rabbit IgG-HRP
diluted 1 : 1000 (Santa Cruz), goat anti-mouse IgG-HRP
diluted 1 : 2000 (Santa Cruz) and goat anti-rat IgG-HRP
diluted 1 : 400 (Boehringer Mannheim). Blots were developed using either the ECL or Pierce Super Signal Ultra
Chemiluminescence signaling systems.
Neurite outgrowth assay
Nunc lux 60 mm dishes were covered with 2 mL of 0.5 mg/
ml poly-DL-ornithine (Sigma in 0.15 M borate bu€er,
pH 8.6) overnight at room temperature. Plates were
washed twice with sterile ddH2O and 16105 cells were
plated in PC12 media. Cells were infected 24 h later as
NGF signals through distinct signaling cascades
LJ Klesse et al
previously described. For the DN adenoviruses, 50 ng/mL
NGF (Regeneron) was added to the media. Cells were
incubated in the NGF containing media for 6 days. Cells
with and without neurites were then counted (a neurite was
de®ned as a process equal to or greater than three cell
bodies in length). Percentages of cells with neurites were
then calculated. Statistics were performed using SigmaStat
statistics program. For wild type and CA expressing
recombinant adenovirus experiments, the cells were
incubated in growth media for 3 ± 6 days.
Proliferation assay
PC12 cells were plated on polyornithine coated Nunclon 6
well dishes. Twenty-four hours after plating, the cells were
infected at a multiplicity of infection of 50. The adenovirus
containing media was replaced 24 h after infection with
fresh growth media. 10 mM BrDU was added to the
cultures 48 h after initial infection and remained in the
culture for the next 48 h. The cells were washed with PBS
and ®xed for 30 min with 4% paraformaldehyde. The cells
were again washed with PBS and incubated with 3% H2O2/
10% MeOH in PBS for 30 min. Cells were then
permeabilized with 200 mg/mL pepsin/0.1 N HCl in PBS
for 10 min, incubated in 2N HCl for 30 min and
neutralized for 10 min with 0.1 M Na Borate pH 8.5. The
cells were blocked with 5% heat inactivated horse serum
alone, with avidin D blocking solution, and then with
Biotin blocking solution (Vector Labs). Anti BrDU
antibody was added overnight at 1 : 50 dilution at 48C.
The cells were incubated with a goat antimouse secondary
antibody (Vector Labs) and then incubated with Vector
lab's Elite Vectastain kit's AB reagents. The immunocytochemistry was developed with DAB, counter stained with
eosin and coverslipped. The number of cells with and
without BrDU incorporation were then counted.
Survival assays
PC12 cells were plated on Nunc lux 60 mm dishes in
growth media. Twenty-four hours after plating the cells
were infected with multiplicities of infection of 100 as
previously described. Seventy-two hours after initial
infection the media was removed, the cells washed three
times with PBS, and serum free media (Gibco DMEM high
glucose, 1% Sodium Pyruvate, 1% Penn/Strep) was added.
NGF was added to indicated plates at 100 ng/mL. In
experiments including LY294002 (Sigma), the inhibitor was
added at 10 mM to the serum free media. PD98059 (UBI)
was added where indicated to serum free media at 50 mM.
Pictures were taken 48 h after the removal of serum from
the dishes. For quantitation of recombinant adenovirus
e€ects, 48 h after the removal of serum, the cells were
washed once in PBS and removed from the plate with
Trypsin-EDTA (Gibco). For quantitation of LY294002
dose e€ects, the cells were counted 24 h after addition of
serum free media. Viable cells were counted under the
microscope with a hemacytometer. Statistics were performed using Sigma Stat.
DNA fragmentation
PC12 cells were plated, infected and switched to serum free
media as described above for the survival assays. Inhibitors
were added where indicated at the same concentrations as
above. Twenty-four hours later, the cells were washed once
with PBS and lysed in DNA lysis bu€er (100 mM NaCl,
10 mM Tris-Cl, pH 8, 25 mM EDTA, pH 8, 0.5% SDS,
0.1 mg/ml proteinase K). DNA was run on a 1.5% agarose
gel, transferred to nitrocellulose, and probed with EcoRI
digested rat genomic DNA.
Akt kinase assay
PC12 cells were plated and infected as described
previously. Seventy-two hours after infection, the cells
were washed three times with PBS and serum free media
was added. One hour later, 100 ng/mL NGF was added to
indicated plates for 5 min. 1 mg/ml insulin was used to
stimulate NIH3T3 cells for 3 ± 5 min as a positive control
(data not shown). LY294002 (100 mM) and PD98059
(100 mM) were used, where indicated. The cells were then
washed with ice cold PBS and lysed in cold bu€er (50 mM
Tris, pH 7.5, 1 mM EDTA, 1 mM EGTA, 0.5 mM Na3VO4,
0.1% 2-mercaptoethanol, 1% Triton X-100, 50 mM sodium
¯uoride, 5 mM sodium pyrophosphate, 10 mM sodium Bglycerol phosphate, 0.1 mM PMSF, 1 mg/ml of aprotinin,
pepstatin, leupeptin and 1 mM microcystin). Lysates wee
snap frozen in liquid nitrogen. Lysates were analysed using
the Akt/PKB Immunoprecipitation Kinase Assay Kit from
Upstate Biotechnology (Lake Placid, NY, USA). Mean
counts per min were determined using Beckman's LS 6500
scintillation counter. Background counts (PC12 cells in the
absence of serum and NGF) were subtracted from each
sample. Each sample is graphed as a percentage of the
activity of PC12 cells in the absence of serum but in the
presence of NGF (100 ng/mL).
Acknowledgements
We wish to thank L Feig for providing the DN ras and wt
ras DNA; M White for the CA ras DNA; U Ra€ for the
raf DNA constructs; and M Cobb and J Frost for the Erk
constructs. Thanks to B Gerard for the pAC-CMV vector
and C Newgard for the pJM17 vector for adenovirus
generation. We would also like to thank J Alcorn for the
B-galactosidase expressing adenovirus and for advice in
recombinant adenovirus generation. This work was supported by NINDS ROI NS 34296. LFP and LJK were
partially supported by DOD grant DAMD17-97-1-7343,
LKJ was supported by NIH T326M08203.
References
Bar-Sagi D and Feramisco JR. (1985). Cell, 42, 841 ± 848.
Borasio GD, Markus A, Wittinghofer A, Barde YA and
Heumann R. (1993). J. Cell. Biol., 121, 665 ± 672.
Brannan CI, Perkins AS, Vogel KS, Ratner N, Nordlund
ML, Reid SW, Buchberg AM, Jenkins NA, Parada LF and
Copeland NG. (1994). Genes Dev., 8, 1019 ± 1029.
Bruder JT, Heidecker G and Rapp UR. (1992). Genes Dev.,
6, 545 ± 556.
Burgering BM and Co€er PJ. (1995). Nature, 376, 599 ± 602.
Cowley S, Paterson H, Kemp P and Marshall CJ. (1994).
Cell, 77, 841 ± 852.
Datta K, Bellacosa A, Chan TO and Tsichlis PN. (1996). J.
Biol. Chem., 271, 30835 ± 30839.
Datta SR, Dudek H, Tao X, Masters S, Fu H, Gotoh Y and
Greenberg ME. (1997). Cell, 91, 231 ± 242.
del PL, Gonzalez-Garcia M, Page C, Herrera R and Nunez
G. (1997). Science, 278, 687 ± 689.
Dikic I, Batzer AG, Blaikie P, Obermeier A, Ullrich A,
Schlessinger J and Margolis B. (1995). J. Biol. Chem., 270,
15125 ± 15129.
2067
NGF signals through distinct signaling cascades
LJ Klesse et al
2068
Dudek H, Datta SR, Franke TF, Birnbaum MJ, Yao R,
Cooper GM, Segal RA, Kaplan DR and Greenberg ME.
(1997). Science, 275, 661 ± 665.
Edwards SN and Tolkovsky AM. (1994). J. Cell Biol., 124,
537 ± 546.
Feig LA and Cooper GM. (1988). Mol. Cell Biol., 8, 3235 ±
3243.
Franke TF, Kaplan DR, Cantley LC and Toker A. (1997).
Science, 275, 665 ± 668.
Franke TF, Yang SI, Chan TO, Datta K, Kazlauskas A,
Morrison DK, Kaplan DR and Tsichlis PN. (1995). Cell,
81, 727 ± 736.
Gomez-Foix AM, Coats WS, Baque S, Alam T, Gerard RD
and Newgard CB. (1992). J. Biol. Chem., 267, 25129 ±
25134.
Greene LA. (1978). J. Cell. Biol., 78, 747 ± 755.
Greene LA and Tischler AS. (1976). Proc. Natl. Acad. Sci.
USA, 73, 2424 ± 2428.
Hagag N, Halegoua S and Viola M. (1986).Nature, 319,
680 ± 682.
Heidecker G, Huleihel M, Cleveland JL, Kolch W, Beck TW,
Lloyd P, Pawson T and Rapp UR. (1990). Mol. Cell. Biol.,
10, 2503 ± 2512.
Hemmings BA. (1997). Science, 275, 628 ± 630.
Herz J and Gerard RD. (1993). Proc. Natl. Acad. Sci. USA,
90, 2812 ± 2816.
Jackson TR, Blader IJ, Hammonds-Odie LP, Burga CR,
Cooke F, Hawkins PT, Wolf AG, Heldman KA and
Theibert AB. (1996). J. Cell. Sci., 109, 289 ± 300.
Jing S, Tapley P and Barbacid M. (1992). Neuron, 9, 1067 ±
1079.
Kaplan DR, Hempstead BL, Martin-Zanca D, Chao MV
and Parada LF. (1991a). Science, 252, 554 ± 558.
Kaplan Dr, Martin-Zanca D and Parada LF. (1991b).
Nature, 350, 158 ± 160.
Kaplan DR and Miller FD. (1997). Curr. Opin. Cell Biol., 9,
213 ± 221.
Kennedy SG, Wagner AJ, Conzen SD, Jordan J, Bellacosa
A, Tsichlis PN and Hay N. (1997). Genes Dev., 11, 701 ±
713.
Klippel A, Kavanaugh WM, Pot D and Williams LT. (1997).
Mol. Cell. Biol., 17, 338 ± 344.
Klippel A, Reinhard C, Kavanaugh WM, Apell G, Escobedo
MA and Williams LT. (1996). Mol. Cell. Biol., 16, 4117 ±
4127.
Korsmeyer SJ. (1995). Trends Genet., 11, 101 ± 105.
Korsmeyer SJ, Shutter JR, Veis DJ, Merry DE and Oltvai
ZN. (1993). Semin. Cancer Biol., 4, 327 ± 332.
Kouhara H, Hadari YR, Spivak-Kroizman T, Schilling J,
Bar-Sagi D, Lax I and Schlessinger J. (1997). Cell, 89,
693 ± 702.
Kremer NE, D'Arcangelo G, Thomas SM, DeMarco M,
Brugge JS and Halegoua S. (1991). J. Cell. Biol., 115,
809 ± 819.
Kulik G, Klippel A and Weber MJ. (1997). Mol. Cell. Biol.,
17, 1595 ± 1606.
Lange-Carter CA and Johnson GL. (1994). Science, 265,
1458 ± 1461.
Marte BM and Downward J. (1997). Trends Biochem. Sci.,
22, 355 ± 358.
Merry DE and Korsmeyer SJ. (1997). Ann. Rev. Neurosci.,
20, 245 ± 267.
Nakamura T, Sanokawa R, Sasaki Y, Ayusawa D, Oishi M
and Mori N. (1996). Oncogene, 13, 1111 ± 1121.
Nobes CD, Reppas JB, Markus A and Tolkovsky AM.
(1996). Neuroscience, 70, 1067 ± 1079.
Noda M, Ko M, Ogura A, Liu DG, Amano T, Takano T and
Ikawa Y. (1985). Nature, 318, 73 ± 75.
Obermeier A, Bradshaw RA, Seedorf K, Choidas A,
Schlessinger J and Ullrich A. (1994). EMBO J., 13,
1585 ± 1590.
Obermeier A, Lammers R, Wiesmuller KH, Jung G,
Schlessinger J and Ullrich A. (1993). J. Biol. Chem., 268,
22963 ± 22966.
Ong SH, Goh KC, Lim YP, Low BC, Klint P, ClaessonWelsh L, Cao X, Tan YH and Guy GR. (1996). Biochem.
Biophys. Res. Commun., 225, 1021 ± 1026.
Qui MS and Green SH. (1992). Neuron, 9, 705 ± 717.
Rabin SJ, Cleghorn V and Kaplan DR. (1993). Mol. Cell.
Biol., 13, 2203 ± 2213.
Raoni S and Bradshaw RA. (1992). Proc. Natl. Acad. Sci.
USA, 89, 9121 ± 9125.
Rinkenberger JL and Korsmeyer SJ. (1997). Curr. Opin.
Genet. Dev., 7, 589 ± 596.
Robbins DJ, Cheng M, Zhen E, Vanderbilt CA, Feig LA and
Cobb MH. (1992). Proc. Natl. Acad. Sci. USA, 89, 6924 ±
6928.
Robbins DJ, Zhen E, Owaki H, Vanderbilt CA, Ebert D,
Geppert TD and Cobb MH. (1993). J. Biol. Chem., 268,
5097 ± 5106.
Rodriguez-Viciana P, Warne PH, Dhand R, Vanhaesebroeck
B, Gout I, Fry MJ, Water®eld MD and Downward J.
(1994). Nature, 370, 527 ± 532.
Rodriguez-Viciana P, Warne PH, Vanhaesebroeck B,
Water®eld MD and Downward J. (1996). EMBO J., 15,
2442 ± 2451.
Rukenstein A, Rydel RE and Greene LA. (1991). J.
Neurosci., 11, 2552 ± 2563.
Segal RA and Greenberg ME. (1996). Ann. Rev. Neurosci.,
19, 463 ± 489.
Snider WD and Silos-Santiago I. (1996). Philos. Trans. R.
Soc. Biol. Sci., 351, 395 ± 403.
Solto€ SP, Rabin SL, Cantley LC and Kaplan DR. (1992). J.
Biol. Chem., 267, 17472 ± 17477.
Szeberenyi J, Cai H and Cooper GM. (1990). Mol. Cell. Biol.,
10, 5324 ± 5332.
Szeberenyi J, Erhardt P, Cai H and Cooper GM. (1992).
Oncogene, 7, 2105 ± 2113.
Tabin CJ, Bradley SM, Bargmann CI, Weinberg RA,
Papageorge AG, Scolnick EM, Dhar R, Lowy DR and
Chang EH. (1982). Nature, 300, 143 ± 149.
Thomas SM, DeMarco M, D'Arcangelo G, Halegoua S and
Brugge JS. (1992). Cell, 68, 1031 ± 1040.
Tischler AS and Green LA. (1975). Nature, 258, 341 ± 342.
Vlahos CJ, Matter WF, Hui KY and Brown RF. (1994). J.
Biol. Chem., 269, 5241 ± 5248.
Vogel KS, Brannan CI, Jenkins NA, Copeland NG and
Parada LF. (1995). Cell, 82, 733 ± 742.
Wang JK, Xu H, Li HC and Goldfarb M. (1996). Oncogene,
13, 721 ± 729.
Wood KW, Qi H, D'Arcangelo G, Armstrong RC, Roberts
TM and Halegoua S. (1993). Proc. Natl. Acad. Sci. USA,
90, 5016 ± 5020.
Xia Z, Dickens M, Raingeaud J, Davis RJ and Greenberg
ME. (1995). Science, 270, 1326 ± 1331.
Yan CYI and Greene LA. (1998). J. Neurosci., 18, 4042 ±
4049.
Yang E, Zha J, Jockel J, Boise LH, Thompson CB and
Korsmeyer SJ. (1995). Cell, 80, 285 ± 291.
Yao R and Cooper GM. (1995). Science, 267, 2003 ± 2006.
Zha J, Harada H, Osipov K, Jockel J, Waksman G and
Korsmeyer SJ. (1997). J. Biol. Chem., 272, 24101 ± 24104.
Zha J, Harada H, Yang E, Jockel J and Korsmeyer SJ.
(1996). Cell, 87, 619 ± 628.
Zhang XF, Settleman J, Kyriakis JM, Takeuchi-Suzuki E,
Elledge SJ, Marshall MS, Bruder JT, Rapp UR and
Avruch J. (1993). Nature, 364, 308 ± 313.