Forebrain-Specific Knockout of B-raf
Kinase Leads to Deficits in Hippocampal
Long-Term Potentiation, Learning,
Adele P. Chen,1Masuo Ohno,1K. Peter Giese,1Ralf Ku ¨hn,2Rachel L. Chen,1
and Alcino J. Silva1*
1Departments of Neurobiology, Psychiatry and Psychology, Brain Research Institute,
University of California, Los Angeles, California
2Department of Genetics, University of Ko ¨ln, Ko ¨ln, Germany
Raf kinases are downstream effectors of Ras and
upstream activators of the MEK-ERK cascade. Ras and
MEK-ERK signaling play roles in learning and memory
(L&M) and neural plasticity, but the roles of Raf kinases in
L&M and plasticity are unclear. Among Raf isoforms, B-raf
is preferentially expressed in the brain. To determine
whether B-raf has a role in synaptic plasticity and L&M,
we used the Cre-LoxP gene targeting system to derive
forebrain excitatory neuron B-raf knockout mice. This
conditional knockout resulted in deficits in ERK activation
and hippocampal long-term potentiation (LTP) and impair-
ments in hippocampus-dependent L&M, including spatial
learning and contextual discrimination. Despite the wide-
spread expression of B-raf, this mutation did not disrupt
other forms of L&M, such as cued fear conditioning and
conditioned taste aversion. Our findings demonstrate that
B-raf plays a role in hippocampal ERK activation, synaptic
plasticity, and L&M.
C 2005 Wiley-Liss, Inc.
Key words: MEK-ERK;Ras; Cre-LoxP; conditionalknock-
out water maze; contextual discrimination; fear con-
ditioning; conditioned taste aversion
The Raf family of Ser/Thr kinases was originally
identified as being protooncogenes. In mammalian cells,
there are three Raf isoforms encoded by different genes:
A-, B-, and C-raf (Bonner et al., 1985; Huleihel et al.,
1986; Ikawa et al., 1988). Raf kinases are best character-
ized as the downstream effectors of Ras and upstream
effectors of MEK-ERK (mitogen-activated protein kinase
kinase-extracellular signal regulated kinase) signaling and
are thought to mediate cell proliferation and differentia-
tion. The three Raf kinases differ in their tissue distribu-
tions and biochemical properties, suggesting divergent
physiological functions for each of these isoforms.
There are numerous studies implicating Ras and
MEK-ERK signaling in learning and memory (L&M)
and neural plasticity (Silva et al., 1997; Giese et al., 2001;
Ohno et al., 2001; Adams and Sweatt, 2002; Dhaka et al.,
2003). In recent years, a large body of evidence has dem-
onstrated that Ras signaling pathways are involved in
mechanisms of cognitive function, including transcrip-
tional and translational controls of synaptic plasticity and
memory (Atkins et al., 1998; Wu et al., 1999; Mazzuc-
chelli et al., 2002; Kelly et al., 2003; Kelleher et al., 2004).
However, it is still unknown whether Raf kinases in gen-
eral, and B-raf in particular, play a role in bridging Ras and
MEK-ERK signaling in plasticity and memory. Indeed,
direct evidence for the involvement of Raf kinases in syn-
aptic and behavioral memory is lacking. It is possible that
different Raf isoforms may be involved in different proc-
esses of memory formation.
Among the three Raf isoforms, A-raf is mostly con-
fined to urogenital organs (Storm et al., 1990). C-raf is
ubiquitously expressed in both brain and peripheral tissues
(Storm et al., 1990) and is the most extensively studied
Raf with a defined role in the classic Ras-Raf-MEK-
ERK pathway. B-raf is found predominantly in neural
tissues, testis, and melanocytic and hematopoietic cells
(Eychene et al., 1995). Compared with C-raf, B-raf is a
stronger MEK-ERK activator. It has the higher MEK
binding affinity and kinase activity and is necessary for
long-lasting and sustained ERK activation (Wixler et al.,
1996; Marais et al., 1997; Papin et al., 1998). In B-raf,
Masuo Ohno’s current address is Department of Physiology, Northwestern
University Medical School, Chicago, IL 60611-3008.
Peter Giese’s current address is Wolfson Institute for Biomedical Research,
University College London, London, WC1E 6BT United Kingdom.
Ralf Ku ¨hn’s current address is ARTEMIS Pharmaceuticals GmbH, Ko ¨ln,
*Correspondence to: Alcino J. Silva, Departments of Neurobiology, Psy-
chiatry and Psychology, Brain Research Institute, University of California,
Los Angeles, CA 90095-1761. E-mail: firstname.lastname@example.org
Received 4 August 2005; Revised 28 September 2005; Accepted 28 September
Published online 7 December 2005 in Wiley InterScience (www.
interscience.wiley.com). DOI: 10.1002/jnr.20703
Journal of Neuroscience Research 83:28–38 (2006)
' 2005 Wiley-Liss, Inc.
but not in C-raf, null mutants, MEK activation is com-
promised in embryonic fibroblasts (Wojnowski et al.,
2000; Huser et al., 2001; Mikula et al., 2001), suggesting
that B-raf is the principal Raf kinase in the brain. Taken
together, among the three Raf isoforms, B-raf is most
likely to play an important role in L&M. Although C-raf
is expressed in the soma of neurons, B-raf is present in
both soma and neurites (Morice et al., 1999), supporting
its potential involvement in plasticity. Indeed, B-raf is
among a subset of hippocampal genes whose expression is
triggered by the induction of long-term potentiation
(LTP) and behavioral training (Morris water maze), sug-
gesting a role for this gene in these processes (Thomas
et al., 1994; Richter-Levin et al., 1998).
Furthermore, B-raf has been found to turn on ERK
following cAMP activation (Vossler et al., 1997) and to
modulate the nerve growth factor (NGF)-induced ERK
pathway (Jaiswal et al., 1994). B-raf may be regulated by
Rap1, another signaling molecule that, like Ras, has
GTPase activity and mediates Rap1-dependent ERK
activation invoked by cAMP and Ca2þ(Grewal et al.,
2000; Morozov et al., 2003). These findings suggest that
B-raf may be a critical nodal point for inputs mediated by
Ras and other signaling molecules, such as cAMP. To
determine the role of B-raf in plasticity and memory, we
used a reverse-genetic approach. Because B-raf is essential
for embryonic development (Wojnowski et al., 1997), we
generated a conditional line of forebrain-restricted B-raf
null mutant mice. Our studies demonstrate that B-raf kin-
ase is involved in hippocampal synaptic plasticity and hip-
MATERIALS AND METHODS
Generation of B-raff/f, Tg(acre), and Tg(acre)B-raf–/–
The floxed B-raf mice (B-raff/f) mice were generated by
flanking a 1.2-kb targeting region including exon 12 with two
LoxP sites. R1 embryonic stem cells (ES; 129/Sv background;
Nagy et al., 1993) were transfected with the linearized targeting
construct and screened by Southern blot analyses for correct
homologous recombination. Mutant lines were established
from male chimeras and bred into the 129/SvEms background.
The Tg(acre) mice (FVB/J background) were generated in the
transgenic facility at the State University of New York at Stony
Brook. The Cre gene with a nuclear localization signal (NLS)
was driven by a 8.5-kb fragment of a-CaMKII promoter
(Mayford et al., 1996a,b). The Tg(acre)B-raf–/–mice were F2
progeny from a cross between the B-raff/fand the Tg(acre)
mice. Genotypes of mice were determined by PCR analysis of
tail DNA samples. It is important to note that, in all of the
experiments mentioned here, we only used Tg(acre)B-raf–/–
mice that were at least 6 months old to ensure the completeness
of the Cre-mediated deletion. Expression data suggested
increasing levels of deletion and Cre expression between 2 and
6 months. Consistently, we also found much milder behavioral
and physiological phenotypes at 2 months in the knockout
mice (data not shown).
Western Blot Analysis
Tissue samples were dissected rapidly from decapitated
amounts of protein, determined using a Pierce BCA protein
assay kit (Pierce, Rockford, IL), were subjected to electropho-
resis in 10% SDS-PAGE and then transferred onto nitrocellu-
lose membranes (Bio-Rad, Hercules, CA). Membranes were
blocked before being incubated with the primary antibodies for
2 hr at room temperature. Membranes were washed and incu-
bated with corresponding HRP-conjugated secondary antibod-
ies (1:2,000; Bio-Rad) for 1 hr at room temperature. Signals
were visualized by enhanced chemiluminescence (ECL Plus) as
recommended by the manufacturer (Amersham, Arlington
Heights, IL). The primary antibodies used were against: 1) B-raf
C-terminal (C-19), 2) B-raf N-terminal (F-7), 3) C-raf (1:800;
Santa Cruz Biotechnology, Santa Cruz, CA), 4) the dual phos-
pho-Thr202/Tyr204 ERK p42/p44, 5) total ERK p42/p44
(1:1,000; New England Biolabs, Beverly, MA), and 6) b-actin
(1:3,000; Sigma-Aldrich, St. Louis, MO). For the activity-
induced ERK activation analysis, mice were placed into a condi-
tioning chamber; 120 sec later, they received three foot shocks
(0.75 mA, 2 sec) with a 58-sec intertrial interval (ITI). Then,
mice were allowed to stay in the chamber for another 2 min
before placed back in their home cages. Brain tissue was isolated
30 min after training.
Mice were anesthetized with chloral hydrate and intra-
cardially perfused with saline (0.9%), followed by 4% parafor-
maldehyde in 0.1 M PBS. Brains were removed and postfixed
with the same fixative overnight and cryoprotected with 30%
sucrose in PBS until sectioning (40-lm coronal). For fluores-
cent immunocytochemistry, free-floating sections were blocked
in PBST buffer [0.2% Triton X-100, 5% bovine serum albumin
(BSA), 1% normal goat serum] for 2–4 hr at room temperature,
then incubated with primary antibodies (b-galactosidase; 1:
1,000; ICN Pharmaceuticals, Worthington, OH), or Cre
recombinase (1:800; Covance, Berkeley, CA) in PBST (2.5%
BSA) overnight at 48C. Sections were then washed in PBST
and incubated with fluorescent-conjugated secondary antibod-
ies (Alexa 594 goat anti-rabbit or mouse IgG, 3 lg/ml; Molec-
ular Probes, Eugene, OR) for 2 hr at room temperature.
Finally, sections were washed and mounted with Vectashield
mounting medium containing DAPI (Vector, Burlingame,
CA). Immunoreaction was visualized using a fluorescence
microscope. To determine whether the mutation produced
gross morphological changes, parasagittal sections from adult
mutant and WT brain were stained with cresyl violet. Regular
light microscopic immunocytochemistry was performed with
an antibody against Cre (1:10,000; Covance) with NiDAB
staining, and an antibody against GAD-65/67 (1:3,500; Chemi-
con, Temecula, CA) with DAB staining.
In all behavioral experiments, nonknockout littermates of
the mutants were used as the controls, and the experimenter
was blind to the genotype of the mice studied. Mice were kept
on a 12:12 hr light-dark cycle, and all behavioral experiments
Forebrain-Specific Knockout of B-raf Kinase29
Journal of Neuroscience Research DOI 10.1002/jnr
were always conducted during the light phase of the cycle. All
the behavioral procedures used were approved by the University
of California at Los Angeles (UCLA) Animal Research Com-
mittee. The UCLA Franz Hall animal facility is fully accredited
by the American Association for the Accreditation of Laboratory
Animal Care, and the animals are maintained in accordance with
the Animal Welfare Act and the Department of Health and
Human Services (DHHS) guide. Statistical analyses of behavioral
data were via one-way ANOVAs or two-way ANOVA with
Morris Water Maze
The water maze apparatus and procedures were previ-
ously described (Bourtchuladze et al., 1994). In the hidden-
platform version of this task, the platform location is fixed for
the training period and is submerged under opaque water. Mice
were trained with 2 trials per day (1 min ITI) for 14 days. Spa-
tial L&M was assessed in 60-sec probe tests conducted at the
end of training on days 10 and 14, in which the platform was
removed from the pool. The search pattern of the mice was
recorded by a CCD camera and analyzed by using VHS soft-
ware. The visible version of water maze was conducted after
the hidden-platform test. Mice were trained to search for a
marked platform for 2 days with three trials per day. The plat-
form was placed at a different location in each trial.
The apparatus and procedures used were previously
described (Frankland et al., 1998). The contextual discrimination
experiment consisted of three stages: preexposure (1 day), train-
ing (1 day), and testing (2–3 days). On each day, mice were
placed in chamber A in the morning and chamber B in the after-
noon (or vice versa); order (chamber A vs. B) for each day was
counterbalanced across groups (WT vs. mutant). On day 1, mice
were preexposed to both chambers for a total of 10 min each,
during which time no foot shocks were delivered. On day 2 and
the following testing days, mice were placed in the chambers for
180 sec each day. In chamber A, a foot shock (2 sec, 0.75 mA)
was delivered 150 sec after placement. No foot shock was deliv-
ered in chamber B. The percentage of time the mice spent freez-
ing (absence of all but respiratory movement) during the first
150 sec (preceding the shock delivery in A) in both chamber A
and chamber B was recorded by using automated procedures
described previously (Anagnostaras et al., 2000).
Contextual and Cued Fear Conditioning
Separate groups of mice were placed in the conditioning
chamber for 2 min before the onset of the 30-sec tone CS
(2,800 Hz and 85 dB). In the last 2 sec of the CS, the US (a
foot shock of 0.75 mA, 2 sec) was delivered. Mice were left in
the conditioning chamber for an additional 2.5 min and were
replaced back in the home cage. A test of contextual condition-
ing was performed 24 hr after training, in which the percentage
of time mice spent freezing in the context in which they were
previously shocked was assessed. Cued conditioning was tested
48 hr after training. Mice were placed in a novel context for
2 min (pre-CS test), after which they were exposed to the CS
for 3 min (CS test). In all experiments described here, delivery
of the CS and US and measurements of freezing were done with
an automated procedure described elsewhere (Anagnostaras
et al., 2000). In the slow-acquisition contextual conditioning
paradigm, mice were given a foot shock (0.40 mA, 1 sec) 4 min
after being placed into the chamber. Then, they were allowed to
stay in the chamber for an additional 2 min before being placed
back in their home cages. The same procedure was carried out
once per day for 4 days. We measured freezing during the first
4 min before the onset of the shock. Extinction was tested on
days 5–8. Each day, mice were placed in the same chamber for
30 min with no shock delivered.
Conditioned Taste Aversion
On days 1–4 of the experiment, mice were given limited
access to water, and conditioning was carried out on day 5.
Forty minutes following access to saccharin (0.1%, novel taste,
for 15 min), mice received an IP injection of either lithium
chloride (LiCl, 1.0 M, 2% of body weight, as malaise-inducing
agent) or PBS as a control. Twenty-four hours later, mice were
tested for conditioned taste aversion (CTA). They were pre-
sented with two bottles, one containing water and the other
saccharin. The placement of water/saccharin bottles during
conditioning and testing was pseudorandom. The bottles were
weighed before and after the test, and the amount of liquid con-
sumed was calculated. An aversion index (¼amount of saccharin
consumed/total amount of fluid consumed) was used to report
CTA. A low score in the aversion index indicates learning.
All electrophysiological experiments were performed
blind to genotype. Transverse hippocampal slices (400 lm
thick) were maintained in a submerged recording chamber per-
fused with artificial cerebrospinal fluid (ACSF) equilibrated
with 95% O2 and 5% CO2 at 308C. The ACSF contained
120 mM NaCl, 3.5 mM KCl, 2.5 mM CaCl2, 1.3 mM
MgSO4, 1.25 mM NaH2PO4, 26 mM NaHCO3, and 10 mM
D-glucose. Extracellular field excitatory postsynaptic potentials
(EPSPs) were recorded with a metallic electrode from the stra-
tum radiatum layer of the area CA1, and the Schaffer collaterals
were stimulated with 100-lsec test pulses via a bipolar elec-
trode. For synaptic input–output curves, different stimulation
strengths (20–100 mA in steps of 10 mA) were applied. In the
following experiments, the intensity of stimulation was adjusted
to give field EPSP approximately 33% of maximum. Paired-
pulse facilitation was tested by measuring the percentage
increase in the slope of the second field EPSP in relation to the
first one (20-, 50-, 100-, 200-, and 500-msec interpulse inter-
vals). LTP was induced by a single tetanus delivered according
to a high-frequency protocol (100 Hz, 1 sec) or a 2-theta-burst
tetanus stimulation protocol (each burst consists of four pulses
at 100 Hz with a 200-msec interburst interval). Before tetaniza-
tion, the responses were monitored for at least for 20 min to
ensure a stable baseline EPSP slope. To determine whether the
magnitude of LTP differed significantly among the groups,
responses from the last 10-min block of recordings (40–50 min
for 2-theta-burst LTP; 80–90 min for 100-Hz LTP) were com-
pared with a one-way ANOVA.
30Chen et al.
Journal of Neuroscience Research DOI 10.1002/jnr
Generation of Conditional B-raf Knockout Mice
Developmental deficits can confound the interpreta-
tion of knockout studies of adult brain function. B-raf is
known to be involved in cell survival during development
and mice with a homozygous global B-raf disruption are
not viable (Wojnowski et al., 1997). To circumvent these
and other related concerns, we used the Cre-LoxP system
to generate a conditional B-raf knockout. Two mouse
lines were generated for this strategy: a B-raff/f‘‘floxed’’
line and a Tg(acre) transgenic ‘‘deleter’’ line expressing Cre
recombinase specifically in postnatal forebrain (Fig. 1A,B).
By crossing these two lines of mice, we were able to derive
a homozygous brain-specific B-raf knockout restricted to
excitatory neurons of the forebrain.
We floxed exon 12, which is the first exon that enc-
odes the B-raf kinase domain. It also contains a critical
GXGXXG motif that is common to all Ser/Thr kinases
(Bossemeyer, 1994). The deletion of exon 12 led to a
shift in the open reading frame (ORF) and resulted in a
null mutation of B-raf, in that no protein was detected by
antibodies against either the N-terminal or the C-terminal
portion of the protein in Western blot analysis (data not
shown). Western blot analysis also demonstrated that the
insertion of loxP sites per se did not affect B-raf expres-
sion, insofar as the B-raf protein level was normal in
B-raff/fmice (data not shown).
To achieve regional and cell-type specificity, we
derived Tg(acre)B-raf–/–mice with the Cre recombinase
gene under the regulation of a postnatally active a-CaM-
KII promoter (Mayford et al., 1996a,b). As previously
reported by others (Tsien et al., 1996), the Cre transgene
driven by the a-CaMKII promoter is expressed specifically
in forebrain postnatal excitatory neurons. Immunocyto-
chemical analysis of line Tg(acre)1557 brain slices with an
anti-Cre antibody showed that Cre expression is confined
to forebrain structures, i.e., cortex, striatum, hippocampus,
amygdala, but absent in other brain regions, such as mid-
brain and cerebellum (Fig. 2A, and data not shown).
Within hippocampal structures, Cre is highly expressed in
the CA1 region and less so in CA3 and dentate gyrus
of line Tg(acre)1557. Furthermore, immunohistochemical
staining showed that Cre is expressed in cells negative for
both glutamic acid decarboxylase (GAD) and glial fibrillary
acidic protein (GFAP), which suggests that Cre is specifi-
cally expressed in excitatory neurons (Fig. 2C–F). This
deleter line was used in subsequent experiments.
To examine Cre-mediated gene deletion in brain,
Tg(acre)1557 mice were crossed to a lacZ reporter mouse
line (cAct-XstopXlacZ; Tsien et al., 1996; Zinyk et al.,
1998). This reporter line carries a transgene in which the
chicken actin promoter drives the expression of a lacZ
gene, whose transcription and translation are prevented
by a stop site floxed by LoxP sites. Cre-mediated recom-
bination deletes the stop site and activates lacZ gene
expression. In mice that possess both the Cre transgene
and the lacZ reporter, the lacZ gene is activated in fore-
brain as a result of Cre-loxP recombination. In agreement
with the Cre expression pattern, immunocytochemical
analysis with a b-galactosidase antibody showed that Cre-
dependent lacZ expression occurred in forebrain, but not
in other brain structures, such as cerebellum and midbrain
(Fig. 2B, and data not shown).
Fig. 1. Generation of the Tg(acre)B-raf–/–knockout (ko) mice.
A: Scheme of the targeting construct. A 1.2-kb fragment containing
exon 12 of B-raf was floxed by loxP sites. A neo gene was used as an ES
cell selection marker and deleted in a late stage of ES cell culture. H,
HindIII; R, EcoRI; S, StuI. B: Scheme of the cross between the floxed
mice (B-raff/f) and Tg(acre) transgenic mice to produce the forebrain spe-
cific B-raf knockout [F2 homozygous Tg(acre)B-raf–/–]. C: PCR analysis
of B-raf deletion in microdissected tissues of the Tg(acre)B-raf–/–knock-
out mouse. The upper band (413 bp) results from the floxed B-raf gene
and the lower band (282 bp) from Cre-mediated deletion. OB, olfactory
bulb; MC, motor cortex; SC, sensory cortex; AMY, amygdala; DG, den-
tate gyrus; ST, striatum; VC, visual cortex; AC, auditory cortex; HTH,
hypothalamus; TH, thalamus; CERE, cerebellum. D: Western blot anal-
ysis of microdissected brain areas from the Tg(acre)B-raf–/–knockouts
with antibodies against B-raf, compared with that from the their non-ko
littermates (B-raff/f). The percentage of B-raf protein level of their non-
ko counterparts is plotted and shown at left. The right panel shows a
Western blot of different brain regions from the ko and control mice (f/f)
with antibodies against B-raf (95 kd) and C-raf (74 kd). b-Actin signal
was used as loading control. E: Western blot analyses of ERK1/2 phos-
phorylation before and after training (conditioning). Mice were trained
in a conditioning chamber with three foot shocks (0.75 mA, 2 sec, 1 min
ITI). Samples from different brain regions were extracted 30 min after
training and subjected to Western blot analysis by using an antibody
against dual phospho-ERK p44/p42. The same membrane was also blot-
ted with an antibody for total ERK1/2 proteins (lower panel). f/f,
B-raff/f; ko, Tg(acre)B-raf–/–; Hip, hippocampus; Str, striatum; Cx, cor-
tex; Cere, cerebellum.
Forebrain-Specific Knockout of B-raf Kinase31
Journal of Neuroscience Research DOI 10.1002/jnr
Homozygous Tg(acre)B-raf–/–knockout mice were
obtained from a F2 cross between the B-raff/fmice and
Tg(acre)1557 mice. A PCR analysis of DNA from micro-
dissected different brain subregions revealed that the dele-
tion took place exclusively in the forebrain, including corti-
cal, hippocampal, and amygdala regions, but not in cerebel-
lum, thalamus, and hypothalamus (Fig. 1C). Western blot
analysis on microdissected brain regions also showed that
B-raf protein level is decreased in forebrain structures but
not in midbrain and cerebellum areas. If anything, it
appears that B-raf protein levels may be slightly increased in
these other areas (Fig. 1D). Among hippocampal areas,
B-raf is largely deleted in CA1 region and less so in other
regions. It is important to note that, in all of the behavioral
and electrophysiological experiments in this report, we used
only mice that were at least 6 months old to ensure the
completeness and stability of the Cre-mediated deletion of
mice used in the experiments described here.
Disruption of B-raf Led to a Down-Regulation
of the ERK Pathway
To examine whether the deletion of B-raf results in
intracellular signaling deficits, we analyzed the activation
of signaling proteins. Western blot analysis with antibod-
ies against the phosphorylated ERK p42/p44 and total
ERK protein was used to analyze ERK activity. Our
results showed that the expression of total ERK protein
was not changed in the Tg(acre)B-raf–/–mutants com-
pared with the control animals (b-actin blotting as load-
ing control; data not shown). Interestingly, the basal levels
of ERK phosphorylation were also not detectably altered
in the mutants (Fig. 1E basal, and data not shown), per-
haps because B-raf was deleted only in excitatory neu-
rons, which constitute a small portion of the tissue sample
examined. We observed that the basal level of ERK phos-
phorylation in hippocampus is much higher than in most
other brain regions and organs (data not shown), which
suggests that this kinase is normally engaged in hippocam-
Although ERK activity is low under basal physio-
logical conditions, behavior training can elevate ERK
phosphorylation (Atkins et al., 1998; Blum et al., 1999).
Hence, contextual conditioning (with three mild foot
shocks) resulted in an increase in ERK phosphorylation
levels in several brain regions 30 min after training.
Importantly, in forebrain structures, but not in cerebel-
lum, this increase in ERK phosphorylation was clearly
suppressed in Tg(acre)B-raf–/–mutants but not in controls
(Fig. 1E, P-ERK, posttraining). This result confirms that
ERK signaling is activated after behavioral training and
indicates that this activation is mediated by B-raf.
Deletion of B-raf Severely Disrupted Spatial
To examine whether B-raf is involved in learning
and memory, we studied the impact of our conditional
B-raf mutation on several learning tasks, some of which are
known to be dependent on hippocampal function. In our
behavioral tests, B-raff/fmice were used as the control for
the Tg(acre)B-raf–/–mice. To exclude the possible impact
of each of the single transgenes on learning, we tested the
Tg(acre)1557 mice, the B-raff/fmice, and their respective
wild-type littermates in the tasks described below. Our
results demonstrated that each of the transgenes alone does
not affect learning and memory (data not shown).
The Morris water maze is a sensitive and reliable
measure of hippocampus-dependent learning (Brandeis
et al., 1989). In this test, mice are placed in a pool with
opaque water and must swim to a submerged platform to
escape the water. In the hidden-platform version of the
water maze test, the escape platform is located in a fixed
location in the pool, and there are no objects directly
indicating the platform location. Mice were trained on
this task for 14 days with 2 trials per day. The results
show that Tg(acre)B-raf–/–mutant mice performed poorly
compared with their B-raff/fcontrol littermates. Whereas
the controls (n ¼ 14) improved during training (F13,169¼
Fig. 2. Cre recombinase expression and activity in forebrain excitatory
neurons. A,B: Fluorescent microscopy of Cre expression in Tg(acre)
mice (A, immunostained with an antibody against Cre) and Cre-medi-
ated gene deletion in Tg(acre)LacZ mice (B, immunostained with an
anti-b-galactosidase antibody). CA1 and cortex regions are shown.
LacZ expression indicated Cre activity. C,D: Cre expression was not
colocalized with GAD expression. C: NiDAB staining of Cre expres-
sion in CA1. D: Double staining of Cre (black nuclear staining) and
GAD (brown cell body and fiber staining; arrows). E,F: Cre expres-
sion was not colocalized with GFAP expression. Brain slices were
immunostained with antibodies against Cre and GFAP that were
visualized with different fluorescent-conjugated secondary antibodies.
Cre is shown as red and GFAP as green; the same field is shown in E
and F. Sacle bar ¼ 250 lm in A,B; 40 lM in C,D; 125 lM in E,F.
32Chen et al.
Journal of Neuroscience Research DOI 10.1002/jnr
3.64, P < 0.05), mutants (n ¼ 14) showed poor acquisi-
tion of the task (F13,169¼ 1.53, P > 0.05). The acquisi-
tion for mutants is significantly slower than that for con-
trols (F1,26¼ 5.80, P < 0.05; Fig. 3A). Because the time
to find the platform during training is known to be a poor
measure of spatial learning (Brandeis et al., 1989), we
used probe trials to evaluate the performance of the mice.
In these probe tests, the platform was removed and the
mice were given 60 sec to search for it. Probe tests were
given at days 10 and 14 of training. In the probe trial at
day 10, B-raff/fcontrols spent a significantly higher per-
centage of time searching within the target quadrant
compared with other quadrants (F3,39¼ 7.12, P < 0.05),
but the Tg(acre)B-raf–/–mutants did not (F3,39¼ 0.63,
P > 0.05; Fig. 3B). Similarly, analysis of the average
proximity to the exact platform location during the probe
trial (Gallagher et al., 1993) showed that, whereas control
mice searched significantly closer to the target platform
position than to other, nontarget positions (F3,39¼ 13.0,
P < 0.05), the mutant mice showed no preference for the
target location (F3,39¼ 1.13, P > 0.05; Fig. 3C), con-
firming the deficit of the mutants.
To determine whether abnormalities in motivation,
motor coordination, or vision account for the deficits in
spatial learning, the same animals were tested in the visible
platform of the water maze, a hippocampus-independent
task in which mice have to locate a platform marked with
a visible cue (Cho et al., 1999). The time taken to reach
the visible platform was not different between the groups
(F1,19¼ 1.40, P > 0.05; Fig. 4A), and no differences in
swimming speed, floating, or thigmotaxic behavior were
observed between the Tg(acre)B-raf–/–mutants and the
B-raff/fcontrols (day 10 probe trial shown in Fig. 4D).
Additionally, knockout animals also did not show deficits
in either motor coordination in a rota-rod test (F4,80¼
0.14, P > 0.05; Fig. 4B) or exploratory activity in an
open-field test (F1,21¼ 0.41, P > 0.05; Fig. 4C). These
data suggest that the poor performance of the conditional
knockouts in the hidden-platform version of the water
maze was due to deficient spatial L&M, rather than
impaired motivation, motor coordination, or vision.
B-raf Deletion Impaired Contextual Discrimination
We used the contextual discrimination task to address
the possibility that the water maze deficits of the mutants
are due to abnormal hippocampal function. Contextual
discrimination tests a mouse’s ability to distinguish
between two similar contexts, and, as with the water maze,
it is very sensitive to hippocampal lesions (Frankland et al.,
Fig. 3. Impaired spatial learning in conditional B-raf knockouts. The
Tg(acre)B-raf–/–knockouts and B-raff/flittermate controls were trained
for 14 days with two trials per day in the Morris water maze. The
average latency to reach the hidden platform is plotted for each day in
A. Results from the probe trial at day 10 are shown in B (% time spent
in each quadrant) and C (proximity to the platform). TQ, target quad-
rant. The other three bars represent the results for the adjacent left,
right, and opposite quadrants (in that order).
Fig. 4. Normal performance in the visible water maze, rota-rod, and
open-field tests in conditional B-raf knockouts. A: After the hidden-
platform water maze, mice were tested in the visible-platform version
of this test for 2 days, with three trials per day. Average latencies to
reach the platform are plotted. B: Rota-rod test for motor coordina-
tion. Mice were tested with 4–40 rpm accelerating speeds (maximal
duration 300 sec) for five trials with 30-min intertrial interval. Average
duration for each group is presented. C: Open-field test for explora-
tory activity. Mice were allowed to explore an open field for 5 min
(diameter 1 m). Percentage time animals spent exploring the innermost
50% vs. outmost 50% areas of the field is shown. D: Performance
measures from the water maze day 10 probe trial are presented: path
length (path; cm), swimming speed (speed; cm/sec), percentage time
spent swimming slower than 5 cm/sec (%slow), and percentage time
spent hugging the outmost 10% area of the pool (%hug). The F values
of one-way ANOVA analysis for each group are shown. None of the
measures above are significantly different between mutants and con-
trols (P > 0.05).
Forebrain-Specific Knockout of B-raf Kinase33
Journal of Neuroscience Research DOI 10.1002/jnr
1998). In this test, animals are required to discriminate
between two similar chambers, one in which they receive
mild foot shocks (chamber A) and the other in which they
do not (chamber B). The two contexts consisted of both
unique and common cues. Contextual discrimination is
assessed by measuring the time mice spent freezing (i.e.,
the absence of bodily movements aside from respiration) in
each chamber. The data in Figure 5A demonstrate a clear
impairment of the knockout mice in the contextual dis-
crimination task. The B-raff/fcontrol mice (n ¼ 19)
learned to discriminate between the two contexts. On test
day 2, they showed significantly less freezing in the no-
shock context, B (46.5% 6 7.2%), than in the shock con-
text, A (70.1% 6 4.7%; F1,18¼ 20.3, P < 0.05). In con-
trast, Tg(acre)B-raf–/–mutants (n ¼ 13) responded similarly
to both contexts (i.e., 80.7% 6 8.3% freezing in context A
vs. 76.3% 6 5.7% in context B; F1,12¼ 0.44, P > 0.05;
test day 2).
It is noteworthy that, although the Tg(acre)B-raf–/–
mice could not discriminate between contexts, shock reac-
tivity and freezing responses were normal in these mutants.
For example, when the mice were brought back to con-
text A 1 day after being given the foot shock (test day 1 in
context A), knockout mice and their control littermates
exhibited comparable levels of contextual conditioning
(F1,30¼ 0.02, P > 0.05; Fig. 5B). Additionally, in a cued
conditioning test with naı ¨ve mice (one tone-shock pairing;
0.5 mA, 2 sec), Tg(acre)B-raf–/–mice showed normal lev-
els of response compared with wild-type controls (F1,17¼
0.45, P > 0.05; Fig. 5C). This result is consistent with pre-
vious findings indicating that the hippocampus is not
always essential for contextual conditioning (Frankland
et al., 1998), and it also suggests that amygdala function is
intact in the knockouts. Importantly, contextual discrimi-
nation is more sensitive to hippocampal lesions than con-
textual conditioning (Frankland et al., 1998), a result con-
sistent with deficient hippocampal function in the mutant
B-raf Deletion Did Not Disrupt the Formation
and Extinction of Contextual Conditioning
To test whether learning is generally affected in the
Tg(acre)B-raf–/–knockouts, we tested them in a sensitive
fear-conditioning protocol. In this procedure, animals
received a very mild foot shock (0.4 mA, 1 sec) every day
for 4 days. The mice exhibited gradually increasing freez-
ing responses. As shown in Figure 6A, both the Tg(acre)
B-raf–/–knockouts (n ¼ 12) and the controls (n ¼ 11)
showed low baseline freezing and gradually higher freez-
ing responses over the 4 training/test days. Consistently
with the results presented above, we found no differences
between the freezing responses of the mutants and con-
trols (F1,21¼ 0.25, P > 0.05). This gradual increase in
freezing responses over the 4 days of training confirmed
that the lack of differences between mutants and controls
was not due to a ceiling effect. It is important to note
that, although acquisition of contextual discrimination is
Fig. 5. Impaired contextual discrimination in conditional B-raf
knockouts. A: Contextual discrimination. Percentage time the mice
spent freezing in chamber A (shock context, solid squares) vs. chamber
B (no-shock context, open squares) is shown for each of the 3 test
days. *Significant difference, P < 0.05; T1–3, test days 1–3. B: Con-
textual conditioning in chamber A measured 24 hr after the mice were
initially shocked. C: Cued conditioning test with naı ¨ve mice. Mice
were trained with a shock-tone pairing (0.5 mA, 2 sec) and tested for
cued freezing 24 hr later.
Fig. 6. Normal acquisition and extinction of contextual conditioning
as well as intact conditioned taste aversion in conditional B-raf knock-
outs. A: Slow acquisition of contextual conditioning. Mice were
trained with a milder (0.4 mA, 1 sec) foot shock each day for 4 days.
Average freezing during the first 4 min before the onset of shock is
presented (no shock on day 5). B: Extinction of contextual condition-
ing. Starting upon completion of the fifth day of acquisition, mice
were allowed to stay in the conditioning context for 30 min (no shock
was delivered). Extinction was continued on days 5–8. Average freez-
ing during the first 4 min in each day is presented. C: Conditioned
taste aversion. Aversion index is shown for the B-raf mutants and their
littermate controls (B-raff/f) treated with malaise-inducing LiCl. PBS
was used as a treatment control. Aversion index ¼ saccharin con-
sumed/total water and saccharin consumed.
34 Chen et al.
Journal of Neuroscience Research DOI 10.1002/jnr
always dependent on hippocampal strategies, contextual
conditioning is not, insofar as mice can use nonhippo-
campal strategies in this task (Frankland et al., 1998).
To test whether the B-raf disruption affected extinc-
tion, the same group of contextually conditioned mice
was reexposed to the conditioning context for 30 min
each day for 4 days. The results show that the rate of
extinction of freezing responses was also normal in the
knockouts (F1,21¼ 0.02, P > 0.05; Fig. 6B).
B-raf Knockouts Showed Normal CTA Task
Since tone and contextual conditioning was unal-
tered by the deletion of B-raf, it is possible that the fore-
brain excitatory neuron-specific B-raf knockout that we
derived does not affect either amygdala- or cortex-
dependent learning. To test this possibility further, we
examined the mutant mice in the CTA test, a paradigm
known to be dependent on amygdalar and cortical func-
tion, but not on the hippocampus (Welzl et al., 2001).
Previous studies in rats suggest that novel tastes activate
ERK activity and that MEK inhibition can block CTA
(Berman et al., 1998). In the CTA task, mice were trained
to associate saccharine flavored water with an IP injection
of malaise-inducing LiCl. The aversion index (I ¼ saccha-
rin consumed/total fluid consumed) reflects a mouse’s
choice between saccharin and water after conditioning. A
low index indicates good learning, in that it reflects a
strong association between saccharin and LiCl. Consistently
with a specific role for B-raf in hippocampal-dependent
learning, we found that B-raf conditional knockouts did
not disrupt the acquisition of CTA (Fig. 6C).
As expected, both knockout (n ¼ 3, I ¼ 0.67 6
0.08) and control mice (n ¼ 4, I ¼ 0.71 6 0.09) injected
with PBS (instead of LiCl) showed no avoidance of saccha-
rin (Fig. 6C, PBS treatment). In contrast, both Tg(acre)
B-raf–/–(n ¼ 8, I ¼ 0.37 6 0.07) and control mice (n ¼
13, I ¼ 0.39 6 0.03) injected with LiCl exhibited an
equally strong avoidance of saccharin (F1,19¼ 0.40, P >
0.05; Fig. 6C, LiCl treatment). Taken together, the behav-
ioral data presented above indicate that our B-raf mutation
specifically disrupts hippocampus-dependent learning.
Hippocampal CA1 LTP Is Impaired in B-raf
Because the B-raf mutants showed clear hippocam-
pus-dependent learning deficits, we determined whether
these mutations also disrupted hippocampal LTP, an
experimental model of the synaptic changes thought to
underlie L&M. Studies in aplysia demonstrated the
involvement of MAPK signaling in later stages of long-
term synaptic facilitation (Martin et al., 1997), and a num-
ber of mammalian studies implicated MAPK signaling in
early stages of LTP. For example, previous studies reported
ERK activation during LTP induction in various brain
structures, including the hippocampal CA1 region, and
inhibition of the ERK pathway results in impairments in
the early phase of LTP (e-LTP; English and Sweatt, 1996,
1997). Thus, we focused our LTP studies on e-LTP.
The relationship between evoked fiber volleys and
field EPSPs (fEPSPs) was indistinguishable between the
Tg(acre)B-raf–/–mutants (n ¼ 12) and their B-raff/fcontrols
(n ¼ 13; Fig. 7A). Additionally, paired-pulse facilitation
was also indistinguishable between the knockouts (n ¼ 11)
and the control mice (n ¼ 12), suggesting that this fore-
brain- and neuron-specific knockout did not disrupt either
synaptic transmission or presynaptic plasticity (Fig. 7B).
Altogether, these results indicate that the conditional muta-
tion of the B-raf gene does not compromise the general
health of hippocampal slices, synaptic transmission, or cer-
tain measures of presynaptic plasticity.
e-LTP was tested with extracellular field recordings
in stratum radiatum of hippocampal slices and was studied
with a 2-theta-burst tetanus, because this pattern of stimula-
tion is thought to reflect normal hippocampal patterns of
activation during learning (Ranck, 1973; Otto et al., 1991).
Our results demonstrate that e-LTP was deficient in the
mutants. Between 40 and 50 min after the tetanus, the
mutants (n ¼ 6 mice) showed 123.8% 6 6.5% potentia-
tion, whereas control mice (n ¼ 6 mice) showed 143.6%
6 3.8% (F1,10¼ 6.21, P < 0.05; Fig. 7C). Furthermore,
LTP induced with a 100-Hz (1-sec) tetanus was also
impaired in the mutants. Between 70 and 80 min after the
tetanus, the knockouts (n ¼ 9 mice) showed 128.1% 6
Fig. 7. Impaired hippocampal CA1 LTP in conditional B-raf knock-
outs. A: The B-raf mutants showed normal input–output curves. Field
EPSP slopes are plotted with respect to their presynaptic fiber volleys
induced by different strengths of stimulation (20–100 lA in steps of
10 lA). B: The mutants also showed normal paired-pulse facilitation.
Percentage facilitations, calculated as the ratio of the slopes of the sec-
ond vs. the first field EPSPs, are plotted for interpulse intervals of 20–
500 msec. C,D: LTP induced by either a 2-theta-bursts (C) or by a
100-Hz (D) tetanus. Each point indicates the field EPSP slope (mean
6 se) normalized to the average baseline response before the tetanus
delivered at time 0; traces (control, left; mutant, right) are the average
of fEPSPs recorded during baseline and 40–50 min (for 2-theta-burst
LTP) or 70–80 min (for 100-Hz LTP) after tetanization. Calibration
bars ¼ 10 msec, 0.5 mV.
Forebrain-Specific Knockout of B-raf Kinase35
Journal of Neuroscience Research DOI 10.1002/jnr
6.7% potentiation, and the control mice (n ¼ 10 mice)
showed 150.1% 6 7.5% (F1,17¼ 4.6, P < 0.05; Fig. 7D).
In this study, we used forebrain conditional gene
targeting to investigate the involvement of the B-raf kin-
ase in synaptic plasticity and L&M. Our results demon-
strate that, in the adult brain, B-raf is required for hippo-
campal synaptic plasticity and for hippocampus-dependent
learning. Our findings also suggest that B-raf dependent
activation of the ERK pathway in excitatory neurons is
required for LTP and learning.
Our biochemical studies show that the region-
restricted disruption of B-raf affects the levels of ERK acti-
vation in hippocampal neurons. Accordingly, just as phar-
macological disruptions of ERK activation interfere with
LTP (English and Sweatt, 1997), the B-raf null mutation
in excitatory CA1 pyramidal neurons also impairs LTP.
These results are consistent with the hypothesis that B-raf
kinase is one of the upstream regulators of ERK activation
and that this is important for synaptic plasticity. Basal syn-
aptic transmission and paired-pulse facilitation, however,
were normal in the B-raf mutants, suggesting that the loss
of B-raf affects only certain aspects of synaptic function.
Furthermore, LTP induced with different stimulation pro-
tocols were all affected in the conditional mutants, suggest-
ing that B-raf is an universal component of LTP in the
CA1 region of the hippocampus.
Consistently with a role for B-raf in hippocampal
function, we found that B-raf knockout mice were
impaired in contextual discrimination. In contrast, contex-
tual and cued conditionings were normal in these mutants.
Previous studies suggest that the contextual discrimination
task is a more faithful measure of hippocampal function
than contextual fear conditioning (Frankland et al., 1998),
in that animals with hippocampal deficits can use nonhip-
pocampal strategies to recognize the conditioning context.
Instead of creating an integrated polymodal representation
of context, they are thought to make several independent
associations between shock and different cues in the con-
text. The B-raf mutants showed seemingly normal freezing
responses in the context in which they were trained, but,
unlike their controls, they were unable to distinguish
between similar contexts. These results strongly suggest
that the disruption of B-raf specifically affects a hippocam-
pal component of fear memory formation.
Cre-mediated disruption of B-raf protein was not
restricted to the hippocampus. Previous studies have sug-
gested that the inhibition of ERK activation affects neuro-
nal function not only in the hippocampus but also in the
cortex, striatum, and amygdala. For example, fear condi-
tioning activates ERK in the amygdala, and amygdala-spe-
cific inhibition of ERK activation also blocks fear condi-
tioning (Schafe et al., 2000). Loss of ERK1 leads to
enhanced locomotor activity and enhanced active-avoid-
ance L&M (Mazzucchelli et al., 2002) but does not affect
fear conditioning or passive avoidance (Selcher et al.,
2001). Microinjection of a MEK inhibitor into the insular
cortex decreases ERK phosphorylation and impairs CTA
(Berman et al., 1998), and infusion of a MEK inhibitor
into medial prefrontal cortex no later than 4 hr after
extinction training resulted in the return of conditioned
fear response (Hugues et al., 2004). However, our studies
show that the forebrain-specific B-raf disruption does not
affect either acquisition or extinction of fear conditioning,
nor does it disrupt CTA. These learning tasks are known
to be sensitive to functional lesions of the amygdala and
cortex. Therefore, these data demonstrate that B-raf is crit-
ical to L&M processes in the hippocampus but not amyg-
dala or cortex. However, our findings do not contradict
previous findings, insofar as these other manipulations tar-
geted different subsets of cells and pathways. The apparent
lack of an effect of the B-raf disruption on cortex and
amygdala does not necessarily indicate that B-raf is not
involved in amygdala and cortical function. For example,
it is possible that B-raf activity in inhibitory neurons (but
not in excitatory neurons, where the kinase was deleted in
the experiments described here) is crucial for cortical and
amygdala L&M. It is also possible that other pathways
compensate for the loss of B-raf in the amygdala and cor-
tex, but not in the hippocampus. Furthermore, previous
studies demonstrated a specific activation of phosphatidyli-
nositol 3-kinase (PI-3 kinase) in the amygdala after LTP
induction and after fear conditioning (Lin et al., 2001).
Importantly, PI-3 kinase inhibitors blocked ERK activa-
tion, LTP, and fear conditioning, suggesting that PI-3 kin-
ase-dependent activation of the ERK pathway is critical
for amygdalar synaptic plasticity and learning. Neverthe-
less, it is also possible that the specificity of our behavioral
results reflects evolutionary specializations, with different
variants of signaling pathways subserving unique functional
requirements of individual brain regions.
What signaling processes activate the B-raf/ERK
cascade during LTP and learning in the hippocampus?
Ras is likely a strong upstream regulator, in that it has
been shown that defective Ras-ERK signaling leads to
learning deficits (Brambilla et al., 1997; Silva et al., 1997;
Giese et al., 2001; Ohno et al., 2001). However, it is
unknown whether Ras signals through B-raf during
learning. Interestingly, the loss of neurofibromin (the
product of the Nf1 gene), a negative Ras regulator, causes
synaptic plasticity and learning deficits, perhaps because of
an increase in g-aminobutyric acid (GABA)-mediated
inhibition (Costa and Silva, 2002). Interestingly, although
K- and N-Ras mutations were able to rescue the synaptic
dysfunction and learning deficits of the Nf1 mutation
(Costa and Silva, 2002), the B-raf mutation was not
(Chen and Silva, unpublished results). This raises the pos-
sibility that neurofibromin and B-raf may be important
for regulating Ras signaling in inhibitory and excitatory
Previous studies of L&M have demonstrated a uni-
versal role for cAMP-dependent signaling (Bailey et al.,
1996). It is important to note that Rap1 couples cAMP
signaling to ERK in synaptic plasticity and L&M (Moro-
zov et al., 2003) and that Rap1 can regulate B-raf activity.
It is, therefore, possible that B-raf contributes to LTP and
36 Chen et al.
Journal of Neuroscience Research DOI 10.1002/jnr
learning by mediating the activation of ERK in a cAMP/
Rap1-dependent manner (Vossler et al., 1997; Dugan
et al., 1999; Sweatt, 2001). Although further investigation
is required, it is also possible that B-raf is a critical nodal
point between Ras- and cAMP-dependent signaling in
the ERK cascade during synaptic plasticity and learning.
Our results demonstrate an important role for the B-raf
kinase in hippocampal synaptic plasticity and L&M. They
also suggest that, despite the widespread conservation of
mechanisms of L&M among brain structures and species,
there are molecular specializations that reflect the unique
functional requirements of each memory system.
We thank L. Kaczmarek, R.M. Costa, P.W. Frank-
land, S. Josselyn, G. Murphy, A. Matynia, and S. Kushner
for discussions that helped to shape the manuscript. We
thank T. Rosenquist for producing transgenic mice. This
work was funded by a grant from the NIH (RO1NS38480)
Adams JP, Sweatt JD. 2002. Molecular psychology: roles for the ERK
MAP kinase cascade in memory. Annu Rev Pharmacol Toxicol 42:135–
Anagnostaras SG, Josselyn SA, Frankland PW, Silva AJ. 2000. Computer-
assisted behavioral assessment of Pavlovian fear conditioning in mice.
Learn Mem 7:58–72.
Atkins CM, Selcher JC, Petraitis JJ, Trzaskos JM, Sweatt JD. 1998. The
MAPK cascade is required for mammalian associative learning. Nat Neu-
Bailey CH, Bartsch D, Kandel ER. 1996. Toward a molecular definition of
long-term memory storage. Proc Natl Acad Sci U S A 93:13445–13452.
Berman DE, Hazvi S, Rosenblum K, Seger R, Dudai Y. 1998. Specific
and differential activation of mitogen-activated protein kinase cascades
by unfamiliar taste in the insular cortex of the behaving rat. J Neurosci
Blum S, Moore AN, Adams F, Dash PK. 1999. A mitogen-activated pro-
tein kinase cascade in the CA1/CA2 subfield of the dorsal hippocampus
is essential for long-term spatial memory. J Neurosci 19:3535–3544.
Bonner TI, Kerby SB, Sutrave P, Gunnell MA, Mark G, Rapp UR. 1985.
Structure and biological activity of human homologs of the raf/mil
oncogene. Mol Cell Biol 5:1400–1407.
Bossemeyer D. 1994. The glycine-rich sequence of protein kinases: a mul-
tifunctional element. Trends Biochem Sci 19:201–205.
Bourtchuladze R, Frenguelli B, Blendy J, Cioffi D, Schutz G, Silva AJ.
1994. Deficient long-term memory in mice with a targeted disruption of
the cAMP-responsive element-binding protein. Cell 79:59–68.
Brambilla R, Gnesutta N, Minichiello L, White G, Roylance AJ, Herron
CE, Ramsey M, Wolfer DP, Cestari V, Rossi-Arnaud C, Grant SG,
Chapman PF, Lipp HP, Sturani E, Klein R. 1997. A role for the Ras sig-
nalling pathway in synaptic transmission and long- term memory. Nature
Brandeis R, Brandys Y, Yehuda S. 1989. The use of the Morris Water
Maze in the study of memory and learning. Int J Neurosci 48:29–69.
Cho YH, Friedman E, Silva AJ. 1999. Ibotenate lesions of the hippocam-
pus impair spatial learning but not contextual fear conditioning in mice.
Behav Brain Res 98:77–87.
Costa RM, Silva AJ. 2002. Molecular and cellular mechanisms underlying
the cognitive deficits associated with neurofibromatosis 1. J Child Neurol
17:622–626 [discussion 627–629, 646–651].
Dhaka A, Costa RM, Hu H, Irvin DK, Patel A, Kornblum HI, Silva AJ,
O’Dell TJ, Colicelli J. 2003. The RAS effector RIN1 modulates the for-
mation of aversive memories. J Neurosci 23:748–757.
Dugan LL, Kim JS, Zhang Y, Bart RD, Sun Y, Holtzman DM, Gutmann
DH. 1999. Differential effects of cAMP in neurons and astrocytes. Role
of B-raf. J Biol Chem 274:25842–25848.
English JD, Sweatt JD. 1996. Activation of p42 mitogen-activated protein
kinase in hippocampal long term potentiation. J Biol Chem 271:24329–
English JD, Sweatt JD. 1997. A requirement for the mitogen-activated
protein kinase cascade in hippocampal long term potentiation. J Biol
Eychene A, Dusanter-Fourt I, Barnier JV, Papin C, Charon M, Gisselbrecht
S, Calothy G. 1995. Expression and activation of B-Raf kinase isoforms in
human and murine leukemia cell lines. Oncogene 10:1159–1165.
Frankland PW, Cestari V, Filipkowski R, McDonald RJ, Silva AJ. 1998.
The dorsal hippocampus is essential for context discrimination but not
for contextual conditioning. Behav Neurosci 112:863–874.
Gallagher M, Burwell R, Burchinal M. 1993. Severity of spatial learning
impairment in aging: development of a learning index for performance
in the Morris water maze. Behav Neurosci 107:618–626.
Giese KP, Friedman E, Telliez JB, Fedorov NB, Wines M, Feig LA, Silva
AJ. 2001. Hippocampus-dependent learning and memory is impaired in
mice lacking the Ras-guanine-nucleotide releasing factor 1 (Ras-GRF1).
Grewal SS, Fass DM, Yao H, Ellig CL, Goodman RH, Stork PJ. 2000.
Calcium and cAMP signals differentially regulate cAMP-responsive ele-
ment-binding protein function via a Rap1-extracellular signal- regulated
kinase pathway. J Biol Chem 275:34433–34441.
Hugues S, Deschaux O, Garcia R. 2004. Postextinction infusion of a mito-
gen-activated protein kinase inhibitor into the medial prefrontal cortex
impairs memory of the extinction of conditioned fear. Learn Mem 11:
Huleihel M, Goldsborough M, Cleveland J, Gunnell M, Bonner T, Rapp
UR. 1986. Characterization of murine A-raf, a new oncogene related to
the v-raf oncogene. Mol Cell Biol 6:2655–2662.
Huser M, Luckett J, Chiloeches A, Mercer K, Iwobi M, Giblett S, Sun
XM, Brown J, Marais R, Pritchard C. 2001. MEK kinase activity is not
necessary for Raf-1 function. EMBO J 20:1940–1951.
Ikawa S, Fukui M, Ueyama Y, Tamaoki N, Yamamoto T, Toyoshima K.
1988. B-raf, a new member of the raf family, is activated by DNA rear-
rangement. Mol Cell Biol 8:2651–2654.
Jaiswal RK, Moodie SA, Wolfman A, Landreth GE. 1994. The mitogen-acti-
vated protein kinase cascade is activated by B-Raf in response to nerve
growth factor through interaction with p21ras. Mol Cell Biol 14:6944–6953.
Kelleher RJ 3rd, Govindarajan A, Jung HY, Kang H, Tonegawa S. 2004.
Translational control by MAPK signaling in long-term synaptic plasticity
and memory. Cell 116:467–479.
Kelly A, Laroche S, Davis S. 2003. Activation of mitogen-activated pro-
tein kinase/extracellular signal-regulated kinase in hippocampal circuitry
is required for consolidation and reconsolidation of recognition memory.
J Neurosci 23:5354–5360.
Lin CH, Yeh SH, Lu KT, Leu TH, Chang WC, Gean PW. 2001. A role
for the PI-3 kinase signaling pathway in fear conditioning and synaptic
plasticity in the amygdala. Neuron 31:841–851.
Marais R, Light Y, Paterson HF, Mason CS, Marshall CJ. 1997. Differen-
tial regulation of Raf-1, A-Raf, and B-Raf by oncogenic ras and tyrosine
kinases. J Biol Chem 272:4378–4383.
Martin KC, Michael D, Rose JC, Barad M, Casadio A, Zhu H, Kandel
ER. 1997. MAP kinase translocates into the nucleus of the presynaptic cell
and is required for long-term facilitation in Aplysia. Neuron 18:899–912.
Mayford M, Bach ME, Huang YY, Wang L, Hawkins RD, Kandel ER.
1996a. Control of memory formation through regulated expression of a
CaMKII transgene. Science 274:1678–1683.
Forebrain-Specific Knockout of B-raf Kinase 37
Journal of Neuroscience Research DOI 10.1002/jnr
Mayford M, Baranes D, Podsypanina K, Kandel ER. 1996b. The 30-untrans- Download full-text
lated region of CaMKII alpha is a cis-acting signal for the localization and
translation of mRNA in dendrites. Proc Natl Acad Sci U S A 93:13250–
Mazzucchelli C, Vantaggiato C, Ciamei A, Fasano S, Pakhotin P, Krezel
W, Welzl H, Wolfer DP, Pages G, Valverde O, Marowsky A, Porrazzo
A, Orban PC, Maldonado R, Ehrengruber MU, Cestari V, Lipp HP,
Chapman PF, Pouyssegur J, Brambilla R. 2002. Knockout of ERK1
MAP kinase enhances synaptic plasticity in the striatum and facilitates
striatal-mediated learning and memory. Neuron 34:807–820.
Mikula M, Schreiber M, Husak Z, Kucerova L, Ruth J, Wieser R,
Zatloukal K, Beug H, Wagner EF, Baccarini M. 2001. Embryonic lethal-
ity and fetal liver apoptosis in mice lacking the c-raf-1 gene. EMBO J 20:
Morice C, Nothias F, Konig S, Vernier P, Baccarini M, Vincent JD, Barnier
JV. 1999. Raf-1 and B-Raf proteins have similar regional distributions but
differential subcellular localization in adult rat brain. Eur J Neurosci 11:
Morozov A, Muzzio IA, Bourtchouladze R, Van-Strien N, Lapidus K, Yin
D, Winder DG, Adams JP, Sweatt JD, Kandel ER. 2003. Rap1 couples
cAMP signaling to a distinct pool of p42/44MAPK regulating excitability,
synaptic plasticity, learning, and memory. Neuron 39:309–325.
Nagy A, Rossant J, Nagy R, Abramow-Newerly W, Roder JC. 1993.
Derivation of completely cell culture-derived mice from early-passage
embryonic stem cells. Proc Natl Acad Sci U S A 90:8424–8428.
Ohno M, Frankland PW, Chen AP, Costa RM, Silva AJ. 2001. Inducible,
pharmacogenetic approaches to the study of learning and memory. Nat
Otto T, Eichenbaum H, Wiener SI, Wible CG. 1991. Learning-related
patterns of CA1 spike trains parallel stimulation parameters optimal for
inducing hippocampal long-term potentiation. Hippocampus 1:181–192.
Papin C, Denouel-Galy A, Laugier D, Calothy G, Eychene A. 1998. Mod-
ulation of kinase activity and oncogenic properties by alternative splicing
reveals a novel regulatory mechanism for B-Raf. J Biol Chem 273:24939–
Ranck JB Jr. 1973. Studies on single neurons in dorsal hippocampal forma-
tion and septum in unrestrained rats. I. Behavioral correlates and firing
repertoires. Exp Neurol 41:461–531.
Richter-Levin G, Thomas KL, Hunt SP, Bliss TV. 1998. Dissociation
between genes activated in long-term potentiation and in spatial learning
in the rat. Neurosci Lett 251:41–44.
Schafe GE, Atkins CM, Swank MW, Bauer EP, Sweatt JD, LeDoux JE.
2000. Activation of ERK/MAP kinase in the amygdala is required for
memory consolidation of pavlovian fear conditioning. J Neurosci
Selcher JC, Nekrasova T, Paylor R, Landreth GE, Sweatt JD. 2001. Mice
lacking the ERK1 isoform of MAP kinase are unimpaired in emotional
learning. Learn Mem 8:11–19.
Silva AJ, Frankland PW, Marowitz Z, Friedman E, Lazlo G, Cioffi D, Jacks
T, Bourtchuladze R. 1997. A mouse model for the learning and memory
deficits associated with neurofibromatosis type I. Nat Genet 15:281–284.
Storm SM, Cleveland JL, Rapp UR. 1990. Expression of raf family proto-
oncogenes in normal mouse tissues. Oncogene 5:345–351.
Sweatt JD. 2001. The neuronal MAP kinase cascade: a biochemical signal
integration system subserving synaptic plasticity and memory. J Neuro-
Thomas KL, Laroche S, Errington ML, Bliss TVP, Hunt SP. 1994. Spatial
and temporal changes in signal transduction pathways during LTP. Neu-
Tsien JZ, Chen DF, Gerber D, Tom C, Mercer EH, Anderson DJ, May-
ford M, Kandel ER, Tonegawa S. 1996. Subregion- and cell type-
restricted gene knockout in mouse brain. Cell 87:1317–1326.
Vossler MR, Yao H, York RD, Pan MG, Rim CS, Stork PJ. 1997. cAMP
activates MAP kinase and Elk-1 through a B-Raf- and Rap1-dependent
pathway. Cell 89:73–82.
Welzl H, D’Adamo P, Lipp HP. 2001. Conditioned taste aversion as a
learning and memory paradigm. Behav Brain Res 125:205–213.
Wixler V, Smola U, Schuler M, Rapp U. 1996. Differential regulation of
Raf isozymes by growth vs. differentiation inducing factors in PC12
pheochromocytoma cells. FEBS Lett 385:131–137.
Wojnowski L, Zimmer AM, Beck TW, Hahn H, Bernal R, Rapp UR,
Zimmer A. 1997. Endothelial apoptosis in Braf-deficient mice. Nat
Wojnowski L, Stancato LF, Larner AC, Rapp UR, Zimmer A. 2000.
Overlapping and specific functions of Braf and Craf-1 protooncogenes
during mouse embryogenesis. Mech Dev 91:97–104.
Wu SP, Lu KT, Chang WC, Gean PW. 1999. Involvement of mitogen-
activated protein kinase in hippocampal long-term potentiation. J
Biomed Sci 6:409–417.
Zinyk DL, Mercer EH, Harris E, Anderson DJ, Joyner AL. 1998. Fate
mapping of the mouse midbrain–hindbrain constriction using a site-spe-
cific recombination system. Curr Biol 8:665–668.
38Chen et al.
Journal of Neuroscience Research DOI 10.1002/jnr