Overexpression of the Cell Adhesion Protein Neuroligin-1 Induces
Learning Deficits and Impairs Synaptic Plasticity by Altering the
Ratio of Excitation to Inhibition in the Hippocampus
Regina Dahlhaus,1Rochelle M. Hines,1,2Brennan D. Eadie,3,4,5Timal S. Kannangara,2,4,5
Dustin J. Hines,1Craig E. Brown,1,5,6,7Brian R. Christie,4,5,6,7*and Alaa El-Husseini1,6
cated in regulating CNS synaptogenesis. Among these, the Neuroligin
(NL) family (NLs 1–4) of postsynaptic adhesion proteins has been shown
to promote the development and specification of excitatory versus
inhibitory synapses. NLs form a heterophilic complex with the presynap-
tic transmembrane protein Neurexin (NRX). A differential association of
NLs with postsynaptic scaffolding proteins and NRX isoforms has been
suggested to regulate the ratio of excitatory to inhibitory synapses (E/I
ratio). Using transgenic mice, we have tested this hypothesis by overex-
pressing NL1 in vivo to determine whether the relative levels of these
cell adhesion molecules may influence synapse maturation, long-term
potentiation (LTP), and/or learning. We found that NL1-overexpressing
mice show significant deficits in memory acquisition, but not in memory
retrieval. Golgi and electron microscopy analysis revealed changes in
synapse morphology indicative of increased maturation of excitatory
synapses. In parallel, electrophysiological examination indicated a shift
in the synaptic activity toward increased excitation as well as impair-
ment in LTP induction. Our results demonstrate that altered balance in
the expression of molecules necessary for synapse specification and
development (such as NL1) can lead to defects in memory formation
and synaptic plasticity and outline the importance of rigidly controlled
synaptic maturation processes. V V
Trans-synaptic cell-adhesion molecules have been impli-
C 2009 Wiley-Liss, Inc.
plasticity; learning and memory
The development of synaptic networks depends on the synchronized
actions of several guidance and cell-adhesion proteins during synaptogene-
sis (Piechotta et al., 2006) as well as synaptic activity later in development
(Waites et al., 2005). Although somewhat distinct,
some proteins play critical roles in both synaptogenesis
and synapse maturation (Luthl et al., 1994; Tang et al.,
1998; Contractor and Heinemann, 2002). For exam-
ple, the postsynaptic density (PSD) protein PSD95 can
drive glutamate synapse maturation (El-Husseini et al.,
2000a), while also playing a critical role in experience
dependent synapse stabilization (Ehrlich et al., 2007)
and long-term potentiation (LTP) of synaptic efficacy
(Migaud et al., 1998; Ehrlich and Malinow, 2004).
The Neuroligin (NL) family of postsynaptic trans-
membrane proteins (NL1–4) is active early in develop-
ment (Lise and El-Husseini, 2006; Craig and Kang,
2007). NLs were first identified as binding partners of
Neurexins (NRXs), presynaptic cell-adhesion molecules
(Rowen et al., 2002; Tabuchi and Sudhof, 2002). In a
Ca21-dependent manner, the extracellular domain of
NLs binds to that of NRXs, thereby bridging the synap-
tic cleft and linking NLs to the exocytotic machinery
(Rao et al., 2000; Dean et al., 2003). The NL1 intracel-
lular PDZ-binding domain interacts with PSD95, which
clusters N-methyl-D-aspartate receptors (NMDARs) and
controls AMPA (a-amino-3-hydroxy-5-methyl-4-isoxa-
zolepropionic acid) receptor numbers (El-Husseini et al.,
2000b; Schnell et al., 2002; Beique and Andrade, 2003;
Stein et al., 2003; Ehrlich and Malinow, 2004; Prange
et al., 2004; Ehrlich et al., 2007). In addition, the NL1/
PSD95 complex is able to modulate neurotransmitter
release probability via trans-synaptic protein–protein
interactions (Futai et al., 2007).
NLs have also been implicated in regulating the ratio
of excitatory to inhibitory synapses (E/I ratio) (Prange
et al., 2004; Levinson et al., 2005; Nam and Chen,
2005; Chih et al., 2006; Graf et al., 2006; Kang et al.,
2008). In vitro studies have shown that the expression
of individual NLs is sufficient to alter the density of
excitatory and inhibitory synapses (Scheiffele et al.,
2000; Graf et al., 2004; Prange et al., 2004; Chih et al.,
2005; Levinson et al., 2005; Nam and Chen, 2005;
Gerrow et al., 2006), biasing the ratio of excitation to
inhibition. Specifically, NL1 has been shown to affect
excitatory synaptic responses, whereas NL2 has been
demonstrated to influence inhibitory synaptic responses
(Chubykin et al., 2007). Because chronic inhibition of
NMDARs or CaM-Kinase II blocks NL1 activity, while
1Department of Psychiatry, University of British Columbia, Vancouver,
ver, BC;3MD/PhD Program, University of British Columbia, Vancouver,
British Columbia, Vancouver, BC;5Division of Medical Sciences, Univer-
sity of Victoria, Victoria, BC;
British Columbia, Vancouver, BC;7Department of Biology, University of
Victoria, Victoria, BC
Additional Supporting Information may be found in the online version of
R.D., R.M.H., and B.D.E. contributed equally to this work.
Grant sponsor: CIHR; Grant number: MT-12675; Grant sponsor: FXRFC.
*Correspondence to: Brian R. Christie, Division of Medical Sciences,
University of Victoria, Victoria, BC, V8P5C2 Canada. E-mail: brain64@
Accepted for publication 20 March 2009
Published online 12 May 2009 in Wiley InterScience (www.interscience.
2Neuroscience Program, University of British Columbia, Vancou-
4Department of Cellular and Physiological Sciences, University of
6Brain Research Centre, University of
HIPPOCAMPUS 20:305–322 (2010)
C2009 WILEY-LISS, INC.
chronic inhibition of synaptic activity suppresses the effects of
NL2, NLs have been suggested to determine synapse type in an
activity-dependent manner (Chubykin et al., 2007).
Despite the body of research implicating NLs as key mole-
cules regulating synapse maturation and specificity, a role for
NLs in synaptic plasticity, outside of the developmental setting,
has not been established. The present experiments examine the
influence of NL1 expression on synapse maturation, synaptic
plasticity, synapse specificity, and learning using transgenic mice
Generation and Maintenance of Transgenic Mice
The NL1 transgene was expressed under control of the Thy1
promoter to drive neuron-specific expression (Caroni, 1997).
The sequence for NL1 was generated by insertion of a Sal I site
into the 30Spe I site of mouse NL1 cDNA (Hines et al., 2008).
Briefly, the HA tag was inserted following the signal sequence.
NL1 was then amplified using Sal I and Xho I-anchored primers.
Following restriction digestion, these fragments were then cloned
into the Xho I site of the Thy1.2 plasmid. The NL1-Thy1.2 plas-
mid was cut using Pme I and Pvu I enzymes to remove the bacte-
rial backbone, and the resulting Thy1.2 NL1 fragments were
used to generate NL1 transgenic mice (TgNL1). Distinct strains
were generated from multiple founders via backcrossing to the
C57BL/6 strain. Two primary strains with similar NL1 expression
levels (TgNL1.6 and TgNL1.7) have been maintained on the
C57/BL/6 background for eight generations.
Mice were maintained at the University of British Columbia
Animal Resource Unit, the Centre for Molecular Medicine and
Therapeutics, or the Division of Medical Sciences at the Univer-
sity of Victoria according to protocols approved by the University
of British Columbia/University of Victoria Animal Care Commit-
tees. Mice were group housed with a 12-h light–dark cycle with
constant temperature. Behavioral assessments were conducted
during the light phase. All experiments were conducted genotype-
blinded with wild-type (WT) littermates used as control animals.
Germ line transmission of the TgNL1 transgene was
detected using PCR, with primers spanning from Thy1 into
the NL1 signal sequence. The primer sequences are as follows:
Assessments made in the preliminary screen were based on
the modified SHIRPA protocol from the EMPReSS screen and
designed to evaluate the basic phenotype of transgenic mouse
strains (Rogers et al., 2001). The tasks of this screen were
designed to rule out obvious alterations in sensory, motor, and
reflexive behaviors while evaluating the possibility of neurologi-
cal dysfunction through several assessments.
Open field exploratory behavior of mice was assessed using a
LogitechTMQuickCam webcam and Any-MazeTMsoftware for
video tracking. Mice were allowed to freely explore the novel
environment for a period of 30 min. Parameters measured
include total distance traveled, average speed, number of line
crossings between edge, and field zones as well as mean speed
on the open field or at the edge, respectively. Further parame-
ters analyzed for the different zones include duration of mean
visit, traveled distance, time spent, and the number of defeca-
tion boli. Results from tracking analysis were compared using
one-way ANOVA (WT, n 5 12; TgNL1, n 5 14).
Elevated plus maze
Exploratory behavior of WT (n 5 12) and TgNL1 (n 5
10) mice was recorded in the elevated plus maze for 10 min
using a LogitechTMQuickCam and Any-MazeTMsoftware.
Measurements included: average distance, speed, line crossings,
time spend in the open, risk, and safe zones as well as latency
to first entry of both open and risk zones. Statistical signifi-
cance was tested using two-tailed Student’s t tests.
WT (n 5 10) and TgNL1 (n 5 10) mice were initially
tested in a plus-shaped water maze for their ability to locate a
visible platform in two sessions consisting of three consecutive
trials. Mean latency, distance, speed, time immobile, and the
number of line crossings were calculated for each animal, and
statistic significance was tested by two-tailed Student’s t test.
Learning and memory experiments
To prevent excessive floating behavior, WT (n 5 10) and
TgNL1 (n 5 10) mice were initially trained in the smaller
plus-shaped water maze, followed by the standard Morris water
maze. Training procedures for both water mazes have been
described in detail elsewhere (Brandeis et al., 1989; van Praag
et al., 1999; Van Dam et al., 2000; Vloeberghs et al., 2006).
Briefly, animals were trained daily in two sessions consisting of
three consecutive trials for 6–9 days followed by the probe test
and reverse as well as working memory training in the same
manner. The maximum time given for each trial was 60 s in
the plus-shaped and 90 s in the Morris water maze, thereafter
animals were guided to the platform. Each animal was allowed
to remain on the platform for 15 s, before it was taken back to
the home cage. Tracking was done by a LogitechTMQuickCam
webcam, and datawererecorded
Any-MazeTMand Graph Pad Prism 5 software. Acquisition
curves were compared with two-way analysis of variance
DAHLHAUS ET AL.
(ANOVA) for repeated measurements (RM ANOVA), while
the probe trials were assessed by one-way ANOVA. Pair-wise
comparisons were performed by Bonferroni post hoc testing.
Lysate Preparation and Western Blotting
Rapidly extracted tissues were homogenized in 3 ml/g
HEPES-buffer [10 mM HEPES, pH 7.5; 1 mM EGTA; 0.1
mM MgCl2; 0.15 M NaCl; 13 protease inhibitor cocktail
(Roche Applied Science)], and cell debris was removed by cen-
trifugation at 16000g at 48C for 1 h.
Supernatants were obtained from 14 WT and 14 tgNL1 ani-
mals, subjected to SDS-PAGE and analyzed by immunoblot.
Staining for Coomassie and actin was performed to determine
the protein content of each sample. Protein levels were quanti-
fied using Image J software, and statistical significance of nor-
malized measurements was tested by two-tailed Student’s t tests.
Golgi Cox Impregnation and Spine
Two and a half month old mice were anesthetized with an
overdose of sodium pentobarbital and perfused transcardial
with 0.9% saline. Brains were then quickly removed from the
skull, submersed in Golgi Cox solution (208 mM K2Cr2O7,
260 mM HgCl2, and 43 mM K2CrO4in dH2O), and stored
for 7 days in the dark. After replacement of Golgi Cox solution
with 30% sucrose buffered in phosphate [phosphate buffered
saline (PBS)], brains were stored in the dark for three more
days. Thereafter, 150-lm thick vibratome sections were col-
lected, mounted onto 2% gelatinized slides, and developed by
a modified Golgi Cox technique (Glaser and Van der Loos,
1981). Following rapid dehydration, sections were embedded
in PermountTMand coverslipped.
The density (spines per lm) as well as the head and neck
size of pyramidal cells from cortical layer 5 and the CA1 region
were analyzed using Image J software. Results are shown as per-
cent change normalized to WTmeasurements. Statistical signifi-
cance was analyzed by two-way ANOVA and subsequent Bon-
ferroni post hoc.
For immunohistochemistry (IHC), animals were perfused
transcardial with a 4% solution of paraformaldehyde in PBS
(for nonfluorescent IHC), or brains were harvested fresh and
flash frozen with OCT embedding medium in liquid nitrogen
(for fluorescent synaptic protein IHC). Sections were cut on a
cryostat at thicknesses of 10 lm (synaptic protein IHC) or 30
lm (HA localization).
Sections were incubated in blocking solution [2.5% bovine
serum albumin, 10% normal goat serum, 0.1% triton-x 100,
0.02% sodium azide in PBS] for 60 min, followed by primary
antibody incubations (diluted in blocking solution) overnight
at 48C. Following washing with 13 PBS-triton (1%), sections
were incubated with secondary antibodies. Nonfluorescent im-
munostaining was visualized using avidin–biotin conjugation
(Vectastain Elite Standard Kit; Vector) to the secondary anti-
body, followed by diaminobenzene (Vector). Fluorescent immu-
nostaining (VGluT and VGAT immuno: WT n 5 6; TgNL1.7
n 5 6 as well as HA with synaptic markers: WT n 5 6;
TgNL1.7 n 5 6) were imaged using a Zeiss confocal laser
scanning microscope LSM 510 with a 633 water immersion
objective and examined in Northern eclipse, Axio Vision
(Zeiss), ImageJ, and Adobe Photoshop software. Statistical sig-
nificance was analyzed using two-tailed Student’s t tests.
Primary antibodies used were rat anti-HA (Boehringer
Mannheim); rabbit anti-PSD95 (generated by ABR); mouse
anti-PSD95 (ABR); rabbit anti-NL1 (Synaptic Systems); rabbit
anti-NL2 (generated using a synthetic peptide by ABR); rabbit
anti-NL3 (generated by ABR); rabbit anti-Synaptophysin
(Zymed); mouse anti-Syntaxin 1 (Chemicon); rabbit anti-Syn-
aptotagmin 1 (Synaptic Systems); guinea pig anti-VGluT 1 &
2 (Chemicon); rabbit anti-GluR2/3 and anti-GluR1 (Chem-
icon); mouse anti-VGAT (Synaptic Systems); rabbit anti-
Gephyrin (Alexis); mouse anti-CaMKII (ABR); mouse anti-
include peroxidase-coupled donkey anti-rabbit and sheep anti-
mouse IgG (WB; GE Healthcare), biotinylated secondary anti-
bodies (nonfluorescent IHC; Vector), or Alexa conjugated second-
ary antibodies (fluorescent IHC; Invitrogen Molecular Probes).
Tissue for electron microscopy (EM) was harvested fresh and
cut using a vibratome in room temperature ACSF, followed by
rapid fixation in 6% glutaraldehyde, 1% paraformaldehyde, 2
mM CaCl2, 4 mM MgCl2in 0.1 M cacodylate buffer. Sections
were examined under a dissecting microscope and CA1 stratum
radiatum regions of the hippocampus were isolated (bregma:
21.64 mm). Tissue blocks were washed in 0.1 M cacodylate,
postfixed in 2% osmium tetroxide, 1.5% potassium ferrocya-
nide in 0.1 M cacodylate buffer for 2 h, followed by en bloc
staining with 2% uranyl acetate for 45 min. Samples were then
dehydrated by immersion in increasing concentrations of alco-
hol before being transferred into propylene oxide and gradually
embedded in eponate resin. Sections (1 lm) were taken from
prospective tissue blocks and examined under a light micro-
scope to ensure a consistent location was selected. All tissue
samples were coded with respect to genotype before electron
microscopic imaging and analysis of number and morphology.
The size of synaptic elements was measured using Image
J on digital micrographs taken across six serial sections in alter-
NL1 OVER-EXPRESSION IMPAIRS LEARNING AND SYNAPTIC PLASTICITY
nation (n 5 4 animals for each group). These numbers were
then averaged to ensure that measurements were not influenced
by the two-dimensional position within the synapse. Synaptic
vesicles were manually counted in high magnification images
and classified according to their location within the presynaptic
bouton (docked or reserve pool).
For the assessment of synapse number, synapses were
counted in fields of 100 lm2. Within these fields, synapses
were classified as either Type I (asymmetric excitatory) or Type
II (symmetric inhibitory) on the basis of synaptic vesicle and
PSD morphology. Type I synapses were identified by the pres-
ence of a PSD and round synaptic vesicles within the presynap-
tic terminal, whereas Type II synapses were identified by the
presence of at least three oval or flattened synaptic vesicles and
the absence of a PSD (Uchizono, 1965). The data was analyzed
using ANOVA, with the LSD post hoc test applied to signifi-
cant ANOVA comparisons.
Adult TgNL1 mice (TgNL1.6 n 5 6; TgNL1.7 n 5 9) and
their WT littermates (WT; n 5 12) were used for electrophysi-
ology experiments as previously described (Christie et al.,
2005). Briefly, brains were rapidly removed while submerged in
chilled sucrose artificial cerebrospinal fluid (sACSF; pH 7.2)
containing (in mM) 110.00 sucrose, 60.00 NaCl, 3.00 KCl,
1.25 NaH2PO4, 28.00 NaHCO3, 0.50 CaCl2, 7.00 MgCl2,
5.00 dextrose and 0.60 ascorbate, and saturated with 95% O2–
5% CO2. Transverse slices containing the hippocampal forma-
tion were sectioned at 400 lm using a vibratome kept at 48C
(Vibratome 1,500, Ted Pella). For recovery, each slice was
sequentially placed in a beaker containing warm (308C) normal
ACSF (nACSF; pH 7.2) containing (in mM) 125.00 NaCl,
2.50 KCl, 1.25 NaH2PO4, 25.00 NaHCO3, 2.00 CaCl2, 1.30
MgCl2, and 10.00 dextrose, continuously bubbled with 95%
O2–5% CO2for ?60 min. Individual slices were transferred to
the recording chamber where they were perfused (2 ml/min)
with warm (308C), bubbled (95% O2–5% CO2) nACSF for
the duration of each recording. Responses were obtained using
a 1–3 MX recording electrode (filled with nACSF) and a Mul-
tiClamp 700B (Axon Instruments) amplifier. All data were
acquired at 100 kHz. Responses were evoked with a concentric
bipolar electrode (FHC) using monophasic negative current
pulses (120 ls, 10–80 lA) and a digital stimulus isolation unit
(Getting Instruments). An Olympus BX50wi (103 objective)
was used to visually position both the recording and stimulat-
ing electrodes for each experiment 50–100 lm apart in the
stratum radiatum of the CA1 region of the hippocampus.
Stimulation intensity was adjusted to yield response amplitudes
40% of the estimated maximum. All evoked responses were ini-
tially tested with paired-pulse (PP) stimuli (50-ls interpulse
interval). During experiments, evoked responses were continu-
ously elicited at 15-s intervals, except during the application of
the conditioning stimulation. To assess LTP, a stable baseline
was obtained for a minimum of 15 min and then conditioning
stimuli were administered (four bursts of 50 pulses at 100 Hz;
30-s interburst interval). Single-pulse stimulation was again ini-
tiated immediately following the conditioning stimuli and con-
tinued for a minimum of 60 min. The initial slope of the nega-
tive going waveform was used to assess changes in synaptic
efficacy. PP responses are presented as the normalized difference
between the slopes of the two fEPSP responses and as a percent-
age change. LTP was quantified as the percentage change in slope
for responses collected following the application of conditioning
stimuli. Assessment of disinhibition of the population spike (PS)
was obtained by bath application bicuculline (5 lM). The
amplitude of the PS was used to assess excitability of the popula-
tion of cells in the immediate vicinity of the recording electrode.
After obtaining a stable response for 15 min, bath application
the GABAA antagonist bicuculline (5 lM) was initiated and
maintained for a minimum of 30 min. All data acquisition and
analysis were performed using ClampFit 10.2 software (Axon
Instruments), Excel (Microsoft), and Statistica 7.0 for PC. Data
were analyzed using unpaired t tests and presented as the mean
6 standard error of mean (SEM).
Similar to slices used for field recordings, for whole-cell
recordings, transverse slices (400 lm) were obtained from adult
TgNL1 mice (n 5 4) and their WT littermates (n 5 4) and
submerged in nACSF and maintained at 308C. Individual slices
were transferred to the recording chamber where they were per-
fused (2 ml/min) with warm (308C), bubbled (95% O2–5%
CO2) nACSF for the duration of each recording.
Whole-cell currents were obtained using a 5–7 MX recording
electrode filled with (in mM) 140.00 KMeSO4, 10.00 HEPES,
4.00 NaCl, 0.10 EGTA, 4.00 ATP, 0.30 GTP, 14.00 Phospho-
creatine, and an Axopatch 200B (Axon Instruments) amplifier.
All data were acquired at 100 kHz. Responses were evoked with
concentric bipolar electrodes (FHC) using monophasic negative
current pulses (120 ls, 10–80 lA) and a digital stimulus isola-
tion unit (Getting Instruments). An Olympus BX50wi (103
objective) was used to visually position electrodes. One stimulat-
ing electrode was placed in the Stratum radiatum to evoke
EPSCs, whereas a second stimulating electrode was placed
directly in the pyramidal cell layer to evoke IPSCs. An input–
output curve for excitatory currents (EPSCs) was obtained in
CA1 pyramidal neurons by obtaining responses with increasing
pulse widths while holding the cells at the reversal potential
(Vrev) for GABA receptors (291 mV). IPSCs were isolated by
holding the cells at the Vrev for AMPA and NMDA receptors
(12 mV) and administering identical stimuli. The identity of
EPSCs and IPSCs was confirmed using CNQX (20 lM), (50
lM) APV, and (10 lM) bicuculline methiodide, respectively. All
data acquisition and analysis were performed using ClampFit
10.2 software (Axon Instruments), Excel (Microsoft), and Statis-
tica 7.0 for PC. Data were analyzed using unpaired t tests and
presented as the mean 6 SEM.
DAHLHAUS ET AL.
The NL1 Transgene Is Expressed Throughout the
Brain and Is Localized at Excitatory Synapses
Mice expressing Neuroligin 1 (TgNL1) were generated using
a construct composed of an HA tagged version of full length
NL1 (details of mouse generation are described in Hines et al.
(2008). There were no significant differences in the expression
level of HA-NL1 in either of the transgenic mouse strains
(TgNL1.6 and TgNL1.7; Figs. 1A,B). Western blots revealed
that HA-NL1 was specifically expressed in brain and to a lesser
extend in spinal cord, but not in nonneuronal tissues such as
liver or heart (Fig. 1C). In the brain, DAB IHC and WB show
that HA-NL1 has a broad distribution, with highest levels of
expression in the cortex and hippocampus, intermediate in the
striatum, and lowest in the cerebellum and olfactory bulb (Fig.
1D and Supp. Info. Fig. 1). Specifically, HA-NL1 was found to
be localized in neurons that were positive for CamKII and in
neurons positive for parvalbumin. During brain development,
an increase in the expression level of HA-NL1 can be observed
during early postnatal development, in particular between post-
natal day 10 and 14, while thereafter the expression level remains
constant throughout adulthood (Fig. 1E). Interestingly, the
observed increase in HA-NL1 expression is similar to the devel-
opmental expression pattern of endogenous NLs (Jamain et al.,
2008) and correlates with the development of dendritic spines.
To further address the localization of HA-NL1, the colocaliza-
tion of HA-NL1 with postsynaptic marker proteins for excitatory
(PSD95) and inhibitory (Gephyrin) synapses was studied. The
analysis revealed that 73% of HA-NL1 puncta are localized with
puncta positive for the excitatory synapse marker PSD95. Con-
versely, 2% are found to be localized with puncta positive for the
inhibitory synapse marker Gephyrin (Figs. 1F,G). Moreover,
61% of puncta positive for PSD95 and 8% of puncta positive for
Gephyrin also contain HA-NL1, reflecting that not all cells
express the transgene under the Thy1 promoter. Further HA-
NL1 puncta observed may also represent dendritic transport clus-
ters and/or potential new contact sides lacking PSD95 (Scheiffele
et al., 2000; Dresbach et al., 2004; Rosales et al., 2005).
Mice Expressing NL1 Do Not Show
Abnormalities in Sensory, Autonomic, or Basic
A preliminary phenotype screen, based on the modified
SHIRPA protocol from the European Mouse Phenotyping
Resource of Standardized Screens (EMPReSS; (Rogers et al.,
2001), was used to assess the overall health and basic auto-
nomic, sensory, and motor functions of TgNL1 mice. No alter-
ations were observed in indicators of autonomic function,
including palpebral closure, piloerection, and tail position. Sim-
ilarly, sensory systems were found to be intact with no deficits
in transfer arousal and touch escape (somatosensory system);
corneal reflex, visual placing, and visible platform water maze
performance (visual system); pinna reflex and acoustic startle
(auditory system), and buried food retrieval (olfactory system).
In addition, no changes were observed in body position, body/
limb tone, righting reflex, or basal activity levels, indicating no
gross motor impairments in TgNL1 mice.
Exploration in the open field was also compared between
TgNL1 mice (n 5 10) and their WT littermates (n 5 10). In two
independent experiments, no difference was observed in the aver-
age speed or distance traveled in the open field session (Table 1).
Also, no indications of thigmotaxis were observed in TgNL1 mice.
In the elevated plus maze paradigm, there were also no indications
of the TgNL1 animals showing altered anxiety (Table 1). Thus,
TgNL1 mice do not show impairments in basic sensory, auto-
nomic, or motor behaviors and display exploration activity and
anxiety-related behaviors similar to WTmice.
Specific Learning Impairments Are Observed
in Mice Expressing NL1
Because no deficit was detected in the basic autonomic, sensory,
or motor performance of TgNL1 mice, we became interested in
the possibility of more subtle defects on complex functions such
as learning and memory. To test hippocampal-dependent spatial
learning and memory, the hidden platform version of the plus-
shaped and Morris water maze tests were applied to compare
TgNL1 mice (n 5 10) and their littermate controls (n 5 10). In
both tests, animals have to learn to escape from the water maze by
locating a hidden platform using spatial cues, which requires
regular hippocampal function (Morris et al., 1982).
Both, WT and TgNL1 mice, improved their learning per-
formance in the plus-shaped water maze (Figs. 2A–C; Supp.
Info. Fig. 2) as well as in the Morris water maze during their ac-
quisition training (data not shown, but see also Fig. 6a), as indi-
cated by a significant within groups effect of training session
(plus-shaped water maze training: distance F(8,464)5 5.08, P <
0.001 and latency F(8,464)5 9.62, P < 0.001; reversal training:
distance F(5,290)5 12.88, P < 0.001; latency F(5,290)5 16.42,
P < 0.001; Morris water maze training: distance F(5,290)5 6.39,
P < 0.001; latency F(5,290)5 6.91, P < 0.001; reversal training:
distance F(3,174)5 4.60, P < 0.001; latency F(3,174)5 2.37,
P < 0.001). However, during forward training in the plus
shaped water maze, TgNL1 mice show a significant increase in
the distance traveled (Fig. 2B; F(1,464)5 63.05, P < 0.001), the
escape latency (Fig. 2C; F(1,464)5 79.66, P < 0.001) and the
number of lines crossed (Supp. Info. Fig. 2; F(1,464)5 67.27, P
< 0.001), as well as a significant reduction in the number of cor-
rect attempts (Supp. Info. Fig. 3; F(1,464)5 26.13; P < 0.001)
when compared with WT littermates. Moreover, additional
training sessions (session 5–9) did not improve the performance
of TgNL1 mice any further nor allow them to reach the per-
formance level of WT littermates. As no significant difference
between TgNL1 and WT control swim speed could be observed
(data not shown; F(1,464)5 0.7669, P > 0.05), these differences
cannot be explained by locomotor deficits in TgNL1 mice.
Instead, these results are interpreted as disrupted acquisition of
memory for the platform location. The difference in perform-
ance observed in TgNL1 mice and their WT littermates during
the first training session might relate to visual testing which
NL1 OVER-EXPRESSION IMPAIRS LEARNING AND SYNAPTIC PLASTICITY
immediately preceded water maze testing. It is possible that due
to a learning deficit in TgNL1 mice, WT mice benefit more
from the experience of the vision test. In support of this explana-
tion, TgNL1 mice and their WT littermate controls show an
equal performance during the first session of all subsequent water
maze tasks (reversal training, two Morris water maze paradigms)
when the TgNL1 mice had more time to become familiar with
the hidden platform condition. Consistent with forward train-
ing, the reverse training (Figs. 2D–F and Supp. Info. Fig. 2)
demonstrates a significant impairment in new memory forma-
tion in TgNL1 mice (Latency F(1,290)5 95.74; P < 0.001; dis-
tance F(1,290)5 48.0, P < 0.001; line crossings F(1,290)5 34.68;
P < 0.001; number of correct tries F(1,290)5 13.43, P 5 0.006;
mean swimming speed F(1,290)5 0.15, P > 0.05).
After completion of training, a probe trial lacking the plat-
form was performed (Figs. 2G–J). Surprisingly, both, wild-type
and TgNL1 mice, showed a significant preference for the area of
the former platform location after training (F(3,79)5 21.34, P <
0.001; post hoc t 5 7.57, P < 0.001 for WTand t 5 2.59, P <
0.05 for TgNL1 mice, data not shown) as well as after reversal
training (F(3,79)5 18.07, P < 0.001 post hoc t 5 6.18, P <
0.001 for WT and t 5 3.97, P < 0.01 for TgNL1 mice, data
not shown), and no difference in the time spent in the former
platform area was detected between TgNL1 mice and their litter-
mate controls (P > 0.05; Figs. 2G,H,I,J), suggesting that despite
the poorer learning performance of TgNL1 mice, once learned,
the memory for the platform position remains intact. However,
TgNL1 mice did show a significant reduction in the number of
entries to the former platform area compared to their WT litter-
mates (F(1,18)5 10.16, P 5 0.005; reverse training: F(1,18)5
55.49, P < 0.001), implying the use of a different search strategy
or a lack of motivation to continue searching in TgNL1 mice.
Similar results were obtained from the Morris water maze
training paradigms (Fig. 3). Although no difference in the
ern blots showing representative samples from WT, TgNL1.6, and
TgNL1.7 whole brain lysate probed for HA, actin as a protein
loading control. B: Quantification of HA signal, shown as the per-
centage signal increase over background. C: Expression of HA-
NL1 in different tissues of TgNL1.7 mice. D: Representative sagit-
tal sections from WT, TgNL1.7 animals stained for HA by DAB
IHC. E: Expression of HA-NL1 during development. (d 5 days, m
5 months). F: Confocal analysis of HA-NL1 localization in
Expression of the Neuroligin 1 transgene. A: West-
TgNL1.7 mice. Top panel: colocalization of HA-NL1 with a
marker of excitatory synapses, PSD95, in the CA1 region of the
hippocampus. Bottom panel: colocalization of HA-NL1 with a
marker of inhibitory synapses, gephyrin, in the CA1 region of the
hippocampus. Scale bar: 2 lm. G: Quantification of HANL1
colocalization in TgNL1.7 mice. White bars: total area of HA-NL1
positive synapses. [Color figure can be viewed in the online issue,
which is available at www.interscience.wiley.com.]
DAHLHAUS ET AL.
water maze in mice expressing Neuroligin 1. A–C: Plus shaped
water maze training. D–F: Reverse training. A,B,D,E: Each data
point represents mean of summed results per training session.
Error bars indicate confidence interval (95%). ***P < 0.001. Black
symbols represent TgNL1.7 mice, while white symbols indicate
WT animals. A,D: Path length in meters. B,E: Escape latency in
seconds. C,F: Representative track plots from the last training ses-
sion. S, start. T, target (hidden plat form). G–J: Probe trial. G,I:
Impaired memory acquisition in the plus-shaped
White bars indicate the mean for WT, black bars for TgNL1 mice.
G: Forward training; I: reverse training; G,I (time) P > 0.05.
G(number of entries) F 5 10.16, **P 5 0.005; I (number of
entries) F 5 55.49, ***P < 0.001. H,J: Representative track plots
from the probe trial. S, start; T, target (hidden plat form); H, For-
ward training; J, reverse training. Note the cumulated paths at the
bottom of the target area in (H) resulting from ongoing for and
back swimming by the TgNL1 animal, whereas the WT animal dis-
plays repetitive tries for the former platform location.
NL1 OVER-EXPRESSION IMPAIRS LEARNING AND SYNAPTIC PLASTICITY
swim speed of TgNL1 and WT mice was detected (F(1,290)5
0.53, P > 0.05), TgNL1 mice show a significant increase in
the distance traveled (F(1,290)5 37.56, P < 0.001), the num-
ber of lines crossed (F(1,290) 5 46.68, P < 0.001), and the
escape latency (F(1,290)5 51.62, P < 0.001) when compared
with their littermate controls. During reverse training, the trend
toward an increase in the number of lines crossed was not sig-
nificantly different (F(1,174)5 2.67, P > 0.05), while both the
distance and escape latency of TgNL1 mice show a significant
increase in comparison with WT control animals (distance
F(1,174)5 21.68, P < 0.001; latency F(1,174)5 54.69, P <
0.001; Fig. 3A). These findings further confirm the deficits in
memory acquisition in TgNL1 mice.
Interestingly, the subsequent probe tests in the Morris water
maze (Figs. 3C–E) revealed that only the WT mice developed
a significant preference for the area of the former platform
location, while TgNL1 mice did not show any difference in
the time spent in the different quadrants of the Morris water
maze during both training paradigms (training: F(3,79)5 4.41,
P 5 0.0096; post hoc t 5 3.64, P < 0.01 for WT, but t 5
0.10, P > 0.05 for TgNL1 mice; reversal training: F(3,79)5
3.07, P 5 0.0403; post hoc t 5 3.02, P < 0.05 for WT, but
t 5 0.35, P > 0.05 for TgNL1 mice). These findings might
reflect the increased difficulty of the Morris water maze when
compared with the plus-shaped maze and this may have more
profoundly, and negatively, impacted the learning in the
TgNL1 mice were also assessed for spatial reference working
memory (Fig. 4) by changing the platform position after six
(Morris water maze) or eight (plus shaped water maze) consec-
Morris water maze in mice expressing Neuroligin 1. A: Representa-
tive track plots from the last training session. S, start; T, target
(hidden plat form). B: Forward (light bars) and revers (dark bars)
training in the morris water maze. Each bar displays the mean per-
centual change compared to wild-type animals (wt 5 0%), while
the error bars show the confidence interval (95%). ***P < 0.001,
Similar impairment in memory acquisition in the
ns: not significant. C–E: Probe trial. C: Representative track plots
from probe trial. S, start; T, target (hidden plat form). Number of
entries wild-type animal:12, TgNL1.7:7. D: Forward training. C:
Reverse training. C,D: Bars represent the average time spent in the
former platform area or the other areas, respectively. P 5 former
platform area. O 5 other areas.
the search strategy of mice expressing Neuroligin 1. A–C: Spatial
working memory training in the plus-shaped water maze. A,B:
Each data point represents mean of summed results per training
session. Error bars indicate confidence interval (95%). ***P <
0.001; **P < 0.01. Black symbols represent TgNL1.7 animals,
while white symbols indicate WT mice. C: Representative track
plots from the last training session. S, start; T, target (hidden plat
form); D–F: Spatial working memory training in the Morris water
maze. D,E: Each data point represents mean of summed results
per training session. Error bars indicate confidence interval (95%).
***P < 0.001; **P < 0.01. Black symbols represent TgNL1.7ani-
Working memory impairment and alterations in
mals, whereas white symbols indicate WT mice. F: Representative
track plots from the last training session. S, start; T, target (hidden
plat form). G: Comparison of the search performance of WT,
TgNL1 animals for wall-platform positions (dark bars) and center-
platform positions (light bars). Each bar displays the mean percent
change compared to WT animals, while the error bars show the
confidence interval (95%). ***P < 0.001; **P < 0.01; *P < 0.05,
ns: not significant. H,I: Representative track plots from a final trial
for a center position. S, start; T, target (hidden plat form). Punc-
tate circles: potential platform positions. Note that occasionally,
TgNL1 animals, due to the displayed searching pattern, never find
the platform if it is hidden at a wall position (data not shown).
DAHLHAUS ET AL.
NL1 OVER-EXPRESSION IMPAIRS LEARNING AND SYNAPTIC PLASTICITY
utive trials, respectively, in a daily manner. In both mazes,
TgNL1 mice as well as their WT littermates demonstrate a reg-
ular learning curve as indicated by a significant effect of train-
ing session [latency (plus-shaped water maze) F(7,384)5 4.89,
P < 0.001; latency (Morris water maze) F(5,290)5 5.86, P 5
0.006; distance (plus-shaped water maze] F(7,384)5 5.07, P <
0.001; distance (Morris water maze) F(5,290) 5 5.87, P 5
0.006). But when compared with their WT littermates, TgNL1
mice show a significant increase in the distance traveled
(F(1,384)5 9.07 and P 5 0.003 in the plus-shaped water maze
and F(1,290) 5 12.08 and P 5 0.005 in the Morris water
maze) and the latency (F(1,384)5 46.43 and P < 0.001 in the
plus-shaped water maze and F(1,290)5 62.56 and P < 0.001
in the Morris water maze) to escape from the water maze,
reflecting deficits in working memory acquisition in TgNL1
mice. The track plots (Figs. 4A,D) of the final trial illustrate
the lack of spatial reference memory formation in TgNL1 mice
and furthermore point to the use of a different search strategy
in these mice.
To investigate the possibility of an alternative search strategy,
the ability of TgNL1 mice to locate wall or center platform
positions, respectively, was analyzed in the Morris water maze
(Figs. 4G–I). A significant increase in the latency (F(3,23) 5
34.18, P < 0.001; post hoc t 5 8.56 and P < 0.001), distance
(F(3,23)5 7.75, P 5 0.002; post hoc t 5 4.22 and P < 0.01)
and number of lines crossed (F(3,23)5 7.61, P 5 0.002; post
hoc t 5 4.25 and P < 0.01) can be observed during the search
for wall positions by TgNL1 mice when compared to their
WT littermates (Fig. 4I). By contrast, no significant difference
(F(3,23)5 7.75, P 5 0.002; post hoc t 5 0.90 and P > 0.05)
between WT and TgNL1 mice can be detected in the distance
and number of lines crossed to escape from the water maze
(F(3,23)5 7.61, P 5 0.002; post hoc t 5 0.62 and P > 0.05),
when the platform is hidden at one of the center positions,
although the escape latency of TgNL1 mice is still significantly
increased (F(3,23)5 34.18, P < 0.001; post hoc t 5 4.98 and
P < 0.001) compared to WT animals. The track plots (Figs.
4G,H) illustrate that the circular search pattern typically used
by TgNL1 mice favors the detection of a platform hidden at a
Mice Expressing NL1 Show Changes in the
Expression Levels of Related Synaptic Proteins
Western-blot analysis (Fig. 5) of protein expression levels
showed that increasing the expression of NL1 by 112% results
in a significant upregulation of marker proteins for excitatory
(PSD95: 16% 6 7%; VGluT: 30% 6 16%) and inhibitory
(Gephyrin: 26% 6 11%; VGAT: 29% 6 6%) synapses in
TgNL1 mice. The expression level of GluR2/3 subunits was
significantly decreased (230% 6 10%) in TgNL1 animals,
whereas expression of the GluR1 subunit was significantly
increased (122% 6 35%), thereby indicating altered AMPA re-
ceptor properties and/or numbers at synapses. Further analysis
by IHC confirmed these findings (GluR1: 146% 6 14%, P <
0.000; GluR2/3: 258% 6 8%, P < 0.000). However, there
were no changes in endogenous NL2, synaptophysin, or synap-
totagmin expression levels. Using IHC staining for the synaptic
marker proteins, similar changes in VGluT (82% 6 18%) and
VGAT (75% 6 24%) were observed (Figs. 5C,D) in TgNL1.7
mice. These findings are also in line with recent in vitro studies
in mice expressing Neuroligin 1. A: Representative western blot
strips illustrating expression levels of several synaptic marker pro-
teins from WT, TgNL1.7 mice. B: Quantification, showing the per-
cent change in expression compared to WT expression level
(middle line). Proteins showing significant changes are indicated
by stars: P (NL1) < 0.0001, P (NL3) 5 0.0052, P (gephyrin) 5
0.0198, P (PSD95) 5 0.0116, P (VGluT) 5 0.0289, P (VGAT) 5
0.0026, P (GluR2/3) < 0.0001. Synphys: Synaptophysin, Syntag:
Synaptotagmin. C: Representative immunofluorescence staining for
excitatory (VGluT1 1 2) and inhibitory (VGAT) presynaptic
marker proteins in WT (left panel) and TgNL1.7 (right panel)
hippocampus (CA1). Scale bar: 2 lm. D: Quantification of puncta
areas positive for excitatory or inhibitory synaptic marker proteins,
respectively, by confocal analysis of immunofluorescence labeling.
White bars: WT; black bars: TgNL1.7, ***P < 0.0001 and **P <
0.001. [Color figure can be viewed in the online issue, which is
available at www.interscience.wiley.com.]
Alterations in the expression of synaptic proteins
DAHLHAUS ET AL.
showing that NLs can influence both excitatory and inhibitory
synapses (Levinson et al., 2005; Chih et al., 2006).
Alterations in Spine and Synapse Morphology in
Mice Expressing NL1
Because of prior in vitro results showing that NL expression
increases synapse maturation and stabilization (Scheiffele et al.,
2000; Graf et al., 2004; Prange et al., 2004; Chih et al., 2005;
Levinson et al., 2005; Nam and Chen, 2005), we wanted to
assess synapse morphology in TgNL1 mice. Assessment of spine
morphology using Golgi impregnation revealed an increase in
the head size (21% 6 4.4%, P < 0.001) of mushroom-shaped
spines, with a decrease in the neck length (226% 6 8.8%, P <
0.001) in CA1 of TgNL1.7 (Figs. 6A,B). No change was
observed in the overall length of these spines in TgNL1.7 mice
(n 5 4 animals) compared with littermate controls (n 5 4 ani-
mals; Fig. 6B). It was also observed using Golgi impregnation
that the spines of CA1 pyramidal neurons in the TgNL1.7 mice
had complex or lobed morphologies (Fig. 6A, arrow heads).
The dramatic changes observed in spine
prompted a more detailed assessment of synapse morphology
via EM. Quantitative assessment of synaptic elements (Fig. 6D)
revealed a significant increase in the average length (nm) of
asymmetric synaptic contacts (WT 5 649.73 6 40.01,
TgNL1.6 5 785.40 6 46.77, TgNL1.7 5 878.39 6 74.22;
ANOVA F(2,121) 5 4.21 P 5 0.017; LSD post hoc: P 5
0.048 WT vs. TgNL1.6, P 5 0.005 WT vs. TgNL1.7) in
TgNL1.6 and TgNL1.7 stratum radiatum (Fig. 6F), with no
change in the average length (nm) of the PSD (Fig. 6E; WT
5 363.47 6 20.75, TgNL1.6 5 36.83 6 14.89, TgNL1.7 5
394.26 6 20.67; ANOVA F(2,121)5 0.797 P 5 0.453). The
increased synaptic contact length seen in TgNL1.6 and
TgNL1.7 mice is in agreement with the Golgi findings that
in mice expressing Neuroligin 1. A: Representative images of den-
dritic spines in the CA1 region of the hippocampus of WT,
TgNL1.7 mice. Scale bar: 3 lm. B: Quantification of spine mor-
phology in the CA1 region of the hippocampus. Bars show percent
change compared to WT for total spine length, head and neck
size, ns 5 not significant, ***P < 0.001. C: Representative micro-
graphs of asymmetric synapses in WT, TgNL1 stratum radiatum.
Scale bar: 500 nm. D: Schematic diagram of the two representative
synaptic contacts showing the morphological parameters assessed.
Changes in the morphology of spines and synapses
E: Quantification of the average length of the postsynaptic density
and length of contact between pre and postsynaptic compartments.
F: Quantification of the curvature of synaptic contacts, calculated
by dividing the length of contact by the straight line length across
the contact. G: Quantification of the average number of docked
and reserve pool vesicles of asymmetric synapses in WT, TgNL1.6,
and TgNL1.7 mice. [Color figure can be viewed in the online
issue, which is available at www.interscience.wiley.com.]
NL1 OVER-EXPRESSION IMPAIRS LEARNING AND SYNAPTIC PLASTICITY
these spines have increased head size. We also assessed the cur-
vature of synaptic contacts by dividing the total contact length
by the straight line length across the contact (Fig. 6D). This
analysis revealed an increase in the curvature of synaptic con-
tacts (WT 5 1.11 6 0.03, TgNL1.6 5 1.44 6 0.09,
TgNL1.7 5 1.64 6 0.11; ANOVA F(2,121) 5 10.16 P <
0.001; LSD post hoc: P 5 0.006 WT vs. TgNL1.6, P <
0.001 WT vs. TgNL1.7) in both TgNL1.6 and TgNL1.7 syn-
apses compared to WT controls (Fig. 6F). It is unclear why we
were able to observe an increase in PSD95 expression via WB
while no change in the length of the PSD was detected using
EM. The change observed via WB may instead reflect the
increase in the number of excitatory synapses. It is also interest-
ing to consider the lack of change in PSD length in light of
the change observed in synaptic contact length, as increases in
synapse size commonly parallel one another. This may suggest
that although expression of NL1 can result in altered synapse
morphology, NL1 may not be sufficient to recruit the full
complement of proteins that make up the PSD. However,
because increased head to neck ratio, synaptic contact length,
and curvature of synapses are hallmarks of stable or mature
spines (Marrone and Petit, 2002; Bourne and Harris, 2007),
together these findings suggest that TgNL1 mice display a
greater proportion of spines with a mature morphology com-
pared to WT littermates.
Mice expressing NL1 Show a Shift in the Ratio
of Excitation to Inhibition
Because NLs have been indicated in regulating the ratio of
excitatory to inhibitory synapses (Prange et al., 2004; Levinson
et al., 2005; Nam and Chen, 2005; Chih et al., 2006; Graf
et al., 2006; Kang et al., 2008), we wanted to examine in more
detail whether a change in the E/I ratio can be observed in
TgNL1 mice. Using Golgi impregnation, we observed that hip-
pocampal dendrites of CA1 in TgNL1.7 mice show an increased
density of spines (27% 6 8.4%, P 5 0.0303) compared to den-
drites of littermate controls (Figs. 7A,B). We did not observe a
change in the overall branching of dendrites (28%, 6 15%,
P 5 0.730) when comparing TgNL1.7 with littermate controls.
Because the increase in the number of spines observed in
TgNL1 CA1 may suggest an increase in synapses, we wanted
to assess the number of synapses in the stratum radiatum of
the hippocampus, adjacent to CA1, using EM (Fig. 7C). EM
assessments revealed an increase in the number of asymmetric
(typically excitatory) synapses (WT 5 26.74 6 2.15, TgNL1.6
5 45.28 6 3.34, TgNL1.7 5 49.55 6 4.74; ANOVA F(2,44)
5 11.55 P < 0.001; LSD post hoc: P 5 0.001 WT vs.
TgNL1.6, P < 0.001 WT vs. TgNL1.7) in both TgNL1.6 and
TgNL1.7 compared to WT controls. This finding is consistent
with the increased spine density, and parallels increases in
and synapse number, and a shift in the ratio of excitation to inhi-
bition. A: Representative images of dendrites in the CA1 region of
the hippocampus of WT, TgNL1.7 mice. Scale bar: 5 lm. B:
Quantification of dendritic spine number (spines/lm) in the CA1
region of the hippocampus. White bar: WT animals and black bar:
Mice expressing Neuroligin 1 show increased spine
TgNL1.7 mice, *P (number) 5 0.0293. C: Quantification of the
number of asymmetric (typically excitatory) and symmetric (typi-
cally inhibitory) synapses (per 100 lm2) using EM in the stratum
radiatum of WT, TgNL1.6, and TgNL1.7 mice. D: Calculation of
the number of excitatory (asymmetric) to inhibitory (symmetric)
synapses in WT, TgNL1.6 and TgNL1.7 mice.
DAHLHAUS ET AL.
excitatory synaptic markers (VGluT and PSD-95) observed
using WB. In contrast to asymmetric synapses, no change was
observed in the number of symmetric (typically inhibitory;
WT 5 7.14 6 1.70, TgNL1.6 5 8.56 6 1.54, TgNL1.7 5
7.84 6 1.94; ANOVA F(2,44)5 0.168; P 5 0.846) synapses.
It is unclear why the increases in the inhibitory markers
VGAT and gephyrin observed are not paralleled by an increase
in symmetric synapses. It is possible that protein expression is
increased as a means of compensation for the dramatic
increase observed in excitatory synapses, but the formation and
alteration of symmetric synapses might fail in vivo due to reg-
ulatory mechanisms. To obtain an estimate of the excitatory to
inhibitory ratio, we divided the number of asymmetric synapses
(excitatory) by the number of symmetric synapses (inhibitory;
Fig. 7D). This calculation reveals a shift in the ratio toward
increased excitation (WT 5 4.95 6 0.86, TgNL1.6 5 8.55 6
0.77, TgNL1.7 5 8.11 6 0.82; ANOVA F(2,44)5 5.87; P 5
0.008; LSD post hoc: P 5 0.004 WT vs. TgNL1.6, P 5 0.013
WT vs. TgNL1.7) in stratum radiatum of both TgNL1.6 and
Impairments in the Induction of LTP in the
Hippocampus of Mice Expressing NL1
To further elucidate how these structural alterations might
alter synaptic communication and thus exert their effects on
behavior, we also examined synaptic plasticity in the CA1 sub-
field of the hippocampus (Fig. 8). We did not find any evi-
dence for alterations in presynaptic transmitter release in the
TgNL1 animals (P 5 0.36), who showed PP facilitation
(PP50: 15.98% 1 1.04%; n 5 10) that was equivalent to that
observed in WT littermates (PP50: 17.54% 1 1.45%; n 5 9).
However, when 100 Hz conditioning stimuli were applied, the
TgNL1 animals showed significantly (P < 0.05) less short-term
potentiation (STP: 22.22% 1 9.15%, n 5 11) and LTP (LTP:
13.99% 1 4.69%; n 5 11) than their WT littermates (STP:
115.5% 1 17.5%; LTP: 50.1 1 5.71; n 5 8; Fig. 8A). This
reduction in synaptic efficacy did not appear to be due to a
reduction in the threshold for LTP induction, as even adminis-
tering the conditioning stimuli multiple times (33), did not
induce significant LTP in TgNL1 animals (1.05% 1 20.7%;
n 5 9; Fig. 8B) even though they did exhibit robust STP
(97.47% 1 36.56%). Thus, overexpression of NL1 reduces the
capacity of the CA1 region of the hippocampus to exhibit
long-lasting alterations in synaptic efficacy.
Increased Basal Excitation in the Hippocampus
of Mice Expressing NL1
To address the functional relationship between excitation and
inhibition, we examined the response of the population of cells
to the removal of inhibition using bicuculline and examined
the evoked excitatory and inhibitory responses of individual
pyramidal neurons. PSs were recorded in the pyramidal cell
layer and induced by stimulation of the Schaffer’s collaterals
(Pavolov et al., 2004). After a minimum of 15-min stable base-
line responses, bicuculline methiodide (5 lM) was applied for
30 min. Following blockade of inhibition, the PS increased
significantly less in TgNL1 mice (152% 1 18%; n 5 5) com-
pared to controls (515% 1 116%; n 5 5) (Figs. 9A,B). This
may suggest that the hippocampus is basally more excitable due
to an increase in excitation or a decrease in inhibition as a
result of NL1 expression (consistent with our observation that
PSs were more easily obtained from TgNL1 mice). To test this
more directly, we performed whole-cell recordings and evoked
EPSCs and IPSCs at increasing stimulation strengths. We
observed larger magnitude EPSCs (evident at multiple stimula-
tion strengths) in TgNL1 mice (2151.2 pA; n 5 4) relative to
nificantly impaired in mice expressing Neuroligin 1. A: Application
of high frequency stimuli (100 Hz) reliably induces long-term
potentiation of synaptic efficacy in WT animals (open circles).
These same stimuli are far less efficacious when applied to hippo-
campal slices taken from mice overexpressing NL1 (closed circles).
Insets show example evoked responses taken from WT (left) and
TgNL1 (right) CA1 subfields. Note that the initial size of evoked
responses in both animals is equivalent, but that only the TgNL1
response fails to enhance. B: Comparison of the percentage change
in the EPSP slope at one hour postconditioning for WT animals
(white bar at left); TgNL1 animals administered one bout of high-
frequency stimulation (HFS) (black bar in center); and TgNL1
administered three bouts of HFS (gray bar at right). In all instan-
ces, the TgNL1 animals exhibited significantly less LTP than WT
animals. (*P < 0.05).
Long-term potentiation in the CA1 region is sig-
NL1 OVER-EXPRESSION IMPAIRS LEARNING AND SYNAPTIC PLASTICITY
controls (259.6; n 5 5) (Figs. 9C,D). In contrast, we did not
observe a difference in the magnitude of IPSCs between geno-
types (TgNL1: 184.8 pA; n 5 4) (WT: 161.0 pA; n 5 5)
(Figs. 9E,F). At a functional level, it appears that overexpres-
sion of NL1 leads to an increase in basal excitation in the
hippocampus without altering inhibition.
The present experiments focused on NL1, a postsynaptic
cell-adhesion molecule preferentially localized at excitatory syn-
apses and examined NL1’s function in synapse development,
synaptic plasticity and learning in vivo. Using mice overexpress-
obtained from the stratum pyramidale in response to stratum radi-
atum stimulation in age-matched WT, TgNL1 slices. B: Time
course of changes in population spike amplitude in field record-
ings obtained following bath application of bicuculline methio-
dide. C: Increasing evoked synaptic stimulation (pulse width) to
the Schaffer’s collaterals induces larger excitatory postsynaptic cur-
rents (EPSCs) in mice expressing Neuroligin 1 (TgNL1; closed
circles) relative to wild-type (WT; open circles) mice. D: Averaged
A: Representative traces showing population spikes
traces from one set of IO EPSCs from WT mice (gray lines) and
TgNL11 mice (black lines). E: Increasing evoked synaptic stimula-
tion (pulse width) to the Schaffer’s collaterals induces similar in-
hibitory postsynaptic currents (IPSCs) in mice expressing Neuroli-
gin 1 (TgNL1; closed circles) relative to wild-type (WT; open
circles) mice. F: Averaged traces from one set of IO IPSCs from
WT mice (gray lines) and TgNL11 mice (black lines). Vertical
scale bar represents 50 pA and horizontal scale bar represents 10
ms. Scale bars are identical for D and F.
DAHLHAUS ET AL.
ing NL1 as a model, we found that increased NL1 expression
leads to striking deficits in memory acquisition, paralleled by
impairment in the induction of LTP, altered spine morphology,
and enhanced excitatory synaptic transmission.
NL1 Is Involved in Memory Acquisition
NLs have been primarily implicated in autism spectrum dis-
orders (ASD) as mutations in the genes encoding NL1, NL3,
and NL4 have been associated with some rare cases of inher-
ited ASD (Jamain et al., 2003; Hines et al., 2008; Jamain
et al., 2008; Lawson-Yuen et al., 2008; Talebizadeh et al.,
2005; Yan et al., 2005; Ylisaukko-oja et al., 2005). Although
the involvement of NLs in synapse formation has also been
addressed in several studies (Lise and El-Husseini, 2006; Craig
and Kang, 2007), a role for NLs in synaptic plasticity and
memory formation is currently being established. Tabouchi
et al. (2007) found that the Arg451 > Cys451 substitution in
NL3 that has been linked to ASD, not only induced social
deficits in transgenic mice, but also increased their spatial
learning performance. However, as a deletion of NL3 did not
cause corresponding changes, these findings represent a specific
gain of function mutation. By contrast, NL1 has just recently
been demonstrated to be necessary for the expression of LTP
in the amygdala as well as the development of associative fear
memory in adult animals (Kim et al., 2008). In line with
these findings, our data now indicate that NL1 expression lev-
els can also influence spatial learning and that, moreover,
increased expression levels result in the blockage of LTP in the
hippocampus. Importantly, neither of the TgNL1 strains dis-
played deficits in basic sensory, reflexive, and motor function
or exploratory motivation, thereby confirming TgNL1 mice as
candidates for detailed behavioral characterization.
In the plus-shaped and Morris water maze tasks, TgNL1
mice showed a striking reduction of their performance during
memory acquisition training. However, as indicated by the
swim paths and the strategy analysis, TgNL1 mice also devel-
oped an alternative search strategy, which might contribute to
their improvement in performance. Interestingly, the probe tests
conducted after memory acquisition revealed different results in
the plus shaped and the Morris water maze, respectively:
Although TgNL1 mice displayed a significant preference for
the area of the former platform location in the plus-shaped
water maze, they failed to do so in the Morris water maze.
These results indicate that the ability of TgNL1 mice to de-
velop true spatial memory is limited. In this context, the
observed reduction in the number of entries into the former
platform area by TgNL1 mice might reflect the use of a differ-
ent search strategy or indicate a lack of motivation to continue
searching after the platform is not found. On the other hand,
these results also suggest that the formation and/or employment
of true spatial memory by TgNL1 mice is limited as the
increased difficulty mice experience in the Morris water maze,
which was in contrast to the plus-shaped maze originally
designed for rats, is sufficient to prevent TgNL1 mice from the
formation of true spatial memory, but not their WT litter
Memory acquisition in the water maze is thought to be
largely determined by the ability of mice to develop appropri-
ate strategies. Therefore, acquisition performance is thought to
reflect not only the capacity for an animal to learn, but also
the degree of behavioral flexibility. The circular, or chaining
response strategy displayed by TgNL1 mice, usually a transi-
tional search behavior, has been described earlier (Janus, 2004;
Brody and Holtzman, 2006), but is less commonly found than
a thigmotactic search pattern, which is often indicative of
increased anxiety (Sakic et al., 1993; Oitzl et al., 1997; Gass
et al., 1998; Bjorklund et al., 1999; Champagne et al., 2002;
Lang et al., 2003; Schmitt et al., 2006; Wilcoxon et al., 2007;
Zhu et al., 2007). Interestingly, the chaining strategy was not a
useful alternative for our mice as it fails to detect five of nine
possible positions. Although more efficient strategies can be
adopted by the WT mice if chaining does not provide a solu-
tion (Brody and Holtzman, 2006), TgNL1 mice failed to adapt
their strategy, indicating impaired procedural learning and a
lack of flexibility. Similar results have been obtained with rats
when the septohippocampal cholinergic system was impaired
(Sutherland et al., 1982).
NL1 Expression Results in Increased Ratio
of Excitation to Inhibition
Western blot and IHC revealed an increased expression of
marker proteins for excitatory and inhibitory synapses. Closer
examination using EM revealed an increase in the number of
asymmetric (excitatory), but no change in the number of sym-
metric (inhibitory) synapses. Furthermore, we observed an
increase in the contact length of asymmetric synapses that was
not observed with symmetric synapses. This data suggests that
in vivo, NL1 expression mainly enhances excitatory synapse
formation, although it also elevates the expression of proteins
resident at inhibitory synapses. The finding that excitatory
activity was not enhanced as much in the TgNL mice as the
WT when inhibition was blocked indicates that inhibition is
playing a lesser role in these animals. This could be because (1)
there is already greater excitatory input at activated synapses, so
removing inhibition has less of an effect, or (2) because there
are fewer inhibitory synapses in TgNL1 mice. As no differences
were observed in inhibitory synapse numbers, it seems likely
that there has been a shift toward excitation. This seems likely
given the shift in the E/I ratio observed in the whole-cell
recordings made from the TgNL mice.
It remains unclear why no changes were observed in the
number or function of inhibitory synapses in the hippocampus
despite observations of increased inhibitory synaptic markers
via Western blotting. One possibility may be that NL1 expres-
sion can alter the recruitment of select inhibitory synaptic pro-
teins but is not sufficient to alter the morphology, number, or
function of inhibitory synapses. This effect could occur due to
direct effects of NL1, but most likely reflects an indirect action
of NL1 by altering other NLs such as NL3, which has been
NL1 OVER-EXPRESSION IMPAIRS LEARNING AND SYNAPTIC PLASTICITY
identified at inhibitory synapses (Budreck and Scheiffele,
2007), and is found to be increased in TgNL1 mice. It is also
possible that the increased VGAT staining observed via Western
blotting and IHC reflects changes in nonsynaptic or even non-
neuronal clusters of VGAT (Minelli et al., 2003), because post-
synaptic markers were not used to confirm that the VGAT
clusters quantified were synaptic.
A previous paper has shown that NL2 expression results in
an increase in inhibitory synaptic markers, the maturation, and
transmission of inhibitory synapses, but also increases the
expression of VGluT (Hines et al., 2008). Thus, in vivo studies
of NL1 and NL2 function suggest that activity of NLs is not
fully restricted to effects on either excitatory or inhibitory syn-
apses but may act more subtly to sway the ratio of excitation
to inhibition. Furthermore, these two papers taken together
demonstrate the influence of altering the ratio of excitation to
inhibition on brain function and behavior, as the two strains of
mice (expressing either NL1 or NL2) display differences in the
behavioral phenotype observed.
Synapse Morphology and Regulation of Synaptic
Plasticity by NL1
The hypothesis that the physical substrate of memory in the
brain resides in alterations of synaptic efficacy has been pro-
posed frequently (Bliss and Collingridge, 1993; Bliss et al.,
2007). The most common long-term alteration in synaptic effi-
cacy applied to study memory is LTP. LTP is an activity-
dependent strengthening of synaptic connections induced by
calcium influx through NMDA receptors leading to insertion
of AMPA receptors (Andrasfalvy and Magee, 2004; Ehrlich and
Malinow, 2004). LTP has also been reported to be accompa-
nied by alterations in spine morphology (Marrone and Petit,
2002; Bourne and Harris, 2007).
Interestingly, TgNL1 mice show a shift toward spines with a
more mature appearance, which is paralleled by impairment in
LTP induction and learning, but not memory deficits. This
supports the idea that larger spines are less capable of adapta-
tion in response to synaptic activity and that maintaining a
population of more flexible, thin spines, which can form, dis-
appear or reshape more rapidly (Zuo et al., 2005; Holtmaat
et al., 2006), may be essential for plasticity and learning (Kasai
et al., 2003; Bourne and Harris, 2007). In addition, spine mor-
phology can also regulate calcium handling (Rusakov et al.,
1996; Majewska et al., 2000; Holthoff and Tsay, 2002; Nogu-
chi et al., 2005), with spines with longer or narrower necks
retaining more calcium in their heads following activation. Fur-
thermore, it has also been shown that protein diffusion is
regulated by spine morphology (Bloodgood and Sabatini,
2005). In particular, proteins such as inositol 1,4,5-triphos-
phate (Santamaria et al., 2006) and PSD95 (Gray et al., 2006)
can be more effectively trapped in spines with narrower or lon-
ger necks, further impacting calcium dynamics and/or synaptic
efficacy. Therefore, it is possible that the larger spines with
shorter necks observed in the hippocampus of TgNL1 mice
may demonstrate aberrant calcium handling and thus impair-
ments in signaling cascades that follow synaptic activation.
Under normal conditions, the volume of the spine head is
directly proportional to the number of postsynaptic receptors
(Nusser et al., 1998; Matsuzaki et al., 2001) and the number
of docked vesicles in the presynaptic bouton (Schikorski and
Stevens, 1999). Although Golgi and EM analysis showed an
increase in the head size of spines/synapses in TgNL1 mice,
EM studies revealed no change in the number of docked
vesicles or the size of the PSD. This may suggest that the
morphologically mature-appearing spines in TgNL1 mice are
in fact lacking corresponding alterations in functional compo-
nents, resulting in a discrepancy between morphological
properties and functional capabilities of asymmetric synapses.
This possibility may indicate that NL1 is not sufficient to
induce the formation of fully functional synapses in vivo,
but instead might influence the composition and/or numbers
of AMPA receptors. This idea is supported by the Western-
blot analyses and the fact that these animals show impaired
LTP expression. These findings also reflect that the coordi-
nated action of multiple proteins or pathways is required to
build fully functional synapses. By interrupting the coordi-
nated process of synapse formation, we have reduced the
capacity for processes like LTP. Taken together, our data sug-
gest that neither excessively thin nor excessively large spines
are adaptive and that spine morphology and functional sig-
naling machinery must be appropriately balanced for plastic-
ity and learning.
In conclusion, we have detailed synapse and behavioral
abnormalities in mice with altered expression of NL1 that dif-
fer from those observed with mice expressing NL2 (Hines
et al., 2008). These data demonstrate the overlapping, yet dis-
tinct roles for specific NL family members, and further sheds
light on the importance of the ratio of excitation to inhibition
for brain function and behavior. Our new findings demonstrate
the influence of NL1 expression in vivo on synapse morphol-
ogy, LTP, and acquisition of spatial learning in the water maze.
Furthermore, these findings draw a link between aberrant syn-
apse morphology, impaired synaptic plasticity, and deficits in
BRC and AEH are Michael Smith Foundation for Health
Research Senior Scholars. RSH and BDE are MSFHR and CIHR
graduate awardees. RD is supported by CIHR and FXRFC. This
work is dedicated to the memory of Dr. Alaa El-Husseini, a men-
tor and friend that we all miss dearly.
Andrasfalvy BK, Magee JC. 2004. Changes in AMPA receptor currents
following LTP induction on rat CA1 pyramidal neurones. J Physiol
Beique JC, Andrade R. 2003. PSD-95 regulates synaptic transmission
and plasticity in rat cerebral cortex. J Physiol 546(Pt 3):859–867.
DAHLHAUS ET AL.
Bjorklund M, Sirvio J, Sallinen J, Scheinin M, Kobilka BK, Riekkinen
P Jr. 1999. a2C-Adrenoceptor overexpression disrupts execution of
spatial and non-spatial search patterns. Neuroscience 88:1187–
Bliss TV, Collingridge GL. 1993. A synaptic model of memory: Long-
term potentiation in the hippocampus. Nature 361:31–39.
Bliss T, Collingridge G, Morris R. 2007. Synaptic plasticity in the hip-
pocampus. In: Andersen P, Morris R, Amarakl D, Bliss T, O’Keefe
J, editors. The Hippocampus Book. New York: Oxford University
Bloodgood BL, Sabatini BL. 2005. Neuronal activity regulates diffu-
sion across the neck of dendritic spines. Science 310:866–869.
Bourne J, Harris KM. 2007. Do thin spines learn to be mushroom
spines that remember? Curr Opin Neurobiol 17:381–386.
Brandeis R, Brandys Y, Yehuda S. 1989. The use of the Morris
Water Maze in the study of memory and learning. Int J Neurosci
Brody DL, Holtzman DM. 2006. Morris water maze search strategy
analysis in PDAPP mice before and after experimental traumatic
brain injury. Exp Neurol 197:330–340.
Budreck EC, Scheiffele P. 2007. Neuroligin-3 is a neuronal adhesion
protein at GABAergic and glutamatergic synapses. Eur J Neurosci
Caroni P. 1997. Overexpression of growth-associated proteins in the
neurons of adult transgenic mice. J Neurosci Methods 71:3–9.
Champagne D, Dupuy JB, Rochford J, Poirier J. 2002. Apolipopro-
tein E knockout mice display procedural deficits in the Morris
water maze: Analysis of learning strategies in three versions of the
task. Neuroscience 114:641–654.
Chih B, Gollan L, Scheiffele P. 2006. Alternative splicing controls
selective trans-synaptic interactions of the neuroligin-neurexin com-
plex. Neuron 51:171–178.
Christie BR, Li AM, Redila VA, Booth H, Wong BK, Eadie BD,
Ernst C, Simpson EM. 2005. Deletion of the nuclear receptor
Nr2e1 impairs synaptic plasticity and dendritic structure in the
mouse dentate gyrus. Neuroscience 137:1031–1037.
Chubykin AA, Atasoy D, Etherton MR, Brose N, Kavalali ET, Gibson
JR, Sudhof TC. 2007. Activity-dependent validation of excitatory
versus inhibitory synapses by neuroligin-1 versus neuroligin-2.
Contractor A, Heinemann SF. 2002. Glutamate receptor trafficking in
synaptic plasticity. Sci STKE 2002:RE14.
Craig AM, Kang Y. 2007. Neurexin-neuroligin signaling in synapse
development. Curr Opin Neurobiol 17:43–52.
Dean C, Scholl FG, Choih J, DeMaria S, Berger J, Isacoff E,
Scheiffele P. 2003. Neurexin mediates the assembly of presynaptic
terminals. Nat Neurosci 6:708–716.
Dresbach T, Neeb A, Meyer G, Gundelfinger ED, Brose N. 2004.
Synaptic targeting of neuroligin is independent of neurexin and
SAP90/PSD95 binding. Mol Cell Neurosci 27:227–235.
Ehrlich I, Malinow R. 2004. Postsynaptic density 95 controls AMPA
receptor incorporation during long-term potentiation and experi-
ence-driven synaptic plasticity. J Neurosci 24:916–927.
Ehrlich I, Klein M, Rumpel S, Malinow R. 2007. PSD-95 is required
for activity-driven synapse stabilization. Proc Natl Acad Sci USA
El-Husseini AE, Craven SE, Chetkovich DM, Firestein BL, Schnell E,
Aoki C, Bredt DS. 2000a. Dual palmitoylation of PSD-95 medi-
ates its vesiculotubular sorting, postsynaptic targeting, and ion
channel clustering. J Cell Biol 148:159–172.
El-Husseini AE, Schnell E, Chetkovich DM, Nicoll RA, Bredt DS.
2000b. PSD-95 involvement in maturation of excitatory synapses.
Futai K, Kim MJ, Hashikawa T, Scheiffele P, Sheng M, Hayashi Y.
2007. Retrograde modulation of presynaptic release probability
through signaling mediated by PSD-95-neuroligin. Nat Neurosci
Gass P, Wolfer DP, Balschun D, Rudolph D, Frey U, Lipp HP, Schutz
G. 1998. Deficits in memory tasks of mice with CREB mutations
depend on gene dosage. Learn Mem 5:274–288.
Gerrow K, Romorini S, Nabi SM, Colicos MA, Sala C, El-Husseini
A. 2006. A preformed complex of postsynaptic proteins is involved
in excitatory synapse development. Neuron 49:547–562.
Glaser EM, Van der Loos H. 1981. Analysis of thick brain sections by
obverse-reverse computer microscopy: Application of a new, high
clarity Golgi-Nissl stain. J Neurosci Methods 4:117–125.
Graf ER, Kang Y, Hauner AM, Craig AM. 2006. Structure function
and splice site analysis of the synaptogenic activity of the neurexin-
1 b LNS domain. J Neurosci 26:4256–4265.
Gray NW, Weimer RM, Bureau I, Svoboda K. 2006. Rapid redistribu-
tion of synaptic PSD-95 in the neocortex in vivo. PLoS Biol
Hines RM, Wu L, Hines DJ, Steenland H, Mansour S, Dahlhaus R,
Singaraja RR, Cao X, Sammler E, Hormuzdi SG, Zhuo M, El-Hus-
seini A. 2008. Synaptic imbalance, stereotypies, and impaired social
interactions in mice with altered neuroligin 2 expression. J Neurosci
Holthoff K, Tsay D. 2002. Calcium dynamics in spines: Link to syn-
aptic plasticity. Exp Physiol 87:725–731.
Holtmaat A, Wilbrecht L, Knott GW, Welker E, Svoboda K. 2006.
Experience-dependent and cell-type-specific spine growth in the
neocortex. Nature 441:979–983.
Jamain S, Quach H, Betancur C, Rastam M, Colineaux C, Gillberg
IC, Soderstrom H, Giros B, Leboyer M, Gillberg C, Bourgeron T;
Paris Autism Research International Sibpair Study. 2003. Mutations
of the X-linked genes encoding neuroligins NLGN3 and NLGN4
are associated with autism. Nat Genet 34:27–29.
Jamain S, Radyushkin K, Hammerschmidt K, Granon S, Boretius S,
Varoqueaux F, Ramanantsoa N, Gallego J, Ronnenberg A, Winter
D, Frahm J, Fischer J, Bourgeron T, Ehrenreich H, Brose N. 2008.
Reduced social interaction and ultrasonic communication in a
mouse model of monogenic heritable autism. Proc Natl Acad Sci
Janus C. 2004. Search strategies used by APP transgenic mice during
navigation in the Morris water maze. Learn Mem 11:337–346.
Kang Y, Zhang X, Dobie F, Wu H, Craig AM. 2008. Induction of
GABAergic postsynaptic differentiation by a-neurexins. J Biol
Kasai H, Matsuzaki M, Noguchi J, Yasumatsu N, Nakahara H. 2003.
Structure-stability-function relationships of dendritic spines. Trends
Kim J, Jung SY, Lee YK, Park S, Choi JS, Lee CJ, Kim HS, Choi YB,
Scheiffele P, Bailey CH, Kandel ER, Kim JH. 2008. Neuroligin-1
is required for normal expression of LTP, associative fear memory
in the amygdala of adult animals. Proc Natl Acad Sci USA
Lang UE, Lang F, Richter K, Vallon V, Lipp HP, Schnermann J,
Wolfer DP. 2003. Emotional instability but intact spatial cognition
in adenosine receptor 1 knock out mice. Behav Brain Res
Lawson-Yuen A, Saldivar JS, Sommer S, Picker J. 2008. Familial dele-
tion within NLGN4 associated with autism and Tourette syn-
drome. Eur J Hum Genet 16:614–618.
Levinson JN, Chery N, Huang K, Wong TP, Gerrow K, Kang R,
Prange O, Wang YT, El-Husseini A. 2005. Neuroligins mediate
excitatory and inhibitory synapse formation: Involvement of PSD-
95 and neurexin-1b in neuroligin-induced synaptic specificity.
J Biol Chem 280:17312–17319.
Lise MF, El-Husseini A. 2006. The neuroligin and neurexin families:
From structure to function at the synapse. Cell Mol Life Sci
Luthl A, Laurent JP, Figurov A, Muller D, Schachner M. 1994. Hip-
pocampal long-term potentiation and neural cell adhesion mole-
cules L1 and NCAM. Nature 372:777–779.
NL1 OVER-EXPRESSION IMPAIRS LEARNING AND SYNAPTIC PLASTICITY
Majewska A, Tashiro A, Yuste R. 2000. Regulation of spine calcium Download full-text
dynamics by rapid spine motility. J Neurosci 20:8262–8268.
Marrone DF, Petit TL. 2002. The role of synaptic morphology in neu-
ral plasticity: Structural interactions underlying synaptic power.
Brain Res Brain Res Rev 38:291–308.
Matsuzaki M, Ellis-Davies GC, Nemoto T, Miyashita Y, Iino M, Kasai
H. 2001. Dendritic spine geometry is critical for AMPA receptor
expression in hippocampal CA1 pyramidal neurons. Nat Neurosci
Migaud M, Charlesworth P, Dempster M, Webster LC, Watabe AM,
Makhinson M, He Y, Ramsay MF, Morris RG, Morrison JH,
O’Dell TJ, Grant SG. 1998. Enhanced long-term potentiation and
impaired learning in mice with mutant postsynaptic density-95
protein. Nature 396:433–439.
Minelli A, Alonso-Nanclares L, Edwards RH, DeFelipe J, Conti F.
2003. Postnatal development of the vesicular GABA transporter in
rat cerebral cortex. Neuroscience 117:337–346.
Morris RG, Garrud P, Rawlins JN, O’Keefe J. 1982. Place navigation
impaired in rats with hippocampal lesions. Nature 297(5868):681–683.
Nam CI, Chen L. 2005. Postsynaptic assembly induced by neurexin-
neuroligin interaction and neurotransmitter. Proc Natl Acad Sci
Noguchi J, Matsuzaki M, Ellis-Davies GC, Kasai H. 2005. Spine-neck
geometry determines NMDA receptor-dependent Ca21 signaling
in dendrites. Neuron 46:609–622.
Nusser Z, Lujan R, Laube G, Roberts JD, Molnar E, Somogyi P. 1998.
Cell type and pathway dependence of synaptic AMPA receptor num-
ber and variability in the hippocampus. Neuron 21:545–559.
Oitzl MS, de Kloet ER, Joels M, Schmid W, Cole TJ. 1997. Spatial
learning deficits in mice with a targeted glucocorticoid receptor
gene disruption. Eur J Neurosci 9:2284–2296.
Pavolov I, Rauvala H, Taira T. 2004. Enhanced hippocampal gabaergic
inhibition in mice overexpressing heparin-binding growth-associ-
ated molecule. Neuroscience 139:505–511.
Piechotta K, Dudanova I, Missler M. 2006. The resilient synapse:
Insights from genetic interference of synaptic cell adhesion mole-
cules. Cell Tissue Res 326:617–642.
Prange O, Wong TP, Gerrow K, Wang YT, El-Husseini A. 2004. A bal-
ance between excitatory and inhibitory synapses is controlled by
PSD-95 and neuroligin. Proc Natl Acad Sci USA 101:13915–13920.
Rao A, Harms KJ, Craig AM. 2000. Neuroligation: Building synapses
around the neurexin-neuroligin link. Nat Neurosci 3:747–749.
Rogers DC, Peters J, Martin JE, Ball S, Nicholson SJ, Witherden AS,
Hafezparast M, Latcham J, Robinson TL, Quilter CA, Fisher EM.
2001. SHIRPA, a protocol for behavioral assessment: Validation
for longitudinal study of neurological dysfunction in mice. Neuro-
sci Lett 306:89–92.
Rosales CR, Osborne KD, Zuccarino GV, Scheiffele P, Silverman MA.
2005. A cytoplasmic motif targets neuroligin-1 exclusively to den-
drites of cultured hippocampal neurons. Eur J Neurosci 22:2381–
Rowen L, Young J, Birditt B, Kaur A, Madan A, Philipps DL, Qin S,
Minx P, Wilson RK, Hood L, Graveley BR. 2002. Analysis of the
human neurexin genes: Alternative splicing and the generation of
protein diversity. Genomics 79:587–597.
Rusakov DA, Stewart MG, Korogod SM. 1996. Branching of active
dendritic spines as a mechanism for controlling synaptic efficacy.
Sakic B, Szechtman H, Denburg S, Carbotte R, Denburg JA. 1993.
Spatial learning during the course of autoimmune disease in MRL
mice. Behav Brain Res 54:57–66.
Santamaria F, Wils S, De Schutter E, Augustine GJ. 2006. Anomalous
diffusion in Purkinje cell dendrites caused by spines. Neuron
Scheiffele P, Fan J, Choih J, Fetter R, Serafini T. 2000. Neuroligin
expressed in nonneuronal cells triggers presynaptic development in
contacting axons. Cell 101:657–669.
Schikorski T, Stevens CF. 1999. Quantitative fine-structural analysis of
olfactory cortical synapses. Proc Natl Acad Sci USA 96:4107–
Schmitt U, Hiemke C, Fahrenholz F, Schroeder A. 2006. Over-expres-
sion of two different forms of the a-secretase ADAM10 affects
learning and memory in mice. Behav Brain Res 175:278–284.
Schnell E, Sizemore M, Karimzadegan S, Chen L, Bredt DS, Nicoll
RA. 2002. Direct interactions between PSD-95 and stargazin con-
trol synaptic AMPA receptor number. Proc Natl Acad Sci USA
Stein V, House DR, Bredt DS, Nicoll RA. 2003. Postsynaptic density-
95 mimics and occludes hippocampal long-term potentiation and
enhances long-term depression. J Neurosci 23:5503–5506.
Sutherland RJ, Kolb B, Whishaw IQ. 1982. Spatial mapping: Defini-
tive disruption by hippocampal or medial frontal cortical damage
in the rat. Neurosci Lett 31:271–276.
Tabuchi K, Sudhof TC. 2002. Structure and evolution of neurexin
genes: Insight into the mechanism of alternative splicing. Genomics
Talebizadeh Z, Bittel DC, Veatch OJ, Kibiryeva N, Butler MG. 2005.
Brief report: Non-random X chromosome inactivation in females
with autism. J Autism Dev Disord 35:675–681.
Tang L, Hung CP, Schuman EM. 1998. A role for the cadherin family
of cell adhesion molecules in hippocampal long-term potentiation.
Uchizono K. 1965. Characteristics of excitatory and inhibitory synap-
ses in the central nervous system of the cat. Nature 207:642–643.
Van Dam D, D’Hooge R, Hauben E, Reyniers E, Gantois I, Bakker
CE, Oostra BA, Kooy RF, De Deyn PP. 2000. Spatial learning,
contextual fear conditioning and conditioned emotional response
in Fmr1 knockout mice. Behav Brain Res 117:127–136.
van Praag H, Christie BR, Sejnowski TJ, Gage FH. 1999. Running
enhances neurogenesis, learning, and long-term potentiation in
mice. Proc Natl Acad Sci USA 96:13427–13431.
Vloeberghs E, Van Dam D, D’Hooge R, Staufenbiel M, De Deyn PP.
2006. APP23 mice display working memory impairment in the
plus-shaped water maze. Neurosci Lett 407:6–10.
Waites CL, Craig AM, Garner CC. 2005. Mechanisms of vertebrate
synaptogenesis. Annu Rev Neurosci 28:251–274.
Wilcoxon JS, Nadolski GJ, Samarut J, Chassande O, Redei EE. 2007.
Behavioral inhibition and impaired spatial learning and memory in
hypothyroid mice lacking thyroid hormone receptor alpha. Behav
Brain Res 177:109–116.
Yan J, Oliveira G, Coutinho A, Yang C, Feng J, Katz C, Sram J,
Bockholt A, Jones IR, Craddock N, Cook EH Jr, Vicente A,
Sommer SS. 2005. Analysis of the neuroligin 3 and 4 genes in au-
tism and other neuropsychiatric patients. Mol Psychiatry 10:329–
Ylisaukko-oja T, Rehnstrom K, Auranen M, Vanhala R, Alen R,
Kempas E, Ellonen P, Turunen JA, Makkonen I, Riikonen R,
Nieminen-von Wendt T, von Wendt L, Peltonen L, Ja ¨rvela ¨ I.
2005. Analysis of four neuroligin genes as candidates for autism.
Eur J Hum Genet 13:1285–1292.
Zhu H, Lee M, Agatsuma S, Hiroi N. 2007. Pleiotropic impact of
constitutive fosB inactivation on nicotine-induced behavioral altera-
tions and stress-related traits in mice. Hum Mol Genet 16:820–
Zuo Y, Lin A, Chang P, Gan WB. 2005. Development of long-term
dendritic spine stability in diverse regions of cerebral cortex. Neu-
DAHLHAUS ET AL.