Background
The Notch signaling pathway is a signaling module that is evolutionarily conserved from nematodes to human, which plays essential roles in pattern formation and cell fate determination through local cell-cell interactions [
1]. Notch signaling is initiated by the interaction of the Notch receptors with their ligands, Deltalike (Dll) and Jagged (Jag) [
2,
3]. These interactions induce proteolytic cleavages of the Notch receptors, and generate a soluble intracellular domain (Nicd) that translocates to the nucleus to form a transcriptional activator complex with Su(H)/CBF1/RBP-Jκ. This complex activates the basic helix-loop-helix (bHLH) repressors, such as Hes1 and Hes5 [
4]. Notch signaling is implicated in brain development by regulating cell-fate decisions and proliferation of progenitors [
5,
6]. In addition, Notch signaling is also involved in structural maturation of postmitotic neurons, stimulating neurite branching but inhibiting neurite growth in primary cultured neurons [
7,
8] and in adult-born neurons in the early stages of maturation in the dentate gyrus (DG) [
9]. It has been suggested that Notch signaling plays an important role in cognitive functions, such as long-term memory and synaptic plasticity [
10]. Mice heterozygous for Notch1 or RBP-Jκ display deficits in the formation of long-term spatial memory, but not in the acquisition of new information or in the formation of short-term memory [
10,
11]. In addition, mice overexpressing Notch1 antisense mRNA (NAS mice) showed impaired early-phase long-term potentiation (LTP) and enhanced long-term depression (LTD) at the CA3-CA1 synapses in the hippocampus [
12]. In these genetic models, however, Notch signaling could have been previously altered during development as well as during functional maturation of postmitotic neurons. Moreover, it has been reported that activity-induced Notch signaling in neurons requires
Arc/Arg3.1 and is essential for synaptic plasticity in hippocampal networks [
13]. However, it is still unclear whether these impaired cognitive functions are due to defective Notch signaling in mature neurons or structural changes of postmitotic neurons during development.
Mib1 regulates the endocytosis of Notch ligands to promote Notch activation in the signal-receiving cells [
14‐
16]. Since Mib1 functions in the signal-sending cells and is required for both Deltalike- and Jagged-mediated Notch signaling in mammalian development [
17],
mib1 conditional knockout mice were proved to be an excellent model to elucidate the requirement of Notch signaling in diverse processes of various tissues [
18‐
20]. Especially,
mib1 ablation in the developing brain resulted in complete blockage of Notch signaling and the premature differentiation of radial glial cells, suggesting that Mib1 is essential for Notch signaling during embryonic neurogenesis [
21].
Here we have generated conditional knockout mice of mib1 gene in the differentiated excitatory neurons of the adult brain using CaMKII-cre transgenic mice. These CaMKII-Cre; mib1f/f (mib1 cKO) mice displayed the marked reduction of Notch signaling in the adult brain, but did not exhibit changes in neuronal morphology or structural synaptic connectivity. However, hippocampus-dependent long-term memories, such as object recognition memory, contextual fear memory, and spatial memory in Morris water maze task, were severely impaired in mib1 cKO mice. Moreover, acute hippocampal slices from mib1 cKO mice showed impaired late-phase LTP and LTD. Interestingly, L-LTP impairment in mib1 cKO mice was totally recovered by expression of a constitutively active form of Notch1 (NICD). These results suggest that Mib1-mediated Notch signaling between excitatory neurons is essential for long-lasting synaptic plasticity and memory formation in the hippocampus.
Discussion
It is known that Notch signaling is important for long-term memory and synaptic plasticity in
Drosophila and mammals [
11,
12,
43,
44]. However, the previous studies did not demonstrate clearly: (1) whether deficits in memory and synaptic plasticity were caused by disrupted Notch signaling during maturation or after maturation of the brain; (2) what types of cells send and receive Notch signaling in the adult brain. In this study, we demonstrated that ablation of Mib1 after brain development causes the deficits of both hippocampus-dependent cognitive functions and synaptic plasticity. In addition, since
CamKII-cre–mediated gene ablation is restricted only in excitatory neurons in the forebrain [
23], our data show that Notch signaling responsible for long-term memory and synaptic plasticity functions between excitatory neurons in the hippocampus.
Numerous evidences have demonstrated that Notch signaling is important for structural changes in developing neurons [
7‐
9,
28]. In addition, overexpression of active Notch1 in differentiated neurons can alter neuronal morphology and structural connectivity of pyramidal neurons in the visual cortex [
29]. Consistently, inactivation of Notch1 in CA1 pyramidal neurons resulted in reduced spine density [
13]. In our study, however, no structural abnormalities were observed in 3-months-old
mib1 cKO brains despite reduced Notch activity. Since Mib1 acts as an E3 ubiquitin ligase that ubiquitinates Notch ligands [
45], its inactivation in CA1 pyramidal neurons does not affect the expression of Notch1 itself. Thus, inconsistency in structural abnormality between models suggests that Notch1 itself may play a role in structural integrity or cleavage-independent non-canonical Notch signaling, although high doses of exogenous Notch signaling have the ability to change structural characters of differentiated neurons.
The L-LTP requires
de novo transcription and translation [
46], while E-LTP is mediated by the potentiation of glutamate receptors response at synapses without
de novo transcription [
47]. In our study, the expression levels of each glutamate receptor subunits were not significantly altered in the synaptoneurosome of the
mib1- deficient hippocampus (Figure
6A). Considering the role of Notch as a transcription coactivator, it is plausible that L-LTP, not E-LTP, is regulated by Notch signaling. In line with this hypothesis, we observed that only L-LTP was impaired in
mib1 cKO mice and this deficit was recovered by overexpression of activated Notch1 (Figure
7F). Moreover, in our biochemical data, PKMζ, a well known protein in hippocampal L-LTP [
37], was decreased in
mib1 cKO mice compared to wild-type mice, suggesting that the PKMζ expression in the hippocampus may underlie
de novo transcription and translation by Notch signaling.
In apparent contrast to our observations, a previous report has shown that ubiquitous transgenic expression of Notch1 antisense RNA (NAS) abolished E-LTP in the hippocampus [
12]. Since Notch signaling is implicated in brain development as well as in structural maturation of postmitotic neurons, it is possible that defects of E-LTP in NAS transgenic mice might be caused by developmental or structural abnormalities. On the other hand, Alberi,
et al. showed that inactivation of Notch1 in CA1 pyramidal neurons lead to abnormalities in both E-LTP and L-LTD without any deficits in basal synaptic transmission [
13]. As mentioned above, the conditional deletion of Notch1 affected spine density. Thus, we cannot exclude the possibility that reduced spine density might influence E-LTP.
Lastly, in this study, specific deletion of Mib1 in excitatory neurons using a
CamKII-cre transgenic line caused decreased Notch signaling in the hippocampus, which was accompanied by hippocampus-dependent memory deficits and impaired L-LTP and LTD. Considering the nonautonomous role of Mib1 in signal-sending cells for the proper transduction of Notch signaling [
15,
21], both signal-sending cells and signal-receiving cells of Notch signaling are excitatory neurons in the hippocampus. In addition, coexistence of Mib1 and Jagged1 proteins in the synaptosome (Figure
1D) suggests that Notch-Notch ligand interaction might occur at excitatory synapses. Careful electron microscopy analysis to identify the detailed localization of Notch receptor and ligand proteins will help in probing this hypothesis in future studies.
Methods
Mice
The floxed (f) allele of
mib1 was generated previously [
17]. The
CaMKII-Cre transgenic mice [
23] were obtained from Artemis Pharmaceuticals (Cologne, Germany). The
Rosa-Notch1 mice were kind gifts from Dr. Douglas Melton (Harvard University, Cambridge, MA).
CaMKII-Cre;mib1f/f mice were generated by mating the
mib1f/f mice with
CamKII-cre;mib1+/f or
CamKII-cre;mib1f/f mice. Rosa
NICD/+;
CamKII-cre;
mib1f/f mice were generated by mating
CamKII-cre;mib1f/f mice with Rosa
NICD/+;
mib1f/f mice. The mice used for this study were backcrossed at least 10 generations into the C57BL/6 N background from the original genetic background. All experiments were conducted with the approval of the Animal Care and Use Committee of Seoul National University (Approval No. 081001-3).
Electrophysiology
The fEPSPs were recorded from transverse-sectioned acute hippocampal slices (400 um thick) from mice aged 2–3 months. Mice were anesthetized with ether just before decapitation, and the hippocampal tissues were isolated from the brain and sectioned by using the Vibratome 800-Mcllwain Tissue Chopper (Vibratome, Bannockburn, IL). Acute hippocampal slices were maintained in oxygenated (95% O2, 5% CO2) artificial cerebrospinal fluid (aCSF; 119 mM NaCl, 2.5 mM, KCl, 2 mM MgSO4, 1.25 mM NaH2PO4, 26 mM NaHCO3, 10 mM Glucose, 2.5 mM CaCl2 [pH 7.4]) at 25°C for at least 1 h. The fEPSPs were recorded in the striatum radiatum of the CA1 subfield with 3 M NaCl-filled microelectrodes (3–5 MΩ) after delivering stimulation pulses (200 μs in duration) with a bipolar concentric electrode (World Precision Instruments [WPI], Sarasota, FL) to the Schaffer Collateral (SC) afferent fiber. Test fEPSPs were evoked by a stimulation intensity that yielded one third of the maximal fEPSP responses in a aCSF bath solution containing 100 μM Picrotoxin (Tocris Bioscience, Bristol, UK), and the data were acquired with an Axopatch 200A amplifier and Digidata 1200 (Axon Instrument Inc., Foster City, CA) interface. The basal responses were collected at a frequency of 0.033 Hz for 20 min. Early-phase LTP was then induced by a single train of high-frequency stimulation (HFS, 100 Hz stimulus for 1 s), and late-phase LTP was induced by the TBS protocol (five episodes of TBS at 0.1 Hz, which were composed of 10 trains [4 pulses at 100 Hz] at 5 Hz or 4 trains of HFS at 0.1 Hz).
Behavioral tests
Adult
mib1 cKO and wild-type mice (3-month-old littermates) were used throughout all behavioral tests and all animals were managed as previously described [
48]. The same mice underwent various tests in the following order: object recognition task, water maze test, and fear conditioning. Student’s
t-tests or repeated measures ANOVA with
post hoc pairwise comparisons (Bonferroni
t-test) were used to determine effects of the genotype. All data are reported as mean ± standard error of the mean.
Open-field test
The exploratory behavior of mib1 cKO and wild-type mice was assessed in an open-field test. On the day of the experiment, the mice were transferred to a test room dimly lit by indirect red lighting and were allowed to acclimate for at least 30 min prior to testing. The apparatus consisted of a gray rectangular box (50 × 50 × 25 cm: length × width × depth), the floor of which was illuminated to approximately 60 lux. The open field was divided into a central area (30 × 30 cm) and a peripheral area. Each mouse was placed in the center of the test box and then allowed to explore the novel environment for 10 min. Their behavior was recorded and analyzed using an automated tracking system (SmarTrack). Recording parameters included the total distance traveled, the time spent in the central and peripheral areas, and the frequency of rearing and grooming. After each test, the apparatus was cleaned with a 70% ethanol solution to remove any olfactory cues.
Rotarod test
Motor coordination and motor learning were evaluated using two different modes of the rotarod test, a fixed speed and an accelerating speed. For the fixed-speed mode, each mouse was placed on a bar (3.8 cm diameter; IITC Life Science, Woodland Hills, CA). After a 1-min adaptation period, the bar was rotated at 10 rpm for up to 300 s. Two trials were conducted and the mean latency to fall off was recorded. On the following day, the accelerating-speed rotarod test was conducted. The bar was accelerated from 4 to 40 rpm over 5 min. Each mouse performed three trials.
Acoustic startle response test
Acoustic startle responses were measured using a standard startle reflex system, which consisted of four ventilated, sound attenuating startle chambers (50 × 50 × 50 cm), a rack-mounted operating station, and a personal computer. In each chamber, two wideband speakers (1–16 kHz) provided the audio source for the startle stimuli and background noise (60 dB), respectively, whereas a startle sensor platform, signal transducer, and load cell amplifier served to measure the animal’s startle response. The presentation and ordering of all stimuli were controlled by LabView software (National Instruments, Austin, TX). Prior to startle testing, each mouse was acclimated to a cylindrical acrylic restrainer (5 × 10 cm) for 30 min on 3 consecutive days. On the test day, the animal was placed in the restrainer, which was attached to the sensor platform. The chamber was then sealed and the mouse was given a 5-min habituation period. Once the habituation period had elapsed, 11 different intensities of acoustic stimuli (white noise, 70–120 dB for 30 ms/stimulus, in 5 dB increments) were randomly presented with an interstimulus interval of 30 s, and the amplitude of the acoustic startle response, defined as the peak voltage that occurred during the 250-ms recording window, was recorded. The startle response to each stimulus intensity was calculated as the difference in amplitude from the response following the presentation of the 70 dB (baseline) stimulus.
Object recognition task
The apparatus was a gray rectangular box (50 × 50 × 25 cm). Each mouse was habituated to the test box for 10 min per day for 3 consecutive days. No objects were presented during the habituation period. On the sample test day, two identical objects (A and A′; two pyramids, 5 × 4 × 5 cm) were located symmetrically 10 cm away from the wall and separated 30 cm from each other. Each mouse was placed in the center of the test box and allowed to explore the objects for 10 min. The animal was then returned to its cage. Two retention tests were performed 24 h and 48 h after the sample test. During the first recognition test, the mouse was placed in the test box for 5 min in the presence of one familiar (A) and one novel (B; wood block, 4 × 4 × 5 cm) object. On the next day, the second recognition test was given with object A and another novel object (C; lego block, 5 × 5 × 4 cm). The objects and box were washed with 70% ethanol solution between mice. The time spent exploring the objects was recorded, and relative time spent exploring each object was calculated by dividing by the total time spent exploring the two objects. Exploration of an object was defined as directing the nose to the object at a distance of <2 cm and touching it with the nose.
Water Maze test
The water maze is a circular metal pool (100 cm in diameter, 40 cm in height) that was filled with water (27 ± 1°C) made opaque by adding powdered milk. Detailed training procedures were provided in a previous study [
49]. Briefly, each mouse was habituated to the water maze for 90 s on two consecutive days. For acquisition of spatial memory, a hidden platform (10 cm in diameter) was placed in one of the quadrants, and three trials per day were given over a period of 8 days. If the mouse did not find the hidden platform within 90 s, the animal was guided by an experimenter. After a period of 30 s on the platform, the next trial was begun. To evaluate the retrieval of spatial memory, a 90-s probe trial was performed in the absence of the platform 2 h after the daily training on day 8. The swimming path of the mice was monitored by an overhead video camera connected to a personal computer and analyzed by a tracking system (SmarTrack; Smartech, Madison, WI). The same mice were further tested in the visible platform task, which had been modified to incorporate a black plastic ball (4 cm in diameter, 7 cm high) was added to the raised platform above the water level. Four trials were given with a 40-min intertrial interval (ITI), and the location of the cued platform was moved to a different quadrant between trials.
Contextual and cued fear conditioning tests
A Plexiglas chamber (17 × 20 × 30 cm) was used for fear conditioning. The unconditional stimulus (US) was a 0.6-mA scrambled footshock, 2 s in duration, and the conditional stimulus (CS) was an 80-dB sound at 4 KHz 30 s in duration. On the conditioning day, each mouse was located in the chamber for 3 min to measure the initial freezing level, followed by two paired presentations of the US that coterminated with the CS (ITI = 2 min). To investigate contextual fear memory, the mouse was exposed to the chamber on the next day, and the freezing response was recorded for 5 min. Subsequently, the animal was placed in a novel chamber, and the baseline freezing level was measured for 3 min before the onset of the tone. Then, the CS was presented for 3 min, and freezing behavior was analyzed (cued fear memory). Freezing was defined as no movement except for breathing.
Immunohistochemistry
Immunohistochemistry was performed as previously described [
21]. Mice were anesthetized by intraperitoneal injection with avertin and perfused with 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS). Brains were dissected, postfixed overnight at 4°C, and cryoprotected in 4% PFA/30% sucrose in PBS overnight. Next, the brains were embedded in optimal cutting temperature (OCT) compound and sectioned (14 μm in thickness) on a freezing microtome. Slices underwent antigen retrieval in 0.01 M citric acid, pH 6.0, at 100°C for 15 min.
The sections were stained with the following antibodies: mouse anti-NeuN (Chemicon, Temecula, CA), rabbit anti-cleaved-Notch1 (Cell Signaling Technology, Beverly, MA), mouse anti-MAP2 (Sigma, St. Louis, MO), rabbit anti-GFAP (Dako Cytomation, Glostrup, Denmark), and mouse anti-synaptophysin (Sigma). Alexa 488- and Alexa 594-labeled secondary antibodies (Molecular Probes, Eugene, OR) were used for secondary antibodies. For immunostaining after X-gal staining, frozen sections were soaked in X-gal staining buffer overnight at 37°C. After postfixation and washing, primary antibodies were incubated with the sections overnight at 4°C, and the secondary detection was performed using the Vectastain Elite ABC Kit (Vector Laboratories, Burlingame, CA). Images were taken using a Zeiss Axioskop 2 Plus microscope (Carl Zeiss, Göttingen, Germany).
Neurobiotin labeling and dendritic spine counting
Hippocampal slices (400 μm in thickness) were transferred to a submerged recording chamber continuously oxygenated with aCSF. Cell bodies were visualized by infrared-differential interference contrast (IR-DIC) video microscopy using an upright microscope (Axioskop 2 FS, Carl Zeiss) equipped with a × 40/0.80 W objective (Zeiss IR-Acroplan). Negative pressure was used to obtain tight seals (2–10 GΩ) onto identified pyramidal neurons. The membrane was disrupted with additional suction to form the whole-cell configuration. Pyramidal neurons with membrane potentials below −55 mV were excluded from the analysis. Cells were held at −70 mV for about 20 min. Neurobiotin was injected through glass pipettes with 3–5 MΩ resistances containing the standard pipette solution: K–MeSO4, 120 mM; KCl, 20 mM; HEPES, 10 mM; EGTA, 0.2 mM; ATP (magnesium salt), 2 mM; phosphocreatine (disodium salt), 10 mM; GTP (Tris-salt), 0.3 mM; and 3 mg/mL neurobiotin (Vector Laboratories). Neurobiotin injection lasted for about 20 min. Thereafter, the patch pipette was carefully withdrawn from the membrane, and the slice was fixed with 4% PFA in PBS overnight at 4°C. After washing, nonspecific binding of antibodies was prevented by incubating the sections for 1 h with 5% goat serum in PBS and 0.3% Triton X-100. Subsequently, slices were incubated with streptavidin Alexa 488 conjugate (Molecular Probes) overnight at 4°C. Spine density on CA1 pyramidal neurons was expressed as spines per 10-μm length on secondary dendrites that were located 150–200 μm away from the cell body. All protrusions, irrespective of their morphological characteristics, were counted as spines if they were in direct continuity with the dendritic shaft. A total of 9 neurons from three wild-type animals and 11 neurons from seven cKO animals were subjected to spine density analysis. Images were taken using Olympus FV1000 confocal microscopy.
Western blotting and Cdk5 kinase assay
For the Western blotting, the hippocampi were homogenized in lysis buffer (50 mM HEPES [pH 7.5]; 80 mM NaCl; 3 mM EDTA; 1% Triton-X 100; 1 mM dithiothreitol; 0.1 mM phenylmethylsulfonyl fluoride, 0.1 mM NaVO
4; and 2 μg/mL each of aprotinin, leupeptin, and pepstatin). The lysates were incubated for 15 min on ice and centrifuged for 15 min at 15,000 ×
g and 4°C. The supernatant was collected as cytosolic protein extract. Generally, 10 ~ 20 μg of protein-containing supernatants were separated by size, blotted with primary and secondary antibodies, and visualized with ECL Plus (Amersham Biosciences, Uppsala, Sweden). The primary antibodies were as follows: rabbit anti-DIP-1/Mib1 (kindly provided by Dr. Patricia J. Gallagher, Indiana University, Indianapolis, IN), mouse anti-actin (MP Biomedicals, Irvine, CA), and rabbit anti-activated Notch1 (Abcam, Cambridge, MA). For the Cdk5 kinase assay, immunoprecipitated endogenous Cdk5 from hippocampi of wild-type and
mib1 cKO mice were mixed with 8 μg histone H1 peptide as a substrate in a kinase reaction buffer containing 25 mM HEPES, pH 7.4; 25 mM beta-glycerophosphate; 25 mM MgCl
2; 100 μM Na
3VO
4; 500 μM DTT; and, 1 mM [γ-
32P]ATP. The reaction was allowed to proceed at 30°C for 30 min, as described previously [
50], and radioactivity was measured by autoradiography. The Cdk5 antibody (C-8) and histone H1 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA) and Calbiochem (La Jolla, CA), respectively.
Quantitative real-time PCR
For the quantitative real-time PCR, total RNA was extracted from isolated forebrains using an RNeasy Micro Kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. Aliquots of 1 or 2 μg of RNA were used for the RT (Omniscript RT, Qiagen) with oligo-dT priming. Real-time PCR reactions were set up with each cDNA preparation in an Applied Biosystems 7300 real-time PCR system (Applied Biosystems, Carlsbad, CA) using a master mix of SYBR green I premix ExTaq (Takara Bio, Shiga, Japan) according to the manufacturer’s instructions. The levels of mRNA expression were normalized to that of β-actin. The sequences of the synthesized oligonucleotides are as follows:
mib1-forward: 5′-CCTACGACCTGCGTATCCTG-3′
mib1-reverse: 5′-ACCTTTCCTCTACGCCCATT-3′
nrarp-forward: 5′-TTTGCCACGATTAAATGTCA-3′
nrarp-reverse: 5′-GGGTACACAACAGCCTTCAC-3′
hes1-forward: 5′-ACACCGGACAAACCAAAGAC-3′
hes1-reverse: 5′-GTCACCTCGTTCATGCACTC-3′
hes5-forward: 5′- TACCTGAAACACAGCAAAGC-3′
hes5-reverse: 5′- GCTGGAGTGGTAAGCAG-3′
NICD- forward: 5′-CGTACTCCGTTACATGCAGCA-3′
NICD- reverse: 5′- AGGATCAGTGGAGTTGTGCCA-3′
actin-forward: 5′-AAGGAAGGCTGGAAAAGAGC-3′
actin-reverse: 5′-AAATCGTGCGTGACATCAAA-3′
Competing interests
The authors declare that they have no competing financial interests.
Authors' contributions
KJY and YYK conceived and designed the experiments. HRL and KA performed the electrophysiological analysis. YSJ performed the behavioral analysis. SYJ performed the neurobiotin labeling. MWJ performed the Cdk5 kinase assay. SKK, NSK and HWJ provided essential reagents and helped data analysis. HRL, SHA and KL performed unpulished behavioral analysis. KTK, EK, JHK JSC BKK and YYK supervised and coordinated the works. KJY, HRL, BKK and YYK wrote the manuscript. All authors read and approved the final manuscript.