Mice lacking the PSD-95–interacting E3 ligase, Dorfin/Rnf19a, display reduced adult neurogenesis, enhanced long-term potentiation, and impaired contextual fear conditioning

Protein ubiquitination has a significant influence on diverse aspects of neuronal development and function. Dorfin, also known as Rnf19a, is a RING finger E3 ubiquitin ligase implicated in amyotrophic lateral sclerosis and Parkinson’s disease, but its in vivo functions have not been explored. We report here that Dorfin is a novel binding partner of the excitatory postsynaptic scaffolding protein PSD-95. Dorfin-mutant (Dorfin−/−) mice show reduced adult neurogenesis and enhanced long-term potentiation in the hippocampal dentate gyrus, but normal long-term potentiation in the CA1 region. Behaviorally, Dorfin−/− mice show impaired contextual fear conditioning, but normal levels of cued fear conditioning, fear extinction, spatial learning and memory, object recognition memory, spatial working memory, and pattern separation. Using a proteomic approach, we also identify a number of proteins whose ubiquitination levels are decreased in the Dorfin−/− brain. These results suggest that Dorfin may regulate adult neurogenesis, synaptic plasticity, and contextual fear memory.

with PDZ domains from PSD-95 relatives (PSD-93/Chapsyn-110, SAP97, SAP102), but not with those from S-SCAM, GRIP2, or Shank1. A mutant Dorfin, in which the last amino acid residue was mutated (I840A), failed to interact with PSD-95, indicative of a canonical PDZ interaction. These PDZ interactions were verified by GST-pull down assays, which showed that full-length PSD-95 family proteins expressed in heterologous cells were pulled down by GST-Dorfin, although PSD-93 and SAP102 showed weaker interactions relative to PSD-95 and SAP97 (Fig. 1C). In vitro coimmunoprecipitation experiments independently confirmed the PDZ-dependent interactions of Dorfin with PSD-95 and SAP97 (Fig. 1D,E).
Generation of Dorfin −/− mice. In order to explore the in vivo functions of Dorfin, we generated Dorfin −/− mice using a gene-trap approach in which the intron between exons 2 and 3 of the Dorfin gene was targeted ( Fig. 2A). The trapped allele was verified by genotyping polymerase chain reaction (PCR) (Fig. 2B). Reverse transcription PCR (RT-PCR) and quantitative PCR (qPCR) revealed that Dorfin mRNA was undetectable in both the cortex and hippocampus of Dorfin −/− mice (Fig. 2C-E).
In situ hybridization experiments on wild-type (WT) and Dorfin −/− brains indicated that Dorfin mRNA is widely expressed in various brain regions of WT mice, including the cortex and hippocampus, whereas signals were almost undetectable in the Dorfin −/− brain (Fig. 2F). Widespread Dorfin mRNA signals were also observed in rat brains (Fig. 2G). Notably, Dorfin mRNA signals in the hippocampus were stronger in the DG region than in the CA1 region, especially in the mouse brain. Our results are largely in agreement with data from the Allen Brain Atlas (http://www.brain-map.org/) for mouse Dorfin mRNA. Despite repeated attempts, we could not characterize expression/distribution patterns of Dorfin protein because suitable antibodies are lacking.
General characterization of Dorfin −/− mice. Immunohistochemical staining of WT and Dorfin −/− brains using antibodies to NeuN, a neuronal marker, revealed that Dorfin −/− brains have normal gross morphology at postnatal weeks 4 and 8 (Fig. 3A,B). In addition, a Sholl analysis of biocytin-injected neurons indicated that Dorfin −/− principal neurons in the DG and CA1 regions of the hippocampus have normal dendritic morphology ( Fig. 3C-G). Dorfin −/− mice grew normally during postnatal weeks 4-8, with both males and females displaying body weights comparable to those of WT mice (Fig. 3H). In addition, immunoblot analyses of hippocampal lysates, the hippocampal crude synaptosomal fraction, and microdissected DG lysates revealed that levels of synaptic scaffolding and receptor proteins were not different between the genotypes (Fig. 3I).
Dorfin −/− mice show increased mEPSC amplitude and enhanced LTP in the hippocampal DG region. Because Dorfin interacts with PSD-95, an abundant excitatory synaptic scaffolding protein, we explored possible alterations in excitatory synaptic transmission in Dorfin −/− mice. Granule cells in the Dorfin −/− DG region showed an increase in the amplitude, but not frequency, of miniature excitatory postsynaptic currents (mEPSCs) compared with WT neurons (Fig. 4A). In contrast, both the frequency and amplitude of miniature inhibitory postsynaptic currents (mIPSCs) were normal in these neurons (Fig. 4B). In addition, the input-output curve at Dorfin −/− medial perforant pathway synapses on DG granule cells (MPP-DG synapses) was normal, as measured by the slopes of field EPSPs (fEPSPs) plotted against fiber volley amplitudes (Fig. 4C). Lastly, paired-pulse ratio plots against inter-stimulus intervals were normal (Fig. 4D), suggestive of unaltered presynaptic release probability.
Intriguingly, LTP at MPP-DG synapses induced by four trains of high-frequency stimulation (HFS; 100 Hz) was significantly enhanced in Dorfin −/− mice compared with WT mice (Fig. 4E). This does not seem to be attributable to an increase in NMDAR function because the ratio of NMDAR-and AMPAR-mediated currents at these synapses was not changed (Fig. 4F), although LTP at MPP-DG synapses is known to be dependent on NMDARs 41,42 .
Because the CA1 region of the hippocampus also expresses Dorfin mRNA, although to a lesser extent than the DG, we tested possible changes in mE/IPSCs and LTP in this region. We found that neither mEPSCs nor mIPSCs were altered in Dorfin −/− CA1 pyramidal neurons (Fig. 4G,H). In addition, LTP induced by HFS (100 Hz) was unaltered at Dorfin −/− Schaffer collateral-CA1 pyramidal (SC-CA1) synapses (Fig. 4I). These results suggest that Dorfin deletion leads to an enhancement in LTP at MPP-DG, but not SC-CA1, synapses. Dorfin −/− mice are hyperactive in a familiar, but not in a novel, environment. We next explored whether Dorfin −/− mice show any behavioral abnormalities. Dorfin −/− mice exhibited enhanced locomotor activity during the light-off phase in a familiar environment, as measured by 48-hour continuous monitoring of movements in a familiar environment in which mice had been habituated for 3 days before the test (Fig. 5A). In contrast, Dorfin −/− mice showed normal locomotion in a novel environment, measured by the open field test (Fig. 5B), suggesting that Dorfin −/− mice are hyperactive in a familiar, but not a novel, environment.
We next tested whether Dorfin −/− mice display anxiety-like behaviors by three different tests. Dorfin −/− mice spent a normal amount of time in the center region of the open field arena (Fig. 5B), in the open/ closed arm of the elevated plus-maze (Fig. 5C), and in the light-dark chamber of the light-dark box (Fig. 5D). In the novelty-suppressed feeding test, WT and Dorfin −/− mice showed comparable levels of Scientific RepoRts | 5:16410 | DOI: 10.1038/srep16410 latency to feed in a novel, open field (Fig. 5E). These results indicate that Dorfin −/− mice do not display anxiety-like behaviors.
Lastly, Dorfin −/− mice showed normal levels of self-grooming behavior (Fig. 5F), pentylenetetrazole (PTZ)-induced seizure (Fig. 5G), social interaction and social novelty recognition in the three-chamber test (Fig. 5H), and motor coordination in the rotarod test (Fig. 5I). These results collectively indicate that Dorfin −/− mice show a specific deficit in contextual fear conditioning, but not in other types of learning and memory. Because Dorfin −/− mice show enhanced LTP in the hippocampal DG region, we subjected Dorfin −/− mice to a battery of learning and memory behavioral tests. In the Morris water maze test, which measures spatial learning and memory 43 , Dorfin −/− mice performed normally during the learning, probe, and reversal phases (Fig. 6A). Dorfin −/− mice also performed normally in the T-maze rewarded alteration and T-maze spontaneous alteration tests, which measure spatial working memory 44 (Fig. 6B,C). In the novel object recognition test, which measures object recognition memory and novel object preference, WT and Dorfin −/− mice showed similar preference for the novel object (Fig. 6D).
In the contextual fear-conditioning test, Dorfin −/− mice showed reduced freezing in the same shock chamber 24 hours after conditioning compared with WT mice (Fig. 6E). In contrast, Dorfin −/− mice showed contextual fear extinction and cued fear conditioning comparable to those of WT mice (Fig. 6F,G). These results collectively suggest that Dorfin −/− mice have a specific deficit in contextual fear conditioning, but not in other types of learning and memory. Dorfin −/− mice display suppressed neurogenesis in the adult DG but normal pattern separation. The increases in mEPSC amplitude and LTP observed in the DG region of Dorfin −/− mice might be associated with alterations in the adult neurogenesis, known to regulate synaptic plasticity in the DG 45 . When cells in the inner border region of the DG granule cell layer were labeled with BrdU, a marker of dividing cells, the number of BrdU-positive cells was significantly reduced in the adult Dorfin −/− DG (6 weeks), compared with that in WT mice (Fig. 6H). In addition, the number of cells doubly positive for BrdU and doublecortin, a marker of newly generated neurons, was significantly reduced in the Dorfin −/− DG. These results indicate that Dorfin deletion leads to the suppression of neurogenesis in the adult DG.
Neurogenesis in the adult DG has been associated with spatial pattern separation in addition to spatial/ contextual learning and memory [45][46][47] . We thus subjected Dorfin −/− mice to the contextual discrimination test and found that Dorfin −/− mice could normally learn to discriminate two similar contexts, in which one is associated with foot shock (Fig. 6I).
Proteomic identification of under-ubiquitinated proteins in the Dorfin −/− brain. If Dorfin acts as a ubiquitin ligase in the brain, we should be able to observe decreases in the ubiquitination levels of Dorfin substrate proteins in the Dorfin −/− brain. To this end, we immunoprecipitated hippocampal lysates using an antibody that targets the K-GG peptide in ubiquitinated proteins, followed by liquid chromatography-tandem mass spectrometric (LC-MS/MS) analyses of the precipitates (Fig. 7A).
We sorted positive proteins in order of greater decreases in ubiquitination, generating a list of 24 proteins whose ubiquitination levels were substantially decreased (> 2.5 fold) in the Dorfin −/− brain relative to WT controls (Supplemental Table S1). These proteins could be classified into several functional groups, including adhesion molecules/extracellular matrix, adaptor/scaffolds, G protein-related proteins, phosphatases, proteases, enzymes, and cell cycle/chromatin regulatory proteins (Fig. 7B).
Using coimmunoprecipitation experiments in heterologous cells, we verified the association of some of these proteins with Dorfin ( Fig. 7C-G). These proteins included Rab11b (a small GTPase), neurofilament M, the α and β subunits of PP1 phosphatase, and histone H2A. Unexpectedly, none of these five proteins were found to be ubiquitinated when coexpressed with Dorfin in heterologous cells ( Supplementary  Fig. 1). In addition, their protein levels were unaltered in Dorfin −/− mice, as measured by immunoblot analysis of hippocampal lysates, the hippocampal crude synaptosomal fraction, and microdissected DG lysates (Fig. 7H). In addition, there were no changes in the expression levels of other proteins in the list, including ATP6V1A, Plexin A1/A4, NPEPPS, 14-3-3 η /ζ , and gelsolin (Fig. 7H). Therefore, Dorfin deletion in vivo does not seem to cause detectable decreases in the levels of the tested proteins.

Discussion
Our study indicates that Dorfin deletion in mice leads to decreased neurogenesis in the DG, enhanced LTP in the MPP-DG pathway, and a specific behavioral deficit in contextual fear conditioning, but not in other types of learning and memory. We also identified Dorfin as a novel binding partner of PSD-95, and identified a number of proteins whose ubiquitination levels are decreased in Dorfin −/− mice.
The functional significance of the interaction between Dorfin and PSD-95 was not directly explored in the present study. However, given that PSD-95 is an abundant scaffolding protein enriched at excitatory synapses, it is possible that PSD-95 may promote synaptic targeting of Dorfin, as demonstrated for many other PSD-95-interacting proteins 48,49 . Alternatively, because PSD-95 is a multi-domain protein that interacts with diverse membrane, signaling, and scaffolding/adaptor proteins 49 , and E3 ligases often interact with adaptor/scaffolding proteins to broaden the spectrum of substrate proteins 1-5 , PSD-95 might function as a gateway through which Dorfin interacts with diverse substrates. This is reminiscent of the interaction between the synaptic PDZ protein CASK/LIN-2 and Parkin 50 , an E3 ubiquitin ligase known to negatively regulate excitatory synapse number and strength 51   from excitotoxicity 52 . Another example is the interaction of Piccolo and Bassoon, two large active zone proteins controlling presynaptic protein assembly 53 , with Siah1 (an E3 ubiquitin ligase), which is thought to promote the maintenance of synaptic integrity by modulating presynaptic protein degradation 30 .
Dorfin −/− mice displayed two synaptic phenotypes in the hippocampus: increased mEPSC amplitude in DG granule cells and enhanced HFS-induced LTP at MPP-DG synapses. These results suggest that Dorfin may have negative influences on the synaptic AMPARs as well as on LTP. Behaviorally, Dorfin −/− mice showed a specific impairment in contextual fear conditioning, but not in other types of learning and memory, including spatial learning and memory (Morris water maze), novel object recognition memory, spatial working memory (T maze), fear extinction, cued fear conditioning, and pattern separation.
It is unclear how Dorfin deletion leads to increases in mEPSC amplitude and LTP in the DG, although they likely involve increases in the number or function of AMPARs under basal conditions or during LTP. Phosphorylation of two sites on the GluA1 subunit of AMPARs, S831 and S845, have been implicated in the regulation of the biophysical properties of AMPARs and AMPAR-mediated synaptic plasticity 54,55 . However, our data indicated that the levels of these phosphorylations were unaltered in the Dorfin −/− brain, excluding this possibility.
Our data provide additional support for the notion that E3 ubiquitin ligases regulate synaptic plasticity and learning and memory. It has been shown that a maternal deficiency of Ube3A (ubiquitin protein ligase E3A) suppresses LTP and impairs contextual fear conditioning in mice 56 . In addition, Praja2 (praja ring finger 2, E3 ubiquitin protein ligase) knockdown in rats suppresses late LTP 57  Our results are unique in that an E3 deletion in the DG (not CA1) region of the hippocampus alters LTP and learning and memory. This is in line with the reported association of hippocampal DG granule cells with spatial behavior and contextual fear conditioning [63][64][65] . In addition, this suggests that E3 ligases in diverse brain regions in addition to the hippocampal CA1 region are important for the regulation of synaptic plasticity, as previously observed in the amygdala of APC/C-Cdh1 −/− mice 66 , striatum of Parkin −/− mice 67 , and visual cortex of Ube3a m−/p+ mice 68 .
A notable finding was that enhanced-not reduced-LTP in Dorfin −/− mice was associated with impaired contextual fear conditioning, contrary to our expectations. However, previous studies on mice lacking, for example, Fmr2, Gα 1, PDE4D, IRSp53, and SCRAPPER proteins, have linked enhanced hippocampal LTP in the Schaffer collateral pathway with decreased contextual fear conditioning 61,62,69-73 . These results, together with ours, suggest that deviation of hippocampal LTP in either direction in the DG or CA1 region may cause similar impairments in learning and memory.
Although the mechanisms leading to enhanced LTP and impaired contextual fear conditioning in Dorfin −/− mice remain to be determined, the list of proteins whose ubiquitination levels were reduced by > 2.5-fold in Dorfin −/− mice might provide some starting points. Indeed, many of these proteins are enriched in the postsynaptic density (PSD) and have been implicated in the regulation of excitatory synaptic transmission, synaptic plasticity, and learning and memory (Supplemental Table S1). For example, Rab11a/b promotes translocation of recycling endosomes into dendritic spines and local exocytosis of GluA1 during LTP [74][75][76][77] . Therefore, Dorfin-dependent ubiquitination and degradation of Rab11a/b proteins would negatively regulate LTP. In addition, transgenic mice in which the signaling adaptors 14-3-3η and 14-3-3ζ are functionally inhibited by genetic expression of an inhibitor protein display reductions in synaptic content of NMDARs, NMDAR-mediated synaptic transmission, LTP, and contextual fear conditioning 78 , suggesting that 14-3-3η and 14-3-3ζ positively regulate synaptic plasticity and learning and memory. Therefore, these functions may be suppressed by Dorfin-dependent 14-3-3 protein degradation.
Contrary to our initial expectation, our data indicate that some of the proteins that exhibited decreased ubiquitination in Dorfin −/− mice were expressed at normal levels in the mutant brain, despite our demonstration that they associate with Dorfin in vitro. However, protein ubiquitination has been shown to have functions other than protein degradation, such as modulation of protein function, endocytosis and subcellular localization, as well as protein-protein interactions [1][2][3][4][5] . Our data indicate that at least the five proteins that we demonstrated to form a complex with Dorfin (Rab11b, neurofilament-M, PP1-α catalytic subunit, PP1-β catalytic subunit, and histone H2A.Z) are not directly ubiquitinated in heterologous cells. Given that these proteins were found to be less ubiquitinated in the Dorfin −/− hippocampus through an unbiased screen, their ubiquitination might require certain conditions unique to neuronal/brain environments, or some proteins other than these five proteins can be ubiquitinated by Dorfin in heterologous cells.
Lastly, our results indicate that neurogenesis in the Dorfin −/− adult DG is suppressed. Because adult neurogenesis in the DG has been implicated in the regulation of synaptic plasticity at DG synapses and hippocampus-dependent behaviors including pattern separation and spatial/contextual learning and memory [45][46][47]79 , the reduced neurogenesis in the Dorfin −/− adult DG might be associated with the enhanced LTP and impaired contextual fear conditioning in these mice. However, because reduced adult neurogenesis has been shown to suppress both LTP and LTD 80 , our results, where reduced neurogenesis was associated with enhanced (but not suppressed) LTP, differ from the previous results. However, reduced neurogenesis in the adult DG has been associated with reduced performance in contextual fear conditioning in previous studies 45,81 , being in line with our results.
In conclusion, our study identifies the E3 ubiquitin ligase Dorfin as a novel ligand of PSD-95, and provides in vivo evidence supporting the hypothesis that Dorfin may regulate adult neurogenesis, excitatory synaptic transmission, synaptic plasticity, and contextual fear conditioning.
Antibodies. The  Yeast two-hybrid constructs and assays. The last 7 amino acids of mouse Dorfin (aa 834-840, WT and I840A) were generated by annealing oligonucleotides and subcloning them into pBHA. The PDZ domains of various proteins were subcloned into pGAD10 (a prey vector: Clontech) as described previously 86 . Yeast two-hybrid assays were performed as previously described 87 . In situ hybridization. In situ hybridization was performed as previously described 82 . Hybridization probes for mouse Dorfin mRNA were prepared using nt 2161-2820 of mouse Dorfin (NM_013923.2) and subcloned into pGEM-7Zf vector. Hybridization probes for rat Dorfin mRNA were prepared using nt 2609-3187 of rat Dorfin (NM_001130560.1) and subcloned into pGEM-7Zf vector.

Animals. Dorfin
BrdU labeling. 5-Bromo-2′ -deoxyuridine (BrdU, Sigma) was injected intraperitoneally (50 mg/kg) to mice (6 weeks) for 3 days. Five days after the last injection, mice were perfused with 4% paraformaldehyde. Fixed brain slices (50 μ m) obtained by vibratome were incubated in 1N HCl for 10 min on ice, followed by 2N HCl for 10 min at RT before moving them to 37 °C incubator for 20 min. Borate buffer (0.1 M) was added to the slices for 12 min at RT. Slices were then incubated with BrdU and DCX antibodies at 4 °C for 12 hours, followed by secondary antibody staining and mounting using Vectashield and DAPI. Slice images were acquired by confocal microscopy (LSM780, Zeiss).

Sholl analysis.
Pyramidal neurons in the CA1 region and granule cells in the DG region of the hippocampus were randomly chosen and infused with biocytin (0.5%) for 5 min. After injection, the slices were collected and fixed with 4% paraformaldehyde for 24 hours. The slices were stained by streptavidin antibody. Images were captured using 20× objectives (LSM 780, Carl Zeiss). Dendritic branch intersections were analyzed manually in a blind manner.
Hippocampal/DG lysates and subcellular fractions. The crude synaptosomal (P2) fraction of adult mouse hippocampi (2-3 months) was prepared as described previously 88 . Whole hippocampal/DG lysates were prepared using homogenization buffer (0.32 M sucrose, 10 mM HEPES, 2 mM EDTA). For DG microdissection, we used vibratome to cut brain slices (400 μ m) and dissect DG. Electrophysiology. Sagittal slices of hippocampus (300 μ m for whole cell recordings, 400 μ m for field recordings) were prepared using vibratome (VT1200S, Leica) in ice-cold sectioning buffer (212 mM sucrose, 25 mM NaHCO 3 , 5 mM KCl, 1.25 mM NaH 2 PO 4 , 10 mM D-glucose, 2 mM sodium pyruvate, 1.2 mM sodium ascorbate, 3.5 mM MgCl 2 , and 0.5 mM CaCl 2 bubbled with 95% O 2 /5% CO 2 ) and recovered at 32 °C in aCSF (125 mM NaCl, 25 mM NaHCO 3 , 2.5 mM KCl, 1.25 mM NaH 2 PO 4 , 10 mM D-glucose, 1.3 mM MgCl 2 , and 2.5 mM CaCl 2 ). After 30 min, slices were maintained at room temperature. Whole-cell patch clamp recordings and field potential recordings were performed as previously described 72  Field potential recordings were made with recording pipettes filled with aCSF. For input/output and paired pulse ratio recordings, stable fEPEPs were recorded for 10 min before main experiments began. For input/output recordings, stimuli were gradually increased to induce fiber volley amplitudes of 0.05-0.3 mV/ms, and 3 traces per stimulus were recorded. For paired pulse ratio recordings, inter-stimulus intervals were 25, 50, 100, 200, and 300 ms. Recordings were repeated 3 times. For high-frequency stimulation LTP, picrotoxin (100 μ M) was added to aCSF. After obtaining stable baseline responses for 20 min, a single stimulus of 100 Hz (1 sec) was given to the SC-CA1 pathway, and four trains of 100-Hz stimuli (each 1 sec; 20 sec inter-train interval) were given to the MPP-DG pathway. The responses were recorded for an hour.

48-hour continuous monitoring of movements.
For long-term continuous monitoring of mouse movements, we used the LABORAS TM system (Metris), designed to detect vibrations originating from the movements of a mouse in the cage. Mice were single isolated and habituated in LABORAS cages in the recording room for 3 days. After habituation, the LABORAS cage with a mouse was placed on top of a vibration-sensitive recording platform for 72 hours. The data from the last 48 hours were analyzed, and averaged to 24-hour cycles. Open field test. The mice were placed in a 40 × 40 × 40 cm box and recorded of their horizontal locomotor activity under complete darkness (~0 lux) for an hour. The center zone line was 10 cm apart from the edge. Movements were analyzed using Ethovision XT 10 (Noldus).
Elevated plus-maze. The maze, elevated to a height of 50 cm from the floor, consisted of two open arms, two closed arms, and a center zone. Mice placed on the center zone were recorded of their movements for 7 min and analyzed of the time spent in closed or open arms manually.
Light-dark box test. The box had a dimension of 12 × 30 × 20 cm for the light chamber (~250~300 lux) and 14 × 13 × 20 cm for the dark chamber (~0~5 lux). A mouse was placed in the center of the light chamber, followed by movement monitoring for 10 min. Time spent in the light or dark chamber was analyzed manually.
Novelty-suppressed feeding test. Mice were starved for 24 hours and placed in novel a 40 × 40 × 40 cm box. A food pellet was place in the center of the box, and the latency to approach and eat the food was measured.
Self-grooming behavior. Mice were placed in an empty home cage and recorded of self-grooming behavior for 10 min. Self-grooming behavior was defined as stroking or scratching of face, head, or body areas with forelimbs, or licking body parts. Time of self-grooming was measured in a blind manner.
PTZ-induced seizure. Mice were injected of 50 μ g/weight (g) pentylenetetrazole (PTZ) into the intraperitoneal cavity and recorded of seizures for 10 min. Seizure was defined as convulsive movements including clonic and/or tonic-cloninc seizures while lying on the side and/or wild jumping.
Three-chamber test. This test was performed to measure social interaction and social novelty recognition 89 . The apparatus consisted of 3 chambers with the dimension of 12 × 20 × 26 cm for the center chamber, and 14 × 20 × 26 cm for side chambers. Both side chambers had plastic cages in the corner to place objects or mice. The experiment consisted of 4 sessions, including habituation in the center camber (10 min), habituation in all chambers (10 min), object/mouse exploration (10 min), and old/novel mouse exploration (10 min). In the third session, a WT stranger mouse was gently confined in a plastic cage, and an inanimate object was placed in the opposite plastic cage. Mice were allowed to freely move and explore, and sniffing and chamber times were measured for the analysis of social interaction. Sniffing was defined as the nose of mouse facing towards and getting close to the target object or mouse. In the last session, the object was replaced with a new WT stranger mouse. The positions of object and stranger 1 were alternated in a random manner to prevent side preference.

Rotarod test.
Mice were placed gently on the rotating rod of the rotarod apparatus (Stoelting), followed by an increase in the speed of the rod from 4 to 40 rpm over 5 min. The experiments were performed for 5 consecutive days, measuring the latency to fall from the rod.
Morris water maze test. Mice in a circular tank (100 cm diameter) filled with white opaque water (22-24 °C) were trained to find a hidden plat form (10 cm diameter). The learning phase consisted of 3 trials per day for 5 days. For the probe test (1 min), the platform was removed, and the time a mouse spent exploring each quadrant was measured. In the reversal test (3 days), the platform was relocated to the opposite quadrant position. Mouse movements were analyzed using EthoVision XT 10 (Noldus) Rewarded alteration T maze. This test was performed as previously described 44 . In brief, underfed mice with the body weight of 80-85% of a normal range were habituated to the T maze arena for 5 min for 2 days, followed by training with 20 μ l of 70% milk reward for 10 days.
Spontaneous alteration T maze. Mice placed in the start point were allowed to freely explore the two arms of the maze. When a mouse enters into one of the two arms, it was brought back to the start point after it spent 5 sec in the arm for the next round, which was repeated 10 times.
Novel object recognition test. Mice, habituated to an open-field box for 60 min on day 1, were presented with two identical objects for 11 min on day 2. On day 3, one of the objects was replaced with a novel one. Object recognition was measured by the amount of time during which the nose of the mouse pointed the object.
Contextual fear conditioning and extinction. On day 1, mice were habituated to the shock chamber (Coulbourn Instruments) for 5 min. On day 2, a mouse spent 3 min in the chamber before it received foot shocks (0.8 mA) at 3-, 4-, and 5-min time points, and freezing levels during 3-6 min were quantified. On day 3, mice were re-exposed to the chamber without foot shock for 5 min, and freezing levels during this 5 min were quantified. Freezing behavior was analyzed using FreezeFrame 3 software (Coulbourn Scientific RepoRts | 5:16410 | DOI: 10.1038/srep16410 Instruments). For contextual fear extinction, mice were exposed to the same shock chamber for 5 min for 10 consecutive days.
Cued fear conditioning. Mice without habituation spent 3 min in the context-A shock chamber before they received an auditory cue (4 kHz, 75 dB) for 30 sec that co-terminated with 2 sec foot shock (0.8 mA). This conditioning was repeated 3 times with one minute intervals. Twenty four hours later, a mouse was placed in the context B where it spent 3 min before it received the same auditory cue for 3 min. Freezing behavior during this 3 min period was analyzed.
Contextual discrimination. We subjected mice to a contextual fear conditioning test using a less distinct pair of contexts (A and B) that shared an identical metal grid floor, but had unique wall paper, bottom color, and lighting. In this protocol, fear conditioning took place incrementally over several days in order to investigate the effects of repeated experiences. The experiments were performed in the same condition of time, temperature, and humidity: 2-9 pm, 20-22 °C, and 45%. Each trial was performed after habituation in their cage for 30-60 min and twice a day. During days 1-3 of the experiment, mice were placed in a shock chamber for 180 s, given a single foot-shock of 0.8 mA for 2 s, and taken out after 60 s. On days 4 and 5, for generalization, mice of each genotype were divided into two groups, one visiting chamber A and the other visiting chamber B (one trial/day) for 4 min on day 4, and visiting the unvisited chambers on day 5. Mice did not receive foot-shock during the generalization, and freezing was assessed. During days 6-13, for discrimination training, mice visited the two chambers every day for 8 days, always receiving a foot-shock (0.8 mA for 2 s) at 180 s after being placed in chamber A but not chamber B, followed 1 more min stay. Freezing during the first 3-min exposure in each chamber was used to calculate the daily discrimination ratio.
Analysis of behavioral results. All the behavioral results were analyzed in a blind manner.