Abstract Mitotic arrest deficient 2-like protein 2 (MAD2B) is not only a DNA damage repair agent but also a cell cycle regulator that is widely expressed in the hippocampus and the cerebral cortex. However, the functions of MAD2B in hippocampal and cerebral cortical neurons are poorly understood. In this study, we crossed MAD2B^flox/flox and calcium/calmodulin-dependent protein kinase II alpha (Camk2a)-Cre mice to conditionally knock out MAD2B in the forebrain pyramidal neurons by the Cre/loxP recombinase system. First, RNA sequencing suggested that the differentially expressed genes in the hippocampus and the cerebral cortex between the WT and the MAD2B cKO mice were related to learning and memory. Then, the results of behavioral tests, including the Morris water maze test, the novel object recognition test, and the contextual fear conditioning experiment, suggested that the learning and memory abilities of the MAD2B cKO mice had improved. Moreover, conditional knockout of MAD2B increased the number of neurons without affecting the number of glial cells in the hippocampal CA1 and the cerebral cortex. At the same time, the number of doublecortin-positive (DCX^+) cells was increased in the dentate gyrus (DG) of the MAD2B cKO mice. In addition, as shown by Golgi staining, the MAD2B cKO mice had more mushroom-like and long-like spines than the WT mice. Transmission electron microscopy (TEM) revealed that spine synapses increased and shaft synapses decreased in the CA1 of the MAD2B cKO mice. Taken together, our findings indicated that MAD2B plays an essential role in regulating learning and memory. Keywords: MAD2B, learning and memory, dendritic spine, synapse, hippocampus Introduction Learning and memory are vital foundations of life activities. Neurodegenerative diseases such as Alzheimer's disease (AD) and Parkinson's disease (PD), which are accompanied by a gradual decline in learning and memory ability, are increasing every year. AD is the most common neurodegeneration disease with eventual memory decline (Hu et al., [33]2016). Some neuronal pathological changes, including neuronal loss (Spangenberg et al., [34]2016), synapse loss (Duncan and Valenzuela, [35]2017), and decrements in dendritic arborization and spine density (Maiti et al., [36]2021), occur in AD. Abnormal gene expression in some brain areas can also lead to changes in learning and memory ability (Konopka et al., [37]2010; Brault et al., [38]2021; Kar et al., [39]2021). For example, Bin1 deficiency in neurons leads to the impairment of spatial learning and memory due to abnormal presynaptic regulation (De Rossi et al., [40]2020). In mice, the brain-specific knockout of follistatin results in declines in spatial learning and working memory, as well as deficits in LTP and neurogenesis in the hippocampus (Chen et al., [41]2021). Mitotic arrest deficient 2-like protein 2 (MAD2B, also known as MAD2L2 or REV7) is a component of the mitotic spindle assembly checkpoint, which plays an important role in regulating cell cycle, gene transcription, DNA repair, and carcinogenesis. It has been reported that MAD2B is mainly expressed in neuronal-like cells with pyramidal shapes in the hippocampus and cerebral cortex (Meng et al., [42]2012). In addition, MAD2B is related to the pathophysiology of diabetic encephalopathy (Meng et al., [43]2014). Under high glucose treatment, primary cultured neurons with high MAD2B expression undergo apoptosis. MAD2B is also overexpressed in the cerebral cortex of diabetic rats. As one of the risk factors for AD (Baglietto-Vargas et al., [44]2016; Khan et al., [45]2021; Zhang et al., [46]2021; Athanasaki et al., [47]2022), diabetes encephalopathy shares similar pathogenesis with AD. In diabetes encephalopathy, the proliferation and neuronal differentiation of newborn cells in the dentate gyrus (DG) and neuronal density in the hilar region (Beauquis et al., [48]2009) is reduced. Moreover, diabetic rats have low dendritic spine density and downregulated synaptic-related protein expression in the hippocampus (Jin et al., [49]2018). Transmission electron microscope indicates that pathological changes in synaptic ultrastructure such as the thickness of postsynaptic density and the synaptic gap of the hippocampal CA1 fields as well as decreased survival of CA1 neurons occur in diabetes encephalopathy (Xu et al., [50]2021). These abnormal changes in synaptic plasticity have become the main reasons for the decline in learning and memory in diabetes encephalopathy. Although MAD2B is closely related to diabetes encephalopathy (Meng et al., [51]2014), its direct role in the nervous system remains to be clarified. The conditional knockout (cKO) of a gene in specific neurons has become a valuable method for studying gene function. The calcium/calmodulin-dependent protein kinase II alpha (Camk2a)-Cre mouse line, in which forebrain pyramidal neurons in postnatal mice are targeted (Tsien et al., [52]1996; Wang et al., [53]2021a), is widely used in this field. Therefore, in the present study, we generated the MAD2B cKO mice under the control of Camk2a gene regulatory elements by using the Cre/loxP system and investigated the direct role of the MAD2B gene in the nervous system. We first analyzed the RNA sequencing (RNA-seq) results of the hippocampus and the cerebral cortex of the MAD2B cKO mice and found several differentially expressed genes (DEGs) that might be involved in learning and memory regulation. Next, we conducted behavioral tests on learning and memory. The MAD2B cKO mice demonstrated enhanced hippocampus-dependent learning and memory in the Morris water maze (MWM) test, the novel object recognition test (NORT), and the contextual fear conditioning (CFC) experiment. Finally, we found that the number of doublecortin-positive (DCX^+) cells in the DG, mushroom-like spines, and spine synapses in the CA1 region increased. These findings provide a foundation for revealing the role of MAD2B in the central nervous system and for treating cognitive dysfunction. Materials and methods Animals All procedures for the care and use of laboratory animals were approved by the Institutional Guidelines and Animal Care and Use Committee of Huazhong University of Science and Technology, Wuhan, China. In addition, animal experiments were performed in accordance with the Guidelines for the Care and Use of Laboratory Animals from the National Institutes of Health, USA. Mice were housed in a 21 ± 1 °C temperature room with constant humidity under a 12 h light/dark cycle (8:00 AM−8:00 PM). Food and water were freely available. All male mice (7–8 weeks) shared the same C57/BL6J genetic background (Jackson Laboratory stock). MAD2B was conditionally knocked out in the forebrain pyramidal neurons under the control of the Camk2a promoter with the Cre/loxP system. MAD2B^flox/flox mice, which have loxP sites flanking exons 3 and 4 of the MAD2B gene, were generated and Camk2a Cre mice were purchased from Shanghai Model Organisms Center, Inc. MAD2B^flox/flox and Camk2a-Cre mice were intercrossed to generate MAD2B^flox/flox; Camk2a Cre^+ mice (cKO mice) and MAD2B^flox/flox; Camk2a Cre^− mice (WT mice) as previously described (McGill et al., [54]2018). All mice were handled by the experimenter for a week before the behavioral experiments to reduce the unnecessary influence of stress. All mice were acclimated to the behavioral room for 2 h before experiments. All behavioral experiments were conducted between 8:00 AM−8:00 PM with dim light and performed in a double-blinded manner. Genotyping Mice were genotyped through polymerase chain reaction (PCR) by using tail DNA as previously described (Sadick et al., [55]2022). In brief, a 3 mm piece of tissue was cut from the end of the tail of a 1-month-old mouse and then digested in 75 μl of alkaline lysis regent (25 mM NaOH and 0.2 mM EDTA in DNase- and RNase-free water) at 98 °C for 1 h. Subsequently, the mixture was centrifuged for 1 min at 1,000 rpm, and 75 μl of neutralizing reagent (40 mM Tris-HCl, pH 5.5) in DNase- and RNase-free water) was added to each sample on ice. Then, the solution was centrifuged at 4,000 rpm for 3 min at 4 °C, and the supernatant containing genomic DNA (gDNA) was collected for later genotyping. The PCR system comprised 10 μl of 2× Taq Plus Master Mix (Vazyme, code number: P212); 0.5 μl of forward primer (10 pmol/μl); 0.5 μl of reverse primer (10 pmol/μl); and 9 μl of gDNA. The cycling reaction was step1: 94 °C for 3 min; step 2 (35 repeats): 94 °C for 30 s, 58 °C for 30 s, and 72 °C for 30 s; step 3: 72 °C for 5 min; and step 4: holding at 12 °C. The primer sequences were MAD2B forward: 5′-TCTTCCCTTAGATTGGGTTTCTC-3′ and MAD2B reverse: 5′-GCACGAATAGGACAAACAGCAA-3′ and Camk2a-Cre forward: 5′-GGGGAGGTAGGAAGAGCGATGA-3′ and Camk2a-Cre reverse: 5′-ATCGACCGGTAATGCAG-3′. Open field test The open field test was performed as previously described (Yoshizaki et al., [56]2020). Briefly, each mouse was placed in the center of an open-field apparatus (50 × 50 × 50 cm with four black walls and a white floor) and allowed to move freely in the apparatus, which was divided into 25 small squares. The nine middle squares were defined as the central zone, and the four squares in four corners were defined as the corner zone. Behaviors were automatically recorded using a camera connected to a digital tracking device (Xinruan Information Technology Co. Ltd, Shanghai, China) for 10 min. Once finished, the mouse was moved back to its home cage immediately. The equipment was cleaned with 70% ethanol between mice. A computer with the Super maze software (Xinruan Information Technology Co. Ltd, Shanghai, China) was then used to process the tracking information. Data on total moving distance and speed and the time in the corner zone or central zone were then analyzed. Elevated plus maze The elevated plus maze was 50 cm in length, 10 cm in width, and 40 cm in height with two open arms, two closed arms, and a central zone. The elevated plus maze test was performed as previously described (Qi et al., [57]2022). Briefly, each mouse was placed in the central zone facing the open arm, and its behaviors were recorded for 10 min. Once finished, the mouse was moved back to its home cage immediately. The equipment was cleaned with 70% ethanol between mice. Behaviors were automatically recorded by a camera connected to a digital tracking device (Xinruan Information Technology Co. Ltd, Shanghai, China). A computer with the Super maze software (Xinruan Information Technology Co. Ltd, Shanghai, China) then processed the tracking information. The time that the mice spent in the open arms and the closed arms was then analyzed. Morris water maze The MWM was performed as previously described (Qu et al., [58]2021). The setup for MWM consisted of a circular pool (diameter of 1.2 m and height of 50 cm) filled with water that was maintained at 24–25 °C and whitened with titanium dioxide. The maze was located in a room with numerous extra-maze visual cues and divided into four equal quadrants. A platform with a 10 cm diameter was placed in a fixed quadrant (target quadrant) of the pool. On the first day, the platform was placed 1–2 cm above the water surface (Bromley-Brits et al., [59]2011; Luo et al., [60]2020) to enable the mice to see the platform and cues. The influence of mouse vision on the experimental results was thus excluded. Each mouse was trained for three trials with a 30 min inter-trial interval. From the second to the 5th day, each mouse was trained for three trials to find the platform that was hidden 1–2 cm below the water surface. The mouse was placed into the pool from the other three quadrants in each trial, except for the target quadrant, and given 60 s to find the platform. If the mouse failed to find the platform within 60 s, it was guided to the platform by the experimenter. The mouse was allowed to stay on the platform for 20 s. On the 6th day (probe test), the mouse was placed in the pool from the opposite side of the target quadrant for 60 s. Once finished, the mouse was moved back to its home cage immediately. Behaviors were automatically recorded by a camera (fixed to the device 1 m from the water surface) connected to a digital tracking device (Xinruan Information Technology Co. Ltd, Shanghai, China). A computer with the Morris water maze software (Xinruan Information Technology Co. Ltd, Shanghai, China) then processed the tracking information. Escape latency, escape length on training days, velocity, the number of platform crossings, and total distance traveled in the target quadrant on the probe test day was automatically recorded for later analysis. Novel object recognition test The NORT was conducted in an open field apparatus (50 × 50 × 50 cm) with four black walls and a white floor as previously described (Li et al., [61]2021). Briefly, the mice were acclimated to the conditions of the experimental room for 2 h. Then, each mouse was presented with two identical objects (familiar objects; 3 × 3 × 6 cm blue plastic cuboids) placed at the rear left and right corners (fixed onto the floor with adhesive tape so that the mouse cannot push them down) in the open field apparatus. The mouse was allowed to explore the objects freely for 5 min. The apparatus and objects were thoroughly cleaned with 70% ethanol between mice. One h later, one of the familiar objects was randomly replaced with a new object (a 3 × 3 × 7 cm red wooden object with a mushroom-like shape), and the mouse was again placed in the open field for 5 min. Once finished, the mouse was moved back to its home cage immediately. The apparatus and objects were thoroughly cleaned with 70% ethanol between mice. Behaviors were automatically recorded by a camera connected to a digital tracking device (Xinruan Information Technology Co. Ltd, Shanghai, China). A computer with the Super maze software (Xinruan Information Technology Co. Ltd, Shanghai, China) then processed the tracking information. The time mice spent exploring each object (within 2 cm of the object) with their noses or paws was calculated. The discrimination index (DI) was defined as the percentage of time spent exploring the novel object to the total time spent exploring both objects. Contextual fear conditioning The CFC was performed as previously described (Zhao et al., [62]2019). On the training day, each mouse was acclimated to the fear conditioning chamber for 5 min. Then, the mouse received four trials of electric foot shock (0.5 mA, 2 s duration) with a 1 min intertrial interval and 1 min of no stimulation followed by the last foot shock. Once finished, the mouse was moved back to its home cage immediately. Twenty-four h later, the mouse was placed in the same chamber for 5 min. The equipment was cleaned with absolute ethanol between mice. Behaviors were automatically recorded by a camera connected to a digital tracking device (Xinruan Information Technology Co. Ltd, Shanghai, China). A computer with the fear conditioning software (Xinruan Information Technology Co. Ltd, Shanghai, China) then processed the tracking information. The freezing time was recorded and analyzed. Freezing behavior was defined as immobility for more than 2 s. Rotarod test The rotarod test was carried out with a rotarod machine (YLS-4C, Yiyan, Jinan, China) as previously described (Genc et al., [63]2021). Before the test, each mouse was habituated to staying on a stationary rod for 2 min. Then, each mouse was placed on the machine rotating from 5 to 40 rpm for 5 min and tested for three trials with a 30-min interval between each trial. Once falling, the mouse was moved back to its home cage immediately, and the latency of the mouse falling was automatically recorded by the machine. If the mouse insisted on the apparatus for more than 5 min of a trial, the latency was counted as 5 min, and the mouse was then moved from the apparatus. The equipment was cleaned with 70% ethanol between mice. The average latency of three trials represented the motor function of a mouse. Immunofluorescence Mice were sacrificed immediately after the last behavioral task. Immunofluorescence was performed as previously described (Jiang et al., [64]2021). Briefly, after fixation and dehydration, brains were sliced into 30 μm coronal sections by using a cryostat microtome. The brain slices immediately adhered to the slides. Then, brain slices at the same level were randomly selected and subjected to the following procedures: Membranes were lysed with 0.5% Triton-X for 10 min. Subsequently, non-specific protein binding sites were blocked with 5% donkey serum for 1 h at room temperature. The slices were then incubated overnight at 4 °C with mouse monoclonal anti-neuron-specific nuclear protein (NeuN) (1:100, MAB377, Millipore), rabbit polyclonal anti-glial fibrillary acidic protein (GFAP) (1:200, 16825-1-AP, Proteintech), rabbit monoclonal anti-induction of brown adipocytes 1 (Iba1) (1:100, ab178847, Abcam), rabbit polyclonal anti-doublecortin (DCX) (1:200, 13925-1-AP, Proteintech), and rabbit polyclonal anti-MAD2B (1:200 ab180579, Abcam) antibodies diluted with 5% bovine serum albumin (BSA). After being rinsed three times for 5 min each with PBS, the brain slices were incubated with fluorochrome-conjugated secondary antibodies (1:200, DyLight-488-labeled or DyLight-594-labeled donkey anti-mouse or donkey anti-rabbit, Jackson, USA) for 1 h at room temperature in the dark. Then, the brain slices were incubated with Hoechst 33258 (Thermo Fisher Scientific, Shanghai, China) for 5 min. Finally, the brain slides were automatically scanned by a fluorescence microscope (Olympus, SV120, Japan) for imaging. The VS-ASW-S6 software was used to acquire the images of MAD2B, NeuN, GFAP, and Iba1. A Carl Zeiss LSM780 laser scanning confocal microscope (Zeiss Microsystems, Jena, Germany) was used to acquire images of DCX. The number of NeuN-, GFAP-, and Iba1-positive cells were determined in a 100 × 100 μm square and the number of DCX^+ cells was obtained from the DG zone in each brain slice. All positive puncta were counted by Fiji software (National Institutes of Health, Bethesda, MD, USA). The relative number of positive cells was normalized to that in the WT group by the following formula: the relative number of positive cells of each mouse = the number of positive cells of each mouse/the average number of positive cells in the WT group. Protein extraction and Western blot Western blot was conducted as previously described (Zhong et al., [65]2020). According to the references (McGill et al., [66]2018), the