Important roles of Vilse in dendritic architecture and synaptic plasticity

Vilse/Arhgap39 is a Rho GTPase activating protein (RhoGAP) and utilizes its WW domain to regulate Rac/Cdc42-dependent morphogenesis in Drosophila and murine hippocampal neurons. However, the function of Vilse in mammalian dendrite architecture and synaptic plasticity remained unclear. In the present study, we aimed to explore the possible role of Vilse in dendritic structure and synaptic function in the brain. Homozygous knockout of Vilse resulted in premature embryonic lethality in mice. Changes in dendritic complexity and spine density were noticed in hippocampal neurons of Camk2a-Cre mediated forebrain-specific Vilse knockout (VilseΔ/Δ) mice. VilseΔ/Δ mice displayed impaired spatial memory in water maze and Y-maze tests. Electrical stimulation in hippocampal CA1 region revealed that the synaptic transmission and plasticity were defected in VilseΔ/Δ mice. Collectively, our results demonstrate that Vilse is essential for embryonic development and required for spatial memory.

mutants claim that Vilse transduces signals downstream of Robo receptor to regulate Rac-dependent midline repulsion in Drosophila 15,16 . In mice, Vilse interacts with connector enhancer of KSR-2 (CNK2) in the postsynaptic membrane also via WW domain to modulate Rac/Cdc42 signaling during spine morphogenesis in hippocampal neurons 17 . Structurally, MyTH4 domain has a conserved helical bundle and is often located N-terminal to a FERM domain to associate with microfilaments and microtubules 18 . However, there is no FERM domain in Vilse and the role of Vilse MyTH4 domain in regulating Rho activity and cytoskeleton remains to be elucidated.
In this report, we characterized the structural and functional phenotypes of Vilse Δ/Δ mice with anatomical, physiological and behavioral means. In Vilse Δ/Δ mice, the dendritic architecture in the hippocampal neurons were altered while hippocampus-mediated spatial learning/memory function was impaired. Electrophysiological recordings in the hippocampal CA1 region also revealed defects in synaptic transmission and plasticity in mutant mice. Our results led to a conclusion that Vilse is essential for embryonic development and required for dendritic structure and synaptic function.

Results
Generation of Vilse knockout mice. To address the function of Vilse in vivo, embryonic stem (ES) cells with LacZ reporter insertion in Vilse allele were obtained from International Mouse Phenotyping Consortium (IMPC). The fragment containing exons 1 to 4 of Vilse is designated to be spliced to the lacZ trapping element in this targeting cassette (Fig. 1a). Chimeras capable of germ-line transmission were backcrossed to C57BL/6 mice to generate Vilse heterozygous mice (Vilse +/− ). Male and female Vilse +/− mice were intercrossed to produce Vilse homozygous mice (Vilse −/− ). Of 102 progeny analyzed, Vilse +/+ and Vilse +/− mice were approximately in a ratio of 1:2 without noticeably phenotypic differences (data not shown). However, Vilse −/− mice died at early embryonic stages. Dead embryos could be found at E13.5 with two interrupted alleles (Fig. 1b), indicating that deficiency of Vilse might result in prematurely embryonic death. Whole embryo extracts were collected and immunoblotting analysis revealed the loss of Vilse protein in Vilse −/− embryos at E10.5 (Fig. 1c). We further isolated embryos at E9.0 for H&E staining. At this stage, prospective lungs, brain, forelimb and hindlimb buds appeared in Vilse +/+ and Vilse +/− embryos. By contrast, Vilse −/− embryos did not display normal organogenesis, which had a severely impaired embryo morphology (Fig. 1d), indicating that the ablation of Vilse deteriorated the development and survival of Vilse −/− embryos Since Vilse is important for cell viability, we set out to generate Vilse's conditional knockout mice. To determine which Cre-transgenic mice could be applied for Vilse ablation, we examined the expression of Vilse in different tissues. The exons 1-4 of Vilse are expected to be fused to β -galactosidase in the original gene-trap construct (Fig. 1a). Therefore, the Vilse-LacZ fusion protein could be visualized by the addition of X-gal as substrates. The appearance of embryos at E13.5 displayed blue colors with ubiquitously spatial patterns (Fig. 1e). Immunoblotting analysis also revealed a ubiquitous expression of Vilse in a variety of adult tissues. More interestingly, the expression of Vilse in the brain is much higher than other tissues (Fig. 1f). The double labeling experiment revealed that Vilse predominantly colocalized with MAP2 in the pyramidal cells of hippocampal CA1 neurons. By contrast, little colocalization of Vilse with Iba-1, a microglia marker, and glial fibrillary acidic protein (GFAP), an astrocyte marker, were found in the hippocampal CA1 region, indicating that Vilse mainly expressed in CA1 neurons but not in glia cells. We therefore included Camk2a-driven Cre expression strategy to generate the forebrain-specific conditional knockout of Vilse to explore the function of Vilse in the brain.
Generation of Vilse Δ/Δ mice by Camk2a-Cre mediated recombination. There are two FRT sequences flanking the gene-trap LacZ cassette. This LacZ cassette could be removed by FLP recombinase to leave exon 5 (a.a. 197-648) flanked by two loxP sequences (Fig. 2a), which can be further eliminated by Cre recombinase to produce Vilse Δ/Δ mice. Therefore, Vilse +/− mice were first mated with FLP transgenic mice to produce Vilse fl/+ and Vilse fl/fl littermates. Both alleles containing lacZ cassettes were successfully removed by FLP recombinase shown as fl/fl (Fig. 2b). These Vilse fl/fl pups were viable, supporting the notion that ablation of Vilse resulted in embryonic lethality. Subsequently, Vilse fl/fl mice were crossed with Camk2a-Cre transgenic mice to produce Vilse Δ/Δ mice. Vilse Δ/Δ mice were viable with no noticeably developmental defects. Immunohistochemistry of adult Vilse Δ/Δ mice confirmed a significant reduction of Vilse in the cerebral cortex and hippocampus (Fig. 2c) and immunoblotting analysis revealed the loss of Vilse in the hippocampal extracts (Fig. 2d).
Dendritic structures of CA1 pyramidal neurons were altered in Vilse Δ/Δ mice. To estimate the synaptic function in the structural aspect, we examined the dendritic architecture of CA1 neurons. Golgi-Cox impregnated CA1 pyramidal neurons (30 neurons from control and 30 neurons from Vilse Δ/Δ mice) were reconstructed and analyzed using Neurolucida software (Fig. 3a). The dendritic complexity was evaluated using the concentric-ring method of Sholl in a three-dimensional manner. Compared with controls, the numbers of intersections between dendritic branches and the concentric balls were significantly reduced in both apical and basilar dendrites of CA1 neurons in Vilse Δ/Δ mice (Fig. 3b). The numbers of dendritic segments (Fig. 3c), bifurcation nodes and terminal ends (Fig. 3e) were also reduced, especially in the apical dendrites of CA1 neurons in Vilse Δ/Δ mice. These results indicated a reduction in dendritic complexity in mutant mice. We further analyzed the length of dendritic segments. Compared with those in controls, the length between two bifurcation nodes (intermodal length) was shorter in the basilar dendrites of CA1 neurons in mutants, while the length of terminal segments was reduced in both apical and basilar dendrites of these cells (Fig. 3d), leading to the shorter dendritic length in Vilse Δ/Δ mice (Fig. 3e). These results revealed that the elongation and bifurcation of dendrites are affected in the absence of Vilse. We also characterized the density of dendritic spines, the protrusions on the dendrites where the excitatory synapses reside. Interestingly, CA1 pyramidal neurons in Vilse Δ/Δ mice had greater spine density in both apical and basilar dendrites than those in control mice (Fig. 3f). We further examined the features of dendritic spines (Fig. 3g). Compared with those in control mice, greater portion of immature spine (thin spines)  Fig. 1). DG granule cells in Vilse Δ/Δ mice also had more immature spines compared with those in control mice (2834 protrusions counted in control group and 2302 in Vilse Δ/Δ mice, Supplementary Fig. 1g). Together, the morphological alterations of hippocampal neurons in the absence of Vilse suggested the changes of hippocampal function in Vilse Δ/Δ mice. Spatial memory impaired in Vilse Δ/Δ mice. To evaluate the function of the hippocampus in Vilse Δ/Δ mice, hippocampus-mediated spatial learning/memory tests were performed. In training phase of Morris water maze test, Vilse Δ/Δ and control mice behaved similarly in finding the hidden platform on Day 1, suggesting the comparable visual and mobile capabilities among control and Vilse Δ/Δ mice. However, during day 2 to day 5, a remarkably longer latency to find the hidden platform was exhibited in Vilse Δ/Δ mice compared with control groups, suggesting a poorer spatial learning ability in Vilse Δ/Δ mice (Fig. 4a). To evaluate the spatial memory of mice, a probe test was performed on day 6 in which the hidden platform was removed and swim paths of mice were traced. As anticipated, control mice swam to the removed platform areas faster than Vilse Δ/Δ mice did ( Supplementary Fig. S2). Male and female mice displayed similar patterns ( Supplementary Fig. S2), indicating that gender differences in Vilse Δ/Δ mice were not associated with impaired spatial memory. Furthermore, in the Y-maze test, Vilse Δ/Δ mice also displayed a defective spatial working memory compared with control mice (Fig. 4b). However, the fear memory measured by passive avoidance test did not imply a significant difference between control and Vilse Δ/Δ mice (Fig. 4c). Together, these results indicated the hippocampus-mediated spatial learning/memory function is impaired in Vilse Δ/Δ mice.
Reduced long-term potentiation in hippocampal CA1 area of Vilse Δ/Δ mice. To bridge the structural defects in hippocampal neurons and impaired hippocampal function in Vilse Δ/Δ mice, the features of synaptic transmission in the hippocampal CA1 neurons were examined. We recorded field excitatory postsynaptic potential (fEPSP) in hippocampal CA1 slices taken from control and Vilse Δ/Δ mice. Input-output curves showed that the slope of the field excitatory postsynaptic potential (fEPSP) was significantly lowered in Vilse Δ/Δ mice compared with control mice as the stimulus intensity increased (5-7 V, P = 0.002; 8-10 V, P < 0.001) (Fig. 5a), reflecting a dampened basal synaptic transmission in Vilse Δ/Δ mice. We next measured the responses to paired-pulse stimuli. Paired-pulse facilitation (PPF) is a character at the Schaffer collateral-CA1 synapse and  considered as a form of short-term synaptic plasticity in which the second response to the closely spaced paired stimuli is increased due to residual Ca 2+ in the presynaptic nerve terminal from the first stimulus 19 . No difference in the PPF between the control and Vilse Δ/Δ mice was noticed, indicating that Vilse deficiency did not affect short-term presynaptic plasticity (Fig. 5b). Long-term potentiation (LTP), a use-dependent change in synaptic strength, has been well established as a cellular substrate for information storage in the brain 20 . We compared the properties of LTP of fEPSP in hippocampal CA1. After a period of at least 10 min baseline fEPSP recording, application of high frequency-stimulation immediately induced a post-tetanus potentiation (PTP), followed by a long-term potentiation (LTP) of fEPSP in slices. In wild-type mice, the PTP and LTP were 233 ± 27% and 161 ± 8% of the baseline, respectively. In contrast, the PTP and LTP were 175 ± 12% (p < 0.05, paired t test) and 91 ± 8% of the baseline (p = 0.450, paired t-test), respectively. The significant difference between control and Vilse Δ/Δ mice was found in LTP (p < 0.01, Mann-Whitney U-test) but not in PTP (p = 0.073, Mann-Whitney U-test), indicating a deficit in long-term synaptic plasticity in Vilse Δ/Δ mice (Fig. 5c).

Discussion
To avoid the premature lethality, forebrain neuron-specific conditional Vilse knockout mice were generated. Camk2a-mediated Vilse ablation in mice did not result in premature embryonic death, suggesting that Vilse might not be essential for cell viability in neuronal cells. We then evaluated the consequences of Vilse knockout in the aspects of neuronal morphology and synaptic transmission in the hippocampal neurons and hippocampus-mediated spatial learning and memory function. Results from water maze and Y-maze tests revealed the defective spatial memory in Vilse Δ/Δ mice. Defective spatial memory in mutant mice could be reasoned by reduced synaptic transmission and impaired synaptic plasticity such as LTP as well as the morphological alterations in dendritic complexity and spine properties in hippocampal CA1 neurons. The structural and functional changes of neurons in Vilse Δ/Δ mice are likely due to the dysregulation of actin cytoskeletons resulted from the changes of Rho GTPases downstream of Vilse.
Rho GTPases not only regulate actin filament reorganization, but also participate in CTL-and Fas-induced apoptosis. Dominant-negative mutants of Rho GTPases and Clostridium difficile toxin B, an inhibitor of all Rho GTPases, inhibit cellular susceptibility to Cytotoxic T lymphocyte-and Fas-induced apoptosis 21 . Furthermore, Rho GTPase signaling pathways has been shown to be associated with apoptosis and the killing of superfluous cells is important during embryonic development 22 , indicating that loss of Vilse may potentiate the over-activation of Rho GTPases in embryos and consequently result in embryonic lethality. Intriguingly, Camk2a-mediated Vilse ablation in the forebrain did not result in a significant increment of RhoA GTPase activity (data not shown), suggesting that the loss of Vilse might be compensated by other RhoGAPs to modulate RhoA GTPase activity.
It is always a contention that there are approximately 70 RhoGAPs, from yeast to human, in contrast to 22 Rho proteins 3 . One plausible explanation is that some GAPs have preferential tissue expression and tissue specific functions. Their specific spatial and temporal expressions during development enable them to be tightly coped with Rho GTPases and modulate synaptic function through the interaction with diverse effectors 23 . Several GAPs have been shown to uniquely regulate synaptic development and plasticity [24][25][26] . We used forebrain-specific Vilse conditional knockout mice model to demonstrate that Vilse is involved in the modulation of dendritic architecture including dendritic length, branching property as well as the density, shape and length of dendritic spines. In Figure 4. Impaired spatial memory in Vilse Δ/Δ mice. (a) Individual mouse in control and Vilse Δ/Δ groups was placed in a water maze and allowed to swim to the hidden platform four times a day for five days. The cutoff time was 90 seconds. Swimming path and time spent in quadrants of the water pool were recorded and the latency to locate the platform was averaged and analyzed in each day. *p < 0.05, **p < 0.01, ***p < 0.001 by student's t-test. (n = 18; 6 males, 12 females). (b) In Y-maze test, each mouse was placed at the end of one arm and allowed to move freely through the maze during an 8-min session. Comparison of spontaneous alternation between control and Vilse Δ/Δ mice in Y-maze test was performed. Columns represent the mean ± S.E.M, n = 7 mice for each genotype, **p < 0.01, unpaired t-test. (c) Passive avoidance test. Mice were placed into the light side of the chamber and the latency to enter the dark side was measured. A mild foot shock is delivered upon entry to the dark side of the chamber. Memory of passive avoidance was assessed at 24 h and 48 h (n = 7). n.s., not significant.
Scientific RepoRts | 7:45646 | DOI: 10.1038/srep45646 CA1 and DG neurons of Vilse Δ/Δ mice, greater spine density was noticed. It seemed odd to correlate this result with the reduced synaptic transmission. Since the profiles of dendritic spines are important for the transmission and integration of neural signals 27 , altered dendritic architectures in CA1 and DG cells of Vilse Δ/Δ mice may result in defects of the synaptic transmission and integration of neural information. Compared with the controls, greater amounts of immature and unstable spines were observed in CA1 and DG neurons of Vilse Δ/Δ mice, in line with the in vitro study of interfering the binding between Vilse and CNK2 17 . Removal of Vilse would affect the structural integrity of postsynaptic scaffold which might in turn influence the distribution and function of postsynaptic receptors and ion channels and result in impaired synaptic transmission. The increased density of dendritic spines could be a compensatory change. Alternatively, these excess dendritic spines might be spared from synaptic activity-mediated spine pruning. Indeed, this notion is partly supported by electrophysiological examinations on basal synaptic transmission and long term potentiation. Little change in paired pulse facilitation suggests that short-term presynaptic plasticity at CA1 synapse is not altered in Vilse Δ/Δ mice. Nevertheless, the capability for CA1 synapses to undergo the use-dependent change is limited in Vilse Δ/Δ mice as LTP expression is impaired. Since the hippocampus is specific to spatial memory, these observations, including dendritic architecture and electrophysiological examinations, echoed the results of water maze and Y maze tests and provide evidence that Vilse is an important player in synaptic plasticity for spatial memory. Together, the structural changes of dendritic architecture and disruption of postsynaptic integrity might account for the hippocampal LTP deficiency and behavioral deficits in hippocampus-related learning and memory tasks in Vilse conditional KO mice.
Immunofluorescence. Mice were perfused with PBS followed by paraforaformaldehyde fixation. The fixed brains were cryoprotected in 30% sucrose and cut to 30 mm thick sections. The sections were blocked in 10% normal goat serum and 2% bovine serum albumin in PBST (phosphate-buffered saline and 2% Triton X-100) and incubated with individual primary antibodies against Vilse, MAP2, Iba-1, and GFAP in blocking buffer overnight. Subsequently, DyLight 488-and Alexa fluor 594-conjugated secondary antibodies (Jackson ImmunoResearch Laboratories, West Grove, PA) were applied. Sections were coverslipped with antifade mounting medium (Southern Biotech, Birmingham, AL) and examined on a confocal microscope (Leica TCS SP5, Wetzlar, Germany).
Generation of Vilse conditional knockout mice. All animal studies were performed in compliance with the guidelines of the Institutional Animal Use and Care Committee, Academia Sinica. An ES cell containing LacZ reporter-tagged insertion in Vilse/Arhgap39 allele were obtained from International Mouse Phenotyping Consortium (IMPC) using a combination of both a reporter-tagged and a conditional mutation 28 . Based on the original construct, exon1 to exon4 of Vilse is spliced to a lacZ trapping element in the targeting cassette containing the mouse En2 splice acceptor and the SV40 polyadenylation sequences 19 . This ES cell was injected into C57BL/6 blastocysts to generate chimeric mice. Germ-line transmission of the mutant allele was tested by PCR while two independent male chimeric mice were back-crossed with female C57BL/6 mice to generate Vilse heterozygous F1 offsprings. Conditional alleles were further generated by removal of the gene-trap LacZ cassette by FLP recombinase, which reverts the mutation to wild type allele, but leaving loxP sites on either side of exon 5 (a.a 197-648). Floxed homozygous male and female mice (Vilse fl/fl ) were mated with the mouse calcium/calmodulin-dependent protein kinase II alpha promoter driving Cre recombinase transgenic mice (Camk2a-Cre mice, a gift from Dr. Che-Kun Shen, Academia Sinica) to produce Vilse Δ/Δ mice. All animal studies were performed in compliance with the guidelines of the Institutional Animal Use and Care Committee, Academia Sinica.

Golgi-Cox impregnation and morphometric analyses. After transcardiac perfusion with
phosphate-buffered saline and fixative (4% paraformaldehyde in phosphate buffer, pH 7.4), whole brains were taken and immersed in the impregnation solution from the FD Rapid Golgi Stain kit (NeuroTechnologies, Ellicott City, MD) for Golgi-Cox impregnation as previously described 29 . In brief, after 3 weeks of impregnation, brain samples were sectioned at a thickness of 150 μ m and incubated with a mixture of developer and fixer solutions (FD Rapid Golgi Stain kit), washed and mounted. The pyramidal neurons in the hippocampal CA1 region and granule cells in the dentate gyrus (DG) were examined under a light microscope with a 20x objective lens for dendritic morphology and 100x objective lens for spine analysis, respectively. Series of pictures were taken by a CCD camera with the aid of computer-controlled motorized stage using the Stereo Investigator system (MBF Bioscience, Williston, VT). The morphology of selected neurons was reconstructed and analyzed with Neurolucida software (MBF Bioscience). Data were expressed as the mean ± SEM. Two-tailed unpaired Student's t-test was used for statistical analysis.

Morris water maze.
All experiments regarding behavior tests, Golgi stain, and electrophysiology were performed in accordance with guidelines approved by National Taiwan University College of Medicine and College of Public Health (20150132). Mice were reared in the animal facility of National Taiwan University under a 12-h light/dark cycle with free access to food and water. The properties of hippocampus-mediated spatial learning and memory were evaluated using Morris water maze test 30 with minor modifications. Briefly, the circular pool was 1.88 m in diameter and contained water (temperature 19 °C) that was made opaque with non-fat milk. A hidden platform was put in a fixed location 1 cm below the water surface. During the training period, individual mouse was placed in the water facing the pool wall at one of four start points (north, south, east, or west). Upon release into the water, the mouse was allowed to swim for 90 secs to locate the hidden platform. An additional 30 secs were then allowed for mice to stand still on the platform and escape from the water before being removed. If the mouse failed to stand on the platform, it was guided to the platform and remained there for 30 sec. Four trials of different start points were given each day for 5 consecutive days. After 5-day training period, another test trial was given in which the platform was removed. Latencies to reach the hidden platform and the swim paths were recorded with an automatic video tracking system. Y-maze test. Mice were placed individually at the end of one arm of the Y-maze and allowed to move within three arms freely for an 8-min period. The total number and series of arm entries were recorded. The number of non-overlapping entrance sequences (e.g. ABC, BCA) was defined as the number of alternations. The spontaneous alternation (%) was calculated as: (number of alternations)/ (total number of arm entries-2) × 100 31 .
Passive avoidance tes. During habituation, mice were placed into the light side of the chamber and the latency to enter the dark side was recorded. On the training day, a mild foot shock is delivered upon entry to the dark side of the chamber. Memory of passive avoidance was assessed at 24 h and 48 h after training and the latency to enter the dark side was analyzed.
Electrophysiology. Vilse Δ/Δ or control mice aged 8 weeks were decapitated and their brains were quickly removed and placed in ice-cold artificial cerebrospinal fluid (ACSF) containing (in mM) NaCl, 119; KCl, 2.5; NaHCO 3 , 26.2; NaH 2 PO 4 , 1; MgSO 4 , 1.3; CaCl 2 , 2.5; glucose, 11; pH was adjusted to 7.4 by gassing with 5% CO 2 : 95% O 2 . Transverse hippocampal slices of 450 μ m thickness were cut with a vibrating tissue slicer (Dosaka microslicer, Japan) and transferred to an interface-type holding chamber at room temperature (25 °C). For extracellular field potential recording, slices were transferred to an immersion-type recording chamber, superfused with ACSF containing 0.1 mM picrotoxin at room temperature. The superfusion rate was controlled at 2 mL/ min. To prevent epileptiform discharge of pyramidal neurons, a surgical cut was made at the border between areas CA1 and CA3. A glass pipette filled with 3 M NaCl was positioned at stratum radiatum of CA1 to record field excitatory postsynaptic potentials (fEPSP) evoked by a bipolar stainless steel electrode (FHC, Bowdoinham, ME) placed in the vicinity of recording pipette. To stimulate Schaffer collateral branches, a constant current pulse of 0.5 ms duration (DS3, Digitimer, UK) was delivered through the stimulating electrodes every 30 s with the intensity adjusted so that 40 ± 50% of the maximal response was elicited. Long-term potentiation (LTP) of fEPSP was induced by high-frequency-stimulation consisting of 3 trains of 100 pulses at 100 Hz with inter-train interval being 10 s. All signals were filtered at 2 kHz by a low-pass Bessel filter provided by the amplifier (Axopatch-1D, Axon Instruments, Foster City, CA) and digitized at 5 kHz using CED micro 1401 interface running Signal software provided by CED (Cambridge Electronic Design, Cambridge, UK). The initial slope of fEPSP was measured for data analysis. Synaptic responses were normalized to average values measured over a baseline period. The averaged slope of fEPSPs recorded between 0-2 and 55-60 min after high frequency stimulation was used for statistical comparisons of post-tetanus potentiation (PTP) and LTP, respectively. All data given were Mean ± SEM, and were statistically compared using either paired t-tests or one-way ANOVA. The criterion for significance was P < 0.05.