The zinc finger/RING domain protein Unkempt regulates cognitive flexibility

Correct orchestration of nervous system development is a profound challenge that involves coordination of complex molecular and cellular processes. Mechanistic target of rapamycin (mTOR) signaling is a key regulator of nervous system development and synaptic function. The mTOR kinase is a hub for sensing inputs including growth factor signaling, nutrients and energy levels. Activation of mTOR signaling causes diseases with severe neurological manifestations, such as tuberous sclerosis complex and focal cortical dysplasia. However, the molecular mechanisms by which mTOR signaling regulates nervous system development and function are poorly understood. Unkempt is a conserved zinc finger/RING domain protein that regulates neurogenesis downstream of mTOR signaling in Drosophila. Unkempt also directly interacts with the mTOR complex I component Raptor. Here we describe the generation and characterisation of mice with a conditional knockout of Unkempt (UnkcKO) in the nervous system. Loss of Unkempt reduces Raptor protein levels in the embryonic nervous system but does not affect downstream mTORC1 targets. We also show that nervous system development occurs normally in UnkcKO mice. However, we find that Unkempt is expressed in the adult cerebellum and hippocampus and behavioural analyses show that UnkcKO mice have improved memory formation and cognitive flexibility to re-learn. Further understanding of the role of Unkempt in the nervous system will provide novel mechanistic insight into the role of mTOR signaling in learning and memory.

through loss-of-function mutations in the genes TSC1 or TSC2, causes the multisystem disorder tuberous sclerosis complex (TSC). Patients with TSC have benign tumours in multiple organs, including the brain, which can result in epilepsy, autism and intellectual disability 7 . Activating mutations in other components of the mTORC1 pathway cause focal cortical dysplasia, hemimegalencephaly and other epilepsy syndromes 6 . mTOR also plays a crucial role in regulating translation at the synapse and as a result in synaptic plasticity. Long term modification of synaptic strength, or long term potentiation (LTP), requires increased local translation. Early studies showed that the mTORC1 substrate 4E-BP1 and its binding partner eIF4E colocalise with post-synaptic markers 8 . Moreover, mTORC1 regulates the local translation of the elongation factor eEF1A in dendrites to promote LTP 9 . More recently, evidence from patients and animal models has shown that increased synaptic translation due to upregulation of mTORC1 activity contributes to epilepsy associated and autism spectrum disorders, such a fragile X syndrome, Angelman syndrome and tuberous sclerosis complex 10,11 .
Unkempt is a highly conserved zinc finger/RING domain protein that was originally identified in Drosophila, where it was shown to be expressed in the developing embryonic nervous system and to be important for tissue patterning. Drosophila null mutants in Unkempt are developmentally lethal, while hypomorphic mutants are viable but have an 'unkempt' phenotype, with roughened eyes, splayed wings and crossed scutellar bristles 12 . We previously showed that Unkempt acts genetically downstream of mTOR to regulate differentiation of photoreceptors in the developing retina in Drosophila [13][14][15] . In nutrient rich conditions, loss of Unkempt does not affect cell proliferation, but alters the timing of Drosophila photoreceptor differentiation 13,16 . Loss of Unkempt causes precocious differentiation of photoreceptor neurons and patterning defects in the adult eye 13 . Recently we also showed that Unkempt is strongly expressed in the larval nervous system in Drosophila, where it negatively regulates the cell cycle in neural progenitor cells 17 .
Mammalian Unkempt is most strongly expressed in cell lines with a neuronal origin and in vivo its expression is strongest in the developing central nervous system 18 . Unkempt expression in the developing brain peaks between embryonic days 12 and 18 and is particularly abundant in Tuj-1 expressing neurons 18 . In vitro experiments have shown that mammalian Unkempt binds mRNAs through its zinc finger domain to regulate their translation 18,19 . Moreover, both Drosophila and mammalian Unkempt physically interact with the mTORC1 component Raptor 16,20 . However, the role of mammalian Unkempt in vivo is largely uncharacterised. To assess the role of Unkempt in vivo we generated a nervous system-specific Unkempt knockout mouse. Loss of Unkempt in the developing nervous system causes a reduction in the expression of the mTORC1 component Raptor but surprisingly does not affect neural progenitor proliferation. The overall development of the nervous system is also unaffected by loss of Unkempt in this model. However, expression studies show that Unkempt is strongly expressed in the adult cerebellum and hippocampus, and behavioural analyses show that Unkempt knockout mice have improved reversal learning. Thus, loss of Unkempt improves cognitive flexibility.

Results
Loss of Unkempt causes reduced expression of Raptor in the developing brain. Unkempt was originally identified in Drosophila as a zinc finger/RING domain protein essential for developmental viability and patterning (Fig. 1A,B) 12 . Unkempt is expressed ubiquitously in Drosophila but enriched in developing nervous system, where it acts as a component of the mTOR pathway to regulate neurogenesis 12,13,17 . The mTOR pathway has important roles in nervous system development and hyperactivation of the mTOR pathway causes neurological diseases associated with aberrant intellectual development, epilepsy and autism 5,6 . Unkempt is conserved in mammals (Fig. 1A,C) and most strongly expressed in developing neurons but its role in the mammalian nervous system is largely unknown 18 . To address this, we generated a conditional allele of Unkempt in which exons 3 and 4 are flanked by loxP sites (Fig. 1D). Removal of exons 3 and 4 is predicted to generate a premature stop codon after 113 amino acids (Supplemental Figure S1). We crossed these mice to Nestin-Cre expressing mice 21 , to generate animals in which Unkempt is knocked-out in neural progenitors (Unk cKO ) from around embryonic day (E) 10.5 ( Fig. 1D-F and Supplemental Figure S2A). qRT-PCR analysis of E16.5 brain tissue using two primer sets, both within the deleted genomic region, showed a dramatic reduction in Unkempt transcript in Unk cKO embryos (Fig. 1G). At E16.5 homozygous Unk cKO embryos also had no detectable Unkempt protein expression in the brain, while heterozygotes had around 50% Unkempt expression levels compared to littermate controls ( Fig. 1H; Supplemental Figure S2A).
In mammals, Unkempt has a paralog called Unkempt like (Unkl), which also contains N-and C-terminal zinc finger and RING domains respectively 22 . To test whether there is cross regulation between Unkempt and Unkempt like we analysed Unkempt like expression in Unk cKO embryos. Unkempt like expression was unaltered in E16.5 Unk cKO embryos (Fig. 1I, J, K; Supplemental Figure S2B), indicating that loss of Unkempt does not affect Unkempt like expression in the developing nervous system. Proteomic analysis of the mTOR pathway identified physical interactions between Unkempt and several mTORC1 components in Drosophila cultured cells 20 . Subsequent co-immunoprecipitation studies have shown that both Drosophila and mammalian Unkempt interact specifically with the mTORC1 component Raptor 16 .
To confirm these findings, we used a Raptor overlay assay as a complementary approach. This assay uses direct association of the purified candidate protein to Raptor or Raptor mutant 4, which retains the ability to interact with mTOR but not does not interact with mTOR substrates 23,24 . Using the Raptor overlay assay we found that, like the mTORC1 substrate 4E-BP1, Unkempt bound avidly to wild-type HA-Raptor but binding to HA-Raptor mutant 4 was greatly reduced ( Fig. 2A; Supplemental Figure S3A). These data confirm that Unkempt directly interacts with Raptor in vitro.
We next tested whether loss of Unkempt affects the expression of Raptor in the developing nervous system. Interestingly, Unk cKO embryos (Fig. 2B,C; Supplemental Figure S3B) had a significantly reduced level of Raptor in the brain (Fig. 2B,D). To determine whether the reduction in Raptor perturbs mTORC1 signaling, we analysed www.nature.com/scientificreports/ the expression and phosphorylation of ribosomal protein S6 (rpS6) and eukaryotic translation initiation factor 4E-binding protein 2 (4E-BP2), the predominant form of 4E-BP in the brain 25 . Total rpS6 and phospho-rpS6   Figure S3C). Total 4E-BP2 and phospho 4E-BP (P-4E-BP) levels in the brain were also not significantly different between  Figure S3C). We also used phosphorylation of AKT at S473 as a readout of mTORC2 activity. Phosphorylation of AKT at S473 was not altered in Unk cKO embryos ( Fig. 2E,J, Supplemental Figure S3D), indicating that loss of Unkempt does not affect mTORC2 activity. These data indicate that Unkempt is necessary to maintain normal levels of Raptor expression during nervous system development. However, the reduced Raptor expression resulting from loss of Unkempt does not negatively impact mTORC1 signaling.

Loss of Unkempt does not affect nervous system development. To investigate whether loss of
Unkempt affects neurodevelopment we analysed neuroanatomy and neurogenesis in Unk cKO embryos. Overall neuroanatomy of the brain of Unk cKO mice at E16.5 appeared normal compared to littermate controls (Fig. 3A,B).
To assess the requirement for Unkempt in early cortical development we stained brains from E16.5 Unk cKO embryos with markers of different neural sub-types. Expression of the neuronal markers NeuN and DCX, as well as the cortical layer specific markers Tbr1, Tbr2 and Ctip2 were largely normal in Unk cKO embryos (Fig. 3C,D). We also stained for and quantified the expression of markers of cell proliferation and mitosis (Ki67, phosphohistone 3 (PH3) and BrdU incorporation), as well as Mash1-expressing transit amplifying cells in the subventricular zone of Unk cKO embryos at E16.5. Expression of all these neurogenic markers were not significantly different from controls in Unk cKO mice ( Fig. 3E-P). Overall expression of Ki67 in the developing olfactory bulb and medial region of the brain in Unk cKO mice at E16.5 was also similar to control (Supplemental Figure S4). Thus, although Raptor expression is reduced in the brain of Unk cKO embryos, neurogenesis and overall brain development appear unimpeded.
Unkempt is strongly expressed in the adult cerebellum and hippocampus. Unkempt is strongly expressed in the developing nervous system in both Drosophila and mice but whether its expression continues in the adult nervous system is unknown 17,18 . Weak Unkempt protein expression was detectable in whole brain lysate at post-natal day (P) 60 and, as expected, was not expressed in Unk cKO mice brain tissue ( Fig. 4A; Supplemental Figure S3E). In situ hybridisation analysis of Unkempt mRNA expression from the Allen Brain Atlas shows that at P56 Unkempt is very weakly expressed in most brain regions ( Fig. 4A) 26 . However, Unkempt mRNA expression is increased in the olfactory bulb, and is strongly expressed in the molecular layer of the cerebellum, as well as the pyramidal layer and dentate gyrus granule cell layer of the hippocampus (Fig. 4B,C).

Unk cKO mice have improved cognitive flexibility.
Hyperactivation of the mTOR pathway in the brain can cause seizures, increased anxiety and memory impairment in animal models, and in patients can cause epilepsy, autism and intellectual disability 5,27 . Conversely, chronic inhibition of mTOR signaling is associated with decreased anxiety, improved learning and memory in animal models 28,29 . Unk cKO mice were born at the expected Mendelian ratios (Table 1) but adult mice weighed slightly less than controls (Fig. 5A). Overall neuroanatomy of the mature Unk cKO brain appeared normal ( Fig. 5B,C, Supplemental Figure S5). We then used a battery of tests to determine whether loss of Unkempt affects behaviour in adult mice. To assess anxiety and locomotor activity open field and elevated plus maze tests were used 30,31 . Unk cKO mice had similar locomotor activity to littermate controls ( Fig. 5D,E). However, Unk cKO mice showed a trend towards decreased anxiety, as they spent more time in the central zone of the open field arena ( Fig. 5F) and in the elevated plus maze test spent more time on the open arms ( Fig. 5G,H), although these differences were not statistically significant. We next assessed the requirement for Unkempt in learning and memory. Unk cKO mice performed similar to controls in the Y maze (Fig. 6A), a test of spontaneous spatial novelty preference that measures rapidly acquired, short-term spatial memory 32 . Analysis of spatial learning and reference memory using the Morris water maze 33 , showed that both controls and Unk cKO mice had a significant reduction in the latency to find the platform over the hidden days 1-6 (H1-H6) and spent a significantly increased time in the platform-containing quadrant at the end of day 6 ( Fig. 6B,C), indicating normal acquisition learning and reference memory.
Activation of mTOR signaling impairs the ability to re-learn in 'reversal learning' behavioural paradigms, which test cognitive flexibility 34,35 . We used the Morris water maze to test reversal learning: after the six (hidden) days ( Fig. 6B,C), the platform was moved and testing continued for a further five days (the reversal task, R1-R5). In the reversal learning test Unk cKO mice showed a significantly reduced latency to find the platform and spent more time in the probe quadrant compared to controls (Fig. 6B,D). Unk cKO mice also turned significantly more acutely towards the probe quadrant (Fig. 6E,F). Unk cKO mice swim speed was not significantly different from controls ( Fig. 6G,H) and so did not confound their ability to find the platform. Thus, spatial learning acquisition and retention of spatial memory over time is unchanged in Unk cKO mice, but they show enhanced memory formation and cognitive flexibility to re-learn.

Discussion
Unkempt is a highly conserved zinc finger/RING domain protein. Drosophila and mouse Unkempt have the same overall primary structure and 60% identity in the zinc finger domain. Unkempt has been characterised genetically in Drosophila as a regulator of neurogenesis and growth control, acting downstream of mTOR 13,16,17 . In mammals, mechanistic studies have revealed the role of the zinc finger domain of Unkempt. Unkempt regulates the translation of several hundred target mRNAs, in both cultured cells and the developing brain, by binding a specific U/A-rich motif through its zinc fingers 18,19 . The requirement for Unkempt in the mammalian nervous system has not been previously determined. Given its essential role in Drosophila development our study shows, surprisingly, that conditional knockout of Unkempt in the developing nervous system is not overtly detrimental. However, loss of Unkempt leads to improved cognitive flexibility in adult mice.    18 . Knockdown of Unkempt perturbed neuronal migration and caused aberrant neuronal morphology, which was rescued by expression of an RNAi-resistant Unkempt construct. Given these findings, and that Unkempt regulates neurogenesis and is essential for viability in Drosophila, it is remarkable that knockout of Unkempt did not affect neurodevelopment. There are are several potential explanations for the absence of a neurodevelopmental phenotype in Unk cKO mice. Firstly, the paralog Unkempt like may compensate for the loss of Unkempt during nervous system development. Unkempt like is expressed in the developing neurogenic niche 13 and we find that Unkempt like is still robustly expressed in the developing nervous system in Unk cKO mice. The function of Unkempt like has not been studied in vivo but it may act redundantly with Unkempt. Future double knockout studies may test this hypothesis. Secondly, expression of Unkempt before E10.5, when Nestin-Cre expression begins, in our knockout model may mask the full requirement for the protein during nervous system development. Unkempt is expressed in the nervous system at E10 18 , but protein levels earlier in neurodevelopment have not been analysed. Interestingly, recent single cell RNA-sequencing shows that Unkempt mRNA is expressed in the developing brain as early as E6-E7, and from E8-E10 in ectoderm, neural crest, and radial glial cells (http:// mouse brain. org/ devel opment) 36 . Thus, perdurance of Unkempt mRNA or protein expressed prior to E10 may partially rescue the neurodevelopmental phenotype in Unk cKO mice. However, redundancy with Unkempt like or perdurance do not necessarily explain the difference between Unk cKO mice and Unkempt knockdown; redundancy with Unkempt like should also apply with the knockdown approach, which was performed at E14.5 18 . Although the phenotypes were rescued by expression of an RNAi-resistant Unkempt 18 , the migration and morphology phenotypes   16,20 . The precise function of the interaction between Unkempt and Raptor is currently unknown. Inhibition of mTOR signaling abrogates the physical interaction between Unkempt and Raptor in Drosophila cultured cells 16 . Moreover, the mTOR pathway negatively regulates Unkempt protein levels in the Drosophila developing retina and so Raptor may regulate the stability of Unkempt 13 . Interestingly, we found a strong reduction in Raptor levels in the embryonic brain of Unk cKO mice, suggesting a reciprocal requirement for protein stability between Unkempt and Raptor. Although Raptor levels were reduced, expression of the canonical readouts of mTORC1 activity, phospho-rpS6 and phospho-4E-BP, were not altered in Unk cKO mice. Therefore, Unkempt function likely diverges downstream of mTORC1, potentially acting as a branchpoint in the mTOR pathway (Fig. 6I).
Knockout of Raptor or mTOR in the developing brain causes microcephaly, reduced neural progenitor proliferation and postnatal lethality, while knockout of Tsc1 or Tsc2 causes macrocephaly and premature neural progenitor differentiation 6,37,38 . This contrasts to knockout of Unkempt in the developing nervous system, which leads to reduced Raptor levels, but does not have an obvious effect on cortical neurogenesis or viability. Similar to Unk cKO mice, mice with a brain-specific knockout of FKBP12, which causes partial activation of mTORC1, and mice overexpressing the mTORC1 effector eIF4E, are viable, healthy and have normal spatial learning but have impaired reversal learning 34,35 . Moreover, inhibition of mTOR signaling by chronic rapamycin treatment has been reported to enhance spatial learning, and chronic dietary restriction inhibits mTORC1 to enhance memory performance in young mice 28,29 . The improved reversal learning phenotype in Unk cKO mice is therefore consistent with models of decreased mTOR pathway activity in the brain. Taken together, these studies show that chronic inhibition of mTOR signaling in the brain improves learning and memory. Moreover, manipulation of non-essential mTORC1 components and downstream factors, such as FKBP12, eIF4E and Unkempt, specifically affect cognitive flexibility.
Regulation of local translation by mTORC1 is crucial for synaptic plasticity 39 . Unkempt has been proposed to be a "regulator of regulators", as it controls the translation of proteins that themselves regulate translation, including mTOR, eIF4 and p70S6K pathway associated proteins, suggesting cross-talk between Unkempt and mTORC1 regulated mRNAs 18 . Loss of Unkempt and the resulting altered translation of its target mRNAs may affect local synaptic translation in hippocampal circuits involved in learning and memory (Fig. 6I). The resulting increased synaptic plasticity in the dentate gyrus may lead to improved cognitive flexibility in Unk cKO mice. Conversely, misregulation of Unkempt by hyperactive mTORC1 may contribute to the neurological manifestations of tuberous sclerosis complex and focal cortical dysplasia. Targeting Unkempt may therefore be a novel therapeutic strategy for neurological diseases associated with activated mTOR signaling. Future studies will elucidate the key role of Unkempt in cognition and its intersection with mTOR signaling in learning and memory.
For dendrograms, sequences were aligned with MUSCLE (MEGA-X also has MUSCLE alignment tool implemented) and dendrograms generated with MEGA-X software using the Maximum Likelihood model. Immunofluorescence and Immunohistochemistry. For immunofluorescence, embryonic brains were fixed in 4% (w/v) paraformaldehyde (PFA) in PBS for 24 h at 4 °C while rotating, then cryoprotected by immersion in PBS 15% sucrose for 24 h at 4 °C, and PBS 30% sucrose for 24 h, then embedded in O.C.T Compound (VWR) and stored at − 80 °C. O.C.T embedded brains were cut using a cryostat into 10-20 μM sections on Superfrost plus slides (Thermo Fisher) and dried overnight at room temperature. For better adhesion of sections, the slides were further dried at 50 °C for one hour. Citric acid (pH 6) was pre-heated to 90 °C in a convection oven, and then slides put inside for 45 min for heat-induced epitope retrieval. Slides were left to cool in the same solution in a fume hood, then rinsed three times in 1X TBS (pH 7.6). Slides were drained and a ring drawn around sections with a liquid blocker pen (Agar Scientific). Sections were covered in blocking solution (2% BSA TBS; Sigma-Aldrich) for 5-10 min. Blocking solution was removed and primary antibody added (diluted in blocking buffer) for 16 h at room temperature. Slides were rinsed three times in TBS and incubated with the appropriate Alexa-conjugated secondary antibody (ThermoFisher, 1:300 in blocking buffer) with DAPI for one hour at room temperature. Slides were rinsed three times in TBS and mounted using an aqueous based fluorescence mounting medium (Sigma-Aldrich).
For 3,3′-diaminobenzidine (DAB) staining, embryonic brains were fixed with 4% PFA for 24 h at room temperature and dehydrated using increasing concentrations (70%, 90% and 100%) of ethanol. Dehydrated brains were cleared in xylene and immersed in molten paraffin wax at 64 °C. Brains were then orientated coronally and embedded in moulds (VWR) of fresh molten wax and left to set at 4 °C. Post-setting, the blocks were removed, trimmed to remove excess wax and serially sectioned (6 μm) on to Superfrost slides (ThermoFisher) using a microtome. Slides were dried overnight, then heated at 60 °C for one hour in a convection oven. Slides were dewaxed with xylene (2 × 5 min) and 100% ethanol (2 × 5 min). Slides were washed under running tap water and endogenous peroxidase activity was blocked with 3% hydrogen peroxide (10 min, room temperature, Sigma). Heat-induced epitope retrieval was performed using preheated citric acid (pH 6.4) in a pressure cooker for five minutes prior to incubation with primary antibodies. Slides were incubated in blocking solution (2% (w/v) BSA in 1 × TBS, sodium azide, pH 7.6) for five minutes and incubated in primary antibody for 16 h at room temperature. Slides were washed in TBS and incubated for one hour at room temperature in biotinylated secondary antibody (Vector Labs, 1:500) diluted in blocking solution. Slides were then incubated with StreptABC-HRP (Vector Labs) for 30 min at room temperature and developed in DAB (Sigma) solution for 10 min with gentle agitation. Slides were washed under running water until clear and dehydrated with methylated spirits, cleared with xylene and mounted using DPX (Sigma).
For staining adult brains, transcardial perfusion of adult mice was performed by injection of 50 ml PBS, 12.5 mM EDTA and then 4% (w/v) PFA in PBS through the left ventricle. The brains were then excised and further fixed for 24 h in 4% PFA and then dehydrated and embedded in paraffin.
For BrdU incorporation in embryos, pregnant dams were injected intraperitoneally with BrdU in 0.9% NaCl (100 mg/kg, Sigma) at E16.5. Mice were culled two hours post-injection and embryonic heads were fixed in 4% (w/v) PFA for 24 h, dehydrated and embedded in paraffin. For adult BrdU incorporation, P20 mice were injected intraperitoneally with BrdU in 0.9% NaCl (100 mg/kg, Sigma) three times at two-hour intervals. 24 h later, the animals were anesthetized and fixed by transcardial perfusion using 4% PFA.
Primary antibodies for immunofluorescence were mouse anti-NeuN (1:2500, ab104224, Abcam), rabbit anti-DCX (1:5000, ab18723, Abcam), rabbit anti-TBR1 (1:1000, ab31940, Abcam), rabbit anti-Tbr2 (1:1000, ab23345, Abcam), rat anti-Ctip2 (1:500, ab18465, Abcam), mouse anti-Ki67 (#556003, BD Pharmingen). Primary antibodies for DAB staining were sheep anti-BrdU (1:1000, ab1893, Abcam), rabbit anti-phospho-Histone H3 Brightfield images were taken on a Zeiss Axioskop microscope and quantified using Image J. For each animal, staining in four consecutive 6 µm sections was quantified and averaged. Colour deconvolution was applied to the images, and the same brightness and contrast settings were used for each experiment. Intermodes thresholding was applied and expression per mm 2 calculated using the analyze particles tool for embryonic Ki67, Mash1 and BrdU. For Ki67 staining, the neocortex was divided into five bins in Image J using the Bin division plugin. For embryonic PH3 numbers of cells expressing these markers per mm 2 were counted manually. Raw numbers were then converted to percentage as a proportion of control.
Immunofluorescence imaging was performed using a Zeiss LSM710 confocal microscope with Zen 2012 LSM software. Imaging of controls and experimental samples was performed using identical confocal microscope settings.  Statistical significance for genotype effects for each session (genotype factor) and between the daily sessions (session factor) of the Morris water maze were calculated using two-way ANOVA. Comparisons between Unk cKO and control mice were made using a Students t-test, *p < 0.05, **p < 0.01 compared to control. (I) A model for mTORC1 signaling in the brain. www.nature.com/scientificreports/ Unkempt in situ hybridisation data is available at https:// mouse. brain-map. org/ exper iment/ show/ 69013 684. The Expression Energy was calculated as follows: Within a given area A (voxel or structure), expression energy = (sum of intensity of expressing pixels in A) / (sum of all pixels in A).
Behavioural testing. For behavioural analysis, only male mice were used, and the experimenter was always blinded to the genotypes. The mice were singly housed one week prior to behavioural testing, and throughout the test period, to avoid any potential confounds from social hierarchies, which could influence the controlled assessment of social behaviours 43 . Sawdust was changed every other week but never on the day before, or the day of testing and the enrichment (nesting material and house) was changed less regularly to minimize the disruption to the animals. Testing was performed when the mice were 12-20 weeks old. Tests were recorded with a camera above the test arenas and the mice were tracked using Ethovision software (Noldus Information Technologies bv, Wageningen, The Netherlands). Urine and boli was removed, and the arena cleaned with 1% Anistel® solution (high level surface disinfectant, Trisel Solution Ltd, Cambridgeshire, UK) between each trial to remove odours. Mice were returned to their home cage after testing.
For the open field test 31 the arena consisted of a circular open field (40 cm diameter) enclosed by walls. The light intensity in the room was set to 25 lx by a light-adjustable floor lamp. The mouse was placed inside the arena next to the wall (always in the same starting location) and left to explore the arena for 10 min. Two virtual zones within the arena were defined on Ethovision; a 'central zone' (20 cm diameter) and the 'outer zone' (remainder of arena). The latency(s) to enter, and the time(s) spent in, the central zone of the arena and the mean velocity (cm/s) of the mice in the outer zone were extracted by Ethovision.
For the elevated plus maze test 30 the arena consisted of four arms; two opposing open arms with a 0.5 cm ledge around, and two opposing closed arms enclosed by a 15 cm high Perspex wall. All arms were 30 × 5 cm, and the whole maze elevated 40 cm above the ground. The light intensity in the open arms was 40 lx, and 20 lx in the closed arms. Each mouse was placed in the centre of the platform, and its movement tracked for five minutes undisturbed before being taken out. The duration(s) and frequency of entry in the open arms, was extracted using Ethovision.
The spontaneous spatial novelty preference test was conducted using a perspex Y-maze 32 . Each arm was 22 cm long, 7 cm wide, with 20 cm-high walls. One entry into an arm was defined as placement of two paws into that arm. For the first trial, one of the arms of the Y-maze was closed, therefore mice could either go left or right according to a pseudorandom sequence (equal numbers of left and right arms were blocked in total sessions), mice could also move in the central arm for five minutes. In the second trial one hour later to assess short term memory, the test was repeated with access to both arms. The time(s) spent in each arm was extracted from Ethovision. For trial 2, a percentage preference for the novel arm [100 x (time spent in the novel arm/(time spent in the novel arm + time spent in the familiar arm))] was calculated.
For the Morris water maze test 33 the pool arena was made of white acrylic, with a diameter of 1 m, and 30 cm deep. The pool was filled with a non-toxic, white aqueous emulsion (Acusol OP301 Opacifier, Rohm & Haas, Landskrona, Sweden). In the middle of one of the quadrants a platform with 10 cm diameter was located 1 cm below the surface. During testing, the room was lit with white light (100 lx) using four lamps pointing upwards at each corner. The pool was surrounded by cream-coloured curtains from which distinct spatial cues were suspended to help the mice navigate around the pool. Four equidistant positions around the pool walls were designated as Target (T), Opposite (O), Left (L) and Right (R), dividing the arena into four virtual quadrants. The mice were placed at these alternate locations for the successive trials, which were run in a pseudorandom manner. Each mouse therefore underwent four trials per day, and the latency to find the platform was calculated as an average of these four trials.
Mice were run in squads of six mice, with four trials each. A new trial started when all mice in the squad had finished. Between trials, mice were returned to their home cages and at the end of each session of four trials, mice were returned to the housing room. The trials were started by placing the mouse into the pool close to the wall in one of the four start locations (each trial was started in a different quadrant). Each mouse was given 60 s to swim to the platform and remained on the platform for 10 s before being removed. The latency to reach the platform was manually recorded from when the mouse was put into the pool to when all four paws of the mouse were on the platform. If the mouse did not reach the platform after 60 s it was guided to the platform and left there for 10 s. The same protocol was used over the subsequent six days of hidden sessions. On the final day of hidden platform training, a probe task was run. This was done by removing the platform and allowing the mice to swim in the pool for 60 s. Then, the platform was moved to the quadrant opposite the target quadrant for five days of reversal training. On the final day of reversal training, another 60 s probe task was run. To assess the retention of spatial memory, the average latencies to find the platform were recorded over sessions (days), and by comparing the time spent in the quadrant that contains the platform (target quadrant) with the time spent in other quadrants. The swim speeds (cm/s) and angular velocity were extracted by Ethovision during the probe trials.
Statistical analyses. Data were analysed using GraphPad Prism version 7 (GraphPad Software) or Statistica software (Version 5.5, StatSoft, Inc., Tulsa, OK). Data distributions were assessed for normality, then the effects of genotypes were analysed using a Student's t-test or two-way ANOVA, as appropriate. For two-way ANOVA, the between-factors were always genotype, and the within-factors were sessions.