Implication of Genetic Deletion of Wdr13 in Mice: Mild Anxiety, Better Performance in Spatial Memory Task, with Upregulation of Multiple Synaptic Proteins

WDR13 expresses from the X chromosome and has a highly conserved coding sequence. There have been multiple associations of WDR13 with memory. However, its detailed function in context of brain and behavior remains unknown. We characterized the behavioral phenotype of 2 month old male mice lacking the homolog of WDR13 gene (Wdr13−/0). Taking cue from analysis of its expression in the brain, we chose hippocampus for molecular studies to delineate its function. Wdr13−/0 mice spent less time in the central area of the open field test (OFT) and with the novel object in novel object recognition test (NOR) as compared to the wild-type. However, these mice didn't show any significant changes in total time spent in arms or in frequency of arm entries in elevated plus maze (EPM). In the absence of Wdr13, there was a significant upregulation of synaptic proteins, viz., SYN1, RAB3A, CAMK2A etc. accompanied with increased spine density of hippocampal CA1 neurons and better spatial memory in mice as measured by increased time spent in the target quadrant of Morris water maze (MWM) during probe test. Parallel study from our lab has established c-JUN, ER α/β, and HDAC 1,3,7 as interacting partners of WDR13. WDR13 represses transcription from AP1 (c-JUN responsive) and Estrogen Receptor Element (ERE) promoters. We hypothesized that absence of Wdr13 would result in de-regulated expression of a number of genes including multiple synaptic genes leading to the observed phenotype. Knocking down Wdr13 in Neuro2a cell lines led to increased transcripts of Camk2a and Nrxn2 consistent with in-vivo results. Summarily, our data provides functional evidence for the role of Wdr13 in brain.


INTRODUCTION
WDR13 belongs to a class of WD (tryptophan-aspartate) repeat-containing proteins. Human WDR13 gene was discovered and characterized by Singh et al. (2003) and its highly conserved mouse homolog (Wdr13) by Suresh et al. (2005). It was independently discovered in mice by D'Agata et al. (2003) as Cmrg1 (Cerebellar memory related gene 1). Both human and mouse homologs of this gene localize on the X chromosome and encode a protein comprised of 485 amino acids. Western blot analysis shows two major isoforms of WDR13, one near the predicted molecular weight of 53 kDa and another smaller 43 kDa band corresponding to a truncated 394 amino acid protein (Singh et al., 2015a). WDR13 is a nuclear protein without any consensus nuclear localization signal (Suresh et al., 2005). To elucidate its function, a gene-knockout mouse was generated in our laboratory. The absence of this gene in mice (Wdr13 −/0 ) resulted in age-dependent mild obesity, pancreatic beta cell hyper-proliferation, subsequent hyper-insulinemia (Singh et al., 2012) and improvement of metabolic phenotype in Lepr (db/db) mice (Singh et al., 2015a).
Wdr13 gene is expressed in most tissues with relatively higher expression observed in the brain, pancreas, ovaries, and testes (Suresh et al., 2005). Research from various groups indicates the possible involvement of Wdr13 in brain function. D'Agata et al. (2003) implicated its function in learning and memory based on the association between expression of this gene to classical conditioning of rabbit nictitating membrane response. Wdr13 transcript has been shown to be enriched following synaptogenic lesion of the hippocampus in rats, suggesting its role as a possible neuro-protective molecule (Price et al., 2003). WDR13 had also been reported to be associated with the phenotype of hyperactivity, learning and visual-spatial difficulties of an 11-year-old boy having approximately 1.3 Mb duplication at locus Xp11.23p11.3 (El-Hattab et al., 2011).
These reports, however, were based on correlation and did not give any direct evidence of brain specific function of WDR13.
In the current work, we have delineated the role of this gene in brain by studying behavioral and molecular changes in Wdr13 knockout male mice (Wdr13 −/0 ).

Animal Experiments and Handling
Mice were procured from the central Animal House facility (CCMB). They were housed in polypropylene cages with shredded corn-cob bedding with 12-h light and dark cycle (6 a.m.-6 p.m. light cycle). The required numbers of mice were generated by crossing of wild-type and mutant (heterozygous, Wdr13 +/− ) mice. Unless mentioned specifically, all behavior and molecular data corresponds to male Wdr13 −/0 and wild-type mice in CD1 genetic background. CCMB Institutional Animal Ethics Committee approved all the animal experiments (Reg. No. CPCSEA 20/1999).

Analysis of Brain Metabolism
Nuclear Magnetic Resonance (NMR) was used to analyze changes in brain metabolites of Wdr13 −/0 and wild-type mice at 10 and 2 months of age. Metabolic measurements of cortex and subcortex were carried out on one group of mice containing wild type and Wdr13 knockout mice of age 10 months (n = 5, 6) by co-infusion of [U-13C6]-Glucose and [2-13C]-Acetate through the tail vein. 13C labeling of amino acids in brain tissue extract was analyzed using 1H-[13C] and 13C-[1H]-NMR spectroscopy. The protocol has been described by previous studies (Patel et al., 2001;Shameem and Patel, 2012). Two month old (n = 4) mice cortex was analyzed for metabolic changes using infusion of [1,6-13C2]Glucose for 10 min using the methodology described earlier (Shameem and Patel, 2012;Tiwari and Patel, 2012).

RNA In situ Hybridization (RISH)
Dig-labeled RNA probes for anti-sense and sense strands of full length Wdr13 cloned in pGMT vector were prepared according to instructions provided by ROCHE. RISH was carried out on cryo sections (30 µM) derived from formaldehyde perfused and fixed brain tissue from wild-type and Wdr13 −/0 mice as described previously (Singh et al., 2015b). Images were taken using AxioImager2 (Zeiss) with Apotome.
Behavioral Analysis of Wdr13 −/0 Mice Mice were tested for anxiety, depression, learning and memory. Video tracking software Noldus Ethovision 3.1 was used to record the movements of the mice in the tests. Litter-mates were utilized as control wild-type for different behavioral experiments. Wdr13 −/0 and wild type mice in outbred CD1 background was utilized for all behavioral studies. Few of the following studies were also performed in inbred C57BL/6J background to rule out possibilities of any strain-specific phenotypes.
(i) Open Field Test: This test was designed to study emotionality in rats (Hall, 1934) and later reproduced in mice (Christmas and Maxwell, 1970). This test has since been performed to assess the degree of anxiety and locomoter activity in mice. The Open Field test (OFT) arena is an open square box 50 × 50 cm. The box is virtually demarcated into a central zone and peripheral zones. The experimental mouse was placed in one corner of the box and allowed to explore the arena for 5 min. The amount of time the mouse spent in the center and periphery zones within 5 min was then noted. Mice with higher anxiety levels tend to spend more time in the periphery and less time in the central area. (ii) Elevated Plus Maze test: This test too was performed to assess the degree of anxiety in mice (Crawley, 2006). The setup consists of a four armed maze kept on an elevated platform. The maze consists of two open arms and two closed arms (walls on the side). The mouse was kept at the center of the maze facing the open arm and allowed to explore the arena for 5 min. The amount of time spent by the mouse in the open and closed arms was then calculated. The frequency of visit to each arm was also calculated. Mice with higher anxiety levels tend to spend more time in the closed arms and less in the open arms (thigmotaxis). (iii) Forced Swim Test: Also known as the Porsolt Swim Test (Porsolt et al., 1977), this test is conducted to analyse depression in rodents. In this test the mice were placed in a beaker filled with water till 20-25 cm and the total duration of immobility was measured. The experiment was videorecorded and analyzed post-recording. Usually a mouse with depression phenotype remains immobile for a longer time (Can et al., 2012).
(iv) Novel Object Recognition Test: This test had been developed to measure cognition, anxiety and preference for novelty in rodents (Antunes and Biala, 2012). In this experiment, the mice were placed in an open field on the first day. On the second day, two similar objects were introduced in the arena and the mice were habituated to them. On day three, a novel object of different shape and size replaced one of the objects. The discrimination by the mice between two objects was noted (measured as ratio of difference between time spent near novel object and familial object to that of total time spent exploring both objects). Normally, a wild type mouse prefers to explore novel objects. (v) Hot Plate test: This test is generally used to analyze proper functioning of neuromuscular junction and pain sensitivity (Minett et al., 2012). The mice were placed on a hot plate at 55 • C and the latency to the first jump was noted. (vi) Morris water maze (MWM): This experiment was developed for learning and memory analysis in rodents by Morris (1981). There are many variants of the method for training and analysis (Vorhees and Williams, 2006). Here the modified version of the Morris Water Maze (MWM) has been followed. The cues such as a tripod stand, a hanging bucket and specific arrangement of curtains were located outside a tub of 90 cm in radius and 40 cm in depth. The mouse was first placed on the submerged platform for 30 s on the first day and then taken out. In the second trial on day one, the mice were placed in water and then guided to the submerged platform. Usually each mouse were subjected to 2-3 trials in 1 day. The mice in subsequent trials were placed in water and time to reach the platform was noted. If it was unable to reach, it was guided at the end of the trial, which lasted 1 min 30 s. At the end of the trial, when the mouse reached the platform, it was left there for 15-30 s. Latency to reach the platform was noted and this was taken as an indication of learning capacity of the mice. After 5-6 days when the mice learnt to reach the platform placed anywhere in the tub, probe test was conducted 24 h post last learning trial. For re-learning experiment, the position of the platform was shifted to a new location after mice learnt to reach in its initial location. The learning ability of the mice to reach the platform in its new location was analyzed over multiple trials.
In the probe test, the platform was removed and the mice were placed in the tub. The time it spends in the quadrant in which the platform was kept was noted. Repeated (extinction) trials were performed 24 h after the final learning trial. In an extinction trial the mice were subjected to probe trials from random location multiple times. Normally the probability for a wild type to search for the platform decreases over trials (Maei et al., 2009;Terry, 2009). For long-term memory test, probe trial was not conducted immediately after the learning trial, but mice were subjected to a probe trial after a period of 20 days using the same setup and location of cues. Time spent by the mice in the target quadrant was noted.

Cohorts Used
Multiple cohorts were used for the afore-mentioned behavioral experiments (Table 1). Multiple experiments were executed in duplicates to validate the phenotype observed. Initially, a cohort of CD1 mice (n = 5) was utilized to perform OFT, Elevated Plus Maze (EPM), Novel Object Recognition test (NOR), and MWM. Between the non-stress experiments of OFT, EPM, NOR and that of MWM, there was an interval of 2 weeks. An independent cohort of CD1 mice (n = 17) was then subjected to OFT, EPM, and Forced Swim Test. A separate cohort of mice (n = 9, 10/8) was used to re-perform NOR and Hot Plate test. For MWM, we repeated learning trials and probe test in a separate cohort of mice (n = 6). Two other independent cohorts were utilized separately for re-learning experiment and long-term retention test in MWM (n = 6 each). We have also repeated FST with a separate cohort of mice (n = 6; data not shown). We used two separate cohorts for performing behavioral tests in C57Bl/6J mice. One cohort of mice (n = 16) was tested in OFT. Another (n = 8) was used to perform EPM and MWM.

Proteomic Analysis
Sample preparation, labeling, running of samples in LC-MS/MS and analysis were performed as described below. The mass spectrometry proteomics data along with the list of proteins quantified have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD002466.

iTRAQ 4 Plex
We pooled hippocampi from three wild-type and three knockout (Wdr13 −/0 ) mice for the present study. Proteins were extracted from tissue using 0.5% SDS and were estimated using Bicinchoninic acid assay (BCA) method. Two hundred micrograms of protein from both groups was taken as a starting amount. The protein from each group was treated with 2 µL of

LC-MS/MS Analysis
Samples were analyzed on UPLC (Dionex The UltiMate R 3000 HPLC) interfaced with Q-Exactive mass spectrometer (Thermo Scientific, Bremen, Germany). Trypsin digested peptides were loaded on a 15 cm long column (EASY-Spray column ES800, 15 cm × 75 µm ID, PepMap C18, 3 µm). Column was heated at 30 • C with integrated temperature control. Peptides were separated using linear gradient from 2 to 98% of buffer B (95% acetonitrile and 0.1% formic acid) at a flow rate of 300 nl/min, which was followed by a column re-equilibration reaching 2% of buffer B for few minutes. Gradient length had been adjusted to 50 min. The acquisition of the data was carried out using Xcalibur 2.1 (Thermo Scientific, Bremen, Germany). MS spectra were acquired in a data dependent manner in the range of m/z 350-1800 at a scan resolution of 70,000 and followed by top 10 precursor ions selected for MS/MS analysis at a scan resolution of 17,000. Normalized collision energy (NCE) was set to 27 for fragmentation. The priority of the precursor ion selection was based on the charge state in the order of 2+, 3+ and >3+. Unassigned and single charge state precursor ions were excluded from fragmentation. The dynamic exclusion was set as 30 s during data acquisition. The nano source was operated with 2.2 KV and the capillary temperature at 300 • C. Isolation width has been adjusted.

Data Analysis
The acquired data was analyzed using Proteome Discoverer 1.3 (Thermo Scientific, Bremen, Germany) software. We used International Protein Index (IPI) (version 3.83, Mouse) database to search for peptides. The workflow created included spectrum files, spectrum selector and Sequest. Search nodes were given as searches including peptide validator for false discovery analysis and used a reporter ion quantifier for quantitation. We set the parameters, which included Methylthio (C), iTRAQ label at N-terminus of the peptide and lysine as fixed modifications.
Oxidation of methionine (M) and deamidation(N/Q) were used as variable modification. The parent and fragment mass error tolerance were set as 20 ppm and 0.2 Da respectively. We acquired a total of 13,059 MS/MS scans. We calculated false discovery rate (FDR) by enabling the peptide sequence analysis using decoy database and top ranked hit based on peptide score, XCorr for Sequest. We applied 1% FDR in our analysis and proteins with a minimum of 1 unique peptide were considered.

Golgi Cox Staining
Golgi cox staining was done on 100 µM brain sections using a protocol described previously (Chakravarty et al., 2015). Six mice, each of wild-type and Wdr13 −/0 were selected for the analysis. A minimum of six CA1 neurons from hippocampal sections for each mouse was analyzed for spine density. Spines from the apical region of CA1 neurons were considered for the analysis. Spine density was analyzed in dendritic sections of 10 µm length. In total, about 90-100 spines were analyzed for each genotype. A semi-automated procedure for calculating spine density was utilized as described previously using FIJI and Image J (Orlowski and Bjarkam, 2012). In brief, highresolution images of spines were converted to 8 bit image using ImageJ and threshold applied. Areas to be analyzed were selected, pasted in a new document and were converted to binary. These were then skeletonized using ImageJ or FIJI and analysis of the skeleton was performed. For analyzing dendritic branching, 6-8 CA1 neurons from five each of wild-type and Wdr13 −/0 mice were traced using NeuronJ plugin and Sholl analysis was performed using Sholl Analysis Plugin (v 1.0) for ImageJ. Images were taken using AxioImager2(Zeiss) with Apotome.

BrDU Labeling and Counting
Mice were injected with 200 mg/kg (body weight) of BrDU, 24 h prior to processing and analysis. BrDU injections were given to mice either in resident conditions or at the final day of learning trials of MWM. Brains were collected while whole body perfusion. Serial coronal sections of 30 µm each were obtained from a single brain encompassing SVZ (Subventricular Zone) and SGZ (Sub Granular Zone) or DG (Dentate Gyrus) and every 6th section (excluding the early and late formations) were processed for immunohistochemistry using a protocol described previously (Becker et al., 2008). Similar sections from Wdr13 −/0 and wild-type were stained for BrDU incorporation[Anti-BrDU (Sigma, B8434) 1:300] by fluorescent labeling or DAB (Invitrogen) staining according to a previously described protocol (Singh et al., 2012(Singh et al., , 2015b or manufacturer's instructions and counting of one hemisphere was carried out. Area of each DG per section was analyzed using ImageJ and BrDU counts were normalized for the median areas of the sections (DG). A minimum of 5 mice and 6 s from each mouse has been used for the experiment.

Cell Culture and Transfections
Neuro2a cells were cultured in DMEM media with 10% Fetal Bovine Serum supplemented with antibiotics. Neuro2a cells grown in 24 well plates were assayed for effect of WDR13 on the expression of luciferase from AP1 and ERE site containing promoters using protocol previously described (Singh et al., 2015b). For analysis of luciferase from AP1 site containing promoter, Flag-WDR13 was co-transfected with c-JUN at (150:150) ng and (200:150) ng along with AP1-Luciferase and in Neuro2a cells in 24 well plate. For analysis of luciferase from ERE site containing promoter, Flag-WDR13 was transfected at 150 ng along with ERE-luciferase and in Neuro2a cells grown in 24 well plate, with or without supplementation of 10 nM Estradiol. Neuro2a cells used for assaying luciferase activity from ERE-promoter were maintained in DMEM media with 10% charcoal stripped serum. Transfection was carried out using Lipofectamine LTX (Life Technologies) as per manufacturer instructions. The results obtained from luciferase activity were normalized to β-galactosidase activity (indicative of transfection efficiency). Wdr13 siRNA (Sc-155258-Santacruz) and control scrambled RNA (Sc-37007) (Santacruz) were introduced in Neuro2a cell line using RNAiMax (Invitrogen) at 100 nM as per manufacturer's instructions and previous reports (Singh et al., 2012). All the experiments were conducted in triplicates.

RNA Isolation, Primer Designing, and Real Time Analysis
The primers were designed using the Primer3 software and PrimerBank. RNA was isolated using Trizol TM method and cDNAs were prepared using the protocol described by Sambrook and Russell (2001), and other previous studies (Singh et al., 2012(Singh et al., , 2015b. Briefly, hippocampi were homogenized in Trizol TM and 1/10th volume of 1-Bromo 3-Chloropropane (Sigma) added to the solution. Aqueous phase was collected after centrifugation and RNA precipitated using equal volumes of isopropanol. After ethanol wash, RNA was air-dried and dissolved in DEPC treated water. Real time PCR was performed using SYBR green 2X mix (Invitrogen and Thermo-Fischer). ABI Prism SDS 7000 and ABI 3900 HT were used to perform real time PCR as per company protocol. Primer sequences:

Western Analysis
Protein was isolated from hippocampi after homogenizing the tissues in SDS lysis buffer (Sambrook and Russell, 2001). Western Blotting was performed using manufacturers protocol or protocol previously described (Singh et al., 2012(Singh et al., , 2015b with antibody against WDR13 [(HPA000913, Sigma) 1:1000

Histological Analysis
We performed Nissl staining and Hematoxylin and Eosin (H&E) staining on brain cryo-sections (30 µm) of 4 month old wild-type and Wdr13 −/0 mice to understand any gross changes in anatomy. For H&E staining (Fischer et al., 2008), the sections were airdried and then immersed in Hematoxylin solution followed by Eosin solution with washes in water. The sections were then dehydrated using a gradient of alcohol concentrations, washed in xylene and mounted. For Nissl staining (Nissl, 1894), air-dried cryo-sections were de-fatted using xylene and absolute alcohol before rehydrating them and immersing in 0.1% Cresyl Violet solution (containing glacial acetic acid). The sections were then washed in water, dehydrated and mounted on slides.

Statistical Analysis
One or Two way analysis of variance (ANOVA) and student's unpaired T-test were conducted with significance at p < 0.05(denoted as * ) and at p < 0.005(denoted as * * ). For samples with n > 5, data are presented as mean ± SEM.

RESULTS
Our objective was to delineate the distribution of WDR13 in the brain. Both the isoforms of WDR13-53 and 43 kDa, showed differential expression in 2 months old mice with Cerebellum (Cr) showing comparatively higher levels of expression, followed by other regions of the brain, namely, Hipocampus (Hip), Cortex (Cx), and Olfactory bud (Ob) ( Figure 1A). Hypothalamus (Hypo) showed lower expression levels as compared to Cerebellum. Previous reports (D'Agata et al., 2003) and, those reported in Allen Institute of Science Mouse Brain Atlas (Lein et al., 2007) showed that Wdr13 transcript prominently localized to Cortex, Hippocampus, Cerebellum, and Olfactory Bulb. We performed RNA in-situ hybridization (RISH) using antisense probe against full length Wdr13 and found similar results ( Figure 1B, Supplementary Figure 1). Considering its expression and previous reports of its association with memory (D'Agata et al., 2003), we selected hippocampus for our subsequent analysis. We also compared whole brains from wild-type and Wdr13 −/0 mice to investigate any notable differences. Analysis of the brains did not reveal any significant differences in weight (Singh et al., 2012) or shape. We also did not find any significant differences (n = 3) in the gross anatomy of the brains (in both coronal and sagittal sections) from wild-type and Wdr13 −/0 mice (Supplementary Figure 2).

Two Months Old Wdr13 −/0 Mice Did Not Show Changes in Brain Metabolism
Wdr13 −/0 mice exhibited changes in general metabolism with progression of age (Singh et al., 2012), which became prominent at an age of 9 months. We performed NMR analysis of brain metabolites in 2 months (n = 4) and 10 months age (n = 5, 6) to find out if there were any changes in brain metabolism, and if these were correlated to changes in systemic metabolism in Wdr13 −/0 mice.
We found no significant changes in cortical brain metabolism in 2 months old Wdr13 −/0 mice (Figures 2A,B). However, there was a significant increase in 13C concentration of glutamate (Glu) from [U-13C6]Glucose (p < 0.05) but not in the levels of metabolites of cortex and subcortex of Wdr13 −/0 mice at 10 months of age (Figures 2C-F). There were also no significant changes (p > 0.05) in the 13C glutamate levels enriched from [2-13C]-Acetate (Figures 2G,H) at this age.
Glucose is utilized by both neurons and astroglia in the brain for their metabolism, whereas acetate is specifically taken up and assimilated by astrocytes (Waniewski and Martin, 1998). Labeling of Glutamate, GABA and Glutamine from [U-13C6]Glucose indicates initial estimate of the glutamatergic, GABAergic TCA cycle and total neurotransmitter cycle fluxes, respectively. Labeling of Glutamate, GABA and Glutamine from [2-13C]Acetate is representative of astroglial metabolism (Shameem and Patel, 2012). Thus, the increase in glutamate metabolism enriched from glucose but not acetate indicated it to be specific to neurons. Hence, the observed results might be primarily because of neuron specific activity of Wdr13. However, since brain metabolism is altered by changes in levels of circulating insulin (Bingham et al., 2002) and in obesity (Yau et al., 2012), studying Wdr13 −/0 mice at 10 months might not distinguish properly between the brain-specific function of Wdr13 and changes in systemic metabolism. Therefore, to understand the primary brain and behavior phenotype of Wdr13 −/0 mice, all behavioral, molecular and histological studies were carried out at 2-3 months of age.
Based on existing literature and the pattern of expression of Wdr13 in the brain, we analyzed Wdr13 −/0 mice for any altered anxiety or cognitive function (s).

Wdr13 −/0 Mice Showed Mild Anxiety
Wdr13 −/0 mice in both CD1 (Figures 3A,B; p < 0.05; n = 17; Cohort 2 and Supplementary Figure 3A; p = 0.07; n = 5; Cohort 1) and C57Bl/6J (Figures 5Ai,ii; p < 0.05; n = 16; Cohort 1 ′ ) genetic background spent less time in the central area and traversed more distance in the OFT indicating anxiety and hyper-activity. There were no significant differences in pain response (Figure 3C; p = 0.26) in hot plate test (n = 8; Cohort 3), indicating that neuro-muscular junctions and motor response were not affected. In the NOR, Wdr13 −/0 mice spent less time ( Figure 3D, p < 0.05 n = 10, 9; Cohort 3 and Supplementary Figure 3B; p = 0.06; n = 5; Cohort 1) in Discrimination Index (DI) exploring the novel object and more time with the familiar object. In other words, the Wdr13 −/0 mice actively discriminated against the novel object. Hence this behavioral phenotype may be associated with novelty driven anxiety, rather than the lack of ability to discriminate as the DI = 0. The mutant mice however, didn't show any significant changes in total time spent and frequency of visit to arms in EPM (Figures 3E,F  Wdr13 −/0 mice traveled significantly (T-test; p < 0.05) more distance in the open field arena than wild-type mice (n = 17; Cohort 2). (C) Hot plate test. There was no significant difference between Wdr13 −/0 and wild-type mice (T-test; p > 0.05) in latency to react to pain sensing when placed on a hot plate (55 • C) (n = 8; Cohort 3). (D) Novel object recognition test. There was a significant difference (T-test; p < 0.05) between Wdr13 −/0 mice and the wild-type mice in discrimination index (n = 10, 9; Cohort 3). The knockout mice spent less duration exploring the novel object as compared to the wild type. (E,F) Elevated plus maze test. Wdr13 −/0 mice did not show any significant differences (T-test; p > 0.05) than the wild-type mice in total time spent and frequency of visit to closed arms of the elevated plus maze (n = 17; Cohort 2) (G). Forced swim test. Wdr13 −/0 mice showed slightly less immobility as compared to the wild-type mice (n = 17; Cohort 2). This difference was however not statistically significant (T-test; p = 0.06). Data represented as ±SEM. Wt, wild-type; Wdr13 −/0 , Wdr13 knockout mice. *p < 0.05.
2 and Cohort 7) showed slightly lesser immobility (Figure 3G; p = 0.06) as compared to wild-type, though this was not statistically significant. All these data collectively indicated that the Wdr13 −/0 mice exhibited mild anxiety.

Wdr13 −/0 Mice Showed Better Performance in Hippocampal Dependent Spatial Memory Task Associated with Upregulated Synaptic Proteins
Hippocampus has been shown to be involved in spatial learning and spatial memory (Vorhees and Williams, 2006;Clark et al., 2007;Inostroza et al., 2011;Barnhart et al., 2015) and also long term retention of memory (Ramos, 2000) as assessed by MWM. Hence we employed MWM as a test to assess hippocampal dependent learning and memory behavior ( Figure 4A).
Wdr13 −/0 mice showed no significant difference (p > 0.05) in the time taken to reach the platform through the learning trials [ Figure 4B (CD1; n = 6; Cohort 4)]. Using another cohort of mice we performed re-learning experiment, when the platform was re-located after the mice were trained to reach the platform in a specific quadrant. In this experiment also Wdr13 −/0 mice (n = 6; Cohort 5) didn't differ significantly (p > 0.05) from the wild-type in the time taken to reach the platform in the new location ( Figure 4C). However, Wdr13 −/0 mice spent more time in the target quadrant during the probe (extinction) trials (E) Wdr13 −/0 mice showed better long-term memory by spending significantly (T-test; p < 0.005) more time in the target quadrant when subjected to probe trial after 20 days of learning phase (n = 6; Cohort 6). Data represented as ±SEM. Wt, wild-type; Wdr13 −/0 , Wdr13 knockout mice. *p < 0.05, **p < 0.005. Figure 4D; ANOVA, F (1, 40) = 14.24; p < 0.005] (CD1; n = 6; Cohort 4)}, which were performed on the same cohort of mice which underwent learning trials. Probe trial after learning trials was also conducted on a separate cohort of mice and similar results [Supplementary Figure 3D; ANOVA, F (1.56) = 5.44; n = 5; Cohort 1; p < 0.05] were obtained, validating the observed phenotype. Interestingly, even in a long term retention test performed 20 days after learning trials on an independent cohort of mice, the mutant (n = 6; Cohort 6) spent significantly more time (Figure 4E; p = 0.003) in the target quadrant. We found similar results using Wdr13 −/0 and wild-type mice in C57Bl/6J inbred background; mutant mice did not show any significant differences in the latency to reach the platform during the learning trials [ Figure 5Ci (C57Bl/6J; n = 8; Cohort 2 ′ )] but spent more time in the target quadrant { [Figure 5Cii; ANOVA, F (1, 36) = 10.24; p < 0.005] (C57Bl/6J; n = 8; Cohort 2 ′ )} during the extinction probe trials. Thus, MWM test indicated that the Wdr13 −/0 mice didn't show any differences in learning ability, but had improved retention of spatial memory in MWM than the wild-type and this phenotype was strain independent. Singh et al. (2012) had previously shown that in Wdr13 −/0 mice, there was an increase in pancreatic beta cell proliferation. We asked whether in the brain of the mutant mice, the proliferation of adult neuronal precursor cells were affected. Since, adult neurogenesis has been implicated in learning and memory (Deng et al., 2010), we also asked if the enhanced spatial memory phenotype of Wdr13 −/0 mice was associated with increased neuronal proliferation induced adult  (Supplementary Figures 4A,A ′ ,B,B ′ ; p > 0.05). Next, we wanted to know whether absence of Wdr13 had any effect on eural proliferation after the mice was subjected to learning tasks. Since learning tasks are shown to increase neurogenesis (Gould et al., 1999), we checked BrDU incorporation in proliferating neurons after MWM learning trials but again failed to note significant changes (Supplementary Figures 4C,D; p > 0.05). It may be noted that we also did not find significant difference of labeled BrDU cells between wild-type controls and those subjected to learning task. This is not a contradiction because Gould et al. (1999) also stated that the increased neurogenesis that they observed was mainly due to stability of neural progenitors labeled with BrDU before the start of the learning trials. The authors also failed to find any changes in number of cells labeled during or after the trials, which was similar to our findings. This indicated that adult neurogenesis might not have any significant role behind the observed phenotype.

{[
Dendritic spines have been associated with synaptic plasticity (Segal, 2005) and spine density and structures have been implicated with long and short term memory (Moser et al., 1994;Leuner et al., 2003;Restivo et al., 2009). Therefore, we analyzed spine density of apical CA1 hippocampal neurons from wildtype and mutant mice. We found that it was significantly higher (14%; p < 0.05) in the Wdr13 −/0 mice as compared to the wild-type (Figures 6A-C). We also analyzed dendritic branching of hippocampal CA1 neurons which revealed no significant differences ( Figure 6D; ANOVA, p > 0.05) between the wild-type and Wdr13 −/0 mice.
To understand the molecular mechanism(s) behind the observed phenotype, iTRAQ based quantitative proteomics was done from hippocampus of Wdr13 −/0 and wild-type mice. Proteomic analysis ( Supplementary Sheet 1; Table 2) revealed that 78 proteins were upregulated at greater than 1.5 folds out of a total of 170 quantified (1% FDR) proteins in the Wdr13 −/0 mice. These upregulated proteins included synaptic proteins, namely, SYN1, RAB3A, CAMK2A, and SV2B, proteins belonging to the 14-3-3 family, tubulins, dynamins, etc. In a separate proteomics experiment (data not shown), we found a decrease in protein levels of neurogranin (NRGN) in Wdr13 −/0 hippocampus.
Transcription analysis of Syn1 (synapsin1) and Rab3a ( Figure 7A) from hippocampus was consistent with the proteomics data (p < 0.05), implying that absence of Wdr13 gene resulted in upregulation of key synaptic genes in mice. Similarly, Nrgn transcription was downregulated (p = 0.03). There were, however, no significant changes in the transcript levels of Grin1, Grin2a (NMDA receptors) and Gria1, Gria2 (AMPA receptors) but changes (p < 0.05) were observed in transcript levels of Camk2a consistent with the proteomics data ( Figure 7B). Interestingly, we also found a significant upregulation in transcript levels of immediate early genes c-Fos and Arc, but not Bdnf (Supplementary Figure 5) in hippocampus of Wdr13 −/0 mice when subjected to a novel context (objects placed in an open field) as compared to the wild-type mice.
We wanted to understand whether the changes observed in the expression of multiple synaptic genes were due to the genetic deletion of Wdr13. We found that downregulation of Wdr13 in Neuro2a cells (Supplementary Figures 6A,C) resulted in upregulation of Camk2a and Nrxn2 (Figure 7C), indicating direct effect of absence/levels of WDR13 on the expression of these genes. Previous work carried out in our laboratory has established cJUN, ERα/β, PHIP, and HDACs as interacting partners of WDR13 (V. P. Singh and Shalu Singh, personal communication). We found that in Neuro2a cell line, WDR13 caused repression in transcription (p = 0.03) Cell signaling 9 (plasma membrane calcium ATPase 1, NEDD8, Isoform 2 of Serine/threonine-protein phosphatase 2B catalytic subunit alpha isoform, Isoform 2 of 14-3-3 protein theta, 14-3-3 protein zeta/delta, 14-3-3 protein epsilon, 14-3-3 protein gamma, Isoform 2 of Nck-associated protein 1, Gamma-enolase) 8 Cell adhesion and migration 1 (neurocan core protein-like) 9 Vesicular trafficking and fusion 2 (Clathrin heavy chain 1, AP-2 complex subunit alpha-2)

The classifications have been done manually based on Refseq and Uniprot annotations.
from promoters containing the Estrogen Receptor Element (ERE) (Figure 7D) in both presence and absence of estradiol. Luciferase assay also indicated that co-expression of WDR13 with c-JUN (Supplementary Figures 6A,B,D,E) resulted in repression of transcription from AP1 element containing promoter ( Figure 7E; one way ANOVA; p < 0.05). We therefore hypothesized that in the absence of Wdr13, any repression (competitive or non-competitive) over key genes caused due to its interaction with its partner(s) (Perissi et al., 2010), might be relieved, resulting in their upregulation leading to the observed phenotype.

DISCUSSION
The current work highlights the action of Wdr13 in the brain. We showed that WDR13 represses transcription from AP1 and ERE elements containing promoters, which harbor c-JUN and ERα/β responsive elements respectively. Absence of Wdr13 led to de-regulated expression of multiple genes. Many of these include synaptic genes like Syn1, Rab3a, Nrxn2, Camk2a, etc.
Interestingly, we showed that absence of Wdr13 caused mild anxiety and improved retention in MWM task, associated with increased spine density. We selected the age window of 2-3 months of age to carry out all behavior experiments. There were no systemic changes in the Wdr13 −/0 mice at this age (Singh et al., 2012) and we did not find any changes in brain metabolism. Though we found increased glutamate metabolism at 10 months of age in mutant mice, we argued that this change was mostly neuronal, considering that there were no changes in concentration of 13C glutamate labeled from acetate-indicative of astroglial metabolism. The changes observed at 10 months in Wdr13 −/0 mice might have resulted from accumulating molecular changes in the brain due to the absence of Wdr13. Since de-regulation of glutamate can lead to changes in memory (McEntee and Crook, 1993;Gecz, 2010), anxiety (Bergink et al., 2004), depression (Sanacora et al., 2012), etc., it would be interesting to study the behavioral phenotype of Wdr13 −/0 mice at 10 months of age. Increase in glutamate in a chronic state can also lead to excitotoxicity (Foran and Trotti, 2009). Therefore, analysis of brain function and anatomy is important to shed light on the effect of genetic deletion of Wdr13 in aged mice brain. A conditional knockout mouse with brain specific deletion of Wdr13 would be more suitable for this age dependent study since systemic metabolic effects because of the absence of this gene could be avoided.
Before analyzing differentially regulated genes, we asked if whether the proliferation of adult neuronal precursor cells was affected using BrDU labeling. Our results showed no significant difference in BrDU positive cells of DG and SVZ between Wdr13 −/0 and wild-type mice unlike pancreatic beta cells. This also indicated that the pathways affected in brain might be different from pancreas or other tissues.
We showed that the absence of Wdr13 resulted in upregulation of multiple synaptic proteins (16 out of 78 upregulated) in-vivo. We showed similar results in downregulation of Wdr13 in Neuro2a cell lines indicating that the changes occurred as a consequence of loss of the gene and not due to any systemic effects. The synaptic proteins found upregulated could be classified into AMPA cycling, Glutamate cycling, GABA metabolism, CAM kinase regulation, Synaptic vesicle proton gradient and synaptic vesicle cycling ( Table 1). We observed notable upregulation in Camk2a, Syn1, and Rab3a levels in Wdr13 −/0 mice at transcript and protein levels. It is known that synapsins control the release of neurotransmitters such as glutamate and important for synaptic plasticity (Nichols et al., 1992;Jovanovic et al., 2000). Synapsins are involved in memory formation and consolidation process in drosophila and in aging-related memory impairment in mammals (Godenschwege et al., 2004;John et al., 2009). Similarly, RAB3A is important for synaptic transmission, learning and memory (Yang et al., 2007). CAMK2A belongs to the CaM-kinase II family of proteins-known for their significant role in synaptic plasticity, long term potentiation (LTP), and learning and memory (Soderling, 1993). Levels of Vesicular glutamate transporter 1 (VGLUT1)-an important molecule for excitatory synapse and for maintaining LTP (Balschun et al., 2010) was also found to be increased in Wdr13 −/0 mice. Upregulation was also recorded in the levels of 14-3-3 proteins in Wdr13 −/0 mice. 14-3-3 proteins are positive regulators of associative learning and memory (Qiao et al., 2014). An upregulation in dynamins and tubulins found in the Wdr13 −/0 mice could be important factors behind the increased spine density observed (Gray et al., 2005;Shirao and González-Billault, 2013). Since spine density has been related to learning and memory, upregulation of afore-mentioned proteins might be of significance regarding the observed phenotype. We also found upregulation of immediate early genes Arc and c-Fos in Wdr13 −/0 mice when exposed to a novel environment as compared to wild-type mice. Arc has traditionally been associated with learning and memory, particularly long term spatial memory (Bramham et al., 2010). C-Fos has also been shown to be upregulated during learning trials (Alberini, 2009). It may be noted that no changes were found in the transcript levels of NMDA and AMPA receptor genes. However, considering that the levels of multiple synaptic proteins including important proteins like CAMK2A and immediate early genes like Arc, were found to be upregulated, an increased LTP might be expected to be associated with the phenotype of better spatial memory in Wdr13 −/0 mice. These findings may be corroborated with the help of further experiments particularly electrophysiology. Thus, taken together, increase in levels of above mentioned synaptic proteins and increased synaptic activity as measured by Arc and c-Fos, might be few of the key factors responsible for better performance in spatial memory task of MWM by Wdr13 −/0 mice.
Interestingly, many of the synaptic proteins along with Clathrin, AP-2, and Dynamin found up-regulated in Wdr13 −/0 mice are also known to be important for synaptic vesicle recycling. Synaptic vesicle recycling is essential for synaptic plasticity, memory, and cognitive ability, and any deleterious changes in it lead to mental retardation, schizophrenia, and defects in spatial memory (Murthy and De Camilli, 2003;Glyvuk et al., 2010;Cottrell et al., 2013). Synaptic recycling affects synaptic transmission (Casillas-Espinosa et al., 2012) and consequently memory through LTP (Hölscher, 1999). In this context it would be interesting to investigate synaptic recycling in these mutant mice.
Though Wdr13 −/0 mice showed better spatial memory, they didn't show any differences in spatial learning as assessed using MWM than that of the wild-type mice. Previous reports show that disruption of a gene can affect memory without significantly changing learning ability (Maguschak and Ressler, 2008;Tsai et al., 2012) though the molecular reasons are not clearly understood.
Wdr13 −/0 mice also showed decreased exploration of central area of OFT and active avoidance of novel object but failed to show any significant difference in EPM, particularly in CD1 background. It is possible that the response of the mutant mice was directed specifically against novel environment (Bailey and Crawley, 2009) and therefore differences were observed in OFT and NOR. Since the mutant mice also traversed significantly more distance, it explored EPM as much as the wild-type mice leading to no observable differences in CD1 background. However, it is to be noted that mutant mice in C57Bl/6J background showed a trend of increased visit to closed arm of EPM, and therefore, it is possible that this phenotype of mild novelty-associated anxiety was affected by the genetic background (CD1).
In our analysis, we found decreased levels of Neurogranin (Nrgn) at protein and transcript levels in Wdr13 −/0 mice. Neurogranin is a brain-specific calmodulin-binding protein expressed particularly in dendritic spines. Nrgn −/− mice exhibit characteristics of anxiety (Miyakawa et al., 2001). Similarly, it has been shown that overexpression of Camk2a leads to increased anxiety in mice (Hasegawa et al., 2009).
It is known through previous studies carried out in our lab that WDR13 interacts with c-JUN, Estrogen receptors ERα/β and HDACs (V. P. Singh and Shalu Singh, personal communication). Consistent with our previous findings, we show that WDR13 represses transcription from promoter containing ERE (ERα/β responsive) and AP1 (c-JUN responsive) elements. The action leading to the observed phenotype and deregulation of the multiple afore-mentioned genes in Wdr13 −/0 mice might have been attained by relieving of competitive repression (Perissi et al., 2010) induced by WDR13 on downstream targets of c-JUN and ERα/β-both shown to be positive regulators of learning and memory genes. Learning trials are known to increase cJun levels which affect learning positively by altering the expression of downstream genes (Alberini, 2009). While estrogen aids synaptic transmission and plasticity by positively regulating synaptic genes (Foy et al., 2010), activation of Estrogen receptors like ERβ has also been shown to improve memory (Liu et al., 2008). Further, estrogen induces synaptic protein CAMK2A activity invivo (Sawai et al., 2002). Additionally, administration of estradiol increases transcription of Syn1 (Pan et al., 2010). Syn1 has also been reported to contain an ERE element upstream (Bourdeau et al., 2004) indicating that the action of estrogen on its transcription is guided through Estrogen receptors. Interestingly, in our analysis we have found both Camk2a and Syn1 to be upregulated in Wdr13 −/0 mice.
Multiple genes that affect memory and learning have been reported in previous studies. Modulation of expression of a number of them had resulted in better learning and memory. These genes affect different processes like excitatory synaptic transmission (Nr2b, Cdk5, p25, Hgf etc.), inhibitory synaptic transmission (GABA receptor, Grpr, Pkr etc.), pre-synaptic function (H-ras, Ncx2, Cbl-b, etc.), transcriptional regulation (CREB, CamkIV, Gcn2, Calcineurin, etc.), miRNA biogenesis (Dicer1), epigenetic regulation (Hdac2) etc. (Lee, 2014). Our results show that absence of Wdr13 in 2 months old mice leads to better spatial memory associated with upregulation of multiple synaptic proteins. While it is exciting to find a molecule, manipulation of which may enhance memory, longterm implications of removal of this gene needs to be studied. As mentioned earlier increased glutamate can lead to excitotoxicity and neuronal death and hence aged mice lacking Wdr13 should be analyzed. Also, the persistent increase in synaptic proteins has been associated with schizophrenia (Li et al., 2015). Hence, it should be investigated whether absence of Wdr13 leads to hallucinogenic or schizophrenia like effect in mice with increasing age.

AUTHOR CONTRIBUTIONS
SM designed and executed the experiments and drafted the manuscript. GS executed with SM proteomics experiments. VT with SM executed NMR experiments. BL helped in animal breeding, maintenance, and genotyping. ST supervised the proteomics experiment providing inputs and helping in design, execute and in acquisition of the results for the experiment. SK gave inputs, aided in interpreting the data, revised the draft manuscript and finalized it.