Crybb2 coding for βB2-crystallin affects sensorimotor gating and hippocampal function

βB2-crystallin (gene symbol: Crybb2/CRYBB2) was first described as a structural protein of the ocular lens. This gene, however, is also expressed in several regions of the mammalian brain, although its function in this organ remains entirely unknown. To unravel some aspects of its function in the brain, we combined behavioral, neuroanatomical, and physiological analyses in a novel Crybb2 mouse mutant, O377. Behavioral tests with male O377 mutants revealed altered sensorimotor gating, suggesting modified neuronal functions. Since these mouse mutants also displayed reduced hippocampal size, we concentrated further investigations on the hippocampus. Free intracellular Ca2+ levels were increased and apoptosis was enhanced in the hippocampus of O377 mutants. Moreover, the expression of the gene encoding calpain 3 (gene symbol Capn3) was elevated and the expression of genes coding for the NMDA receptor subunits was downregulated. Additionally, the number of parvalbumin-positive interneurons was decreased in the hippocampus but not in the cortex of the mutants. High-speed voltage-sensitive dye imaging demonstrated an increased translation of input-to-output neuronal activity in the dentate gyrus of this Crybb2 mutant. These results point to an important function of βB2-crystallin in the hippocampal network. They indicate pleiotropic effects of mutations in the Crybb2 gene, which previously had been considered to be specific to the ocular lens. Moreover, our results are the first to demonstrate that βB2-crystallin has a role in hippocampal function and behavioral phenotypes. This model can now be further explored by future experiments. Electronic supplementary material The online version of this article (doi:10.1007/s00335-013-9478-7) contains supplementary material, which is available to authorized users.

behavioral phenotypes. This model can now be further explored by future experiments.

Introduction
For more than 100 years, crystallins have been considered structural proteins of the ocular lens, accounting for up to 90 % of the entire protein content. However, there is increasing evidence that crystallins have pleiotropic effects in different organs: a-crystallins have been associated with neurodegenerative diseases like Alzheimer's disease, Parkinson's disease, or Alexander's disease (for a review see Graw 2009 and references therein). Moreover, it has been demonstrated that aA-crystallin is involved in maintaining dopaminergic neurons in the olfactory bulb (Ninkovic et al. 2010).
One of the most prominent members of the crystallins is the bB2-crystallin (gene symbol CRYBB2 in humans and Crybb2 in mice). The bB2-crystallin is characterized by four Greek key motifs, each encoded by its own exon. A Greek key motif consists of four adjacent antiparallel b-strands and their linking loops (Wistow 2012). The b-crystallins consist of an N-and C-terminal globular domain, each of which comprises two topologically equivalent Greek key motifs. In the native bB2-crystallin dimer, the N-terminal domain of each monomer interacts with the C-terminal domain of the other forming a bB2-crystallin homodimer (Trinkl et al. 1994). These Greek key motifs define the b/c-crystallin superfamily and are considered to be important for the folding of these proteins, allowing their dense packaging in the lens. Mutations in the CRYBB2/Crybb2 gene lead to cataracts, which are characterized by the loss of lens transparency. In humans, a frequent feature of mutations in the CRYBB2 gene is gene conversion to a closely linked pseudogene (wCRYBB2). In mice, three mutations have been described (Philly, Aey2, and O377), characterized by dominant, progressive cataracts (for review see Graw 2009, and references therein).
The most recently described mutant is the O377 allele; it is characterized as an A ? T substitution at the end of intron 5. The mutation forms an alternative splice site and consequently a 57-bp insertion in the mRNA, and it leads to 19 additional amino acids in the protein, just in front of the fourth Greek key motif. The predicted structure of the mutated bB2-crystallin shows an additional loop near the carboxyl terminus (Ganguly et al. 2008). Moreover, initial experiments in our laboratory with recombinant proteins demonstrated that the O377 bB2-crystallin is insoluble, making it impossible to apply common biochemical or biophysical methods to further characterize this protein (M. Sun, unpublished observation).
Besides being a structural protein in the ocular lens, there is evidence that bB2-crystallin has additional functions. Philly mutants suffer from reduced fertility (Duprey et al. 2007) and Crybb2 is upregulated in the adult regenerating retina and involved in axonal regeneration (Liedtke et al. 2007). bB2-crystallin can be phosphorylated in a cAMP-dependent manner (Kleiman et al. 1988) and is found to be phosphorylated in old lenses at eight different sites, most of which are Ser residues (Huang et al. 2011). Most interesting in the context of our experiments reported here, bB2-crystallin also has moderate ability to bind calcium, suggesting a role as calcium buffer (Jobby and Sharma 2007). This Ca 2? -binding activity of the b/ccrystallin superfamily was originally described in the protein S from Myxococcus xanthus and further validated in several members of the b/c-crystallin superfamily, including bB2-crystallin. Most of the b/c-crystallins have affinities in the micromolar range (4-250 lM) (Aravind et al. 2009).
Recently, we demonstrated Crybb2 expression in different brain regions, including the hippocampus, olfactory bulb, cerebellum, and cerebral cortex (Ganguly et al. 2008), but its role in the brain is still unknown. Based on these expression sites, we hypothesized that the corresponding mouse mutant, O377, may display central nervous system deficits. Here we describe for the first time behavioral alterations in a Crybb2 mouse mutant together with some molecular, cellular, and electrophysiological studies. We obtained converging evidence that the bB2-crystallin mutation interferes with the functional integrity of the hippocampus.

Mice
Mice were kept at the Helmholtz Center Munich, Neuherberg. All animal experiments were carried out in accordance with the regulations of the German Law on Animal Protection and institutional guidelines. The O377 mice were characterized previously on a C3H background; after 20 generations on a C3H background, homozygous mutants were crossed back to C57BL/6J for 8 generations and thereafter maintained as an intercross line (Ganguly et al. 2008). For the experiments described here, only male mice were used.

Behavioral analysis in mice
The open field test and prepulse inhibition (PPI) of the acoustic startle reflex (ASR) were assessed according to the standardized phenotyping screens developed by the Eumorphia partners (Mandillo et al. 2008), available as EM-PReSSslim protocols (see www.eumodic.org). The results were obtained by an experimenter blinded to the genotype conditions.
The open field apparatus consisted of a transparent and infrared light-permeable acrylic test arena with a smooth floor (internal measurements 45.5 9 45.5 9 39.5 cm). Illumination levels were set at *150 lux in the corners and 200 lux in the middle of the test arena. Mice were placed individually in a corner of the arena and allowed to freely explore it for 20 min. Data were recorded and analyzed using the ActiMot system (TSE, Bad Homburg, Germany).
The ASR/PPI protocol was adapted to the specifications of our startle equipment (Med Associates Inc., St. Albans, VT, USA). Background noise [no stimulus (NS)] was 65 dB and trial types for ASR included seven different stimulus intensities (NS,70,80,85,90,100,110,and 120 dB). Trial types for PPI included four different prepulse intensities (67, 69, 73, and 81 dB), with each prepulse preceding the startle pulse (110 dB) by a 50-ms interstimulus interval. Each stimulus type was assessed ten times, and trial types were arranged in blocks of ten in random order.
Social discrimination was assessed as previously described (Feil et al. 2009). The procedure consisted of two 4-min exposures of stimulus animals (ovariectomized 129Sv females) to the test animal in a fresh cage to which the test animal had been moved 2 h prior to testing. During the first exposure, one stimulus animal was exposed to the test animal. After a retention interval of 2 h, this stimulus animal was reexposed to the test animal together with an additional, previously not presented stimulus animal. The duration of investigatory behavior of the test animal toward the stimulus animals was recorded by a trained observer with a hand-held computer. A social recognition index was calculated as time spent investigating the unfamiliar stimulus mouse/time spent investigating both the familiar and unfamiliar stimulus mouse.
Spontaneous alternations were assessed using the Y-maze, which was made of opaque light gray PVC and had three identical arms (30 9 5 9 15 cm) placed at 120°from each other; illumination in the center of the maze was 100 lux (Wall et al. 2003). Each mouse was placed at the end of one arm and allowed to move freely through the maze during a 5-min session. Spontaneous alternations (defined as consecutive entries into all three arms without repetitions in overlapping triplet sets) were scored. The total number of arm entries was collected over the 5 min. Spontaneous alternation performance percentage is defined as the ratio of actual (total) alternations to possible alternations (total number of triplets) 9 100. When placed in the Y-maze, normal mice prefer to explore the least recently visited arm and thus they tend to alternate visits among the three arms. To explore the three arms successively, the mouse must maintain an ongoing record of the most recently visited arm and continuously update the record. Therefore, alternation behavior is a measure of spatial working memory.
The age of the mice was 9 weeks during the Open Field test, 11 weeks at the ASR/PPI measurement, 13-16 weeks during the social discrimination test, and 26 weeks during the Y-maze test. Data are reported as mean ± standard error of the mean (SEM) in bar or line graphs, or are individually shown in scatterplots with a horizontal line indicating the mean. Data were statistically analyzed by two-way analysis of variance (ANOVA) (Fig. 1g, h) or by unpaired Student's t test ( Fig. 1a-f). The chosen level of significance was p \ 0.05.

Morphological and histological analysis
Brains from adult mice (transcardially perfused with 4 % paraformaldehyde in PBS) were transferred to a 30 % sucrose solution for 48 h. Coronal brain sections 40 lm thick were made using a sliding microtome. Cresyl violet staining was used to analyze the differences in hippocampal morphology between mutants and wild-type mice. The volumes of dentate gyrus (DG) and hippocampus were measured under a microscope with Stereo Investigator software (MBF Bioscience, Williston, VT, USA). For neurogenesis assessment, 3-month-old O377 mutants and C57BL6 wild-type mice were injected with bromodeoxyuridine (BrdU; 50 mg/kg body weight i.p.) once a day for five continuous days. Then the animals were perfused 1 month after the first injection.
Hippocampal volume measurement 1-and 3-month-old wild-type and O377 male mice were sacrificed and 40-lm-thick free-floating sections were prepared as described previously (Feil et al. 2009). Every 12th section was selected and stained with Nissl staining. The volumes of the regions of interest (ROIs) [DG, cornu ammonis (CA), and hippocampus] were measured as previously described (Boldrini et al. 2009).

Manufacturing of monoclonal antibody against mouse bB2-crystallin
A new monoclonal antibody against bB2-crystallin was generated by cloning the full sequence of Crybb2 into the pETM11 vector. This construct was then transfected in Escherichia coli BL21/DE3, and after IPTG (isopropyl-1thiogalactopyranoside) induction, a soluble lysozyme extract was purified via Ni resin (GE Healthcare, Piscataway, NJ, USA). Shortly thereafter, the 6His fusion protein  Fig. 1 Behavioral analysis of male O377 mutant mice and wild-type littermate control mice (n = 5-10). a-c Spontaneous exploratory behavior in the open field. No differences in horizontal (a) or vertical (b) locomotor activity. d No difference in working memory as indicated by spontaneous alternation in the Y-maze. e, f Social investigation of the stimulus animal during the sample phase (first exposure) of the social discrimination test. f O377 mutants spent significantly more time in social investigation than control males. g There were significant differences in the acoustic startle curves of O377 mutants in comparison to littermate controls. h Increased PPI in O377 mutants. *p \ 0.05; **p \ 0.01; ***p \ 0.001; ?/? versus -/-was eluted from the column with elution buffer (50 mM Tris, 500 mM NaCl, 250 mM imidazol, and 1 mM TCEP, pH 7.4). Salts were removed using a Sephadex G-25 column (NAP-25 column, GE Healthcare). The antigen was suspended in PBS and then used for immunization. Fifty micrograms of the purified N-His-fusion protein (bB2crystallin) was injected intraperitoneally (i.p.) and subcutaneously (s.c.) into LOU/C rats using incomplete Freund's adjuvant supplemented with 5 nmol CpG 2006 (TIB MOLBIOL, Berlin, Germany). After 6 weeks a final boost with 50 lg of bB2-crystallin and CpG 2006 was given i.p. and s.c. 3 days before fusion. Fusion of the myeloma cell line P3X63-Ag8.653 with rat immune spleen cells was performed according to standard procedures. Hybridoma supernatants were tested in a solid-phase immunoassay with bB2-crystallin or an irrelevant N-His fusion protein coated to ELISA plates. Antibodies from tissue culture supernatant bound to bB2-crystallin were detected with HRP (horseradish peroxidase)-conjugated monoclonal antibodies against the rat IgG isotypes [TIB173 IgG2a, TIB174 IgG2b, and TIB170 IgG1, all from ATCC (Manassas, VA, USA), and R-2c IgG2c (Ascenion, Munich, Germany)], thus avoiding monoclonal antibodies of IgM class. Hypothalamic-pituitary-gonadal (HPG) activity was visualized with ready-to-use TMB (tetramethylbenzidine) (1-step TM Ultra TMB-ELISA, Thermo Fisher Scientific, Waltham, MA, USA). Monoclonal antibodies that reacted specifically with bB2-crystallin were further analyzed; in this study, we used the antibody BB2-4C4 (rat IgG2a).
In situ hybridization Mouse brain tissue was frozen in isopentane at -30°C and stored at -80°C. cRNA probes were generated from cloned inserts into the pCRII-TOPO vector (Invitrogen, Carlsbad, CA, USA). Cryosections were fixed in 4 % paraformaldehyde, and in situ hybridization was processed as described previously (Haupt et al. 2007).
Unbiased stereological cell counting Cells expressing parvalbumin, calretinin, somatostatin, and cleaved caspase-3 were counted using the Stereo Investigator 5.05.4 software on every sixth serial 40-lm coronal free-floating section. All the positive cells in the prefrontal cortex, ventral DG, and ventral CA were counted. At the same time, the volume of the counted regions was also measured using the Stereo Investigator 5.05.4 software. DAPI staining was used as counterstain and to calculate cell density (cells/mm 3 ) for each region.

Calcium concentration measurement
Hippocampal neurons were isolated from the brains of E17.5-E18.5 wild-type and O377 embryos. Hippocampi were dissected free of meninges, cut into small pieces, and digested with trypsin, and the cells were collected. The calcium measurement procedure was based on previous descriptions (Grynkiewicz et al. 1985;Díaz-Hernández et al. 2002). Hippocampal neurons were washed twice in HBSS ? (137 mM NaCl, 5.8 mM KCl, 10 mM HEPES, supplemented with 5 mM glucose, 2.5 mM CaCl 2 , 10 lM glycine, and 1 mM MgCl 2 ) by resuspension and subsequent centrifugation [1,000 rpm, 10 min at room temperature (RT)]. After cell counting and a viability check in a Neubauer chamber using Trypan blue dye exclusion (0.2 %), the cell concentration was adjusted to 1 9 10 6 cells/ml. Then the cells were incubated with 2.5 lM of the calcium dye Fura-PE3/AM for 30 min at 37°C in a humidified 5 % and CO 2 95 %. Afterward, the extracellular dye was removed by HBSS? rinsing and centrifugation (1,000 rpm, 10 min at RT). Before measurements, dye desertification was allowed for a further 30 min. For measurements, 2 ml of the hippocampal neuron suspension at a concentration of 0.5 9 10 6 cells/ml was transferred to a cuvette equipped with a magnetic stirrer and a thermostatted waterjacket kept at 37°C.
Fluorescence ratios (F 380nm /F 340nm ) proportional to Ca 2? concentration were monitored in a LS50B spectrofluorimeter (PerkinElmer, Waltham, MA, USA). NMDA (50 lM) was used to open the NMDA receptor to influx calcium ions from the environment to the cell cytoplasm after a stable Ca 2? level was obtained. Then, 10 lM thapsigargin was used to allow an influx of calcium into the cytosol from the endoplasmic reticulum. The calibration was performed by adding 10 lM ionomycin (R max ) and 5 mM EGTA (R min ). Ca 2? concentrations were calculated from the raw data using the Grynkiewicz formula (Grynkiewicz et al. 1985

Voltage-sensitive dye imaging (VSDI)
Preparation and staining of brain slices, VSDI, and data analysis were performed as previously described (von Wolff et al. 2011) with the following modifications. We used a custom-made monopolar tungsten electrode (Tefloninsulated to the tip of 50 lm in diameter) for electrical stimulation of the perforant path. This electrode allowed for very precise placement into the neuronal tissue and did not interfere with VSDI (e.g., by producing irritating shadows as is seen with other electrodes). A highly localized electrical stimulation was achieved by positioning the indifferent electrode far away from the slice in the recording chamber. The electrode was placed on the visually identified perforant path near its entry zone to the DG (Fig. 6a). ROI1 and ROI2 were created by the polygon-drawing function of the MiCAM02 software.

Behavioral studies in mice
Because of the observed Crybb2 expression in different brain regions, we first investigated the behavior of homozygous male O377 mutants. Analysis of spontaneous activity in a novel environment, as measured by the Open Field test, did not reveal any genotype effects in forward (total distance) or vertical (rearing) exploratory activity in O377 mice (Fig. 1a, b), nor in anxiety-related behavior as measured by the time spent in the aversive center of the open field (Fig. 1c). Assessment of working memory by analysis of spontaneous alternations in the Y-maze also did not reveal any genotype effect in O377 mice (Fig. 1d). When tested for social investigation, male O377 mutants did not reveal any deficit in social recognition (Fig. 1e) but spent significantly more time in social investigation than control littermates ( Fig. 1f; Student's t test, t (13) = 7.601, p \ 0.0001). A simple test for olfactory function (on C3H background), in which soiled bedding from a cage of unfamiliar male mice is presented in a tube together with another tube containing clean, fresh bedding, indicated that homozygous male O377 mutants can smell a social odor as well as the control mice (data not shown). This result, with the lack of a recognition deficit, suggests that the observed increase in social investigation is unlikely to be due to an olfactory deficit. Moreover, we performed social discrimination experiments to assess olfaction-based social memory and tested spontaneous alternation in the Y-maze to assess working memory and could not detect any deficits (data not shown).
The most prominent behavioral phenotype was observed in the PPI of the ASR. Assessment of the ASR revealed a significant reduction of the startle response of the O377 mice compared to the controls at the highest sound intensities, 110 and 120 dB [ Fig. 1g, two-way ANOVA for the factors genotype and startle stimulus intensity (dB), interaction genotype 9 dB: F (1,7) = 5.703, p \ 0.0001; post hoc test for 70-100 dB was not significant, post hoc tests for 110 and 120 dB, p \ 0.001]. Moreover, PPI measurements revealed a clearly increased PPI in comparison to littermate controls in O377 mutants [ Fig. 1h; two-way ANOVA for the factors genotype and prepulse stimulus intensity (dB), genotype effect: F (1,52) = 13.43, p \ 0.01].
Hippocampal volume was reduced in O377 mouse mutants During routine preparation of brain sections, we repeatedly observed smaller hippocampi in O377 compared to wildtype mice (for an overview, see the serial section in Supplementary Fig. 1). To find out whether the smaller size might be associated with Crybb2 expression, we checked the gene expression using RT-PCR and found Crybb2 expressed from at least the embryonic stages to adult (we checked from E14 to 3 months after birth; Supplementary Fig. 2). For a systematic comparison of the hippocampal volume, brain sections were stained with Cresyl violet and the volumes of the CA area, DG, and whole hippocampus were determined in the brains of wild-type and O377 mice at 1 and 3 months. The entire hippocampal volume was significantly decreased by 18.4 % and 19.2 % in 1-and 3-month-old O377 mutants compared to wild-type (for details see Supplementary  Fig. 3; Supplementary Table 1).
Reduced size of the hippocampus of bB2-crystallin mutant mice runs parallel with increased apoptosis A reason for a reduced size of the hippocampus might be an altered balance between neurogenesis and cell death. To determine whether neurogenesis was affected by impaired bB2-crystallin function, we first stained the brain sections with the proliferation marker Ki67. The cell density of Ki67 was not significantly different at the DG of 1-and 3-month-old wild-type and O377 mice ( Supplementary  Fig. 4). To confirm this result, we applied the thymidine analog bromodeoxyuridine (BrdU) to 3-month-old mice and quantified after a 4-week chase the number of labeled cells incorporated into the hippocampus. Again, there was no statistically significant difference between the hippocampus of wild-type and O377 mice ( Supplementary  Fig. 5), suggesting the normal generation and survival of new cells during adulthood. Next, to assess neuronal maturation, in the same paradigm we monitored the proportion of BrdU-positive cells expressing the postmitotic neuroblast marker Doublecortin (DCX) and the neuronal marker NeuN (Neuronal Nuclei). The density of BrdU;DCX;NeuN triple-labeled cells in the hippocampus was comparable in wild-type and O377 mutant mice ( Supplementary Fig. 6). We conclude that neurogenesis parameters are not affected by impaired functioning of bB2-crystallin.
Next, we investigated apoptosis using immunohistochemistry with an antibody against cleaved caspase-3 (Fig. 2). In 2-week-old mice, the density of apoptotic cells was found to be increased by 75.2 % in the ventral DG of O377 mutants in comparison to the wild-type mice; in the ventral CA of the O377 mutants, it was twice that of the wild-type animals. Interestingly, this difference was smaller at 1 month and disappeared at 3 months of age. Immunofluorescence for double labeling of parvalbumin and cleaved caspase-3 was also performed; however, no double-labeled cells were detected. Together, this suggested that the reduced size of the hippocampus (and concomitantly, also the number of parvalbumin-positive neurons) of bB2-crystallin mutants might be an indirect effect of a wave of cell death prominent between 2 and 4 weeks after birth.

Altered calcium homeostasis associated with impaired bB2-crystallin
Since it was suggested that bB2-crystallin has Ca 2? binding ability (Jobby and Sharma 2007), we investigated relative Ca 2? concentrations in hippocampal cells of mutant mice by calcium spectrometry. The calcium-sensitive dye was easily bleached in postnatal hippocampal neurons, rendering it difficult to obtain reliable data (J. Genius; unpublished observation). Since Crybb2 is also expressed at embryonic stages in the hippocampus ( Supplementary  Fig. 2), we chose to measure Ca 2? concentration in the late embryonic hippocampus (E17.5-E18.5), expecting similar effects of the mutated bB2-crystallin as at postnatal stages.
Basal Ca 2? concentration in O377 mutants was significantly increased (129.6 ± 13.5 nM) compared to the wildtype (81.0 ± 11.6 nM). After stimulation with NMDA, the increase in calcium concentration of O377 hippocampal   Fig. 2 Apoptotic cell death is increased in young O377 mutants. a-h Confocal images of anti-caspase-3 staining (red) in 2-week-old wild-type (WT, left panel) and O377 mice (right panel) in the ventral CA (upper panel, a-d) and in the DG (lower panel, e-h). DAPI staining is given in blue. The boxed area is given below in a higher magnification. In i, j, the values are normalized to wild-type mice at the age of 3 months and given in % (T = 100 %). i Density of caspase-3-positive cells in WT and O377 mutants in the ventral DG (WT = 100 %). j Density of caspase-3-positive cells in wild-type and O377 mutants in the ventral CA (WT = 100 %). *p B 0.05, Student's t test, n C 4. Scale bar 40 lm; error bar SEM neurons was also significantly higher (13.6 ± 2.5 nM) than in the wild-type (7.2 ± 1.2 nM). However, after stimulation of O377 hippocampal neurons by thapsigargin (depleting Ca 2? stores), the Ca 2? concentration was not significantly altered compared to the wild-type (Fig. 3). This indicates that the mutation in Crybb2 leads to an increase in basal and NMDA-receptor-mediated calcium concentration in O377 hippocampal neurons.
Previous work demonstrated the upregulation of calpain 3 (gene symbol Capn3) expression in the brain of adult O377 mutants (Ganguly et al. 2008). Calpains are calciumdependent proteases and highly associated with apoptosis (Raynaud and Marcilhac 2006). Given the results above that point to early changes in calcium homeostasis and neuronal death in bB2-crystallin mutants, we analyzed the expression of Capn3 in the brain of wild-type and O377 mutant mice at the age of 1 and 3 months. Capn3 was found to be widely expressed in the brains of both genotypes. We observed positive staining in Purkinje cells and granular cells of the cerebellum, in all layers of the cerebral cortex, in the CA1, CA2, CA3, and DG regions of the hippocampus, and in the glomerular layer, mitral layer, and granule cells of the olfactory bulb pointing to an overlapping expression pattern of Capn3 and Crybb2 (Fig. 4a-g). We also detected Capn3 expression in the hippocampus from E14.5 onward; it reached its maximum level at birth and remained constant at least up to 3 months (Fig. 4h). It might be important to mention that the Capn3 expression pattern in the hippocampus is almost identical to the Cry-bb2 expression pattern in the hippocampus, as described previously (Ganguly et al. 2008).
Since calpain expression has been discussed as a cause of the downregulation of NMDA receptor levels (for review see Wu and Lynch 2006), we tested this hypothesis by quantitative PCR of the corresponding genes. Indeed, we found that the expression of the NMDA receptor genes (Grin1, Grin2a, Grin2b, and Grin2c) was significantly decreased at P14 in the hippocampus of O377 mutant mice compared to wild-type controls (between 37.4 and 50 %, Fig. 4i-l).

Age-dependent loss of parvalbumin-positive neurons in O377 mutants
To better understand the role of bB2-crystallin in the hippocampus, we analyzed bB2-crystallin expression using immunohistochemistry in 1-and 3-month-old wild-type mice. For comparison, we also determined bB2-crystallin expression in the prefrontal cortex. In these structures, interneurons of various chemical subclasses are present, notably expressing markers such as parvalbumin, calretinin or somatostatin (Kubota et al. 2011). We detected bB2-crystallin in the vast majority of parvalbumin-positive cells in the prefrontal cortex (96.7 ± 1.5 % of parvalbumin-positive cells expressed bB2-crystallin), CA1 (94.9 ± 2.0 %), and the DG (97.8 ± 2.6 %). bB2-crystallin expression was also Fig. 3 Calcium concentration is increased in embryonic O377 hippocampal neurons. a Representative calcium recording in hippocampal neurons of wild-type and O377 mutant embryos at E17.5-E18.5. The fluorescence ratio (F 380 /F 340 ) is proportional to the Ca 2? concentration. b Basal calcium concentration and calcium concentration after stimulation by NMDA are significantly higher in O377 embryonic hippocampal neurons compared to the wild type; there is no statistically significant difference in the calcium release after thapsigargin treatment. For each measurement, six to eight embryos were used; the number of measurements was n = 10 for the wild types and n = 12 for the mutants. *p B 0.05, Student's t test. Error bar SEM detected in a subset of calretinin-positive cells in the prefrontal cortex (75 ± 2.6 %), CA (69.6 ± 5.9 %), and the DG (46.3 ± 5.2 %), and in most somatostatin-positive cells in the prefrontal cortex (90.5 ± 3.1 %), CA (92 ± 2.8 %), and the DG (92.6 ± 3.0 %) (Supplementary Figs. 7,8,9). We next asked whether distinct interneuronal subpopulations might be altered in the hippocampus of O377 mutants. Indeed, we found that the cell density of parvalbumin-positive neurons in the ventral CA and DG of 3-month old O377 mutants was significantly lower than that in wild-type mice (Fig. 5). This is likely due to a slow process, with a decreasing trend already visible in the first postnatal weeks leading to a progressively increasing difference between WT mice and mutants (Fig. 5h, i). In contrast, the parvalbumin population was not affected in the prefrontal cortex (Fig. 5g). Likewise, the density of calretinin-and somatostatin-positive neurons in the prefrontal cortex, ventral CA, and ventral DG was not significantly different between 3-month-old WT and O377 mutant mice (Supplementary Figs. 10, 11).
Increased translation of input-to-output neuronal activity in the DG of O377 mutants We investigated whether the phenotypes discussed above correlate with altered hippocampal functionality. For this purpose, we performed electrophysiological measurements in ventral hippocampal brain slices from 13 to 17-week-old male mice. In the first set of experiments, we tested for differences in the frequency of GABA A receptor-mediated mIPSCs in DG granule cells between O377 mutants and wild-type mice; however, these experiments revealed no statistically significant differences in the frequency (and amplitude) of mIPSCs ( Supplementary Fig. 12). Therefore, we considered that functional changes can be uncovered only at the level of neuronal activity of the whole network. To test this, we electrically triggered glutamatergic synaptic transmission at perforant path projections to DG neurons and conducted high-speed voltage-sensitive dye imaging (Airan et al. 2007;von Wolff et al. 2011) of the resultant neuronal activity. By means of a ROI operation, we quantified the neuronal depolarization (''activity'') strength in the region of perforant path synaptic input (outer third of stratum moleculare, ROI1) and the region where the cell bodies of granule cells are located (defined as output region, ROI2). This was done since we hypothesized that a decreased number of inhibitory interneurons should cause an increased translation of input-to-output neuronal activity in the DG of O377 mutants. As the final measure, we divided the neuronal depolarization strength within ROI2 by the neuronal depolarization strength within ROI1. Consistent with our hypothesis, we found this quotient to be markedly enhanced in O377 mutants. This mutation effect was independent of the strength of perforant path electrical stimulation (Fig. 6a-d).

Discussion
Although bB2-crystallin (Crybb2, CRYBB2) was first described as a structural protein of the ocular lens, further roles for this gene are expected given its expression in the brain. In our basic study (Ganguly et al. 2008), we observed increased expression of the gene encoding calpain 3 (Capn3). Therefore, we used this finding, together with an initial behavioral analysis, as a starting point for functional analysis of the effect of the underlying Crybb2 mutation on the brain. The behavioral analysis revealed two significant results: increased activity of the mutants in social investigation and a decreased ASR in combination with an increased PPI (Fig. 1). Since the increased social investigation was not associated with an alteration in olfaction, we did not follow this line any further. However, the increased PPI in the O377 mouse mutants is an intriguing coincidence with results of a previous report that mapped a PPI-QTL to mouse chromosome 5, roughly between 45 and 66 cM (Joober et al. 2002) where the Crybb2 gene is mapped.
The hippocampus is one of the areas of the brain where abundant Crybb2 expression has previously been observed (Ganguly et al. 2008). Moreover, by repeated inspection of brain sections, we realized that the hippocampus of the O377 mutants seemed to be smaller. Detailed measurements confirmed the impression and motivated us to focus our studies on the hippocampus. The smaller size of the hippocampus might be explained at least in part by apoptosis in the hippocampus of young mutants. Differences in apoptosis are not observed in adults, and other possible mechanisms (e.g., decreased neurogenesis or impaired neuronal maturation) are not supported by our data (Fig. 2;Supplementary Figs. 5,6). Increased apoptosis in the mutants might be due to the increased basal concentration of Ca 2? (Fig. 3).
Nevertheless, apoptosis does not explain the specific loss of parvalbumin-positive interneurons in the hippocampus of O377 mutants (Fig. 5); for this feature and its physiological consequences (increased input:output ratio in the DG), we have elaborated on another series of events; however, it may also start from the enhanced level of free Ca 2? in the mutants. The mechanisms by which the expression of Crybb2 translates into a neuronal subtypespecific phenotype might be explained better by the increased Capn3 expression (Fig. 4h) based upon the calcium-binding domains of the mouse Capn3 gene (Herasse et al. 1999). This increased Capn3 expression might lead to the decreased number of NMDA receptors ( Fig. 4i-l) and finally to the loss of parvalbumin-positive interneurons. Even if the data reported here can be organized in a putative series of events, we are not yet in a position to define each step involved in this process. It also includes the caveat that changes in the expression level might not lead immediately to changes at the protein level.
The first and central point in our model is the Ca 2?binding activity of bB2-crystallin, which had been demonstrated earlier (Jobby and Sharma 2007;Aravind et al. 2009). Unfortunately, we could not test the loss of Ca 2?binding activity of the mutated bB2-crystallin, because the corresponding recombinant protein remains insoluble in several buffer systems (M. Sun, unpublished observation). However, the observed increase of free Ca 2? (Fig. 3) is in agreement with the reported Ca 2? -binding properties of bB2-crystallin (Jobby and Sharma 2007). In an evolutionary context, Ca 2? binding by bB2-crystallin might be its original function, and, like other crystallins, bB2-crystallin might have been recruited to the lens to modify its properties (for a recent review see Wistow 2012).
Since Crybb2 is expressed during embryonic development in the brain (from E14.5 onward; Supplementary  Fig. 2), it is very likely that the first alterations can appear early in life. Therefore, the increased concentration of free Ca 2? , which has been observed in extracts from embryonic hippocampi of mutants, may have at least two consequences: increased apoptosis and increased Capn3 expression. Since the increased apoptosis in the hippocampus of the mutant mice during the first 3 months after birth is evident (Fig. 2), the means by which increased Ca 2? levels may lead to an increased expression of Capn3 [as observed previously by Ganguly et al. (2008) in a totalbrain mRNA pool and particularly in the hippocampus as demonstrated here in Fig. 3] is not immediately obvious. This link might be mediated by a protein being able to measure the actual concentration of free Ca 2? . Such a protein is well known as Ca 2? -sensing receptor (CaSR), a G protein-coupled receptor. In situ hybridization, histochemistry, and immunohistochemistry revealed CaSR mRNA and protein in pyramidal cells of all the layers of the hippocampus and in granule cells of the DG of the rat (Chattopadhyay et al. 1997). The highest levels are present within the subfornical organ and in the olfactory bulb. Substantial levels of expression are also evident within the striatum, cingulate cortex, cerebellum, ependymal zones of the cerebral ventricles, and perivascular nerves around cerebral arteries (Yano et al. 2004), but there are no reports dealing with its expression in the cortex. Finally, the involvement of CaSR in the transcription of neuronal genes has also been discussed (Riccardi and Kemp 2012).
The enhanced Capn3 gene expression, as reported previously using expression arrays (Ganguly et al. 2008) and validated here (Fig. 4h), may lead to a downregulation of NMDA receptor subtype gene expression (for review see Wu and Lynch 2006), as observed here 2 weeks after birth ( Fig. 4i-l; P14). Additionally, it is well established that increased calpain activity leads to a degradation of NMDA receptors (Bi et al. 1998;Araújo et al. 2005). In turn, hypofunction of NMDA receptors has been associated with a loss of GABAergic interneurons, especially parvalbuminpositive GABAergic interneurons (Lodge et al. 2009;Romón et al. 2011;Gandal et al. 2012;Gonzalez-Burgos and Lewis 2012). This particular feature-upregulation of Capn3 and downregulation of NMDA receptor gene expression-might explain the specific loss of parvalbumin-positive interneurons observed in the O377 mutants.
The functional cornerstone in our experiments reported here is the increased translation of input-to-output neuronal activity in the DG of the O377 mutants (Fig. 6). It indicates that the morphological changes observed have functional consequences for the neuronal network of the hippocampus and might explain the increase in PPI. PPI is regulated largely by neuronal connections between the limbic cortex (including the entorhinal cortex, hippocampus, and amygdala), ventral striatum, ventral pallidum, and pontine tegmentum. Glutamatergic fibers from the limbic cortex converge at the nucleus accumbens within the PPI regulatory circuitry. GABAergic projections, originating in the nucleus accumbens, then project to the globus pallidus. Regulation of PPI is carried out by GABAergic projections from the globus pallidus to the pedunculopontine nucleus. Increased PPI is associated with a reduction of GABAergic projections from the globus pallidus, while decreases are associated with the reverse (see Geyer et al. 2001 for review;Swerdlow et al. 2001). Thus, the increased translation of input-to-output neuronal activity in the DG of O377 mutants, as determined by the electrophysiological measurements, could lead to enhanced accumbal GAB-Aergic activation, resulting in increased inhibition of the GABAergic neurons of the globus pallidus, thereby disinhibiting the pedunculopontine nucleus, which ultimately increases PPI.
While there is substantial evidence for a functional impairment of GABAergic inhibitory interneurons in the hippocampus in schizophrenia, as reviewed in Powell et al. (2009), there is hardly any evidence in the literature directly linking the loss of parvalbumin-positive neurons in the hippocampus to an increase in PPI. It has been suggested that NMDA receptor hypofunction and alterations in the parvalbumin-positive interneurons are involved in schizophrenia in mice (Lodge et al. 2009) and in humans (Gonzalez-Burgos and Lewis 2012), but PPI was not measured in these studies. Forebrain-specific glutamate receptor B deletion in the mouse resulted in a decrease of parvalbumin-expressing interneurons in the DG and the pyramidal cells in CA3 and decreased neurogenesis in the subgranular zone, altered excitatory synaptic transmission in the hippocampus, and impaired spatial memory, but PPI was also not assessed in this study (Shimshek et al. 2006). In rats, isolation housing induced a reduction in PPI and action potential height in subicular pyramidal neurons as well as an increase in calretinin-positive neurons in DG, but hippocampal parvalbumin immunoreactivity remained unchanged (Greene et al. 2001). However, selective lesioning of septohippocampal GABAergic neurons in rats was associated with reduced parvalbumin immunoreactivity in the septal area and an increase in PPI (Ma et al. 2012).
While PPI deficits are believed to have face, construct, and a high predictive validity for schizophrenia in man (Swerdlow et al. 1994;Geyer et al. 2001), the biological relevance of increased PPI is less clear. One study suggested that in healthy humans, elevated PPI reflects enhanced preattentive perceptual processing and is associated with improved performance in upstream cognitive functions [measured by 5-choice reaction time tests, problem-solving tasks, and a test of spatial working memory (Bitsios et al. 2006)]. This was exemplified by enhanced strategy formation and execution times, possibly due to more efficient early information processing. As problem solving is determined by the integrity of the frontal lobe, increased PPI may signify altered prefrontal lobe function. In contrast, enhancement of PPI was also observed at a prodromal stage in a tauopathy mouse model of Alzheimer's disease, and the authors suggested using the assessment of sensorimotor gating to detect the earliest manifestations of tauopathies (Takeuchi et al. 2011). In clinically healthy humans, no clear adaptive or functional advantage of higher versus lower levels of PPI is known, and it was suggested to be simply used as a ''surrogate measure of neural processes'' (Swerdlow et al. 2008).

Conclusions
Our study demonstrated for the first time an association between sensorimotor gating and hippocampal function with a mutation in Crybb2. These potentially important findings need further validation and additional work from different disciplines to clarify the mechanisms involved. Nevertheless, these results suggest an interesting novel avenue for further exploration and offer a first functional model for the action of bB2-crystallin in the brain.
Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.