FEZ1 (fasciculation and elongation protein zeta 1), a mammalian ortholog of Caenorhabditis elegans UNC-76, interacts with DISC1 (disrupted in schizophrenia 1), a schizophrenia susceptibility gene product, and polymorphisms of human FEZ1 have been associated with schizophrenia. We have now investigated the role of FEZ1 in brain development and the pathogenesis of schizophrenia by generating mice that lack Fez1. Immunofluorescence staining revealed FEZ1 to be located predominantly in γ-aminobutyric acid-containing interneurons. The Fez1−/− mice showed marked hyperactivity in a variety of behavioral tests as well as enhanced behavioral responses to the psychostimulants MK-801 and methamphetamine. In vivo microdialysis revealed that the methamphetamine-induced release of dopamine in the nucleus accumbens was exaggerated in the mutant mice, suggesting that enhanced mesolimbic dopaminergic transmission contributes to their hyperactivity phenotype. These observations implicate impairment of FEZ1 function in the pathogenesis of schizophrenia.
FEZ1 (fasciculation and elongation protein zeta 1) is a mammalian ortholog of the Caenorhabditis elegans UNC-76 protein (1). Worms harboring mutations in unc-76 manifest severe abnormalities in movement and in the elongation of axons along other axonal surfaces (but not in that along non-neuronal surfaces) during development, and these abnormalities are rescued by introduction of human FEZ1 (1). This movement of axons along other axonal surfaces, which is known as fasciculation, allows axons to associate in specific bundles and plays an important role in the determination of neural structures. In Drosophila, Unc-76 associates with kinesin heavy chain and its loss of function results in defects in locomotion and axonal transport reminiscent of those induced by kinesin mutation, suggesting that Unc-76 is required for axonal transport (2). In mammals, FEZ1 mRNA is present during all developmental stages of the mouse brain (3,4) and is abundant in the adult rat brain (5). FEZ1 is phosphorylated by protein kinase Cζ in response to nerve growth factor, and this phosphorylation of FEZ1 triggers axonal elongation. Coexpresssion of FEZ1 and a constitutively active mutant of protein kinase Cζ in PC12 cells enhanced neurite outgrowth in the absence of nerve growth factor (5). We previously showed that FEZ1 interacts with E4B (also known as UFD2a), a U box-type ubiquitin–protein isopeptide ligase, in the yeast two-hybrid system. Furthermore, the association between FEZ1 and E4B was enhanced by coexpression of a constitutively active form of protein kinase Cζ, and phosphorylation of FEZ1 by this kinase and its subsequent ubiquitylation by E4B resulted in neurite extension in PC12 cells (6). FEZ1 also contributes to the polarization of hippocampal neurons by controlling mitochondrial motility (7).
Two-hybrid analysis has shown that FEZ1 interacts with various proteins that function in transcriptional regulation, neuronal cell development, intracellular transport, apoptosis or neurotransmitter release (8). These proteins include necdin, which is inactivated in Prader–Willi syndrome (9), and disrupted in schizophrenia 1 (DISC1) (10). DISC1 is a susceptibility gene for schizophrenia that was originally identified in a large Scottish family with severe mental illness including schizophrenia (11,12). Subsequent genetic and clinical evidence has shown that DISC1 is potentially one of the most important risk genes for schizophrenia and other mental disorders (13). DISC1 forms complexes with a large number of proteins in addition to FEZ1 (10,14–17), suggesting that it may function as a hub in a multidimensional risk pathway for major mental illness and that it plays roles in neurodevelopment, cytoskeletal function, cAMP signaling (18) and adult neurogenesis (19). Two of the most studied proteins that interact with DISC1 are FEZ1 and nuclear distribution gene E homolog-like 1 (NDEL1), and expression of the genes for these two proteins has been found to be altered in postmortem brain specimens from individuals with schizophrenia (20). Studies in primates and rodents have shown that FEZ1 and DISC1 have overlapping spatial and temporal patterns of expression (4,21,22). Moreover, disruption of the DISC1–FEZ1 interaction inhibits the stimulation of neurite outgrowth in PC12 cells by DISC1 (10). Together, these observations are consistent with a possible role for FEZ1 in the pathogenesis of schizophrenia. Whereas association between single nucleotide polymorphisms of FEZ1 and schizophrenia was demonstrated in a Japanese cohort (23), two other studies of FEZ1 polymorphisms and schizophrenia failed to demonstrate an association (24,25).
To explore the physiological role of FEZ1 as well as its relation to schizophrenia, we have generated mice with a disrupted FEZ1 gene by homologous recombination in embryonic stem (ES) cells. Whereas no anatomic abnormality was evident in the mutant mice, several behavioral tests revealed a phenotype characterized by hyperactivity under novel, social or stressful conditions as well as enhanced responsiveness to psychostimulants. These defects were associated with an abnormal increase in dopamine release. Our observations suggest that loss of FEZ1 function may contribute to the pathogenesis of schizophrenia.
Expression of fez1 in mouse brain
In situ hybridization and northern blot analysis have shown that FEZ1 mRNA is abundant in the adult rat brain (5) as well as throughout brain development in mouse embryos (4). We examined the expression pattern of FEZ1 protein in mice. Immunoblot analysis revealed that FEZ1 is expressed exclusively in the central nervous system (Fig. 1A). The low-intensity bands apparent in other organs such as the stomach seem to be attributable to non-specific cross-reaction with the antibodies to FEZ1, given that these bands were also detected in FEZ1-deficient mice (Supplementary Material, Fig. S1). FEZ1 was found to be present in a variety of brain regions (Fig. 1B–D). It was detected from the prenatal period to adulthood; expression peaked at postnatal day 10 and gradually declined thereafter, but was maintained until at least 5 months of age (Supplementary Material, Fig. S1).
We examined the localization of FEZ1 in the adult brain by multilabel immunohistofluorescence analysis. FEZ1 immunoreactivity was detected in a subset of (not all) neurons in various brain regions including the hippocampus (Fig. 1E), neocortex (Fig. 1F), caudate-putamen (Fig. 1G) and nucleus accumbens (Fig. 1H). In all these regions, FEZ1-positive neurons were also labeled with antibodies to glutamic acid decarboxylase 67 (GAD67), the enzyme responsible for synthesis of the inhibitory neurotransmitter γ-aminobutyric acid (GABA), indicating that FEZ1 is expressed mainly in GABAergic inhibitory neurons. One notable exception to this pattern of distribution of FEZ1 was apparent in the dentate gyrus, where granule cells were distinctly labeled with antibodies to FEZ1 along their axon bundles that form the mossy fiber pathway (Fig. 1C and D). Although dentate granule cells are typical excitatory neurons with glutamate as their neurotransmitter, they also express GAD67 along the mossy fiber pathway (26), resulting in merged signals for colocalization of GAD67 and FEZ1 therein (Fig. 1C, D). Tyrosine hydroxylase-positive dopaminergic neurons in the midbrain did not manifest FEZ1 immunoreactivity (Fig. 1I).
Generation of fez1-deficient mice
To probe the physiological function of FEZ1, we generated mice deficient in this protein. Fez1 was disrupted in mouse ES cells by replacement of exon 2, which encodes the N-terminus of FEZ1, with a PGK-lox-neo-poly(A) cassette (Fig. 2A). We verified homologous recombination by Southern blot analysis (Fig. 2B) and polymerase chain reaction (PCR) analysis of genomic DNA (Fig. 2C) from the resulting mutant mice. FEZ1 mRNA was not detected in the homozygous mutant animals by reverse transcription (RT) and PCR analysis (Fig. 2D). Immunoblot (Fig. 2E) and immunohistofluorescence (Fig. 1C–H) analyses also revealed the absence of detectable FEZ1 protein in homozygous mutant mice. Although the abundance of FEZ1 in heterozygous mice was about half of that in wild-type animals (Fig. 2E), the heterozygotes appeared normal and were healthy and fertile. Mice homozygous for the Fez1 mutation were born from heterozygote intercrosses at approximately the frequency predicted by Mendel's law. The Fez1−/− mice were viable and fertile.
Absence of gross anatomic abnormalities in the brain of fez1-deficient mice
Given that FEZ1 has been implicated in neurite elongation (5,6), we performed histological analysis of the brain of Fez1−/− mice at 3 to 4 months of age using hematoxylin and eosin, Klüver-Barrera and Nissl staining. No gross abnormalities, such as disorganization of the cytoarchitecture, demyelination or misorientation of myelinated fibers, were found in the mutant mice (Fig. 3A). Immunohistofluorescence analysis further revealed the absence of obvious morphological changes in cell bodies, dendrites and axons of Fez1−/− mice (Fig. 3B and C).
Hyperlocomotion phenotype of fez1−/− mice under novel, social or stressful conditions
We next subjected Fez1−/− mice and their wild-type littermates to a comprehensive battery of behavioral tests in order to detect possible behavioral effects of FEZ1 deficiency. Fez1+/+ and Fez1−/− mice were generated by mating Fez1+/− heterozygotes that had been backcrossed to the C57BL/6J background for at least six generations. All tests were performed with male mice at 10 weeks of age at test onset. Fez1+/− heterozygotes showed no significant differences relative to Fez1+/+ animals in these tests (data not shown). Fez1−/− mice appeared healthy and showed no obvious differences in physical characteristics, with the exception of a small (but significant) decrease in body weight (Supplementary Material, Fig. S2), relative to wild-type controls. No differences in neuromuscular strength (grip strength and wire hang tests) (Supplementary Material, Fig. S2), sensitivity to a painful stimulus (hot plate test) (Supplementary Material, Fig. S2) or sensorimotor reflexes (eye blink, ear twitch, whisker touch and righting reflex tests) (data not shown) were apparent between mice of the two genotypes.
Examination of the locomotor activity of Fez1+/+ and Fez1−/− mice in several behavioral tasks consistently revealed a hyperlocomotion phenotype of the mutant animals. In the open-field test, Fez1−/− mice exhibited increased locomotor activity compared with wild-type littermates, as revealed by significant increases in distance traveled (Fig. 4A), vertical activity (Fig. 4B) and stereotypic activity (Fig. 4D). The hyperlocomotion phenotype of Fez1−/− mice was also apparent in specialized environments, as revealed by total distance traveled in the elevated plus-maze test (Fig. 5A), in the light–dark transition test (distance traveled in the dark chamber) (Fig. 5B), and in the social interaction test in a novel environment (Fig. 5C). We also measured locomotor activity under a stressful condition by the Porsolt forced swim test. Fez1−/− mice showed a significant decrease in immobility time compared with wild-type animals (Fig. 5D); this difference reflects a hyperactive tendency rather than reduced depression-related behavior. In contrast, measures of anxiety-related behavior, such as the time spent in the central part of the open field (Fig. 4C), in the light box of the light–dark transition test (Supplementary Material, Fig. S3) and in the open arms in the elevated plus-maze test (Supplementary Material, Fig. S3) as well as a measure of social behavior in the social interaction test in a novel environment (Supplementary Material, Fig. S3) did not differ significantly between Fez1+/+ and Fez1−/− mice. These observations indicated that FEZ1 deficiency did not affect emotional behavior.
Effects of fez1 deficiency on learning and memory
To assess whether the loss of FEZ1 was associated with cognitive abnormalities, we performed various memory and learning tests with Fez1−/− mice. In a contextual and cued fear conditioning test, Fez1−/− mice showed a significantly reduced frequency of freezing during conditioning compared with wild-type controls (Fig. 6A). One day after conditioning, a contextual test also revealed a significant decrease in freezing behavior for the mutant mice (Fig. 6B). It was possible, however, that the difference in performance in this test between mice of the two genotypes was attributable to the hyperactive phenotype of Fez1−/− mice. To examine the possible effect of FEZ1 deficiency on spatial working memory, we subjected mice to the Barnes maze test and eight-arm radial maze test. Neither of these tests revealed a significant difference in performance between genotypes (Supplementary Material, Fig. S4), suggesting that Fez1−/− mice are not cognitively impaired. Moreover, in the T-maze test, Fez1−/− mice showed a significantly higher percentage of correct choices than did their wild-type littermates (Fig. 6C). These data were thus also indicative of the lack of a cognitive abnormality in Fez1−/− mice, although they were also suggestive of enhanced food reward and reinforcement in the mutant animals.
Lack of sensorimotor gating defect in fez1−/− mice
Fez1−/− mice were tested for sensorimotor gating deficits in the prepulse inhibition test. The amplitudes of the startle response were similar in wild-type and Fez1−/− mice before testing (Supplementary Material, Fig. S5). Furthermore, the percentage prepulse inhibition did not differ significantly between the two genotypes.
Enhanced responsiveness of fez1−/− mice to psychostimulants
Given that hyperactivity is a prominent endophenotype in many psychiatric conditions, including schizophrenia (27), attention deficit-hyperactivity disorder, Tourette's syndrome and manic disorders (28), we assessed the locomotor-stimulatory effect of the psychostimulant MK-801, which is an inhibitor of the N-methyl-d-aspartate (NMDA)-sensitive glutamate receptor, in Fez1−/− mice. Acute intraperitoneal (i.p.) injection of MK-801 at a dose of 0.25 mg/kg of body weight revealed that the increases in total distance (Fig. 7A) and in its ratio between before and after injection (Fig. 7B) were significantly greater in Fez1−/− mice than in wild-type controls during the 30-min period after drug administration. However, the locomotor activity by MK-801 at a dose of 0.1 mg/kg was not different between genotypes (Fig. 7B).
We also examined whether FEZ1 deficiency affects behavioral changes induced by either bolus or repeated administration of methamphetamine. Acute i.p. injection of methamphetamine at a dose of 1 or 3 mg/kg revealed that the increase in locomotor activity was not significantly different between Fez1−/− mice and in wild-type controls (Fig. 7C and D). Although daily administration of methamphetamine at the same dose progressively increased locomotor activity in both the genotypes, Fez1−/− mice exhibited a greater increase in such activity than did wild-type animals (Fig. 7F). Daily administration of saline had no effect on locomotor activity in mice of either genotype (Fig. 7E). These data thus suggested that Fez1−/− mice show more responsiveness to the activity-stimulating effects of MK-801 and more locomotor sensitization to methamphetamine than their wild-type littermates.
Increased methamphetamine-induced release of dopamine in fez1−/− mice
Locomotor hyperactivity is typically associated with increased dopaminergic transmission in major motor areas of the brain such as the striatum (29). The caudate-putamen and nucleus accumbens play an important role in the control of locomotor activity (30). Given that Fez1−/− mice manifest both hyperactivity and enhanced responsiveness to the locomotor-stimulatory effects of methamphetamine and MK-801, we performed in vivo microdialysis to examine whether FEZ1 deficiency affects the extracellular level of dopamine in the nucleus accumbens after methamphetamine administration. Samples were collected from the nucleus accumbens of freely moving mice, and the extracellular concentration of dopamine was determined by high-performance liquid chromatography both before and after the injection of methamphetamine (1 mg/kg, i.p.). The basal levels of dopamine before drug administration were similar in Fez1+/+ and Fez1−/− mice, although those in Fez1−/− mice tended to be higher than those in Fez1+/+ mice (the difference was not statistically significant possibly because of the substantial level of variability in the data). Whereas methamphetamine administration increased the extracellular level of dopamine in mice of both genotypes, the drug-induced increase in dopamine concentration was significantly greater in Fez1−/− mice than in wild-type controls (Fig. 8). These data thus suggest that FEZ1 deficiency results in an increase in methamphetamine-induced dopaminergic neurotransmission in the nucleus accumbens.
FEZ1 has been implicated in processes that underlie neuronal morphology, including axon fasciculation, axonal transport and neurite extension, as well as in aspects of viral infection such as postentry block of retroviral infection (31) and association with the agnoprotein of the human polyomavirus JC virus (32). FEZ1 may also play a role both in control of transcription, as evidenced by its association with transcriptional factors (SAP30L, DRAP1, BAF60a) (8), as well as in mitochondrial motility (7). Given the predicted function of FEZ1 in the determination of neuronal morphology, our finding that mice deficient in this protein do not exhibit abnormal brain architecture was unexpected. The FEZ1-deficient mice did, however, manifest behavior abnormalities including a hyperlocomotion phenotype and enhanced responsiveness to psychostimulants. These abnormalities are likely attributable to changes in brain-regulatory systems, particularly in the altered control of dopamine release in the mesolimbic pathway.
Our immunohistofluorescence analysis revealed that FEZ1 is expressed in GABAergic inhibitory neurons of various brain regions as well as in granule cells of the dentate gyrus in adult mice. Dentate granule cells are generated from progenitor cells throughout life, and the newly developed cells join and elongate axons along the mossy fiber pathway in the adult hippocampus. Given that, among the various tract systems examined, including the corpus callosum, anterior commissure, internal capsule and fimbria hippocampi, FEZ1 was found to be expressed exclusively in the mossy fiber pathway, the localization of FEZ1 to granule cells might reflect a continuing function in axonal fasciculation and elongation in the adult brain.
The localization of FEZ1 to GABAergic neurons suggests an additional role for FEZ1 in the mature brain. Generally, brain function is executed by the controlled balance of excitatory and inhibitory signals within and between local circuits, with GABAergic neurons playing a prominent role in the regulation of such a balance. Changes in the activity of GABAergic neurons might thus underlie the behavioral hyperactivity and altered dopaminergic transmission of FEZ1-deficient mice.
The nucleus accumbens plays a key role in the control of motor activity (30) and is implicated in reinforcement and addiction associated with drugs of abuse. The increase in the extracellular concentration of dopamine in the striatum induced by psychostimulants enhances the hyperactive phenotype of various gene-targeted mice (33,34). The hyperactive phenotype of FEZ1-deficient mice and the enhancement of this phenotype by psychostimulants are thus consistent with the increased activity of the mesolimbic dopaminergic pathway revealed by the administration of methamphetamine. Basal dopamine levels also tended to be increased (approximately 60%) in the mutant mice compared with those in wild-type mice, although the difference was not statistically significant, likely because of the relatively high variability in the data.
Hyperactivity is a key endophenotype of many human psychiatric conditions including schizophrenia (27), attention deficit-hyperactivity disorder, Tourette's syndrome and manic disorders (28), and it is also a characteristic of rodent models of schizophrenia (27). The acute administration of methamphetamine or MK-801 in humans induces delusions and hallucinations reminiscent of those associated with schizophrenia. Furthermore, sensitivity to these drugs is increased in individuals with schizophrenia (35) as well as in several mouse models of this condition (29,36,37).
Hyperactivity of the mesolimbic dopaminergic system has been implicated in the pathophysiology of schizophrenia (38–41). We have now shown that FEZ1 deficiency results in increased dopaminergic transmission in the nucleus accumbens as well as in behavioral abnormalities. Although the mechanism by which FEZ1 deficiency enhances mesolimbic dopaminergic transmission remains to be elucidated, FEZ1 has been shown to interact in the yeast two-hybrid system with proteins that function in transcriptional regulation, neuronal cell development, intracellular transport, apoptosis or neurotransmitter release (8). Furthermore, in this regard, GABAergic transmission has also been implicated in schizophrenia (37,40,42–46), and primary abnormalities in neurotransmitter systems including cholinergic, glutamatergic and GABAergic pathways are thought to lead to secondary alterations in dopaminergic transmission associated with this disease (30,33,37,44). Mice deficient in the α3 subunit of the GABA type-A (GABAA) receptor also manifest hyperactivity and a schizophrenia-like sensorimotor deficit that are related to an increase in dopaminergic transmission caused by impaired GABAergic inhibitory regulation of dopaminergic neurons (44).
Together with previous observations suggestive of molecular and genetic links of FEZ1 to DISC1 and schizophrenia, our present findings indicate that the function of FEZ1 is related to the pathogenesis of schizophrenia. FEZ1-deficient mice recapitulate some symptoms of schizophrenia, but not others such as cognitive or memory impairments in the prepulse inhibition test and Barnes maze test. However, all individuals with schizophrenia do not necessarily manifest all possible symptoms of the disease. Given that schizophrenia is considered to be a heterogeneous condition, it is unlikely that all behavioral aspects will be reproduced by a single genetic mutation in mice. On the other hand, several mouse models for DISC1 mutation manifest a broad spectrum of schizophrenia-related phenotypes such as impairments in working memory, sensorimotor gating and social interaction as well as morphological abnormalities, such as enlargement of the lateral ventricles and alterations in neuronal architecture (47–50). We therefore propose a model in which the various DISC1-interacting proteins each contribute to the overall phenotype of schizophrenia. Furthermore, factors other than genetic traits, such as viral infection, developmental anomalies and social or other environmental factors, have been proposed to contribute to the pathogenesis of schizophrenia. Although interaction between genetic susceptibility and the environment is thought to play a prominent role in schizophrenia, the precise etiology of this disorder remains unclear. Our results suggest that characterization of the function of FEZ1 may provide insight into the pathophysiology of schizophrenia, and that Fez1−/− mice may thus prove to be a useful animal model for studies of human schizophrenia as well as for the development of new drugs for this disorder.
MATERIALS AND METHODS
Generation of fez1-deficient mice
The Fez1 locus was amplified by PCR from the genome of E14 (embryonic day 14) mouse ES cells with the use of LA-Taq polymerase (TaKaRa). The targeting vector was constructed by the replacement of a 1.4-kb fragment of genomic DNA containing exon 2 of Fez1 with a PGK-lox-neo-poly(A) cassette. The vector thus contained 1.6- and 6.5-kb regions of homology located at 5′ and 3′, respectively, relative to the neomycin resistance gene (neo). A PGK-tk-poly(A) cassette was ligated at the 3′ end of the targeting construct. The maintenance, transfection and selection of ES cells were performed as described previously (51). The recombination event was confirmed by Southern blot analysis with a 0.4-kb fragment of genomic DNA that flanked the 5′ homology region (Fig. 2A). The expected sizes of hybridizing fragments after digestion of genomic DNA with BglII were 6.2 and 9.5 kb for the wild-type and mutant Fez1 alleles, respectively. Mutant ES cells were microinjected into C57BL/6 blastocysts, and the resulting male chimeras were mated with female C57BL/6 mice. The germline transmission of the mutant allele was confirmed by Southern blot analysis. Heterozygous offsprings were intercrossed to produce homozygous mutant animals. For genotyping of mice, DNA was extracted from the tail and analyzed by PCR with the primers NS1 (5′-TTTCCAAGGGACATAGAGTAA-3′), NS2 (5′-AAAGCAGACATTGAGCTTCTC-3′) and PJL (5′-TGCTAAAGCGCATGCTCCAGACTG-3′). Procedures related to the generation of the knockout mice were approved by the animal ethics committee of Kyushu University.
Immunoblot analysis was performed with antibodies to Hsp70 (1 µg/ml) (clone7, Transduction Laboratories, Lexington, KY) and to FEZ1 (1 µg/ml). The rabbit polyclonal antibodies to human FEZ1 were generated by standard procedures using a recombinant full-length protein expressed in and purified from Sf21 cells as described previously (6). Mouse tissues were homogenized in homogenization buffer [50 mm Tris–HCl (pH 7.6), 150 mm NaCl, 0.5% Triton X-100, aprotinin (10 µg/ml), leupeptin (10 µg/ml), 10 mm iodoacetamide, 1 mm phenylmethylsulfonyl fluoride, 0.4 mm ethylenediamminetetraacetic acid, 10 mm NaF, 10 mm sodium pyrophosphate]. Proteins were separated by sodium dodecylsulfate–polyacrylamide gel electrophoresis and transferred to a Hybond P membrane (Amersham Biosciences), which was then exposed for 1 h at room temperature to Tris-buffered saline containing 5% dried skim milk before consecutive incubations for 1 h with primary antibodies and for 30 min with horseradish peroxidase-conjugated secondary antibodies. Immune complexes were detected using Signal West Pico chemiluminescence reagents (Pierce, Rockford, IL).
Male mice (12–15 weeks of age) were anesthetized with pentobarbital (100 mg/kg, i.p.) and perfused through the ascending aorta first briefly with phosphate-buffered saline (PBS) and then with 50 ml of 4% paraformaldehyde in 0.1 m phosphate buffer (pH 7.4). For immunostaining, the brain was excised and cut serially into 40 µm coronal or saggital sections using a vibrating microtome (VT1000, Leica, Heidelberg, Germany). For hematoxylin and eosin, Nissl or Klüver-Barrera staining, portions of the brain were embedded in paraffin and cut coronally into 5 µm serial sections. After cryoprotection with 30% (w/v) sucrose in 0.1 m phosphate buffer (pH 7.4), sections were rapidly subjected to freezing in liquid nitrogen and thawing before processing for multilabel immunostaining as described previously (52), but with slight modifications. In brief, sections were incubated for 7 days at room temperature with the following combinations of primary antibodies in PBS containing 1% bovine serum albumin and 0.3% Triton X-100: (i) rabbit polyclonal antibodies to FEZ1 (1:1000 dilution) (6), mouse monoclonal antibodies to GAD67 (1:5000) (Chemicon, Temecula, CA) and goat polyclonal antibodies to tyrosine hydroxylase (1:5000) (Chemicon); or (ii) rabbit polyclonal antibodies to parvalbumin (1:5000) (kindly provided by C.W. Heizmann) and mouse monoclonal antibodies to calbindin (1:10 000) (SWANT). After washing the sections thoroughly in PBS, they were incubated overnight with biotinylated donkey antibodies to goat immunoglobulin G (Invitrogen, Carlsbad, CA) and then overnight with a mixture of Alexa 488-conjugated donkey antibodies to mouse immunoglobulin G, rhodamine-conjugated donkey antibodies to rabbit immunoglobulin G (Invitrogen) and streptavidin-conjugated Cy5 (Invitrogen). The sections were mounted in Vectashield (Vector) and examined with a laser-scanning confocal microscope (C1 plus, Nikon). Sections from Fez1−/− mice were used as controls for FEZ1 immunostaining. Specimens from Fez1+/+ and Fez1−/− mice were thus prepared and processed at each step simultaneously and were examined under identical microscope settings. The images are presented as digital figures with minimal changes in brightness or contrast.
Fez1+/+ and Fez1−/− mice were generated by mating Fez1+/− heterozygotes that had been backcrossed to the C57BL/6J background for at least six generations. All behavioral tests were carried out with male mice that were 10-weeks-old at the start of the testing. Female mice were excluded because their behavior is influenced by the menstrual cycle. Mice were housed four per cage in a room with a 12 h-light and 12 h-dark cycle (lights on at 0700 h), and they had free access to food and water. Behavioral testing was performed between 0900 and 1800 h in a manner similar to that described previously (29,34,53). All behavioral testing procedures were approved by the Animal Care and Use Committee of Kyoto University Graduate School of Medicine. Detailed descriptions of the behavioral tests are available in Supplementary Methods.
In vivo microdialysis
Data are presented as means±SEM and were analyzed by the two-tailed Student's t-test, Scheffe's test, analysis of variance (ANOVA) or repeated-measures ANOVA using StatView software (SAS Institute, Cary, NC).
This work was supported by Takeda Science Foundation.
We thank Y. Yamada for generating FEZ1 knock-out mice; K. Nakanishi and H. Ougino for technical assistance with behavioral analysis; and A. Ohta for help in preparing the manuscript. This work was supported in part by Grants-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan; by a research grant from the Human Frontier Science Program; by Grants-in-Aid for Scientific Research from the Japan Society for the Promotion of Science; by the BIRD program of the Japan Science and Technology Agency; and by the National Institute of Biomedical Innovation of Japan.
Conflict of Interest statement. The authors declare no competing financial interests.