We used quantitative RT-PCR to confirm the upregulation of KIAA0153 in a larger cohort of subjects with schizophrenia and matched controls, which subsumed cases previously analyzed in the original microarray study (Hakak et al., 2001). There were no differences in age or brain pH between the schizophrenic and control groups. The mean PMI was significantly longer for the schizophrenic group (t=−2.556, df=56, p=0.013). There were more males in the schizophrenic group and more females in the control group, a difference that was nearly significant (χ²=4.55, df=1, p=0.061). There was a significant negative correlation between age and KIAA0513 expression normalized to either ACTB or GAPDH (r≤−0.309, p≤0.019). To account for these differences, PMI, age, and sex were included in an ANCOVA as covariates.#

KIAA0513 was expressed at a significantly higher level in the DLPFC of the schizophrenic subjects, compared to the controls when normalized to either housekeeping gene (ACTB or GAPDH) (F≥6.96; df=1,53; p≤0.011) (Fig. 1). This difference was also significant when the covariates were not included in the analysis. The increase was approximately 1.2- to 1.5-fold, very close to the results of the microarray study. There was no difference in expression of ACTB or GAPDH, calculated as 2^−ΔCt (the amount of target, normalized to a calibrator) between the two diagnostic groups (F≤0.596; df=1,52; p≥0.667).#

Based on the high expression of KIAA0513 in the cerebellum and the evidence for disrupted prefrontal–cerebellar circuitry in schizophrenia, we hypothesized that altered expression might occur in this region as well. We examined expression of KIAA0513 in the cerebellum using a younger cohort of subjects with schizophrenia and matched controls. Data were normalized to the geometric mean of ACTB, GAPDH, B2M, and 18S rRNA to ensure the validity of the results. The means for the two groups were 0.883±0.413 for the schizophrenic subjects and 0.764±0.235 for the controls (mean±SD). Of the four internal control genes, GAPDH and 18S were the most stable based on an analysis using the program GeNorm (Vandesompele et al., 2002). The data were also analyzed using the geometric mean of these two genes. There was no difference in expression between the schizophrenic and control groups when all of the subjects were analyzed as a group (F≤0.330; df=1,52; p≥0.568). However, when the subjects were separated by race, there was a trend level increase in the Caucasian subjects (F=3.485; df=1,20; p=0.077) that was significant when analyzed using a t-test (t=-2.404, df=20.098, p=0.026). The means for the two groups in the Caucasian subjects were 0.931±0.450 for the schizophrenic subjects and 0.618±0.193 for the controls.#

There is very little functional information available for KIAA0513, even several years after it was cloned. Therefore, we carried out a preliminary functional analysis of this gene and its predicted protein product. We determined the normalized expression pattern of KIAA0513 in various tissues semi-quantitatively with a multiple tissue expression array (Fig. 2). KIAA0513 was expressed in every tissue examined but at higher levels in several brain regions than in peripheral tissues. The regions of highest expression included the cerebellum, cortex, hippocampus, pons, putamen, and amygdala. The expression levels in the thalamus and corpus callosum were lower than in other brain regions. Expression in the spinal cord was notably lower than in most other regions of the central nervous system.#

For brain regions of particular interest, we performed in situ hybridization on paraffin sections of human brain and frozen sections of rat brain (Fig. 3). The results were consistent with our findings from the multiple tissue expression array with high expression in the cerebellum, cortex, basal ganglia, and hippocampus. The mRNA was highly expressed in neurons but was also present in glial cells. KIAA0513 was also highly expressed throughout the rat brain, with particularly high expression in the cerebellum, cortex, and hippocampus. Minimal background staining was visible after hybridization with a probe to the sense strand of the cDNA, indicating that the signal was specific to KIAA0513 (Fig. 3K).#

A polyclonal antibody was raised against a GST-fusion protein antigen in order to examine KIAA0513 protein expression. We examined subcellular localization of the KIAA0513 protein by transfecting mouse L cells with an myc-tagged KIAA0513 construct. Immunofluorescence analysis with either an antibody to the myc tag or with the KIAA0513 antiserum revealed that the protein is probably cytoplasmic (Figs. 4A–D). When the cells were labeled with antibodies to KIAA0513 and the myc tag simultaneously, the staining patterns were nearly identical, indicating that the antiserum specifically recognized KIAA0513. Likewise, immunoblots of cells transfected with the same construct produced a single band of the same size with either the KIAA0513 antiserum or anti-myc antibody (Fig. 4E). It is not clear why there is a strong band in the nuclear fraction when KIAA0513 appears to be mainly cytoplasmic by immunocytochemistry.#

KIAA0513 has been regarded as a hypothetical protein because its endogenous expression had not been confirmed due to lack of an available antibody. When the antibody was tested on a western blot with rat hippocampal lysates, a band of approximately 50 kDa was observed (Fig. 4F). This size is very close to the predicted size of 47 kDa. A second band of higher molecular weight was also present, but both bands disappeared when the blot was probed with preimmune serum or with the antibody preadsorbed with an excess of the fusion protein antigen. The endogenous expression in primary hippocampal neurons also appeared to be predominantly cytoplasmic (Fig. 4G).#

As part of the functional characterization of the protein, we performed a yeast two-hybrid screen using the N-terminal portion of the KIAA0513 protein as the bait and a human fetal brain cDNA library. In order to consider a protein as a putative interactor, it had to activate adenine and histidine nutritional reporter genes as well as the α-galactosidase reporter. After isolating the plasmids from these putative interacting proteins, we retested the interactions in a small-scale high-stringency two-hybrid reaction. Proteins that also interacted with a lamin bait were eliminated because the interactions were presumed to be non-specific. Of the proteins that we identified from the initial screen, three appeared to be specific interacting proteins. These proteins were KIBRA/WWC1 (GenBank accession no. NM_015238), HAX-1 (GenBank accession no. NM_006118), and INTS4 (GenBank accession no. NM_033547). KIBRA is involved in synaptic transmission, HAX-1 has an anti-apoptotic function and interacts with components of the actin cytoskeleton as well as with immune-related proteins, and INTS4 is a subunit of the integrator complex that mediates processing of small nuclear RNAs (Baillat et al., 2005). Other potential interacting proteins that passed the specificity test have been identified as common false positives in two-hybrid screens and therefore were not examined further (Serebriiskii and Golemis, 2000). These proteins included a proteasome inhibitor subunit and cytochrome oxidase.#

KIAA0513 and KIBRA are partially colocalized to the perinuclear region in transfected CV1 cells (Figs. 5A–C). The interaction between KIBRA and KIAA0513 was confirmed by co-immunoprecipitation of myc-KIAA0513 with FLAG-KIBRA (Fig. 5D).#

We have confirmed that KIAA0513 is upregulated in DLPFC tissue from individuals with schizophrenia. However, KIAA0513 has not been identified as differentially regulated in subsequent microarray studies, possibly due to different methods and stringency of analysis. Although KIAA0513 is probably not a major susceptibility gene for schizophrenia, it may be part of a signaling pathway that is disrupted in individuals with the disorder.#

The cerebellum is involved in cognitive function independent of its role in the motor system (Schmahmann, 1991), and patients with cerebellar lesions exhibit certain cognitive and behavioral abnormalities characteristic of schizophrenia or prefrontal damage (Schmahmann and Sherman, 1998). Neuroimaging (Andreasen et al. 1996; Marcelis et al. 2003; Szeszko et al. 2003) and gene and protein expression studies (Eastwood et al., 2001, 2003; Guidotti et al., 2000; Kyosseva, 2004; Mukaetova-Ladinska et al., 2002) provide evidence that cerebellar abnormalities may occur in at least a subset of patients with schizophrenia. This information, together with the high expression of KIAA0513 in the cerebellum, prompted us to examine KIAA0513 expression in this brain region from subjects with schizophrenia. However, we did not detect a difference between the two diagnostic groups in the cerebellum.#

Two functional polymorphisms have been identified in the promoter of KIAA0513 that alter expression in a reporter gene assay, suggesting a possible relationship between genotype and gene expression levels (Buckland et al., 2004). However, the minor allele frequencies for these polymorphisms are very low, making it difficult to test whether there is a relationship between genotype and expression level in a small group of subjects.#

The expression profile of the KIAA0513 mRNA is ubiquitous among tissues but is more abundant in the brain than in the periphery. Overall, these results agreed with the microarray data in the Genomics Novartis Foundation's Gene Expression Atlas for KIAA0513, with expression in the cerebellum greater than 10 times the median expression level across all tissues (Su et al., 2002). In the gene expression atlas, KIAA0513 expression was higher than expected in specific populations of myeloid cells and monocytes. This observation suggests that KIAA0513 is important in both the nervous system and the immune system.#

Based on the in situ hybridization studies, KIAA0513 appears to be expressed in both neurons and glial cells. The lack of tissue or regional specificity does not provide any insight into the function of the gene, further complicating the task of characterizing the gene and protein.#

The results of the yeast two-hybrid screen further support our hypothesis that KIAA0513 is involved in both synaptic and apoptotic signaling. The key KIAA0513 interactors are also associated with the cytoskeleton through other binding proteins. By using three reporter genes in the AH109 yeast strain, we were able to eliminate several classes of false positives commonly encountered in two-hybrid screens. The adenine nutritional selection marker provides a strong selection, which is complemented by histidine selection and α-galactosidase expression. We also verified that these interactions did not occur with a non-specific bait to demonstrate the specificity of the interactions.#

KIBRA was initially identified as a WW domain protein that interacts with the postsynaptic protein dendrin (Kremerskothen et al., 2003) and acts as a novel substrate for PKC zeta (Buther et al., 2004). Its C2 calcium-binding domain resembles that of synaptotagmin, suggesting a role in synaptic signaling (Kremerskothen et al., 2003). More recently, KIBRA was shown to interact with dynein light chain 1 and histone H3, leading to activation of estrogen receptor alpha (Rayala et al., 2006). The interaction with dynein light chain 1 indicates that KIBRA may also be involved in transport along the cytoskeleton.#

A second interacting protein, HAX-1, has demonstrated anti-apoptotic function (Matsuda et al. 2003; Sharp et al. 2002) and is degraded during apoptosis (Cilenti et al., 2004). HAX-1 interacts with cortactin, linking it to the actin cytoskeleton (Gallagher et al., 2000). HAX-1 has been identified as an interacting protein in multiple two-hybrid screens, but has been confirmed as a true binding partner in each case. HAX-1 interacts with viral proteins and cytokines, including the Epstein–Barr virus nuclear antigen leader protein (Dufva et al. 2001; Kawaguchi et al. 2000; Matsuda et al. 2003), and mediates a complex between this protein and the anti-apoptotic protein Bcl-2 (Matsuda et al., 2003). HAX-1 also interacts with interleukin-1 (Yin et al., 2001) and other viral proteins (Sharp et al. 2002; Yedavalli et al. 2005), indicating a role in immune function.#

INTS4 is a subunit of the integrator complex that interacts with RNA polymerase II and mediates processing of small nuclear RNAs (Baillat et al., 2005). This protein contains an adaptin N-terminal domain similar to that found in proteins of the adaptor complex that forms clathrin-coated pits and vesicles (Marchler-Bauer et al., 2005), suggesting that it plays a role in endocytosis and vesicle recycling.#

The novel gene KIAA0513 is upregulated in a subset of subjects with schizophrenia. Although studying this gene as a susceptibility gene for schizophrenia is not warranted at the present time, it may prove to be a promising candidate when more functional information becomes available. Although we cannot conclusively determine the function of KIAA0513 at this time, the present study suggests that the protein may be involved in neuroplasticity, cytoskeletal regulation, and apoptosis. This study provides basic information about a gene that has been largely ignored by the scientific community due to difficulties in making functional predictions using standard bioinformatics approaches. The information provided by this study will be valuable to researchers who identify KIAA0513 as a gene of interest in microarrays or other screening experiments.#

Frozen postmortem brain samples from the dorsolateral prefrontal cortex of subjects diagnosed with schizophrenia (n=34) by DSM-IV criteria and normal elderly controls (n=22) were obtained from the Mount Sinai/Bronx Veterans Administration Medical Center Department of Psychiatry Brain Bank. Normal controls had no history of any psychiatric or neurological disorders and no discernible neuropathological lesions (Purohit et al., 1998). Snap frozen tissue from the DLPFC, defined by criteria established by Rajkowska and Goldman-Rakic (1995) was used in this study. All assessment and postmortem procedures were approved by the Institutional Review Boards of Pilgrim Psychiatric Center, Mount Sinai School of Medicine, and the Bronx VA Medical Center.#

Cerebellar tissue from subjects with schizophrenia and controls was from the Clinical Brain Disorders Branch at the National Institute of Mental Health. This analysis included 27 subjects with schizophrenia and 32 matched controls. Diagnoses were based on independent reviews of clinical records and family interviews by two psychiatrists using DSM-IV criteria. Subjects were free of any confounding neuropathological lesions, and toxicology reports were available for all subjects. Demographic characteristics of both cohorts are listed in Table 1.#

Total RNA was isolated by the guanidinium isothiocyanate method using the ToTally RNA kit (Ambion, Austin, TX). To remove genomic DNA contamination, the isolated RNA samples were treated with DNase I (Ambion) in the presence of RNAse inhibitor as described (Dracheva et al., 2001). The yield and the quality of the RNA preparations were analyzed by agarose gel electrophoresis and UV spectrophotometry. RNA yields were similar for schizophrenic and control subjects. First strand cDNA was generated with ThermoScript reverse transcriptase (Invitrogen, Carlsbad, CA) using approximately 2 μg of total RNA and random primers. Reverse transcription reactions were performed at 52 °C for 90 min followed by a 20-minute incubation in RNase H.#

TaqMan probes and primers were designed using PrimerExpress (Applied Biosystems, Foster City, CA) and purchased from Synthegen (Houston, TX) and Integrated DNA Technologies (Coralville, IA). All other reagents were from the TaqMan Core Reagents kit (Applied Biosystems, Foster City, CA). The reactions included the fluorogenic probe, PCR buffer at a 1× concentration, 3 mM MgCl2, 1 unit AmpliTaq Gold DNA polymerase, 200 μM each of dATP, dCTP, dGTP, 400 μM dUTP, 0.3 μM each of the forward and reverse primers, 0.25 units AmpErase UNG, and the reference dye ROX. The GAPDH, B2M, and 18S probes were predesigned assays (Applied Biosystems). Primer and probe sequences for KIAA0513 and ACTB are listed in Table 2. The cDNA obtained by reverse transcription reactions was diluted 1:50 for DLPFC and 1:10 for cerebellum (1:100 for 18S reactions), and 5 μL of the diluted cDNA was used in the PCR reaction. The reactions were performed with an AB7700 sequence detection system (Applied Biosystems) and the cycling conditions were a 2-minute hold at 50 °C, a 10-minute initial denaturation at 95 °C, and 40 cycles of 95 °C for 15 s and 60 °C for 1 min. Each reaction was performed in quadruplicate, and the average threshold cycle was used in statistical analyses. To control for variability between plates, eight wells of each 96-well plate contained a calibrator sample that consisted of a mixture of cDNA pooled from several cases. The relative expression was calculated as 2^−ΔΔCt (the amount of target, normalized to an endogenous control (GAPDH, B2M, or 18S rRNA) and relative to a calibrator) using a relative quantification procedure (Livak and Schmittgen, 2001).#

A multiple tissue expression array was used to determine the relative abundance of the transcript in polyA+ RNA from approximately 70 tissues and cell lines (Clontech, Palo Alto, CA). A cDNA probe was labeled with ³²P-α-dCTP using the Random Prime Labeling Kit (Roche, Indianapolis, IN) and purified on a Sephadex G-50 column (Amersham Biosciences, Piscataway, NJ). The hybridization was carried out overnight at 65 °C using ExpressHyb (Clontech) according to the manufacturer's protocol.#

Female Sprague–Dawley rats were deeply anesthetized with pentobarbital and perfused with 4% paraformaldehyde in PBS. The brains were cryoprotected in sucrose overnight and cut into 10-μm frozen sections. The Institutional Animal Care and Use Committee at Mount Sinai School of Medicine approved the use of vertebrate animals. Paraffin sections of human brain tissue from brain regions of particular interest (5 μm thick) were purchased from Biochain Institute (Hayward, CA).#

Digoxigenin-labeled RNA probes were prepared by in vitro transcription for both the sense and antisense strands using T7 and T3 RNA polymerases. A plasmid containing full-length KIAA0513 was linearized with BspHI. The probe concentration was determined by comparison of the labeled probe with serial dilutions of digoxigenin-labeled control RNA. Sections were fixed in 4% paraformaldehyde, permeabilized with proteinase K, acetylated using 0.25% acetic anhydride in triethanolamine, and dehydrated with a graded ethanol series prior to hybridization. Hybridizations were performed overnight at 53 °C in hybridization buffer (50% formamide, 2× SSC, 1× Denhardt's solution, 10% dextran sulfate, 0.5 mg/mL yeast RNA, 0.5 mg/mL salmon sperm DNA). After washing and blocking with BSA, the slides were incubated in anti-digoxigenin–alkaline phosphatase diluted 1:1000 in blocking buffer (Roche), followed by NBT/BCIP colorization solution, using the Dig Nucleic Acid Detection Kit (Roche). Paraffin sections were cleared with Histoclear (National Diagnostics, Atlanta, GA) prior to fixation and hybridization.#

A fragment of KIAA0513 encoding a 16 kDa portion of the protein from amino acids 17–170 was cloned into pGEX-KG to express a GST-fusion protein in BL21 cells. The resulting fusion protein was purified using the Bugbuster GST purification kit (EMD Biosciences, San Diego, CA). The fusion protein was dialyzed against phosphate buffered saline (PBS) and used as an antigen to produce polyclonal antiserum in New Zealand white rabbits. The antiserum was produced at Pocono Rabbit Farm & Laboratory (Canadensis, PA) according to their standard protocol for fusion protein antigens. The antiserum was purified on a protein A–agarose column to obtain the IgG fraction.#

Mouse L cells were grown in Dulbecco's Minimal Essential Media (DMEM) supplemented with 10% fetal bovine serum and penicillin/streptomycin in 5% carbon dioxide at 37 °C. Cells were transiently transfected with an myc-tagged KIAA0513 fusion protein constructed in pCMV-myc (Clontech) using Lipofectamine 2000 (Invitrogen). Cells were plated the day before transfection in DMEM with serum, and the media was changed to OptiMEM (Invitrogen) approximately 1 h before transfection. The cells were fixed or lysed approximately 24 h after transfection.#

For the colocalization studies with KIBRA, transfection of African green monkey kidney cells (CV1) was done as described previously (Kremerskothen et al., 2003). In brief, CV1 cells were co-transfected with constructs encoding EGFP-KIAA0513 and myc-tagged KIBRA using calcium phosphate. Twenty-four hours after the transfection, cells were fixed in 4% paraformaldehyde in PBS.#

Cultured cells were fixed in 4% paraformaldehyde for 15 min at room temperature. The cells were permeabilized in 0.1% Triton X-100 for 10 min at room temperature, blocked in 10% BSA in PBS for 1 h, and incubated with the KIAA0513 antiserum at 1:100 or anti-c-myc at 1:100 (EMD Biosciences) followed by anti-mouse–rhodamine at 1:100 (EMD Biosciences) or anti-rabbit–FITC at 1:150 (Sigma, St. Louis. MO). The cells were incubated briefly in DAPI (Molecular Probes, Eugene, OR) to visualize the nuclei. Images were obtained using an Axioskop fluorescent microscope (Zeiss, Thornwood, NY).#

For colocalization studies, myc-KIBRA was detected with primary anti-myc antibodies and secondary rhodamine-conjugated antibodies.#

Cell lysates were prepared using Cytobuster (EMD Biosciences). The protein was electrophoresed in an 8% SDS polyacrylamide gel and transferred to a nitrocellulose membrane. The crude antiserum was used at 1:100 for transfected cells, and the purified IgG fraction was used at 1:500 for tissue lysates. The secondary antibody was anti-rabbit–HRP or anti-mouse–HRP (Sigma) diluted 1:20,000. Chemiluminescent detection was performed with the SuperSignal West Pico substrate (Pierce, Rockford, IL).#

A yeast two-hybrid screen was performed using the Matchmaker Two-Hybrid 3 kit (Clontech). The N-terminal portion of the KIAA0513 protein was cloned in the pGBKT7 bait vector. A human fetal brain Matchmaker library was screened by cotransformation of the bait construct and the library plasmid into competent AH109 cells (Clontech). A total of 3 μg of bait plasmid and 1.5 μg of library plasmid were transformed in each 200 μL aliquot of competent yeast cells. The transformations were plated on synthetic dropout plates lacking leucine, tryptophan, histidine, and adenine. The histidine and adenine nutritional selection allowed for two independent reporters. The colonies were restreaked on the same media with x-α-gal, which served as a third reporter of a protein interaction. Because this procedure is so stringent that it may result in false negative results, the screen was repeated on media with adenine. The colonies that grew in these screens were restreaked on media lacking adenine with added x-α-gal. Plasmid DNA was isolated from the positive colonies using the ZymoPrep plasmid isolation kit (Zymo Research Corporation, Orange, CA). The isolated plasmids were transformed into competent E. coli, and plasmid DNA was isolated. Small-scale tests to confirm interactions were performed using 160 ng of bait plasmid and 80 ng of plasmid from the putative interacting protein in 50 μL of competent yeast. The procedure was also performed using a construct with lamin fused to the DNA binding domain as a control for specificity of the interaction. For proteins that showed a positive interaction with KIAA0513 and a negative interaction with lamin, purified plasmid DNA was sequenced to determine the identity of the putative interacting proteins.#

HEK293 cells were co-transfected with expression constructs encoding myc-tagged KIAA0513 and FLAG-tagged KIBRA protein, respectively. Cells were lysed in IP buffer 48 h after transfection (20 mM Tris, pH 7.5; 25 mM NaCl; 50 mM NaF; 15 mM Na4P2O7; 1 mM EDTA, 1% Triton X-100; Complete protease inhibitors (Roche, Mannheim, Germany)). Cell lysates were cleared by centrifugation (10,000×g, 30 min, 4 °C) and subsequently incubated with anti-FLAG antibodies immobilized on agarose beads (Sigma-Aldrich, Deisenhofen, Germany). After intensive washing with IP buffer, immunoprecipitates were eluted from beads and were subjected to western blotting analysis using anti-myc antibodies (Abcam, Cambridge, UK).#

Comparisons of the demographic characteristics of the cohorts between the schizophrenic and control groups were performed using an independent samples t-test. Differences in the sex ratios between the diagnostic groups were compared using a chi square test. Pearson correlation coefficients were calculated to determine whether age, pH, or PMI correlated with KIAA0513 expression. An ANCOVA was used to compare RNA expression between the two groups with age, PMI, and sex entered as covariates. For the cerebellar samples, pH was also included as a covariate in the analysis.#