Skip to main content
PMC home page
American Journal of Human Genetics logoLink to American Journal of Human Genetics
. 2005 Jul 6;77(2):280–296. doi: 10.1086/432518

An Autosomal Dominant Cerebellar Ataxia Linked to Chromosome 16q22.1 Is Associated with a Single-Nucleotide Substitution in the 5′ Untranslated Region of the Gene Encoding a Protein with Spectrin Repeat and Rho Guanine-Nucleotide Exchange-Factor Domains

Kinya Ishikawa 1,,*, Shuta Toru 1,,*, Taiji Tsunemi 1, Mingshun Li 1, Kazuhiro Kobayashi 8, Takanori Yokota 1, Takeshi Amino 1, Kiyoshi Owada 1, Hiroto Fujigasaki 1, Masaki Sakamoto 1, Hiroyuki Tomimitsu 1, Minoru Takashima 1, Jiro Kumagai 2, Yoshihiro Noguchi 3, Yoshiyuki Kawashima 3, Norio Ohkoshi 9, Gen Ishida 10, Manabu Gomyoda 11, Mari Yoshida 12, Yoshio Hashizume 12, Yuko Saito 5, Shigeo Murayama 5, Hiroshi Yamanouchi 6, Toshio Mizutani 7, Ikuko Kondo 13, Tatsushi Toda 8, Hidehiro Mizusawa 1,4
PMCID: PMC1224530  PMID: 16001362

Abstract

Autosomal dominant cerebellar ataxia (ADCA) is a group of heterogeneous neurodegenerative disorders. By positional cloning, we have identified the gene strongly associated with a form of degenerative ataxia (chromosome 16q22.1–linked ADCA) that clinically shows progressive pure cerebellar ataxia. Detailed examination by use of audiogram suggested that sensorineural hearing impairment may be associated with ataxia in our families. After restricting the candidate region in chromosome 16q22.1 by haplotype analysis, we found that all patients from 52 unrelated Japanese families harbor a heterozygous C→T single-nucleotide substitution, 16 nt upstream of the putative translation initiation site of the gene for a hypothetical protein DKFZP434I216, which we have called “puratrophin-1” (Purkinje cell atrophy associated protein-1). The full-length puratrophin-1 mRNA had an open reading frame of 3,576 nt, predicted to contain important domains, including the spectrin repeat and the guanine-nucleotide exchange factor (GEF) for Rho GTPases, followed by the Dbl-homologous domain, which indicates the role of puratrophin-1 in intracellular signaling and actin dynamics at the Golgi apparatus. Puratrophin-1—normally expressed in a wide range of cells, including epithelial hair cells in the cochlea—was aggregated in Purkinje cells of the chromosome 16q22.1–linked ADCA brains. Consistent with the protein prediction data of puratrophin-1, the Golgi-apparatus membrane protein and spectrin also formed aggregates in Purkinje cells. The present study highlights the importance of the 5′ untranslated region (UTR) in identification of genes of human disease, suggests that a single-nucleotide substitution in the 5′ UTR could be associated with protein aggregation, and indicates that the GEF protein is associated with cerebellar degeneration in humans.

Introduction

Autosomal dominant cerebellar ataxia (ADCA) is a clinical entity of heterogeneous neurodegenerative diseases that show dominantly inherited, progressive cerebellar ataxia that can be variably associated with other neurological and systemic features (Harding 1982). Circumscribed groups of neurons in the cerebellum, brainstem, basal ganglia, or spinal cord are selectively involved in different combinations and to varying extents among diseases (Graham and Lantos 2002). ADCA is now classified by the responsible mutations or gene loci. To date, 24 subtypes have been identified: spinocerebellar ataxia type (SCA) 1, 2, 3 (or, Machado-Joseph disease [MJD]), 4–8, 10–19/22, 21, 23, 25, 26; dentatorubral and pallidoluysian atrophy (DRPLA); and ADCA with mutation in fibroblast growth factor (FGF) 14 (Stevanin et al. 2000, 2004; Margolis 2002; van Swieten et al. 2003; Yu et al. 2005). Among these, mutations in SCA1, SCA2, SCA3/MJD, SCA6, SCA7, SCA17, and DRPLA have been identified as the expansion of a trinucleotide (CAG) repeat that encodes the polyglutamine tract, uniformly causing aggregation of polyglutamine-containing causative protein (Ross and Poirier 2004). Expansion of noncoding trinucleotide (CAG or CTG) or pentanucleotide (ATTCT) repeats are involved in SCA8, SCA10, and SCA12 (Holmes et al. 1999; Koob et al. 1999; Matsuura et al. 2000). Very few families are affected by missense mutations in the protein kinase Cγ (PKCγ) (SCA14 [see Chen et al. 2003]) and FGF14 genes (ADCA with FGF14 mutation [see van Swieten et al. 2003]). However, genes or even their loci remain unidentified for >20%–40% of families with ADCA (Sasaki et al. 2003).

We had previously mapped mutations in six Japanese families with ADCA to a 10-cM interval in human chromosome 16q13.1-q22.1, identifying 16q-linked ADCA type III, or spinocerebellar ataxia 4 (SCA4 [MIM 600223]) (Ishikawa et al. 2000). Clinically, our families show cerebellar ataxia without obvious evidence of extracerebellar neurological dysfunction (i.e., “pure cerebellar ataxia,” or “ADCA type III”) (Harding 1982; Ishikawa et al. 2000). The average age at onset of ataxia was >55 years (Ishikawa et al. 1997), which suggests that this disease shows the oldest age at onset among ADCA types with assigned loci. Another important clinical feature of this disease is that a substantial number of patients show progressive sensorineural hearing impairment (Owada et al., in press). Since the hearing impairment can be very mild and of later onset, presence of hearing impairment can be easily overlooked. However, this finding may indicate that the mutated gene could cause hearing impairment as well as ataxia. In this sense, it would be more appropriate to use the term “chromosome 16q22.1–linked ADCA” instead of “ADCA type III” to describe our families. Neuropathological examination showed peculiar degeneration of Purkinje cells that was not described in other degenerative ataxias (Owada et al., in press). Many Purkinje cells undergo shrinkage and are surrounded by amorphous materials composed of Purkinje-cell somato-dendritic sprouts and an increased number of presynaptic terminals. These findings may indicate that certain proteins involved in the cytoskeleton of Purkinje cells are disturbed in chromosome 16q22.1–linked ADCA.

Chromosome 16q22.1–linked ADCA has been assigned to the same locus as another ADCA, SCA4 (Flanigan et al. 1996; Hellenbroich et al. 2003). Although SCA4 and chromosome 16q22.1–linked ADCA may be allelic, SCA4 is clinically distinct from chromosome 16q22.1–linked ADCA, because SCA4 shows prominent sensory axonal neuropathy and pyramidal tract signs, with an age at onset earlier than that of chromosome 16q22.1–linked ADCA (Flanigan et al. 1996; Hellenbroich et al. 2003). Several groups, including ours, have refined the loci of SCA4/chromosome 16q22.1–linked ADCA and have, so far, excluded repeat expansions as mutations (Hellenbroich et al. 2003; Li et al. 2003; Hirano et al. 2004). The minimum candidate region of SCA4 and chromosome 16q22.1–linked ADCA is set at the region between markers D16S3031 and D16S3095. A strong founder effect has been observed for chromosome 16q22.1–linked ADCA (Li et al. 2003), which indicates the need to recruit a large number of families to narrow the critical region.

To discover the causative gene of chromosome 16q22.1–linked ADCA, we embarked on a positional cloning study by recruiting 52 families from diverse regions of Japan. Here, we describe the identification of a strong association between a single-nucleotide change and chromosome 16q22.1–linked ADCA and show the consequence of this genetic change on mRNA and protein levels. The data presented here also suggest that a single-nucleotide change in the 5′ UTR could be associated with aggregation of the gene product.

Patients, Material, and Methods

Recruitment of Families with Chromosome 16q22.1–Linked ADCA

We attempted to include families clinically diagnosed with late-onset ADCA type III from a wide region of Japan. Fifty-two families, including 109 affected individuals and 48 at-risk individuals, were ultimately recruited. These families originated from seven of eight districts of Japan (Hokkaido, Tohoku, Kanto, Chu-bu, Kinki, Chu-goku, and Kyu-shu), which indicates that their origins are widespread. Clinical features of these patients were consistent with those of families described elsewhere (Ishikawa et al. 1997, 2000; Li et al. 2003).

Detailed neuro-otological examinations, including pure-tone audiometry, were performed on 13 families at the Departments of Neurology and Otolaryngology, Tokyo Medical and Dental University. Progressive hearing impairment was assessed when the pure-tone average calculated from thresholds at the frequencies of 0.5, 1, and 2 kHz was more severe than the mean +2 SD of age-matched normal Japanese population (Tsuiki et al. 2002). By that criterion, 6 (42.9%) of 14 families had a hearing impairment other than age-related hearing loss. When we used the recommendations for the description of genetic and audiological data composed by the GENDEAG study group (see The Hereditary Hearing Loss Homepage), all of the patients with the hearing impairment were confirmed to have bilateral sensorineural hearing loss of mild-to-moderate severity. The audiometric configurations of these patients include mid-frequency U-shaped, flat, and high-frequency sloping. These data would suggest that the hearing impairment in our families is not a coincidence.

Preparation of DNA and RNA Samples

After informed consent was obtained, genomic DNA was extracted from peripheral-blood lymphocytes or lymphoblastoid cell lines by use of methods described elsewhere (Ishikawa et al. 1997). All families were excluded for SCA1, SCA2, SCA3/MJD, SCA6, SCA7, SCA8, SCA12, SCA14, SCA17, and DRPLA by testing mutations in the disease genes. For expression analysis on an RNA level, frozen cerebellar tissues of four brains affected with Alzheimer disease (AD) (age at death, range 65–85 years; duration of the disease, range 5–10 years) were used as controls. Clinically, these patients with AD showed a moderate-to-severe degree of dementia, as measured by the Clinical Dementia Rating (Hughes et al. 1982). Neuropathology in the cerebella of these patients showed only a few senile plaques without neuronal losses. For chromosome 16q22.1–linked ADCA, two patients were studied (ages at death, 74 and 78 years). Both control and 16q22.1-linked ADCA brains were obtained at autopsy, with the families’ written consents approved by each institutional ethical committee. These brains were immediately frozen and stored at −80°C until use. Total RNA was extracted from frozen cerebellar tissues as described elsewhere (Ishikawa et al. 1999). Poly-A+ RNA was selected from total RNA by NucleoTrap (Macherey-Nagel).

Restriction of Candidate Interval by Genotyping and Haplotype Analysis

Genotypes were determined for 23 informative markers, including five new markers we identified (GGAA10, TTCC01, TA001, GA001, and AAT01 [GenBank accession numbers AB13610, AB13611, AB13612, , and AB13613, respectively]) (table 1). Standard PCR was performed in a final volume of 20 μl, containing 10 ng of genomic DNA, 3.4 pmol of each primer, 2.5 mM of dNTP, and 0.75 U of Taq polymerase (Takara). Thermal cycling was performed at 94°C for 5 min for initial denaturing, 30 cycles of denaturation (94°C for 30 s), annealing (55°C for 30 s), and extension (72°C for 30 s), followed by a final extension at 72°C for 5 min in an ABI GeneAmp PCR system 9700 (PE Applied Biosystems). The amplified product was separated in the Automated Laser Fluorescent DNA Sequencer II (Pharmacia Biotech), and genotypes were determined with Fragment Manager (Pharmacia Biotech) (Ishikawa et al. 1997). Allele frequencies in the general population were analyzed in 500 unrelated individuals without personal or family history of ataxia or other degenerative diseases.

Table 1.

Microsatellite Markers and Primer Sequences

Primer Sequence
Markera Position in 16q22.1b Forward Reverse Size Range of PCR Products(bp)
D16S3031 64338902–64339180 5′-ACTTACCACTGTGCCAGTTG-3′ 5′-ATACATGGGTCCTTAAACCG-3′ 258–278
GGAA05 64938932–64939196 5′-GTTGCAGTGAGCGGAAATT-3′ 5′-CCTACCTCCAGTCCAGATT-3′ 240–264
GGAA10 65452426–65452828 5′-GAGCTCTGATCACACCATTT-3′ 5′-GCGGCACTTGCATTCATCA-3′ 378–402
TTCC01 65479376–65479551 5′-CCATCTTATCTCTTAGTCCGT-3′ 5′-GTGAGCCAAGATCATGCCAT-3′ 147–175
D16S3086 65493398–65493588 5′-CCTGCTGGGAAGTGACC-3′ 5′-CTAAGGACAAATGTAGATGCTCTGG-3′ 184–190
GATA01 65700158–65700328 5′-GGGTCACTGAGATGTCCAGTA-3′ 5′-CGCTGTGAGCACTGCTCTCA-3′ 150–170
D16S421 65853212–65853436 5′-ACATGAACCGATTGGACTGA-3′ 5′-CCGTTCCCTATATTTCCTGG-3′ 214–224
TA001 65906164–65906326 5′-CTTTCACTTAGCTGACATGATGG-3′ 5′-TGCACCACTGTGTTCCAGCTG-3′ 144–162
GA001 65922378–65922591 5′-GGAGGCTGAGAAGTCCAAGAGC-3′ 5′-CATGCTGTTATCTAGGAGTTAGTTAT-3′ 195–213
TTTA001 65943941–65944165 5′-CCGAAATAGGCGTACCTAGGCTA-3′ 5′-AGTGAGCTGAGTTTGCACCACTG-3′ 204–224
CATG003 66024974–66025174 5′-ACCACAGAGCTGAGGCCT-3′ 5′-AGCAAGTCTGCCTTGGAGAGG-3′ 188–200
17msm 66081379–66081581 5′-ATGACCCCCTGGTCACTATGG-3′ 5′-CCCATGTGCTTCAGGGAAGAT-3′ 184–202
D16S3107 66215715–66216019 5′-CCAGAGTGATGGGGAATA-3′ 5′-TGAGCACTGTCTCAAAAAA-3′ 290–304
D16S3085 66225096–66225302 5′-GCTATGACCGCACCAGT-3′ 5′-CGGAAATGGACCCTTG-3′ 200–206
AAT01 66588293–66588532 5′-CGAATCACAGACACACCCAC-3′ 5′-CAGAGATTGCAGTGAGCCAA-3′ 224–239
GGAA01 66601419–66601639 5′-GCCTAGGCAACATACTAAGA-3′ 5′-GCCCTCTAACTCCTGGTTT-3′ 192–220
D16S3025 67123897–67124008 5′-TCCATTGGACTTATAACCATG-3′ 5′-AGCTGAGAGACATCTGGG-3′ 101–111
CTTT01 67334584–67334848 5′-GCCCCAGATACGCCAGAG-3′ 5′-CCTCTCTACACTCCAGCCT-3′ 226–264
D16S496 67506219–67506439 5′-GAAAGGCTACTTCATAGATGGCAAT-3′ 5′-ATAAGCCACTGCGCCCAT-3′ 208–220
D16S3141 67663582–67663846 5′-GGCGTTCAGATTCTTTATCA-3′ 5′-AAGCCTCAGCTNTTGTGTGT-3′ 242–264
D16S3067 67666169–67666321 5′-GCCACCTCACACTAGCCTG-3′ 5′-TCACTCAAAATGGAGTCACTCTG-3′ 138–152
GT01 67870229–67870422 5′-GGGCACTAAGGACACTTCTT-3′ 5′-GGCTCTGTCATCCATGTTATG-3′ 177–193
D16S3095 68503731–68503891 5′-TCAGTTGGAAGATGAGTTGG-3′ 5′-TATAGTTTGTGTCCCCCGAC-3′ 152–160
Open in a new tab
a

Markers are shown in order from the 16q centromere to the 16q telomere.

b

Positions in chromosome 16q are in accordance with the Ensembl database.

Haplotypes of 16q22.1 markers were determined for 10 families (fig. 1) that were informative for determining the phase of alleles. The remaining 42 families were not sufficiently informative for determining the phase. Paucity of informative families was mainly due to the late age at onset of this disease. For these uninformative families, only combinations of genotypes were assessed, and the region of genotypes shared with the larger 10 families was compared.

Figure 1.

Figure 1

Open in a new tab

Haplotype analysis in the chromosome 16q22.1–linked ADCA critical locus. Ten families (P2, P4, P6, P12, P14, and T1–T5) large enough for haplotype analysis are indicated (yellow boxes at top). A founder haplotype “3-1-4-4-4” was seen in all patients for markers 16cen-D16S421-TA001-GA001-TTTA001-CATG003-16qter, between GATA01 and 17msm. Two alleles are shown (separated by a virgule [/]) for cases in which phase could not be determined. An asterisk (*) indicates that P values show significant linkage disequilibrium by Fisher’s exact probability test. The specific C→T change in the puratrophin-1 gene lies within the founder chromosomal region.

Screening for Mutation in 21 Different Candidate Genes

Within the most critical interval, 21 genes or clusters of ESTs were retrieved from the Ensembl, UniGene, National Center for Biotechnology Information (NCBI), and Celera databases. Primers were designed to amplify individual exon and intron-exon boundaries for each gene, and genomic DNA was amplified by the aforementioned standard PCR protocol (primer sequences are available on request). Amplicons from controls and patients were first separated on 2% agarose gels, to determine by screening whether any aberrant bands were amplified from patient DNA samples. Residual solutions of PCR products were then purified with QIAquick PCR purification kit (Qiagen) and were directly sequenced using Applied Biosystems Model 377 or 3100 Automated Sequencer, as described elsewhere (Li et al. 2003). The mutation was analyzed by comparing sequenced data and annotated databases, with use of the software DNASIS (Hitachi). Genomic rearrangement was screened by Southern-blot analysis by use of cosmid clones for probe synthesis as described elsewhere (Kobayashi et al. 1998). In brief, cosmid clones, tandemly covering the chromosome 16q22.1–linked ADCA critical region, were generated by subcloning from the BAC contig (Li et al. 2003). Then, the radiolabeled (32P) probe was generated from each cosmid clone. Genomic DNA extracted from lymphoblastoid cell lines of control individuals and patients with chromosome 16q22.1–linked ADCA was digested with a restriction enzyme and was subjected to Southern-blot analysis. To screen gene arrangement, analysis was performed with six different restriction enzymes (BamHI, BglII, EcoRI, EcoRV, HindIII, and XbaI).

Twenty-one exons and 20 introns of the puratrophin-1 gene, partially annotated as “Q9H7K4,” were sequenced entirely. The genomic region between the puratrophin-1 gene and SLC9A5 was also screened (fig. 2). The C→T change in the 5′ UTR was detected by amplifying genomic DNA with forward primer UK1-E1F1 (5′-CAGCGCGGTTCACACTGAGA-3′) and reverse primer UK1-E1R1 (5′-GGCCCTTTCTGACAGGACTGA-3′), which yielded a specific product of 360 bp. This amplicon harbors two EcoNI sites, one of which is destroyed by the C→T change.

Figure 2.

Figure 2

Open in a new tab

Positional cloning of the chromosome 16q22.1–linked ADCA gene. A, Genetic and physical maps of the interval in chromosome 16q22.1, showing microsatellite markers used to refine the interval. Twenty-one genes and their direction of transcription (Ensembl) are shown. B, Nucleotide sequences of exon 1 flanking the C→T single-nucleotide change in control and patient DNA samples. The patient harbors a heterozygous C→T substitution on the sense strand. C, RFLP by digestion with EcoNI. PCR was performed with primers UK1-E1F1 and UK1-E1R1. Whereas the normal allele produces digested fragments of 209, 92, and 59 bp, the C→T change in the mutant allele disrupts one EcoNI site, producing fragments of 268 and 92 bp.

Expression Analysis of Puratrophin-1 mRNA by RT-PCR

To characterize puratrophin-1 mRNA in the human brain, poly-A+RNA obtained from control human cerebellar tissues was reverse transcribed with SuperScript II (Stratagene), and the entire coding region was amplified with primers UKI-RT-025F (5′-TTCGCCTGCATTGCCCACTGAG-3′) and UK1-RT-005R (5′-CACACACATCAGAAAGGGTAGTCAAC-3′), which yielded a major product of 3,835 bp. To analyze alternative transcription, the PCR products were subcloned into pCR2.1-TOPO (Invitrogen), and 10 randomly selected clones from each brain mRNA (2 AD and 2 chromosome 16q22.1–linked ADCA cerebella) were sequenced and compared with partially annotated sequences (Q9H7K4, NCBI accession numbers BC054486 and AK024475). Rapid cloning of the 5′ and 3′ ends (5′- and 3′-RACE) was performed using Marathon-ready cDNA kit (Clontech [BD Biosciences]) as described elsewhere (Tsunemi et al. 2002). The entire 5′ UTR was amplified with primers UK1-RT-493F (5′-TGAGACAGTCTCAGTCAGGTCAC-3′) and UK1-RT-021R (5′-GTGGGCACACAGAAGCAGCACTGC-3′), was subcloned into pCR2.1TOPO (Invitrogen), and was sequenced.

Expression of the puratrophin-1 gene in various human tissues was studied by RT-PCR with use of primers UKI-RT-025F and UK1-RT-005R on human Multiple Tissue Panels I and II (Clontech [BD Bioscience]). Expression levels of puratrophin-1 mRNA were compared among control and patient cerebellar cortices by RT-PCR with use of primers UK1-RT-BF001F (5′-TCACGGTTCCCCGCGGCTCG-3′) and UK1-RT-BF019R (5′-GGTTGCATGGCCCTGAGAGTCTGG-3′), which yielded PCR products of 291, 419, and 505 bp, depending on alternative transcription. To compare puratrophin-1 mRNA levels more precisely, real-time RT-PCR analysis was performed on four control (individuals with AD) and two chromosome 16q22.1–linked ADCA cerebella by use of TaqMan technique (Applied Biosystems) on Applied Biosystems 7700 Sequence Detection System. PCR was performed for total puratrophin-1 mRNAs (i.e., full-length and short-form mRNAs [GenBank accession numbers AB197663 and AB197664, respectively]; definitions of these two isoforms are described in the “Results” section) with primers total-taq-F (5′-TGGAGAGATGAGTGTCAAGACTTTG-3′) and total-taq-R (5′-AATGACTTGGGTCTGCCTTGG-3′). The short-form puratrophin-1 mRNA was specifically assessed with primers short-taq-F (5′-ATGCCACCGACTGGAGATTT-3′) and short-taq-R (5′-GCTGCCCTGTAGCTCCTCAT-3′). The TaqMan probe for total puratrophin-1 mRNA was 5′-FAM-CCAGATGCACGTTAAGGACCCAGGTC-TAMRA-3′, and that for short-form puratrophin-1 mRNA was 5′-FAM-TCTCTGACCCTACTCAGGCTGAAGCCC-TAMRA-3′. The experiments were performed three times and were averaged. The level of puratrophin-1 mRNA was assessed relative to the level of G3PDH mRNA amplified with primers supplied from the manufacturer (Applied Biosystems). Since the expression level of G3PDH mRNA was >1,000-fold higher than that of puratrophin-1 mRNA, one control individual was chosen and the value of puratrophin-1 mRNA:G3PDH mRNA ratio (designated “puratrophin-1 mRNA/G3PDH mRNA index”) was standardized as “1”; another three control individuals and two patients were compared relatively. The difference between the control and patient groups was statistically analyzed with the Mann-Whitney U test.

To see the effect of the C→T change in the 5′ UTR of puratrophin-1 gene on expression in vitro, the partial 5′-UTR fragment containing the mutation site was amplified with forward primers UK1-UF10-F2 (5′-CCGGAATTCCAGGCCTGAATTGCAGTTC-3′) and UK1-UF10-R2 (5′-CCAGGCATCCCTGAAACT-3′) and was subcloned into the HindIII-NcoI sites of the luciferase assay vector, pGL3-Control Vector (Promega). The C→T mutation was then generated by site-directed mutagenesis (Stratagene). Three types of vectors (mock, empty pGL3-Control Vector; wild-type, pGL3 with normal puratrophin-1 5′ UTR with allele C; and mutant, pGL3 with mutant puratrophin-1 5′ UTR with allele T) were separately transfected with Renilla vector (Promega) into HEK293 cells to equalize the transfection efficiency. The transfection experiment was performed independently three times. Luciferase activity (F/R: fire-fly luciferase activity against Renilla activity) was assayed in accordance with the manufacturer’s protocol and was statistically analyzed with the Mann-Whitney U test.

Generation of Rabbit Polyclonal Antibodies against Puratrophin-1

On the basis of the deduced amino acid sequence of “full-length” and “short-form” puratrophin-1, the secondary structure of each isoform was predicted, and five synthetic peptides were designed: polypeptide FL01 (aa 1–13), N-MERPLENGDESPD-C; FL02 (aa 421–432), N-MDKADELYDRVD-C; FL03 (aa 570–583), N-EEGQRVLAELEQER-C; FL04 (aa 991–1003), N-RFEIWFRRRKARD-C; and SV01 (aa 13–26), N-REVWEGNGDAWRDE-C. The polypeptide FL01 lies at the amino terminus of the puratrophin-1 (fig. 3). Therefore, the antibody Prtrphn1-Ab FL01 would detect both full-length and short-form puratrophin-1. The polypeptides FL02, FL03, and FL04 are specific for the full-length puratrophin-1 (fig. 3). Notably, FL03 is at the spectrin repeat motif. The polypeptide SV01 lies at the carboxyl-terminus of the short-form puratrophin-1. Therefore, the antibody Prtrphn1-Ab SV01 would specifically recognize the short-form puratrophin-1. Each antigen was immunized into rabbit, and polyclonal antibodies were obtained as described elsewhere (Ishikawa et al. 1999).

Figure 3.

Figure 3

Open in a new tab

Genomic and mRNA structure of the puratrophin-1 gene. A, Intron-exon structure of puratrophin-1. Exons are shown as vertical bars, and coding regions are shown in red. The C→T change at nt 4255 (or, 16 nt upstream of the translation initiation codon) is indicated. Nucleotide numbers (indicated by a number sign [#]) are counted from the 5′ end. B, Puratrophin-1 mRNAs cloned from the human cerebellum. Major transcripts are full-length and short-form puratrophin-1 mRNAs; minor transcripts (B1–B8) were also cloned. The CRL-TRIO, spectrin, Rho GEF/DH, and PH domains are indicated. Epitopes for five rabbit anti–puratrophin-1 antibodies (FL01–FL04 and SV01) are also mapped. Antibodies FL02, FL03, and FL04 are designed to specifically detect full-length puratrophin-1, the antibody SV01 specifically detects short-form puratrophin-1, and the antibody FL01 detects both full-length and short-form puratrophin-1.

The specificity of Prtrhn1-Abs was assessed by detection of recombinant puratrophin-1 protein. For this purpose, an antisense-strand primer was designed to encode hemagglutinin A (HA) in-frame at the carboxyl-terminus of the puratrophin-1. By PCR, full-length puratrophin-1 cDNA with the HA-coding nucleotide sequence at its 3′ end was generated and was then cloned into pcDNA1 expression vector (Invitrogen). The puratrophin-1–HA fusion protein, transiently expressed in HEK293 cells, was first immunoprecipitated with Prtrhn1-Abs and then was detected with rat monoclonal anti-HA High Affinity antibody (3F10 [Roche Mannheim]) (fig. 4A). The fusion protein immunoprecipitated with preimmune sera was not detected with anti-HA antibody.

Figure 4.

Figure 4

Open in a new tab

A, The recombinant, full-length puratrophin-1–HA fusion protein was first immunoprecipitated (IP) with preimmune serum (PIS) (as a negative control), anti-HA antibody (as a positive control), or anti–puratrophin-1 antibodies (FL01, FL02, FL03, and FL04) and was then tested for western blot (WB) detection with the anti-HA antibody. The anti–puratrophin-1 antibodies consistently detected the recombinant fusion protein, with an approximate molecular size of 135 kD. B, Immunoabsorption test. Antiserum against puratrophin-1 (FL01) was completely absorbed with synthetic peptide of 1 mg/liter concentration (FLO1 + peptide). Scale bars = 50 μm.

Specificity of Prtrphn1-Abs on immunohistochemistry was assessed by immunoabsorption test. Solutions containing Prtrphn1-Ab and different concentrations of antigen (synthetic peptide) (0, 0.01, 0.1, and 1 mg/liter) were made. Sections were incubated overnight with these solutions as other immunohistochemical staining. Prtrphn1-Abs were completely absorbed by peptides of higher concentrations (fig. 4B).

Immunohistochemical Analyses of Puratrophin-1, Golgi-Apparatus Protein G58K, and Spectrin in Control Mouse, Human, and 16q22.1-Linked ADCA Tissues

Five polyclonal antibodies against puratrophin-1 (Prtrphn1-Ab types FL01, FL02, FL03, FL04, and SV01) were used for immunohistochemistry. Mouse monoclonal antibody against the microtubule-binding peripheral Golgi-apparatus membrane protein G58K (Sigma) and the mouse monoclonal antibody for α- and β-spectrin (MAB372 [Chemicon International]) were also used for immunohistochemistry.

Formalin-fixed paraffin-embedded tissue sections were prepared. Examined tissues were C57BL/6J wild-type mouse tissue, including cochlea (1 d postnatal), normal-control human tissues (testis, pancreas, prostate gland, lung, liver, heart, kidney, and brain from three individuals who died with nonneurological diseases), disease-control human brains (AD [n=3], SCA6 [n=2], multiple-system atrophy [MSA] [n=3], and SCA3/MJD [n=3]), and three brains with chromosome 16q22.1–linked ADCA. Immunohistochemistry was performed as described elsewhere (Ishikawa et al. 1999). In brief, sections were incubated overnight with Prtrphn1-Abs (diluted 1:200 with PBS), G58K (diluted 1:400), or MAB372 (diluted 1:200) at 4°C, after which the primary antibody was serially detected with avidin-biotinylated peroxidase complex method (Vector), was developed with 3,3′-diaminobenzidine (DAB), and was counterstained with hematoxylin.

Results

Restriction of the Critical Interval to <600 kb within Human Chromosome 16q22.1

Haplotype reconstruction of 10 informative families revealed that all affected individuals in these families were segregated with the haplotype 3-1-4-4-4 for markers 16cen-D16S421-TA001-GA001-TTTA001-CATG003-16qter lying between GATA01 and 17msm (fig. 1). Genotypes of all affected individuals in the remaining 42 families that were not informative enough for haplotype reconstruction were also consistent with the common haplotype. Particularly, a dinucleotide (GA) repeat marker GA001 showed strong linkage disequilibrium: the allele 4 of GA001 was seen in all affected individuals in all families with chromosome 16q22.1–linked ADCA, whereas that allele was seen very rarely (frequency 0.1%) in 1,000 control chromosomes (fig. 1). Whereas only one common haplotype was seen between GATA01 and 17msm, different alleles were seen for GATA01 and other centromeric markers or for 17msm and other telomeric markers (fig. 1). These results indicate that the gene for chromosome 16q22.1–linked ADCA is most likely to exist in the interval between GATA01 and 17msm. This region was within the previously refined intervals (Hellenbroich et al. 2003; Li et al. 2003) and was considered to be <600 kb in size in the database Ensembl.

Identification of a Single-Nucleotide Substitution in the 5′ UTR of the Gene Encoding Puratrophin-1 in the Critical Interval

Within the critical interval between GATA01 and 17msm, 21 different genes were identified (fig. 2A). We screened all 21 genes by direct sequencing of amplified genomic PCR products of all coding exons and flanking intronic sequences. Genomic Southern-blot analyses were also performed but did not detect any rearrangement (fig. 5).

Figure 5.

Figure 5

Open in a new tab

Southern-blot analysis. The probe of this experiment was ∼10 kb in size, covering the Q9H7K4 coding region. The DNA samples from the control individuals and patients (lane 1 = patient 1; lane 2 = control 1; lane 3 = patient 2; lane 4 = patient 3; lane 5 = control 2) were digested with restriction enzyme EcoRV, were separated on a 0.8% agarose gel, were transferred to nylon membrane, and were hybridized with a radio-labeled probe. In all DNA samples, fragments of ∼9 kb were detected, which suggests that there were no genomic rearrangements.

We found that all patients with chromosome 16q22.1–linked ADCA harbored a heterozygous C→T single-nucleotide substitution at the position 16 nt upstream of the putative translation initiation codon of the gene Q9H7K4, encoding hypothetical protein DKFZP434I216 (NCBI accession number BC054486) (fig. 2B). All the at-risk individuals who had the founder haplotype between GATA01 and 17msm also harbored the C→T change. The ages of these individuals were below the maximum age at disease onset in their families, which indicates that they could be in presymptomatic stages. On the other hand, this C→T substitution was not seen in 500 healthy Japanese individuals (1,000 chromosomes), confirmed by EcoNI RFLP analysis (fig. 2C). All intronic sequences between the 21 exons and all the 8-kb sequences upstream of the Q9H7K4 gene were also directly sequenced. Although there was a G→A SNP 44 nt upstream of the 3′ splice site of intron 9 of the Q9H7K4 gene, the allele segregating with the disease was also seen in 75% of control chromosomes. As far as we examined, the C→T change at the position 16 nt upstream of the putative translation initiation codon of the gene Q9H7K4 was the only specific change seen for patients with chromosome 16q22.1–linked ADCA. We renamed this protein “puratrophin-1,” meaning Purkinje cell atrophy associated protein-1. (We reserved the nomenclature “ataxin-4” for families with original SCA4.)

Characterization of the Puratrophin-1 Gene and Its Expression

The puratrophin-1 gene comprises 20 introns and 21 exons spanning a 13.7-kb genomic region (fig. 3A). RT-PCR and RACE experiments revealed that this gene is expressed in the human brain with an 8,561-nt mRNA and a single ORF of 3,576 nt (from nt 4,271 to nt 7,846) (fig. 3B). The first exon was 4,769 nt in length. This “full-length” transcript is predicted to encode a 134-kD protein, consisting of 1,192 aa, that contains four important domains: the cellular retinaldehyde-binding (CRAL)/triple function domain (TRIO) (at codons 304–338), the spectrin-repeat domain in the middle portion (at codons 447–601), the guanine-nucleotide exchange factor (GEF) for Rho/Rac/Cdc42-like GTPases followed by the Dbl-homologous (Rho GEF/DH) domain (at codons 733–907), and the pleckstrin-like homology (PH) domain (at codons 921–1027). The presence of these four domains suggests the role of puratrophin-1 in intracellular signaling and actin dynamics targeted to the Golgi apparatus (Godi et al. 1998; Fucini et al. 2000).

Besides the full-length puratrophin-1 mRNA, we also found several alternative transcripts in human cerebellar cDNA (fig. 3B). The most common one had an unspliced intron 1 between exons 1 and 2 (“short-form” puratrophin-1 mRNA). This intron retention produced a frameshift after exon 1 that resulted in a premature stop codon after 26 aa residues in intron 1. There were several other alternative transcripts (B1–B8 [fig. 3B]) in addition to full-length and short-form puratrophin-1 mRNA. However, expression levels of these transcripts were considered lower than those of full-length or short-form puratrophin-1 mRNA, as judged from cloning frequencies. Puratrophin-1 mRNA was expressed in many tissues with various expression levels (fig. 6). The strongest expression was observed in the testis and pancreas, whereas its expression level was low in the “whole” brain. The expression was very weak on northern-blot analysis, even in the testis and pancreas (data available on request), which indicates that the expression of puratrophin-1 mRNA was not abundant.

Figure 6.

Figure 6

Open in a new tab

Puratrophin-1 mRNA expression in various control human tissues. With this nonquantitative RT-PCR analysis, a relatively stronger expression was observed in the testis and pancreas, whereas mild or moderate expression was seen in the spleen, thymus, prostate gland, heart, placenta, lung, liver, and kidney. Puratrophin-1 mRNA expression was low in the ovary, small intestine, colon, peripheral-blood leukocytes, whole brain, and skeletal muscle.

The Consequence of a C→T Change for Puratrophin-1 mRNA Expression

The consequence of the C→T change in the puratrophin-1 gene was assessed by RT-PCR on frozen cerebellar tissues of two patients. No aberrant puratrophin-1 mRNA was detected, which indicates that the mutation in the 5′ UTR does not affect alternative transcription significantly. Instead, the expression level of puratrophin-1 mRNA in the patient cerebella appeared slightly low compared with that in control cerebella (fig. 7A). On real-time RT-PCR analysis by TaqMan technique, the expression of puratrophin-1 mRNA in the patient cerebella appeared lower than that in the control samples (fig. 7B). The tendency of reduction was seen for both total and short-form puratrphin-1 mRNAs, although the significance was not proven statistically (P=.0641 for total puratrophin-1 mRNA; P=.1649 for short-form puratrophin-1 mRNA).

Figure 7.

Figure 7

Open in a new tab

The consequence of a C→T change in the 5′ UTR of puratrophin-1 mRNA. A, The puratrophin-1 mRNA level assessed relative to G3PDH mRNA levels. The puratrophin-1 mRNA levels in the cerebella of individuals with chromosome 16q22.1–linked ADCA appear slightly decreased compared with levels in control individuals. B, Quantitative analysis of puratrophin-1 mRNA, with the use of the TaqMan technique in cerebellar mRNAs of four control individuals (with AD) and of two individuals with chromosome 16q22.1–linked ADCA. Both total puratrophin-1 mRNA (i.e., full-length and short-form mRNAs) and short-form puratrophin-1 mRNAs tended to be reduced in cerebella of individuals with chromosome 16q22.1–linked ADCA, although the significance was not statistically proven (P=.0641 for total mRNA; P=.1649 for short-form mRNA). C, Histogram of in vivo luciferase assay. “F/R” denotes fire-fly luciferase activity versus Renilla activity. The wild-type construct with allele C increases luciferase expression compared with an empty vector (Mock), whereas the mutant construct with allele T demonstrates significantly reduced luciferase activity (P<.001).

The in vitro study with puratrophin-1 5′-UTR sequences subcloned into luciferase expression vectors showed significantly reduced luciferase activity; the construct had mutant allele T, and the wild-type construct had allele C (fig. 7C). These data suggest that the C→T change in the puratrophin-1 5′ UTR could lead to reduced mRNA expression.

Expression of Puratrophin-1 in Control Human and Mouse Tissues and in Brains of Patients with 16q22.1–Linked ADCA

We next examined puratrophin-1 expression by immunohistochemistry using rabbit polyclonal Prtrhn1-Abs. Puratrophin-1 expression was seen in many control human tissues and mouse cochlea (fig. 8). Interestingly, puratrophin-1 expression was commonly seen in epithelial cells, such as epithelial cells in the human testis and hair cells in the mouse cochlea. As far as we examined in human tissues, the strength of immunoreactivity correlated with the mRNA expression level revealed by RT-PCR experiments (fig. 6).

Figure 8.

Figure 8

Open in a new tab

Expression of puratrophin-1 in control human and mouse tissues. A and B, Human testis. C and D, Human prostate gland. E and F, Human pancreas. G and H, Human kidney. I and J, Mouse cochlea. A, C, E, G, and I, Immunohistochemical analysis with use of the polyclonal anti–puratrophin-1 antibody SV01. B, D, F, H, and J, Hematoxylin and eosin stain of the section adjacent to that shown in A, C, E, G, and I, respectively. Puratrophin-1 is expressed in many tissues, with varying strength. In particular, Leydig cells in the testis (arrow in A), epithelial cells in the prostate gland (C), and Langerhans islet in the pancreas (E) showed strong immunoreactivity. Immunoreactivity in the kidney was weak (G). Expression in the cochlea was particularly intense at the sensory hairlets (arrow in I). Hair cells of the stria vascularis also showed immunoreactivity (arrowhead in I). All scale bars = 50 μm.

In control human brains, both full-length and short-form puratrophin-1 was expressed faithfully in various neurons but most strongly in Purkinje cells (fig. 9A). In chromosome 16q22.1–linked ADCA brains, a striking difference was noted: the presence of microscopic aggregates of puratrophin-1 in the cytoplasm of Purkinje cells (fig. 9B). These aggregates were detected with all Prtrhn1-Abs, which indicates that full-length puratrophin-1 and short-form puratrophin-1 were both involved in the aggregate. On the other hand, puratrophin-1 aggregates were not obvious in neurons other than Purkinje cells. The finding of puratrophin-1 aggregates was specific to chromosome 16q22.1–linked ADCA brains, since the aggregates were not seen in other diseases showing degeneration of Purkinje cells, such as SCA6 or MSA.

Figure 9.

Figure 9

Open in a new tab

Immunohistochemical analysis of puratrophin-1, Golgi-apparatus membrane protein (G58K), and spectrin in control cells and samples from a patient's cerebellum. A, Puratrophin-1 expression in the cell body of Purkinje cells, visualized with use of rabbit polyclonal anti–puratrophin-1 antibody (FL01). B, With use of same antibody FL01, aggregation of puratrophin-1 (arrow), seen in a Purkinje cell of a brain affected with chromosome 16q22.1–linked ADCA. C, G58K expression, seen diffusely in the cell body of control Purkinje cells. D, G58K aggregation (arrow), morphologically quite similar to the puratrophin-1 aggregate, seen in a patient’s Purkinje cell. E, Spectrin expression, seen in the cell body of a control Purkinje cell. F, Spectrin aggregation in a patient’s Purkinje cell. All scale bars = 50 μm.

We next examined the expression of G58K and of the α- and β-spectrin on control and chromosome 16q22.1–linked ADCA brains. Whereas G58K and spectrin were homogeneously stained in control brains (fig. 9C and 9E), these proteins both formed aggregation within the Purkinje cell of chromosome 16q22.1–linked ADCA brains (fig. 9D and 9F). This further suggests that formation of puratrophin-1 aggregates is abnormal and associates with disturbances of Golgi apparatus and spectrin.

Discussion

The single-nucleotide C→T substitution in the 5′ UTR of the puratrophin-1 gene is strongly associated with chromosome 16q22.1–linked ADCA, as evidenced by the following three points. First, this single-nucleotide change was the only specific difference detected in patients, as far as we screened all exons and intron-exon boundaries of the 21 annotated genes lying within the founder chromosome. The change completely segregated with the disease in 52 unrelated families that originated in all sections of Japan, whereas such a change was not seen in 1,000 control chromosomes. We also screened for genomic rearrangement by Southern-blot analysis, but we did not observe any changes. Second, the C→T change was not a mere polymorphism present in the founder chromosome, since it resulted in reduced expression in the in vitro luciferase assay. Consistently, tendency for reduction in mRNA expression was seen in the cerebellum of chromosome 16q22.1–linked ADCA. Third, puratrophin-1 was aggregated in the major target neuron (i.e., the Purkinje cell) of chromosome 16q22.1–linked ADCA. The Golgi-apparatus membrane protein (G58K) and spectrin, both important cytoskeletal proteins, were also aggregated. Since aggregation of mutated protein is a common feature of many neurodegenerative disorders (Ross and Poirier 2004), a single-nucleotide change in the 5′ UTR in the puratrophin-1 gene appears to be the mutation that causes chromosome 16q22.1–linked ADCA. Most mutations that cause ADCAs reside in genes encoding proteins whose functions are not well understood, except for the α1A-calcium channel (Cav2.1) for SCA6 (Zhuchenko et al. 1997), PKCγ for SCA14 (Chen et al. 2003), TATA-binding protein for SCA17 (Koide et al. 1999), and FGF14 for ADCA with FGF14 mutation (van Swieten et al. 2003). Puratrophin-1 is the first known protein related to intracellular signaling and cytoskeleton that is associated with ADCA. Our preliminary results suggest that the frequency of chromosome 16q22.1–linked ADCA is relatively high, since it was ranked the third most frequent ADCA in Japan, after MJD/SCA3 and SCA6. Since chromosome 16q22.1–linked ADCA had a strong founder effect, it will be important to clarify whether this specific C→T change is seen in other ethnic groups. Particularly, it is extremely important to clarify whether original patients with SCA4 reported from North American and German populations harbor mutation within the puratrophin-1 gene (Flanigan et al. 1996; Hellenbroich et al. 2003).

A heterozygous, single-nucleotide substitution in the 5′ UTR that is associated with aggregation of the gene product is a unique feature as a cause of human disease. Generally, a single-nucleotide substitution in the 5′ UTR may result in expression of aberrant mRNA, abnormal trafficking of mutant transcripts due to conformational changes, or reduced transcription of mRNA, as in β thalassemia intermedia (Sgourou et al. 2004). In the present study, no aberrant transcripts were cloned from patient brains. Instead, the result from the in vitro study suggested that the transcription efficiency could be reduced significantly by the C→T substitution in the 5′ UTR of the puratrophin-1 gene. Consistent with this in vitro finding, levels of puratrophin-1 mRNA tended to be lower in cerebella of patients with 16q22.1-linked ADCA than in those of AD-affected controls. Further studies comprising a larger number of samples that include “neurologically normal” controls will be necessary to confirm that reduced puratrophin-1 mRNA expression is associated with the C→T change in the 5′ UTR. Although aggregations of mutated proteins in most neurodegenerative diseases are due to conformational changes produced by amino acid alterations, reduced protein expression may also cause protein aggregation, such as the neurofilament light-chain aggregation in anterior horn cells of patients with amyotrophic lateral sclerosis (ALS) (Bergeron et al. 1994). It is currently not known how the C→T change in the puratrophin-1 gene leads to puratrophin-1 aggregation. However, it is possible that alteration in the stoichiometry or changes in the protein stability, observed in neurofilament proteins in ALS (Ge et al. 2003), also exist for puratrophin-1 in chromosome 16q22.1–linked ADCA. In addition, it should be noted that since reduction of puratrophin-1 mRNA levels in patient cerebella was small, that may not be a sufficient explanation of the formation of aggregation. It will be important to assess the puratrophin-1 mRNA levels specifically in Purkinje cells, where the protein aggregates.

Puratrophin-1 is a novel protein with four important domains: the spectrin repeat, CRAL/TRIO, Rho GEF/DH, and PH domains. Therefore, the puratrophin-1 gene would be the fifth gene that encodes Rho GEF protein to be associated with human disease, after the FGD1GEF gene for faciogenital dysplasia (Pasteris et al. 1994), the ARHGEF6 gene for X-linked nonsyndromic mental retardation (Kutsche et al. 2000), the ARHGEF10 gene for autosomal dominant neuropathy with thin myelinated fibers (Verhoeven et al. 2003), and the ALS2 gene for the autosomal recessive juvenile motor neuron disease ALS2 (Hadano et al. 2001; Yang et al. 2001). GEFs exert diverse functions in intracellular-membrane trafficking and microtubule dynamics and ultimately regulate numerous cellular responses—such as proliferation, differentiation, and movement—by activating small G-protein GTPases (Rossman et al. 2005). Although Rho GEF/DH and PH domains are typical for Rho GEF proteins, the presence of a spectrin repeat suggests that puratrophin-1 could be targeted to the Golgi-apparatus membrane, where it may regulate certain membrane dynamics through modulating actin (Godi et al. 1998; Lemmon et al. 2002). The CRAL/TRIO domain is implicated in the nerve growth factor (NGF) pathway that leads to neurite outgrowth through activating Rho G (Estrach et al. 2002), which indicates that puratrophin-1 may also have a role in cell differentiation signaling through NGF. In the present study, we showed that the Golgi-apparatus membrane protein and spectrin both form aggregation in the Purkinje cells of patient brains, which strongly supports the hypothesis that puratrophin-1 indeed interacts with these essential cytoskeletons. Formation of these aggregations, which seems to be a new phenomenon implicated in pathogenesis of human disease, could be deleterious for cells, since Golgi-apparatus membrane proteins and spectrin are important in maintaining cellular architecture, as noted for autosomal dominant polycystic kidney diseases (Charron et al. 2000). In addition, it also seems rational to speculate that disturbance of puratrophin-1 may affect Purkinje-cell morphology and eventually cause the peculiar Purkinje-cell atrophy that characterizes chromosome 16q22.1–linked ADCA, since Rho GEFs are implicated particularly in neural morphogenesis and connectivity by regulating actin dynamics (Godi et al. 1998).

Finally, the association of progressive hearing impairment in patients with chromosome 16q22.1–linked ADCA may suggest a role of puratrophin-1 in hearing. As far as we were able to examine by auditory tests, 42.8% of study families had hearing impairment. Audiometric configurations showing mid-frequency U-shaped or flat-shaped pattern also suggested that the hearing impairment of these patients was not a simple age-related hearing loss. However, it should be noted that some families had moderate hearing impairment, whereas other families had only mild hearing impairment, which may indicate the presence of other modifying genetic factors for this phenotype. Further clinical analyses with use of detailed auditory tests are needed to ascertain how strongly hearing impairment is associated with ataxia. If hearing impairment is confirmed as being complicated by chromosome 16q22.1–linked ADCA, puratrophin-1 will be the first deafness/hearing-impairment protein related to Rho GEF. Since mutations in the γ-actin gene (ACTG1; DFN20/26) (Zhu et al. 2003), genes coding proteins that interact with actin (e.g., the myosins espin and harmonin), or a gene coding proteins that coordinate actin polymerization (Mburu et al. 2003) are known to cause deafness, it is possible that puratrophin-1 disruption causes hearing impairment by disturbing actin dynamics in the cochlea. In support of this hypothesis, expression of puratrophin-1 was seen in mouse cochlear hair cells, in which actin is also expressed (Mburu et al. 2003). Rho GTPases are also important regulators of actin cytoskeleton in stereocilia development, which is crucial for auditory transduction (Kollmar 2001). We hypothesize that the puratrophin-1 gene mutation ultimately causes hearing impairment by dysregulation of actin in the cochlea. Further studies, such as targeted disruption of puratrophin-1 in mouse, will be important for clarifying whether puratrophin-1 has a role in hearing.

In summary, we have identified that a single-nucleotide C→T substitution in the 5′ UTR of the gene puratrophin-1 is strongly associated with chromosome 16q22.1–linked ADCA. If patients with SCA4 are found to harbor mutations in this gene, it would suggest that the chromosome 16q22.1–linked ADCA in Japan is allelic with SCA4. Identification of the mechanism of puratrophin-1 expression, the upstream signaling cascade that activates puratrophin-1, and the actual Rho GTPase activated by puratrophin-1 would be the next key steps for understanding the molecular mechanisms that underlie cellular degeneration of chromosome 16q22.1–linked ADCA and hearing impairment.

Acknowledgments

We thank the doctors who participated in this work by recruiting families with chromosome 16q22.1–linked ADCA: Drs. Hidenao Sasaki, Department of Neurology, Graduate School, Hokkaido University; Masashi Aoki, Department of Neurology, Graduate School, Tohoku University; Yoshihisa Takiyama, Department of Neurology, Jichi Medical School; Kazuo Yoshizawa, National Mito Hospital; Kazuko Mitani and Yu-ichi Fumimura, Department of Neurology, Tokyo Metropolitan Geriatric Hospital; Hirohiko Murakami, Department of Neurology, Tokyo Women’s Medical School; Satoshi Orimo, Department of Neurology, Kanto Central Hospital; Souichiro Mochio, Department of Neurology, Jikei Medical School, The Third Hospital; Kunihiro Yoshida, Department of Clinical Genetics, Shinshu University; Isao Sahashi, Fourth Department of Internal Medicine, Aichi Medical University; Masanori Nakagawa, Department of Neurology, Kyoto Prefectural Medical College; Akihumi Goto, Department of Neurology, Nagasaki Medical Center of Neurology; Hideki Kida, Kida Hospital; Eiichiro Uyama, Department of Neurology, Kumamoto University of Medicine; Jun Goto and Shoji Tsuji, Department of Neurology, Graduate School, University of Tokyo; and Miho Murata and Ichiro Kanazawa, National Center of Neurology and Psychiatry, Ministry of Health, Labor and Welfare. For human and mouse control specimens, we also thank Professors Morio Koike, Department of Pathology, and Ken Kitamura, Department of Audio-Vestibular Neuroscience, Graduate School, Tokyo Medical and Dental University. We thank Ms. Iku Sudo and Ms. Minori Kono for technical assistance. We deeply acknowledge Professors Ichiro Kanazawa, Director of National Center of Neurology and Psychiatry, Ministry of Health, Labor and Welfare, and Shoji Tsuji, Department of Neurology, Graduate School, University of Tokyo, for supporting this study. This study was supported by Grant-in-Aids for Scientific Research on Priority Areas—Advanced Brain Science Project—from Ministry of Education, Culture, Sports, Science and Technology, Japan (to K.I. and H.M.), as well as grants from Research on Intractable Disorders (to H.M.) and from Human Genome and Regenerative Medicine, Ministry of Health, Labor and Welfare, Japan (to K.I. and H.M.).

Web Resources

Accession numbers and URLs for data presented herein are as follows:

  1. Celera Discovery System, http://www.celeradiscoverysystem.com/index.cfm (for candidate genes)
  2. Ensembl, http://www.ensembl.org/ (for genetic map, genomic sequences, ESTs, and 21 candidate genes, including Q9H7K4 and SLC9A5)
  3. GenBank, http://www.ncbi.nlm.nih.gov/Genbank/ (for microsatellite DNA markers GGAA10 [accession number AB13610], TTCC01 [accession number AB13611], TA001 [accession number AB13612], GA001 [accession number AB197662], and AAT01 [accession number AB13613] and full-length and short-form puratrophin-1 mRNAs [accession numbers AB197663 and AB197664])
  4. NCBI, http://www.ncbi.nlm.nih.gov/ (for DKFZP434I216 [accession numbers BC054486 and AK024475])
  5. Online Mendelian Inheritance in Man (OMIM), http://www.ncbi.nlm.nih.gov/Omim/ (for 16q-linked ADCA type III, or SCA4)
  6. The Hereditary Hearing Loss Homepage, http://webhost.ua.ac.be/hhh/
  7. UniGene, http://www.ncbi.nlm.nih.gov/UniGene/ (for candidate genes and ESTs)

References

  1. Bergeron C, Beric-Maskarel K, Muntasser S, Weyer L, Somerville M, Percy ME (1994) Neurofilament light and polyadenylated mRNA levels are decreased in amyotrophic lateral sclerosis motor neurons. J Neuropathol Exp Neurol 53:221–230 [DOI] [PubMed] [Google Scholar]
  2. Charron AJ, Bacallao RL, Wandinger-Ness A (2000) ADPKD: a human disease altering Golgi function and basolateral exocytosis in renal epithelia. Traffic 1:675–686 [DOI] [PubMed] [Google Scholar]
  3. Chen D-H, Brkanac Z, Verlinde CLMJ, Tan X-J, Bylenok L, Nochlin D, Matsushita M, Lipe H, Wolff J, Fernandez M, Cimino PJ, Bird TD, Raskind WH (2003) Missense mutations in the regulatory domain of PKCγ: a new mechanism for dominant nonepisodic cerebellar ataxia. Am J Hum Genet 72:839–849 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Estrach S, Schmidt S, Diriong S, Penna A, Blangy A, Fort P, Debant A (2002) The human Rho-GEF trio and its target GTPase RhoG are involved in the NGF pathway, leading to neurite outgrowth. Curr Biol 12:307–312 [DOI] [PubMed] [Google Scholar]
  5. Flanigan K, Gardner K, Alderson K, Galster B, Otterud B, Leppert MF, Kaplan C, Ptacek LJ (1996) Autosomal dominant spinocerebellar ataxia with sensory axonal neuropathy (SCA4): clinical description and genetic localization to chromosome 16q22.1. Am J Hum Genet 59:392–399 [PMC free article] [PubMed] [Google Scholar]
  6. Fucini RV, Navarrete A, Vadakkan C, Lacomis L, Erdjument-Bromage H, Tempst P, Stamnes M (2000) Activated ADP-ribosylation factor assembles distinct pools of actin on Golgi membranes. J Biol Chem 275:18824–18829 [DOI] [PubMed] [Google Scholar]
  7. Ge WW, Leystra-Lantz C, Wen W, Strong MJ (2003) Selective loss of trans-acting instability determinants of neurofilament mRNA in amyotrophic lateral sclerosis spinal cord. J Biol Chem 278:26558–26563 [DOI] [PubMed] [Google Scholar]
  8. Godi A, Santone I, Pertile P, Devarajan P, Stabach PR, Morrow JS, DiTullio G, Polishchuk R, Petrucci TC, Luini A, DeMatteis MA (1998) ADP ribosylation factor regulates spectrin binding to the Golgi complex. Proc Natl Acad Sci USA 95:8607–8612 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Graham DI, Lantos PL (eds) (2002) Greenfield’s neuropathology, 7th ed. Arnold, London [Google Scholar]
  10. Hadano S, Hand CK, Osuga H, Yanagisawa Y, Otomo A, Devon RS, Miyamoto N, Showguchi-Miyata J, Okada Y, Singaraja R, Figlewicz DA, Kwiatkowski T, Hosler BA, Sagie T, Skaug J, Nasir J, Brown RH Jr, Scherer SW, Rouleau GA, Hayden MR, Ikeda JE (2001) A gene encoding a putative GTPase regulator is mutated in familial amyotrophic lateral sclerosis 2. Nat Genet 29:166–173 [DOI] [PubMed] [Google Scholar]
  11. Harding AE (1982) The clinical features and classification of the late onset autosomal dominant cerebellar ataxias: a study of 11 families, including descendants of “Drew family of Walworth.” Brain 105:1–28 [DOI] [PubMed] [Google Scholar]
  12. Hellenbroich Y, Bubel S, Pawlack H, Opitz S, Vieregge P, Schwinger E, Zuhlke C (2003) Refinement of the spinocerebellar ataxia type 4 locus in a large German family and exclusion of CAG repeat expansion in this region. J Neurol 250:668–671 [DOI] [PubMed] [Google Scholar]
  13. Hirano R, Takashima H, Okubo R, Tajima K, Okamoto Y, Ishida S, Tsuruta K, Arisato T, Arata H, Nakagawa M, Osame M, Arimura K (2004) Fine mapping of 16q-linked autosomal dominant cerebellar ataxia type III in Japanese families. Neurogenetics 5:215–221 [DOI] [PubMed] [Google Scholar]
  14. Holmes SE, O’Hearn EE, McInnis MG, Gorelick-Feldman DA, Kleiderlein JJ, Callahan C, Kwak NG, Ingersoll-Ashworth RG, Sherr M, Sumner AJ, Sharp AH, Ananth U, Seltzer WK, Boss MA, Vieria-Saecker AM, Epplen JT, Riess O, Ross CA, Margolis RL (1999) Expansion of a novel CAG trinucleotide repeat in the 5′ region of PPP2R2B is associated with SCA12. Nat Genet 23:391–392 [DOI] [PubMed] [Google Scholar]
  15. Hughes CP, Berg L, Danziger WL, Coben LA, Martin RL (1982) A new clinical scale for the staging of dementia. Br J Psychiatry 140:566–572 [DOI] [PubMed] [Google Scholar]
  16. Ishikawa K, Fujigasaki H, Saegusa H, Ohwada K, Fujita T, Iwamoto H, Komatsuzaki Y, Toru S, Toriyama H, Watanabe M, Ohkoshi N, Shoji S, Kanazawa I, Tanabe T, Mizusawa H (1999) Abundant expression and cytoplasmic aggregation of α1A voltage-dependent calcium channel protein associated with neurodegeneration in spinocerebellar ataxia type 6. Hum Mol Genet 8:1185–1193 [DOI] [PubMed] [Google Scholar]
  17. Ishikawa K, Nagaoka U, Takashima M, Yoshizawa K, Yoshizawa T, Ishikawa M, Yamawaki T, Shoji S, Mizusawa H (2000) A gene on SCA4 locus causes dominantly-inherited pure cerebellar ataxia. Neurology 54:1971–1975 [DOI] [PubMed] [Google Scholar]
  18. Ishikawa K, Tanaka H, Saito M, Ohkoshi N, Fujita T, Yoshizawa K, Ikeuchi T, Watanabe M, Hayashi A, Takiyama Y, Nishizawa M, Nakano I, Matsubayashi K, Miwa M, Shoji S, Kanazawa I, Tsuji S, Mizusawa H (1997) Japanese families with autosomal dominant pure cerebellar ataxia map to chromosome 19p13.1-p13.2 and are strongly associated with mild CAG expansions in the spinocerebellar ataxia type 6 gene in chromosome 19p13.1. Am J Hum Genet 61:336–346 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kobayashi K, Nakahori Y, Miyake M, Matsumura K, Kondo-Iida E, Nomura Y, Segawa M, Yoshioka M, Saito K, Osawa M, Hamano K, Sakakihara Y, Nonaka I, Nakagome Y, Kanazawa I, Nakamura Y, Tokunaga K, Toda T (1998) An ancient retrotransposal insertion causes Fukuyama-type congenital muscular dystrophy. Nature 394:388–392 [DOI] [PubMed] [Google Scholar]
  20. Koide R, Kobayashi S, Shimohata T, Ikeuchi T, Maruyama M, Saito M, Yamada M, Takahashi H, Tsuji S (1999) A neurological disease caused by an expanded CAG trinucleotide repeat in the TATA-binding protein gene: a new polyglutamine disease? Hum Mol Genet 8:2047–2053 [DOI] [PubMed] [Google Scholar]
  21. Kollmar R (2001) Who does the hair cell’s 'do? Rho GTPases and hair-bundle morphogenesis. Curr Opin Neurobiol 9:394–398 [DOI] [PubMed] [Google Scholar]
  22. Koob MD, Moseley ML, Schut LJ, Benzow KA, Bird TD, Day JW, Ranum LP (1999) An untranslated CTG expansion causes a novel form of spinocerebellar ataxia (SCA8). Nat Genet 21:379–384 [DOI] [PubMed] [Google Scholar]
  23. Kutsche K, Yntema H, Brandt A, Jantke I, Nothwang HG, Orth U, Boavida MG, David D, Chelly J, Fryns J-P, Moraine C, Ropers H-H, Hamel BCJ, van Bokhoven H, Gal A (2000) Mutations in ARHGEF6, encoding a guanine nucleotide exchange factor for Rho GTPases, in patients with X-linked mental retardation. Nat Genet 26:247–250 [DOI] [PubMed] [Google Scholar]
  24. Lemmon MA, Ferguson KM, Abrams CS (2002) Pleckstrin homology domains and cytoskeleton. FEBS Lett 513:71–76 [DOI] [PubMed] [Google Scholar]
  25. Li M, Ishikawa K, Toru S, Tomimitsu H, Takashima M, Goto J, Takiyama Y, Sasaki H, Imoto I, Inazawa J, Toda T, Kanazawa I, Mizusawa H (2003) Physical map and haplotype analysis of 16q-linked autosomal dominant cerebellar ataxia (ADCA) type III in Japan. J Hum Genet 48:111–118 [DOI] [PubMed] [Google Scholar]
  26. Margolis RL (2002) The spinocerebellar ataxias: order emerges from chaos. Curr Neurol Neurosci Rep 2:447–456 [DOI] [PubMed] [Google Scholar]
  27. Matsuura T, Yamagata T, Burgess DL, Rasmussen A, Grewal RP, Watase K, Khajavi M, McCall AE, Davis CF, Zu L, Achari M, Pulst SM, Alonso E, Noebels JL, Nelson DL, Zoghbi HY, Ashizawa T (2000) Large expansion of the ATTCT pentanucleotide repeat in spinocerebellar ataxia type 10. Nat Genet 26:191–194 [DOI] [PubMed] [Google Scholar]
  28. Mburu P, Mustapha M, Varela A, Weil D, El-Amraoui A, Holme RH, Rump A, Hardisty RE, Blanchard S, Coimbra RS, Perfettini I, Parkinson N, Mallon A-M, Glenister P, Rogers MJ, Paige AJ, Moir L, Clay J, Rosenthal A, Liu XZ, Blanco G, Steel KP, Petit C, Brown SDM (2003) Defects in whirlin, a PDZ domain molecule involved in stereocilia elongation, cause deafness in the whirler mouse and families with DFNB31. Nat Genet 34:421–428 [DOI] [PubMed] [Google Scholar]
  29. Owada K, Ishikawa K, Toru S, Ishida G, Gomyoda M, Tao O, Noguchi Y, Kitamura K, Kondo I, Noguchi E, Arinami T, Mizusawa H. A clinical, genetic and neuropathologic study in a family with 16q-linked ADCA type III. Neurology (in press) [DOI] [PubMed] [Google Scholar]
  30. Pasteris NG, Cadle A, Logie LJ, Porteous ME, Schwartz CE, Stevenson RE, Glover TW, Wilroy RS, Gorski JL (1994) Isolation and characterization of the faciogenital dysplasia (Aarskog-Scott syndrome) gene: a putative Rho/Rac guanine nucleotide exchange factor. Cell 79:669–678 [DOI] [PubMed] [Google Scholar]
  31. Ross CA, Poirier MA (2004) Protein aggregation and neurodegenerative disease. Nat Med Suppl 10:S10–S17 [DOI] [PubMed] [Google Scholar]
  32. Rossman KL, Der CJ, Sondek J (2005) GEF means go: turning on RHO GTPases with guanine nucleotide-exchange factors. Nat Rev Mol Cell Biol 6:167–180 [DOI] [PubMed] [Google Scholar]
  33. Sasaki H, Yabe I, Tashiro K (2003) The hereditary spinocerebellar ataxias in Japan. Cytogenet Genome Res 100:198–205 [DOI] [PubMed] [Google Scholar]
  34. Sgourou A, Routledge S, Antoniou M, Papachatzopoulou A, Psiouri L, Athanassiadou A (2004) Thalassaemia mutations within the 5′ UTR of the human β-globin gene disrupt transcription. Br J Haematol 124:828–835 [DOI] [PubMed] [Google Scholar]
  35. Stevanin G, Bouslam N, Thobois S, Azzedine H, Ravaux L, Boland A, Schalling M, Broussole E, Durr A, Brice A (2004) Spinocerebellar ataxia with sensory neuropathy (SCA25) maps to chromosome 2p. Ann Neurol 55:97–104 [DOI] [PubMed] [Google Scholar]
  36. Stevanin G, Durr A, Brice A (2000) Clinical and molecular advances in autosomal dominant cerebellar ataxias: from genotype to phenotype and physiopathology. Eur J Hum Genet 8:4–18 [DOI] [PubMed] [Google Scholar]
  37. Tsuiki T, Sasamori S, Minami Y, Ichinohe T, Murai K, Murai S, Kawashima H (2002) [Age effect on hearing: a study on Japanese.] Audiol Japan 45:241–250 [Google Scholar]
  38. Tsunemi T, Saegusa H, Ishikawa K, Nagayama S, Murakoshi T, Mizusawa H, Tanabe T (2002) Novel Cav2.1 splice variants isolated from Purkinje cells do not generate P-type Ca2+ current. J Biol Chem 277:7214–7221 [DOI] [PubMed] [Google Scholar]
  39. van Swieten JC, Brusse E, de Graaf BM, Krieger E, van de Graaf R, de Koning I, Maat-Kievit A, Leegwater P, Dooijes D, Oostra BA, Heutink P (2003) A mutation in the fibroblast growth factor 14 gene is associated with autosomal dominant cerebellar ataxia. Am J Hum Genet 72:191–199 (erratum 72:1078) [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Verhoeven K, De Jonghe P, Van de Putte T, Nelis E, Zwijsen A, Verpoorten N, De Vriendt E, Jacobs A, Van Gerwen V, Francis A, Ceuterick C, Huylebroeck D, Timmerman V (2003) Slowed conduction and thin myelination of peripheral nerves associated with mutant Rho guanine-nucleotide exchange factor 10. Am J Hum Genet 73:926–932 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Yang Y, Hentati A, Deng HX, Dabbagh O, Sasaki T, Hirano M, Hung WY, Ouahchi K, Yan J, Azim AC, Cole N, Gascon G, Yagmour A, Ben-Hamida M, Pericak-Vance M, Hentati F, Siddique T (2001) The gene encoding alsin, a protein with three guanine-nucleotide exchange factor domains, is mutated in a form of recessive amyotrophic lateral sclerosis. Nat Genet 29:160–165 [DOI] [PubMed] [Google Scholar]
  42. Yu GY, Howell MJ, Roller MJ, Xie TD, Gomez CM (2005) Spinocerebellar ataxia type 26 maps to chromosome 19p13.3 adjacent to SCA6. Ann Neurol 57:349–354 [DOI] [PubMed] [Google Scholar]
  43. Zhu M, Yang T, Wei S, DeWan AT, Morell RJ, Elfenbein JL, Fisher RA, Leal SM, Smith RJH, Friderici KH (2003) Mutations in the γ-actin gene (ACTG1) are associated with dominant progressive deafness (DFNA20/26). Am J Hum Genet 73:1082–1091 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Zhuchenko O, Bailey J, Bonnen P, Ashizawa T, Stockton DW, Amos C, Dobyns WB, Subramony SH, Zoghbi HY, Lee CC (1997) Autosomal dominant cerebellar ataxia (SCA6) associated with small polyglutamine expansions in the α1A-voltage-dependent calcium channel. Nat Genet 15:62–69 [DOI] [PubMed] [Google Scholar]

Articles from American Journal of Human Genetics are provided here courtesy of American Society of Human Genetics

RESOURCES