Spinocerebellar ataxia type 1 (SCA1) is a neurodegenerative disorder caused by an expanded and unstable (CAG)>40 repeat within a gene of unknown function. We isolated the complete coding region of the rat SCA1 gene (rsca1), the 5'-untranslated region (UTR) and 1.3 kb of the 3'-UTR. The rat sequence exhibits 90% peptide identity to the human counterpart. In comparison to human, the rat (CAG)n block is reduced to two trinucleotide motifs preceded by three different proline codons not present in man. Furthermore, we investigated the expression of rsca1 in different rat tissues. The rsca1 gene is predominantly expressed in brain throughout all developmental stages. In situ hybridizations reveal high levels of expression in various regions of the adult rat brain, including cerebellum, hippocampus and cortex.
The spinocerebellar ataxias (SCA) are a group of clinically, pathologically and genetically heterogenous neurodegenerative disorders. Besides cerebellar manifestations, dementia, pyramidal signs, and ocular motor and spinal cord dysfunction are also present. Usually symptoms develop in the third to fourth decade, and the uncurable disease progressively worsens to cause death within the next 10-20 years. The SCAs are characterized by varying combinations of neuronal cell death in the cerebellar cortex, cerebellar nuclei, brainstem, spinal cord and peripheral nerves.
Autosomal dominant spinocerebellar ataxias can be further divided into different types. Yakura and colleagues (1 ) suggested that SCA maps to the short arm of chromosome 6. The 6p-linked form was subsequently designated as SCA1. Further analysis demonstrated genetic heterogeneity among the dominant ataxias. There is evidence for linkage to chromosome 12q (SCA2; 2), chromosome 14q (SCA3/MJD; 3), chromosome 16q (SCA4; 4), and chromosome 11 (SCA5; 5). Furthermore, ataxia with macular dystrophy maps to chromosome 3 (6 ). However, thus far the disease genes have only been identified for SCA1 and SCA3/MJD. For both ataxias the underlying mutation has been identified as an unstable and expanded (CAG)n repeat (7 ,8 ).
The SCA1 gene has been shown to encode a transcript of 10 660 bp (9 ). The disease causing (CAG)n repeat lies within the coding region and is translated into a polyglutamine tract (10 ). Nothing is known about the function of ataxin-1 nor about its pathobiochemical dysfunction in ataxia patients causing neurodegeneration. The SCA1 gene product, ataxin-1, is a novel protein which does not share homologies with any previously identified molecules. To define regions of potential functional importance, we have cloned and sequenced major parts of the rat SCA1 cDNA (rsca1). Comparison of the deduced protein sequences between human and rat define regions of strong evolutionary conservation. Detailed in situ hybridization and Northern blot analysis demonstrate predominant rsca1 expression in neuronal tissues.
The rat cDNA clones, rsca1-1, rsca1-2 and rsca1-3 (Fig. 1 ), were identified using the human SCA1 cDNA clone 31-5 as hybridization probe by screening a rat hippocampus cDNA library. Additional cDNA clones resulted from repeated screenings of rat whole brain, substantia nigra and hippocampus cDNA libraries with derived rsca1 clones (see Fig. 1 ). The clones obtained cover the human SCA1 cDNA from position 1 to 4621. In total, 16 clones were isolated and sequenced. Of these, four clone pairs (rsca1-5/rsca1-6, rsca1-13/rsca1-16, rsca1-4/rsca1-7, and rsca1-2/rsca1-17) contain each identical sequences. Three clones (rsca1-2, rsca1-11 and rsca1-13) contain additional DNA sequences, not corresponding to the SCA1 gene. These clones are either chimeric, retain introns or represent parts of alternatively spliced exons. The sequence of one clone (rsca1-2) exceeds the described human 5'-untranslated region (UTR) by 905 bp. In total, the length of the rat cDNA amounts to 5588 bp sequence. Repeated screening of additional cDNA libraries using the 3' end-fragment of cDNA rsca1-8, however, did not reveal any additional cDNA clones probably due to limitations of the cDNA libraries.
The predicted protein sequence of the rat gene shows about 90% identity to the human SCA1 sequence (Fig. 2 2). There are regions in the protein sequences of the two species where the degree of similarity is especially high; for example in a stretch of 199 aa between aa # 481 and 679 only one exchange is found. It has been postulated that the first in-frame ATG codon at position 936 of the human sequence represents the translation start site (9 ). At the same position (corresponding to position 1895) an in-frame ATG is also present in the rat. The in-frame TAA stop codon that has been found 57 bp upstream of the start codon in the human SCA1 gene has also been identified in the rsca1 gene at the same position. The open reading frame ends with a TAG stop codon at base 4262. The rsca1 gene is therefore predicted to encode a protein, rat ataxin-1, of 789 amino acids resulting in a slightly smaller protein than in the human (816 aa). Homology searches using both the rat DNA sequence of the coding region and the predicted protein sequence revealed no significant homology with other known nucleotide sequences in the EMBL data base.
Northern blot analysis revealed that, as in humans, a single transcript of approximately 10.6 kb of the rsca1 gene is expressed in rat tissues (Fig. 3 ). The rsca1 gene is identified at high levels in newborn (Fig. 3 , lanes a and b) and adult rat brain (Fig. 3 , lane e). The level of rsca1 expression in the brain appears to be slightly reduced in adult animals (Fig. 3 , lanes a to e). In all investigated body tissues (heart, kidney, thymus, lung, skeletal muscle) a single 10.6 kb transcript has been detected although at significantly lower levels than in the brain (Fig. 3 , lanes f to k).
We have cloned and analyzed the complete coding region of the SCA1 gene of the rat. DNA sequence analysis revealed a high degree of similarity at the aa level with the human sequence (identity 90%). The (CAG)n repeat in the rsca1 gene in rat, however, is markedly reduced to two repeat units compared to an average of 30 repeats in the normal population (12 ,14 ). Reduced (CAG)n repeat lengths are also found in the rodent hd and drpla genes (15 -20 ). This might be one of the reasons that naturally occuring animal models with large (CAG)n stretches are not known for CAG trinucleotide repeat disorders. In all cases, the (CAG)n repeat appears to be localized within the coding regions which codes for glutamine in the protein (10 ,21 ,22 ). The short glutamine stretches in the normal alleles of rodents suggest that a large number of glutamines are not essential to the physiological function of these proteins in the animals. Recently, a transgenic SCA1 mouse model harboring an expanded CAG tract has been created presenting as well with neuronal cell loss of Purkinje cells as with the neurological phenotype of ataxia (23 ). These studies provide evidence that the pathology of trinucleotide disorders is caused by a gain-of-function of the abnormal proteins and not simply the result of a loss-of-function mutation.
To investigate the rsca1 gene in more detail we analyzed its expression using Northern blot and in situ hybridization. The rsca1 gene is predominantly expressed in neuronal tissues although faint hybridization signals exist in all investigated non-neuronal tissues (Fig. 3 ). However, by analyzing equal amounts of total RNA the intensity is reduced more than ten-fold compared to brain (Fig. 3 ). Our in situ hybridization studies confirm that rsca1 mRNA is widely expressed in the brain. The pyramidal cell layer within the hippocampal region is the most prominent site of expression which is concordant with the protein studies (10 ). An intriguing finding is that in the Purkinje cells of the cerebellum, which degenerate in SCA1 (24 ), the mRNA is significantly expressed. The cerebellum is associated with movement coordination, and with learning and memory of motor tasks (25 ). Dysfunction of the cerebellum leads to unbalanced coordination of movements due to irregularities in the timing of onset, rate and force of contraction of synergistic muscle groups (26 ). Thus, impaired function may relate to ataxic symptoms seen in SCA1 patients. Furthermore, in SCA1 patients reduced numbers of neurons are also found in the pontine nuclei (27 ), which are involved in motor coordination. Neuronal cell loss of cranial nerve nuclei found in SCA1 can be correlated with specific clinical features, as ocular motor findings, dysphagia, tongue atrophy and fasciculations (26 ). All of these structures show marked expression of the rsca1 transcript in rat. However, high levels of the rsca1 transcript might not necessarily be related to the clinical phenotype in humans. For instance, predominant expression of rsca1 is found in the bulbus olfactorius, the hippocampus, the cerebral cortex, as well as in the thalamic nuclei but neuronal loss has not been observed in these regions in SCA1 patients (26 ). This raises further questions on the function of ataxin-1 and on the pathobiochemical impact of expanded polyglutamine tracts leading to selective neuronal cell death of specific regions of the brain.
The human cDNA clone 31-5 (kindly provided by H. Zoghbi) covering the entire coding region of the SCA1 gene (9 ) was used as initial probe to screen rat [lambda] ZAPII cDNA libraries constructed from total brain, hippocampus, and substantia nigra RNA. Prehybridization and hybridization were performed in Church buffer at 60oC (28 ). Bluescript phagemids were released from [lambda] ZAPII by in vivo excision utilizing R408 as helper phage (29 ). Using these cDNAs as homologous probes, cDNA libraries were hybridized under stringent conditions at 65oC. Washing procedures included final stringent washes in 0.1 * SSC, 0.1% SDS at 60oC (for heterologous probes) and 65oC (for homologous probes), respectively. Plasmid DNA was prepared as described (30 ). Sequencing was carried out by manual dideoxy sequencing using the sequenase kit (USB).
Rat tissues were homogenized (Ultra Turrax) in guanidinium isothiocyanate and RNA was isolated by centrifugation through a CsCl2 density gradient. A quantity of 10-20 µg of total RNA were separated electrophoretically on an agarose gel (1% agarose, 20% formaldehyde, 1*GRB buffer [20 mM MOPS pH 7.0, 5 mM NaCOOH, 1 mM EDTA]) and RNA was electroblotted for at least 5 h at 700 mA and 4oC (blotting buffer: 25 mM Na2HPO4/NaH2PO4 pH 6.5) onto Hybond-N membranes (Amersham). Methylenblue staining was performed with 0.04% methylenblue in 0.5 M sodium acetate buffer (pH 5.5) for 15 min and final washing in distilled water. Hybridization was performed in Church buffer with 32P-radiolabelled rat cDNA probes rsca1-1 or rsca1-3 at 65oC. For quantitative analysis the filters were rehybridized with a 32P-labelled glyceraldehyde-3-phosphate-dehydrogenase (Gapd; 31) cDNA insert (Fig. 3 ) or [beta]-actin (data not shown).
We are especially grateful to Dr Boulter for providing several rat cDNA libraries and H.Y. Zoghbi for the human SCA1 cDNA probes. We thank M. Kilimann for providing adult rats, W. Mäueler for synthesizing oligodeoxynucleotides, M. Bronzel for assistance on performing in situ hybridizations, K. Meller for help on microscopy, B. Wenzel and E. Günther for cDNA libraries, and Ch. Plehn for photography. The support by members of our laboratories is kindly acknowledged. This article contains parts of the `Diplom' thesis of M.G.
jnl.info{at}oup.co.uk
Human Molecular Genetics
Pages
Introduction
Results
Cloning of the rsca1 cDNA sequence
Analysis of the predicted amino acid (aa) sequence
Northern blot analyses
In situ hybridization
Discussion
Materials And Methods
Screening of cDNA libraries and DNA sequence analysis
RNA preparation and Northern blotting
RNA in situ hybridization
Acknowledgements
References
In situ hybridization of sagittal adult rat brain sections with antisense probes synthesized from clones rsca1-3 and rsca1-14 revealed positive signals in most examined regions, although labeling intensities varied (Fig. 4 4A). In the cerebral cortex, the pyramidal neurons of layers II, III and V were most prominently stained (Fig. 4 4B). In layer I, the interneurons were positive (Fig. 4 4B, C). The stellate cells of layer IV were weakly labeled, as were pyramidal cells of layer VI. In the caudate putamen, the medium-sized spiny stellate neurons were weakly labeled, as were striatal interneurons and cholinergic cells in the striatum (Fig. 4 4B) and the diagonal band of Broca. In the cerebellum, inhibitory Purkinje cells, basket cells, and stellate cells in the molecular layer were intensely labeled (Fig. 4 4F). Golgi cells in the granule cell layer were also labeled. Excitatory granule cells could not be detected as distinct cells due to their small size but were considered positive, when compared to the hybridization for the parvalbumin mRNA (Fig. 4 4G), which is not expressed in granule cells. In the hippocampus, rsca1 mRNA is heavily expressed in excitatory granule cells of the fascia dentata, in inhibitory interneurons in the molecular layer and the hilus region, as well as in excitatory pyramidal cells in the stratum pyramidale, and interneurons in the stratum oriens and stratum radiatum (Fig. 4 4J). In thalamus, the various nuclei were prominently stained (e.g. the laterodorsal thalamic nucleus, lateral posterior thalamic nucleus, and centrolateral thalamic nucleus). In contrast, only a few positive cells were found in the pretectal nuclei (Fig. 4 4K). In the bulbus olfactorius, mitral and external tufted cells, periglomerular cells, and granule cells showed the most intense signals (Fig. 4 4P), whereas interneurons of the external plexiform layer (Van Gehuchten cells) were not stained. A similar labeling pattern delineated the accessory olfactory bulb (Fig. 4 4A), while the anterior olfactory nuclei and the olfactory tubercle displayed labeled pyramidal neurons like the cerebral cortex (Fig. 4 4A). In the periaqueductal gray, groups of neurons were labeled, and most intensely the sensory neurons of the mesencephal trigeminal nucleus (Fig. 4 4N). In the brain stem distinct areas of rsca1 mRNA expressing cells were identified. They include pontine nuclei (Fig. 4 4L), motor neurons of the facial nucleus (Fig. 4 4M), the gigantocellular reticular nucleus, neurons in the pontine reticular formation, and neurons in the nucleus of the trapezoid body (Fig. 4 4O).
Probe generation. RNA riboprobes were generated with a DIG RNA Labeling Kit (Boehringer). Probes were synthesized as run-off transcripts from clones rsca1-3 and rsca1-14, both in antisense and sense direction, utilizing T3- and T7-RNA-Polymerase. The 0.24 kb parvalbumin cDNA (32 ) was transcribed in sense and antisense direction using T3 and T7 promotors. Probe quality was tested on a Northern blot and a test color reaction as described (33 ).Tissue preparation and hybridization. After anesthetizing (60 mg/kg body weight Nembutal i.p.), adult rats [Rattus norvegicus, strain NZR/Mh (34 )] were perfused with 0.9% NaCl in 0.05 M sodium phosphate pH 7.4 and 4% paraformaldehyde and tissues were prepared as described by Wahle (33 ). The brain was cut into 30 µm thick sagittal sections on a cryostat (Jung Frigocut 2800E, Leica) and, after 4 h of prehybridization, hybridized at 45oC overnight in prehybridization solution (33 ) with a mixture of both antisense probes (1:100 each) or a mixture of both sense probes (1:100 each, negative control). Washing was performed at 45oC for 20 min each step. After blocking with Blocking Reagent (Boehringer), anti-DIG-antibodies (Fab fragments, Boehringer) were applied in a 1:1000 dilution. Washing and overnight color reaction were carried out as described (33 ). Sections were mounted on gelatin-coated slides, airdried, coverslipped with Tris-glycerol and analyzed. Alternate sections were stained with thionine dye to identify brain structures according to the atlas of Paxinos and Watson (35 ). Photomicrographs were taken with a Zeiss Axiophot.
REFERENCES
This page is maintained by OUP admin. Last updated Thu Oct 31 15:22:29 GMT 1996. Part of the OUP Journals World Wide Web service.Copyright Oxford University Press, 1996