Autosomal recessive ataxias are a group of rare heterogeneous disorders often affecting both the central and peripheral nervous systems. Ataxias are generally characterised by loss of co-ordination due to damage to those regions of the nervous system that control movement and muscular function. Infantile autosomal recessive cerebellar ataxia (CLA3 or SCAR6, OMIM608029) was first described in 1985 in a Norwegian pedigree (Kvistad et al., 1985). The phenotype of this disorder includes defects in motor co-ordination of the limbs, slow speech, poor eye movement co-ordination, short stature, but without intellectual impairment (Kvistad et al., 1985). More recently, genome-wide homozygosity mapping resulted in the localisation of the gene in this large inbred pedigree to a 19.5-cM region on chromosome 20q11-q13 (Tranebjaerg et al., 2003). The highest evidence of linkage was found between the markers D20S195 and D20S884 at 20q11.2. To our knowledge, this is the only family showing localisation of an autosomal recessive ataxia locus to this chromosomal region. In a screen for genes involved in mammalian craniofacial development, we identified the mouse Tp53inp2/Trp53inp2 (tumor/transformation-related protein p53-induced nuclear protein 2) gene and showed that the human orthologue mapped to the same chromosomal region as the CLA3 locus (Bennetts et al., 2006).
TP53INP2 is an uncharacterised gene that is annotated based on limited homology to a known gene, TP53INP1. TP53INP1 encodes two alternative splice variant proteins, both of which are upregulated in response to cellular stress, interact with the tumor suppressor p53, and regulate p53-dependent apoptosis (Okamura et al., 2001; Tomasini et al., 2002, 2003). The TP53INP1 proteins have also been shown to induce cell cycle arrest by functional association with p73, a p53 homologue (Tomasini et al., 2005). TP53INP1 pro-apoptotic activity is also activated by E2F1 (Hershko et al., 2005), and down-regulation of TP53INP1 expression has been shown to be associated with the development of gastric cancers (Jiang et al., 2006).
Using whole mount in situ hybridisation analysis, we previously showed that the mouse Tp53inp2 gene exhibited a spatio-temporally restricted expression pattern at mouse embryonic day (E)10.5 (Bennetts et al., 2006). Here we report a bioinformatic analysis of the TP53INP2 protein and gene sequences, which suggests that unlike TP53INP1, this is a mammalian specific protein. We have also extended expression analysis using both whole mount and section in situ hybridisation at a range of mid-gestational stages in the mouse, and show expression of this gene primarily in regions of the embryo associated with the developing nervous system. Based on this expression data, combined with the localisation of the human TP53INP2 gene within the CLA3 candidate interval, we analysed the protein-coding region of TP53INP2 for mutations in samples from two affected individuals and one obligate carrier from the CLA3 pedigree previously described (Tranebjaerg et al., 2003). The lack of cosegregating mutations in patient samples indicates that TP53INP2 is unlikely to be the causative gene for the ataxia in this family, although we cannot rule out alterations in non-coding regions of the gene. Given its expression in the mouse, TP53INP2 remains a candidate for other disorders characterised by neuronal dysfunction.
Identification and Analysis of the TP53INP2 Gene and Protein Sequences
We first identified the Tp53inp2 gene from a subtracted cDNA library derived from mouse pharyngeal arch tissue (Fowles et al., 2003; Bennetts et al., 2006). At that time, it had not been annotated, nor had a putative full-length sequence been established in publicly available databases. The insert for the clone corresponding to Tp53inp2 was sequenced and used to BLAST search the GenBank database in order to determine homologies, and was shown to correspond to a 221 amino acid predicted mouse protein, now annotated as Tp53inp2/Trp53inp2 (NP_835212). Further database analysis has detected sequence orthologues in human (NP_067025, 220 aa), cow (XP_588659, 224 aa), rat (CAC82596, 199 aa), and dog (XP_852549, 221 aa), all with greater than 70% sequence identity at the amino acid level (Fig. 1). Additionally, database mining has allowed identification of an EST that represents a partial coding sequence for a likely orthologue in pig (S. scrofa BX672104). Despite extensive searching across public databases for corresponding Tp53inp2 transcripts in lower model organisms such as D. melanogaster, zebrafish, T. nigroviridis, and X. laevis, no orthologous sequences have been confidently predicted in non-mammalian species. Protein sequence analysis using PsortII predicts a nuclear localisation signal (R-R-X-K) at the C-terminus, and SMART predicts a coiled-coil region (Fig. 1). Neither of these predicted domains is found in the TP53INP1 proteins.
By sequence homology searches, TP53INP2 shows partial identity to the stress-induced proteins TP53INP1α and TP53INP1β, which are two nuclear proteins of 18 and 27 kiloDaltons (kDa) respectively, encoded by alternative splice variants of the same gene (Okamura et al., 2001; Tomasini et al., 2001). As shown for the human proteins, sequence homology is observed primarily over the C-terminus (with TP53INP1α) and across a central region (Fig. 2). These molecules do not exhibit any sequence homology to other known proteins, nor do they contain any conserved known protein domains. Unlike TP53INP2, which appears to be a mammalian-specific protein, analysis of sequences present in the GenBank database revealed orthologues for TP53INP1 in a range of lower organisms as well as mammals (data not shown). Alternative splice variants for TP53INP2 were not detected. Based on sequence homology, TP53INP2 may have a related role to those of the TP53INP1 proteins. However, the relatively short regions of homology may represent uncharacterised domains that are potentially not related to TP53INP1 function.
Expression Analysis of Tp53inp2 During Mouse Embryogenesis by Whole Mount In Situ Hybridisation
Our initial analysis by whole mount in situ hybridisation showed that Tp53inp2 is expressed in a restricted manner at E10.5 in the developing mouse embryo, including in components of the developing nervous system, the brain, dorsal root ganglia, and sympathetic chain ganglia, as well as in the pharyngeal arches and frontonasal prominence (Bennetts et al., 2006). We have now extended this by more detailed expression analysis at a range of mid-gestational embryonic stages, on both whole and sectioned mouse embryos, to determine and further characterise the spatio-temporal expression pattern of the Tp53inp2 gene.
At E9.5 and E10.5, Tp53inp2 expression appears to be associated with the neural crest—derived cranial nerve condensations, including the trigeminal (V), facial nerve (VII), glossopharyngeal (IX), and vagal nerve (X) ganglia (Fig. 3A,B,D). This staining is similar to that seen with Sox10, which is known to mark these ganglia (Fig. 3E,F) (Bondurand et al., 1998; Kuhlbrodt et al., 1998). Tp53inp2 expression was detected in neural processes including the dorsal root ganglia, sympathetic chain ganglia, and spinal nerves in the trunk at E9.5–12.5 (Fig. 3A,C,D, G,H,J,K). At E10.5–E14.5, Tp53inp2 was also detected in the developing fore-, mid-, and hindbrain (Fig. 3D,H,I; E13.5 and E14.5 not shown). At later stages including E13.5 and E14.5, staining of neural projections in the trunk was observed (Fig. 3L,M). Tp53inp2 expression was also observed in the otic vesicle at E9.5 and E10.5 (Fig. 3A,B,D). At E9.5–E11.5, Tp53inp2 expression was detected in the developing heart and diffusely in the facial primordia, including the frontonasal, maxillary, and mandibular prominences (Fig. 3A,D,H). At E10.5–E11.5, staining in the limb mesenchyme is diffuse (Fig. 3D,H), at E12.5 and E13.5 this becomes restricted to the interdigital mesenchyme (Fig. 3I,M), and at E14.5 expression is further restricted to sites of joint formation within the digits (Fig. 3N).
Expression Analysis of Tp53inp2 During Mouse Embryogenesis by In Situ Hybridisation of Sectioned Embryos
To investigate the expression of Tp53inp2 in more detail, in situ hybridisations were carried out on sectioned embryos at a range of embryonic stages. At E10.5, sagittal sections and transverse sections through the spinal cord revealed Tp53inp2 expression that appeared to mark both the anterior and posterior nerve roots (sensory; Fig. 4A,C), the dorsal root ganglia, and the postmitotic (differentiated) neurons of the mantle layer of the spinal cord (Fig. 4C). Immunohistochemistry with an anti-βIII-tubulin antibody, which is a marker for differentiated neurons, suggested an overlapping pattern of staining in some of these structures (Fig. 4B,D). Tp53inp2 staining in the spinal cord is particularly concentrated in the ventral region (Fig. 4C). At E10.5, expression within the brain can be seen in the developing neuroepithelium in regions where differentiated neurons reside (surrounding the fourth ventricle and aqueduct of Sylvius), and in the developing medulla (Fig. 4A,E). Expression can also be seen in the dorsal thalamus at E10.5 (Fig. 4E), particularly towards the differentiated region, and was detected in the fasciculus retroflexus (comprising axons that originate in the diencephalon and project ventro-caudally to terminate in the midbrain, and whose function is to convey autonomic signals) at E10.5 and E13.5 (Fig. 4E,F,J). At E10.5–E13.5, expression was detected in the trigeminal ganglion (data for E11.5 and E13.5 shown only; Fig. 4G,N), and at E11.5 expression can be seen in the opthalmic, maxillary, and mandibular divisions projecting into regions of the face, and in the facial nerve ganglion projecting to pharyngeal arch 2 (PA2) (Fig. 4G). At E10.5–E13.5, expression was detected in the spinal cord, dorsal root ganglia, and associated axons of the developing spinal nerves (data not shown for all stages; Fig. 4A,C,H,N). At E13.5, expression in the brain is restricted to the intermediate zone (differentiated neurons) of the developing neocortex surrounding the lateral ventricles as well as the diencephalon and the caudal part of the mesencephalon, the intermediate zone of the pons, the medulla, the cerebellum, and the corpus striatum (Fig. 4J,N,O). Also, expression can be seen in the facial nerve nucleus and pyramidal tract at E13.5 (Fig. 4K,N,O). At E13.5, expression was detected in the muscle mass of the tongue and axons in the craniofacial region (Fig. 4K), in a similar pattern to the neural marker Sox10 (Fig. 4L). When compared with the expression domain of Sox10, which is expressed in the Schwann cells (myelinating oligodendrocytes) surrounding neurons of the peripheral nervous system (Fig. 4I,M) (Bondurand et al., 1998; Kuhlbrodt et al., 1998), it appears that expression of Tp53inp2 may be within or adjacent to subgroups of sensory and motor neurons and their projected axonal tracts within the central and peripheral nervous system (Fig. 4H,N).
The Human TP53INP2 Gene Localises to Chromosome 20q11.2
Our database analysis mapped the human TP53INP2 gene to chromosome 20q11.2, consistent with recent localisation to 20q11.2 by fluorescent in situ hybridisation (FISH) analysis (Nowak et al., 2005). Genome-wide homozygosity mapping has localised the gene for an autosomal recessive ataxia (CLA3 or SCAR6) in an inbred Norwegian pedigree to a region on human chromosome 20q11-q13 (Tranebjaerg et al., 2003). The maximum evidence of linkage was found between the flanking markers D20S195 and D20S884 at 20q11.2. With more detailed investigation, we have established that TP53INP2 localises to the human genomic contig NT_028392, and resides within the region showing the highest linkage to CLA3 (Fig. 5). The TP53INP2 gene consists of five exons, three of which contain coding sequence for the TP53INP2 protein (Fig. 2B). The TP53INP2 transcript is approximately 4 Kb in length and encodes a protein with a predicted molecular mass of 25 kDa.
Mutation Screening of TP53INP2 in CLA3 Patient Samples
Cerebellar ataxia is characterised by motor neuron deficits, and Tp53inp2 is expressed in regions associated with neurons throughout the mouse embryo, including the plexi supplying motor neurons to the limbs. In addition, Tp53inp2 is expressed in the cerebellum, the region of the brain most likely to be affected in ataxic individuals. Based on the correlation between phenotype and our expression data, combined with chromosomal localisation, we considered TP53INP2 an excellent candidate for the CLA3 locus. We, therefore, searched the coding region for causative mutations in DNA samples from two affected individuals (VIII:2 and IX:1) and one obligate carrier (VI:5) from the one known CLA3 pedigree described in Tranebjaerg et al. (2003). The exons and intron-exon boundaries of TP53INP2 were amplified from each sample and sequenced. However, no putative disease-causing mutations were identified in TP53INP2 coding regions. In the exon-flanking non-coding regions, a change, g.IVS4-20G>A, was found. Both patient samples are homozygous for this change, while the obligate carrier sample is heterozygous. While this pattern is consistent with a disease-causing mutation, it also fits with a neutral polymorphism that is tightly associated with the disease locus based on their proximity. In this case, the obligate carrier is heterozygous for the allele associated with the disease haplotype as expected, but carries the reference allele on the remaining chromosome. Given the position and nature of this intronic nucleotide change, it is unlikely to be associated with disease in this family, although this cannot be ruled out definitively at this stage.
Our data suggest that mutations in the coding region of TP53INP2 are unlikely to be responsible for the ataxia phenotype in this family in which the gene has been mapped to chromosome 20.
Tp53inp2 exhibits a distinct spatio-temporal expression pattern in the developing mouse embryo, and by whole mount and section in situ hybridisation analysis appears to be restricted primarily to regions of the developing brain and spinal cord, as well as to sensory and motor neuron tracts of the peripheral nervous system. At later stages, expression in the brain and spinal cord is restricted to the region harbouring differentiated post-mitotic neurons. The expression of Tp53inp2 in some regions of the developing nervous system suggests that this gene may be involved in certain aspects of neuronal development. Expression associated with neural crest–derived ganglia is similar to that seen for the neural crest marker Sox10. Additionally, the expression pattern observed for Tp53inp2 overlaps in some regions with that of the known neuronal protein βIII-tubulin, which marks terminally differentiated neurons. Elucidation of any role of Tp53inp2 in neural development awaits further functional characterisation. However, the fact that we did not detect ubiquitous expression of Tp53inp2 in or adjacent to all neurons suggests that this gene is unlikely to be involved in global neuronal development, but may play more specific and restricted roles. Expression in other regions of the embryo including the limb and facial prominences suggests that the function of Tp53inp2 is unlikely to be confined to neural development.
TP53INP2 is highly conserved at the amino acid level between orthologues identified in mammalian species, although no orthologues have been identified in lower organisms including vertebrates such as zebrafish, or D. melanogaster. This suggests that TP53INP2 may have a novel function in mammalian neurogenesis. TP53INP2 is predicted to contain a coiled-coil and proline-rich region, which suggests a function in protein–protein interactions (Kay et al., 2000), and a C-terminal nuclear localisation signal is also predicted. TP53INP2 exhibits some sequence homology to the stress-induced proteins, TP53INP1α and -β, which are known to be upregulated in response to p53 and are involved in cell cycle control. Cell cycle control is important for neurogenesis, given that for cells to differentiate into mature neurons, they must exit the cell cycle. Certain molecules including p53, Cdk5, and Phox2b control aspects of this process including cell cycle exit and maintenance of a differentiated state (Pattyn et al., 1997; Dubreuil et al., 2002; Miller et al., 2003; Cicero and Herrup, 2005). Whether TP53INP2 is a functional orthologue of TP53INP1 and interacts with cell cycle regulators such as p53 still remains to be investigated. Alternatively, the relatively small regions of homology between the TP53INP1 and TP53INP2 proteins may merely be indicative of conserved novel domains unrelated to the known roles of the TP53INP1 protein isoforms.
Based on the neuronal associated expression of Tp53inp2 in the mouse embryo along with the localisation of the human gene within the CLA3 candidate interval, we considered it a candidate for the causative gene in the family with linkage to chromosome 20. Ataxias often result from damage to the cerebellum and spinal cord, both regions of Tp53inp2 expression in the mouse. The ataxias display marked heterogeneity, and one group, including Ataxia telangiectasia, is due to defects in DNA repair (Savitsky et al., 1995a, b). Given the association of the TP53INP1 genes with p53, which is known to function in DNA repair (reviewed in Sengupta and Harris, 2005), we further considered TP53INP2 a good candidate for causing CLA3. Our failure to detect mutations in the coding region of TP53INP2 in patients from the known CLA3 pedigree implies that this is unlikely to be the causative gene, although the possibility remains that mutations may be present in intronic or regulatory regions not screened in this study.
Through our expression analysis of the Tp53inp2 gene in the mouse embryo, we have uncovered a potential role for this uncharacterised mammalian gene in the function or development of neurons. While the TP53INP2 gene is unlikely to be involved in the ataxia phenotype associated with the CLA3 locus on chromosome 20, it remains an excellent candidate for other neurological disorders. Further analysis of this gene is required to elucidate its precise function and role in neural development.
Whole Mount In Situ Hybridisation
All animal work was conducted at the Institute for Molecular Bioscience, The University of Queensland, according to ethical guidelines and was approved by the relevant UQ Animal Ethics Committee. The Tp53inp2 probe used for in situ hybridization experiments was synthesized from a plasmid containing 485 bp of 3′UTR sequence (nt 2516-3000; NM_178111). Whole mount in situ hybridisation was performed as described previously (Christiansen et al., 1995; Fowles et al., 2003). Briefly, embryos were dissected from pregnant CD1 mice at a range of mid-gestational stages and were fixed overnight in 4% paraformaldehyde/PBS at 4°C, dehydrated, and then rehydrated through a methanol series. Embryos were placed in prehybridisation buffer and incubated overnight at 65°C. Digoxigenin (DIG)-labelled riboprobes were transcribed from linearised clones using T7 or SP6 polymerase and added to prehybridised embryos at a concentration of 0.2–1.0 μg/ml. Both antisense and sense probes were examined. After overnight hybridisation, embryos were washed, blocked, and incubated with anti-DIG antibody conjugated to alkaline phosphatase (Roche Diagnostics). After a further round of washing, embryos were incubated with the colour reagents nitro blue tetrazolium (Roche Diagnostics) and 5-bromo-4-chloro-3-indolylphosphate (Roche Diagnostics) until satisfactory colour development was achieved. Background was reduced by washing in PBS containing 1% Triton X-100 at 4°C for one to three days and the colour then fixed with 4% paraformaldehyde in PBS containing 0.1% Triton X-100 overnight at 4°C. Embryo images were captured digitally using either a MZ8 dissecting microscope (Leica Microsystems, Wetzlar, Germany) with a SPOT camera and SPOT 3.2.6 software (Diagnostic Instruments, Sterling Heights, MI), or an Olympus SZX12 microscope with DP Controller software (Olympus Corporation).
In Situ Hybridisation of Sectioned Mouse Tissue
CD1 mouse embryos were paraffin embedded and sectioned at a thickness of 10–14 μm. In situ hybridisations were performed as previously described (Schaeren-Wiemers and Gerfin-Moser, 1993). Images were captured using an Olympus SZX12 or BX51 microscope with DP Controller software (Olympus Corporation).
Paraffin-embedded mouse embryo tissue was sectioned at a thickness of 10-14 μm. The sections were mounted onto Superfrost Plus slides, which were incubated with anti-βIII-tubulin monoclonal antibody (Promega, Madison WI) diluted at 1:2,000 in PBS with 2% BSA. Slides were washed in PBS, and then incubated with biotinylated anti-mouse antibody (Vector Laboratories, Burlingame, CA) diluted in PBS with 2% BSA, then washed in PBS. Slides were then incubated with VECTASTAIN® ABC reagent (Vector Laboratories, Burlingame, CA) as per the manufacturer's instructions. Slides were washed in PBS and incubated with 3,3′-diaminobenzidine substrate (Vector Laboratories) for 2–10 min, then mounted. Images were captured using an Olympus SZX12 or BX51 microscope with DP Controller software (Olympus Corporation).
Full-length protein and domain sequence (amino acid) multiple alignments were constructed using the ClustalW tool (Thompson et al., 1994) within MacVector 7.2.2 (Accelrys, San Diego, CA). The parameters used for all multiple alignments were an open gap penalty of 10.0–25.0, a gap extension penalty of 0.30, and a gap separation penalty range of 8, and all were aligned using the Dayhoff PAM matrix.
PCR Amplification and Sequencing Analysis of TP53INP2
All mutation searching of human DNA samples was conducted at the Department of Medical Genetics, University Hospital of Tromso, Norway, and Wilhelm Johannsen Centre for Functional Genomics, University of Copenhagen, with approval from the relevant institutional human ethics committee. Genomic DNA was prepared by standard methods from peripheral blood samples from members of the previously described family with infantile autosomal recessively inherited ataxia (Tranebjaerg et al. 2003). Samples from two affected individuals (VIII:2 and IX:1) and one obligate carrier (VI:5) were analysed.
Primers for PCR amplification and sequencing of the TP53INP2 gene were designed using TP53INP2 mRNA sequence (GenBank accession no. NM_021202.1) aligned with chromosome 20 genomic sequence available from the March 2006 human reference sequence in the UCSC Genome Browser (http://genome.ucsc.edu/cgi-bin/hgGateway). The three coding exons and flanking intron sequences of TP53INP2 were amplified as one PCR fragment using Platinum Taq DNA Polymerase (Invitrogen) and the following primers: TP53INP2F: 5′-CCATAGGGCGCCCCCGTG-3′ and TP53INP2R: 5′-CGGCCGGTGGACGCTCAGTAG-3′. Amplifications were performed in a Tetrad 2 thermal cycler (Bio-Rad, Richmond, CA) in 15-μl reactions containing 50–100 ng DNA, 10 pmol of each primer, 0.4 mM of each dNTP, 1 × PCR buffer, 15 mM MgSO4, and 0.5 U Platinum Taq Polymerase, and using the following cycling parameters: 5 min at 98°C, 20 sec at 98°C, 20 sec at 65°C, 1 min at 68°C for 40 cycles followed by 5 min at 68°C. The long-range PCR product was then used as the template for exon by exon sequencing using Big Dye Terminator™ Cycle Sequencing Ready Reaction Kit according to the manufacturer's instructions.
The authors thank Lindsay Fowles for initiating the craniofacial screen. Thanks also to Anne Hardacre and staff of the Queensland Biosciences Precinct (The University of Queensland) animal house for help with mouse husbandry. J.S.B. was the recipient of an Australian Postgraduate Award and C.W. is an Australian National Health and Medical Research Council (NHMRC) Senior Research Fellow. Part of this work was supported by an NHMRC project grant (C.W.). The IMB incorporates the Centre for Functional and Applied Genomics, a Special Research Centre of the Australian Research Council.