- Split View
-
Views
-
Cite
Cite
Manuel Pelé, Laurent Tiret, Jean-Louis Kessler, Stéphane Blot, Jean-Jacques Panthier, SINE exonic insertion in the PTPLA gene leads to multiple splicing defects and segregates with the autosomal recessive centronuclear myopathy in dogs, Human Molecular Genetics, Volume 14, Issue 11, 1 June 2005, Pages 1417–1427, https://doi.org/10.1093/hmg/ddi151
- Share Icon Share
Abstract
Human centronuclear and myotubular myopathies belong to a genetically heterogeneous nosological group with clinical variability ranging from fatal disorder to mild weakness. The severe X-linked form is attributed to more than 200 different mutations in the myotubularin encoding gene (MTM1). In contrast, there are no reports regarding the molecular etiology or linkage studies on the autosomal forms of the disease. Labrador retrievers affected by spontaneous centronuclear myopathy (cnm) have clinical and histological features of the human disorder and represent the first model of recessive autosomal centronuclear myopathy. We previously mapped the cnm locus to the centromeric region of canine chromosome 2. No gene of the MTM1 family maps to the human homologous chromosomal region. Described herein is a disease-associated insertion within PTPLA exon 2, found in both alleles of all affected Labradors and in a single allele in obligate carriers. The inserted tRNA-derived short interspersed repeat element (SINE) has a striking effect on the maturation of PTPLA mRNA, whereby it can be spliced out, partially exonized or involved in multiple exon-skipping. As a result, the amount of wild-type transcripts falls to 1% in affected muscles. This example therefore recapitulates cumulative SINE-associated transcriptional defects that have been previously described as exclusive consequences of independent mutations. Although the function of PTPLA in metazoa remains unknown, the characterization of a hypomorphic mutation in Labradors with centronuclear myopathy provides new clues about the molecular complexity of skeletal myofiber homeostasis. These results also suggest that impaired PTPLA signaling might be implicated in human myopathies.
INTRODUCTION
The nosological group of human centronuclear myopathy (OMIM no. 160150) is defined by heterogeneous forms of inherited muscular disorders that share common clinical and histological features. Hallmarks of the disease include generalized muscle weakness, ptosis, ophthalmoplegia externa, areflexia, muscular atrophy affecting predominantly type 1 myofibers, nuclei centralization and pale central zones with variably staining granules (1–3). Genetic heterogeneity is correlated with the onset and severity of the disease. A significant proportion of infants affected by the severe X-linked form (called myotubular myopathy or XLMTM, OMIM no. 310400) die at a few months of age following respiratory failure (4,5). Eighty percent of survivors become dependent on artificial ventilation and suffer from medical complications (5). Autosomal dominant and recessive forms (OMIM no. 255200) run a milder course, the recessive form being intermediate between the X-linked and the dominant forms (2,6). The XLMTM locus was initially mapped to Xq28 (7) and mutations in the Myotubularin (MTM1) gene were associated with the disease (8). To date, a spectrum of ∼200 disease-associated mutations in MTM1 account for the very severe to mild phenotype observed in XLMTM patients (9). MTM1 encodes myotubularin, which contains the consensus sequence for the tyrosine phosphatase catalytic site (8). However, this protein seems to essentially act in muscles as a phosphoinositide phosphatase, which specifically dephosphorylates the lipid second messengers phosphatidylinositol 3-phosphate (PtdIns3P) (10–12) and phosphatidylinositol 3,5-bisphosphate (PtdIns3,5P2) (13,14). Myotubularin is the first identified member of a large protein family that comprises 13 other known members in human, named myotubularin-related 1 to 13 (MTMR1 to MTMR13). Mutations in MTMR2 and MTMR13 have been associated with Charcot–Marie–Tooth demyelinating diseases (15–17) and reinforce a central role of phosphoinositides in neuromuscular homeostasis (reviewed in 18). The MTM1 protein family encompasses both active and inactive members and recent reports have suggested functional cooperation between these two groups (18–23). Besides the characterized etiology for XLMTM, no significant molecular data for the autosomal forms have been reported so far.
Spontaneous cases of Labrador retrievers suffering from a congenital form of hereditary myopathy (hereditary myopathy of the Labrador retriever, HMLR) were first described in the USA in 1976 (24) and later on by other groups across the world (25–29). The myopathy is transmitted as an autosomal recessive disease (28–30). Clinical features in male and female pups include hypotonia, generalized muscle weakness, abnormal postures, stiff hopping gait, exercise intolerance and increased collapse when exposed to cold (see movie published as Supplementary Material at HMG Online). On examination, there is evidence of skeletal muscle atrophy, particularly involving muscles of the head and tendinous areflexia. Metabolic and shape remodeling of the muscles is observed with type 2 fibers deficiency and predominance of atrophic or hypertrophic type 1 fibers (24,25,28,29). In all reported cases, histopathological evaluation of muscle biopsies has demonstrated a characteristic centralization of myonuclei, often located in areas devoid of myofibrils with mitochondrial aggregation. Centronuclear myopathy (cnm) was therefore proposed as an alternative name for the HMLR (29) and is, to our knowledge, the only spontaneous animal model for human autosomal centronuclear myopathies.
We previously described our experimental pedigree of Labrador retrievers segregating centronuclear myopathy. The cnm locus was mapped with a highly significant LOD score of 9.93 (𝛉=0.00) to a 18.1 cM segment in the centromeric region of canine chromosome 2 (CFA2) (29), a region that forms a shuffled block of synteny with human chromosome segment 10p15–10p11.1 (31). Interestingly, no gene of the MTM1 family maps to human chromosome 10p (12).
The purpose of our study was to identify plausible candidate genes within the cnm interval and to further characterize a disease-causing mutation in one of these genes. We report here a mutation in the gene encoding protein tyrosine phosphatase-like, member A (PTPLA) that perfectly segregates with cnm in Labrador retrievers of our experimental pedigree. Indeed, a tRNA-like short interspersed repeat element (SINE) insertion was identified in exon 2 of PTPLA that leads to multiple transcriptional alterations of the SINE-containing PTPLA allele. Moreover, we disclose here an original and complex disease-associated mutation with multiple transcriptional defects arising from tRNA-like SINE insertion.
RESULTS
Candidate gene selection
Using the human genome resources (NCBI Map Viewer), we identified 208 genes within the cnm region (data not shown). None of these is a member of the MTM1 family. Hence, we focused on genes expressed in skeletal muscles, in a temporal window compatible with the early postnatal symptoms diagnosed in affected dogs. Among these genes, PTPLA was shown to be expressed during embryogenesis in murine myogenic precursors (32) and in adult skeletal muscles of both mouse and human (32,33). Furthermore, PTPLA was annotated as a phosphatase coding gene as PTPLA protein contains a putative phosphatase-like domain, with a proline instead of an arginine in the catalytic site. The motif has characteristics similar to myotubularin which, when mutated, is responsible for human XLMTM.
Mutation identification and segregation in the experimental pedigree
The detailed structure of the canine PTPLA gene and phylogenetic analysis of the PTPLA family members will be described elsewhere (Pelé et al., manuscript in preparation). The following data were taken into consideration when designing the experiments detailed subsequently. Canine PTPLA spans >20 kb and contains seven exons (Fig. 1A), a structure shared with its human (33) and murine orthologs (see Ensembl Gene ID ENSMUSG00000063275). In dog skeletal muscles, PTPLA transcripts exist in two isoforms (Fig. 1B) differing by the retention (PTPLAfl, full-length, EMBL accession no. AJ876904) or the exclusion of exon 5 (PTPLAd5, deleted exon 5, EMBL accession no. AJ876905). PTPLAfl transcript encodes a putative protein of 249 amino acids that contains four transmembrane domains and the putative tyrosine phosphatase-like (HCX2GX2P) catalytic site (Fig. 1C). Exon 5 skipping in the PTPLAd5 transcript leads to frame disruption and premature translation termination. The resulting predicted PTPLAD5 protein contains two transmembrane domains and the putative tyrosine phosphatase-like (HCX2GX2P) catalytic site. PTPLAFL and PTPLAD5 proteins share a common 125 amino acids N-terminus domain and differ by their C-terminus regions, 124 amino acids and 11 amino acids in length, respectively (Fig. 1C).
In this gene, we searched for the specific polymorphism that would segregate with the disease in our experimental pedigree. Genomic DNA of wild-type (wt) (cnm+/+), healthy carriers (cnm+/−) and affected (cnm−/−) Labradors was digested and hybridized with a probe encompassing most of the PTPLA cDNA (shown in Fig. 1B). A BamHI restriction fragment length polymorphism (RFLP) was detected in DNA samples from cnm+/− and cnm−/− dogs; it was not seen in DNA samples from cnm+/+ dogs. Furthermore, this RFLP perfectly segregated with the disease (data not shown).
Using PCR and RT–PCR, we amplified individual exons of PTPLA from genomic DNA and RNAs from skeletal muscles. PCR and RT–PCR amplifications were carried out with samples from a healthy carrier (cnm+/−) and an affected (cnm−/−) dog. Figure 2 illustrates the results obtained for exons 1 and 2. PCR and RT–PCR amplifications of exon 1 using Ex1F1 and Ex1R1 primers (Fig. 2A) led to a single 123 bp fragment corresponding to exon 1 (Fig. 2B, lanes 2–4). The 123 bp signal was weaker when RNAs from cnm−/− muscles were used as a template in the RT–PCR reaction compared with RNAs from cnm+/− muscles (Fig. 2B, compare lanes 3 and 4). PCR amplification of exon 2 using genomic DNA from a cnm+/− dog as a template and Ex2F1 and Ex2R1 as primers yielded two products of 105 and 341 bp, respectively (Fig. 2B, lane 5). The 105 bp product had the normal size of the exon 2 fragment expected with these primers. This 105 bp fragment was also obtained when RNAs from cnm+/− muscles were used as a template for RT–PCR reaction; it was not detected when RNAs from cnm−/− muscles were used (Fig. 2B, lanes 6 and 7, respectively).
To further characterize the 341 bp fragment, PCR amplifications of exon 2 and intron 2 using Ex2F2 and Ex3R1 as primers were performed on genomic DNA from all 61 individuals of the six-generation experimental pedigree (29) and three unrelated healthy (cnm+/+) Labradors. A representative panel of the resulting PCR products is presented in Fig. 2C and complete amplifications from 46 animals, among which 22 are affected dogs, are provided in Supplementary Material, Fig. S1. PCR using DNA from wt cnm+/+ dogs as template yielded a single 282 bp product (lanes 2–4). PCR performed on all affected cnm−/− dogs' DNA contained a single 518 bp product (lanes 8–10). Finally, PCR using cnm+/− dogs' DNA yielded both 282 and 518 bp products (lanes 5–7). To explain the size shift seen when genomic DNA from cnm+/− and cnm−/− dogs were used as templates, we cloned and sequenced the 518 bp product. It contained a 236 bp antisense canine tRNA-like SINE (EMBL accession no. AJ876906), flanked on both sides by a 13 bp direct duplication of the insertion site (Fig. 3). After comparison with RepeatMasker libraries of known vertebrate and carnivore-specific repeat elements, the SINE sequence found within PTPLA best aligned with SINEC_Cf repeats from the SINE/Lys subfamily, first described by Minnick et al. (34). Using tRNAScan, we identified a glutamine anticodon within the tRNA-like sequence of the SINE. Finally, sequence analysis revealed a BamHI restriction site within the insertion, which explained the RFLP previously described.
To strengthen the causative link between this insertion and the centronuclear myopathy, 18 dogs from 14 other breeds, 20 mongrels and six additional unrelated healthy French Labradors were screened for the presence of the SINE. All 44 tested dogs lacked the insertion (data not shown). Finally, BLAST analysis was conducted against the Boxer genome assembly (http://www.ncbi.nlm.nih.gov/genome/seq/CfaBlast.html). No SINE was inserted within PTPLA exon 2 of this genome.
Hence, the SINE insertion seemed highly specific to the Labrador retriever pedigree segregating centronuclear myopathy that was developed at the Alfort School of Veterinary Medicine. Using standard nomenclature system for human mutations (35) and the PTPLA-corresponding genomic sequence from the Boxer genome assembly, this mutation should be called PTPLA *g9459-9460ins236. It will be hereafter referred to as PTPLAalf.
SINE insertion leads to an aberrant splicing pattern in skeletal muscles
Alu insertions have frequently been associated with human diseases. Most of these insertions have occurred within introns causing either close-at-hand exon-skipping (e.g. 36–39) or incorporation of Alu sequences into mature mRNA, a splicing-mediated process called exonization (36). Few SINE insertions within coding sequences have been reported. They may lead to specific skipping of the mutated exon (e.g. 36,40) or to frame disruption with premature translation termination signal (e.g. 36,41,42).
To test for putative transcriptional defects resulting from the SINE insertion within PTPLA exon 2, PTPLA transcripts were amplified by RT–PCR using sequences in exon 1 and exon 7 as primers and RNAs from skeletal muscles of cnm+/+, cnm+/− and cnm−/− dogs as templates. Both PTPLAfl and PTPLAd5 transcripts were amplified from cnm+/+ and cnm+/− samples (Fig. 4A, lanes 1 and 2). In contrast, a ladder-like profile with seven RT–PCR products was obtained from cnm−/− samples (Fig. 4A, lane 3). All seven amplicons were cloned and sequenced to precisely determine acceptor and donor splicing sites used for the synthesis of the corresponding transcripts in cnm−/− muscles (Fig. 4B). Figure 4C proposes a classification of the PTPLA transcripts based on their splicing profile and the predicted proteins. As far as cnm−/− dogs are concerned, RT–PCR products corresponding to PTPLAfl and PTPLAd5 wt isoforms were observed, which indicates that the SINE was occasionally spliced out. Furthermore, no point mutation could be detected in these transcripts, indicating that wt and cnm-associated PTPLA alleles only differ by the SINE insertion. RT–PCR product v1 corresponds to a splicing variant that carried a partial (25 bp) deletion of PTPLA exon 2 as well as 154 bp exonization of the 3′-region of the inserted SINE. Further analysis of the SINE sequence confirmed the presence of cryptic splicing donor and acceptor sites, which could lead either to the proper splicing of the insertion (transcripts PTPLAfl and PTPLAd5) or to variant v1 (Fig. 3 and 4B). The remaining variants (v2 to v5) lacked internal exon(s). These variants arose from exon-skipping involving, respectively, exon 2 (v2); exons 2 and 3 (v3); exons 2, 3 and 4 (v4) and exons 2, 3, 4 and 5 (v5) (Fig. 4B).
To assess the biological impact of the SINE-induced splicing defects, amino acid sequences were inferred from each transcript. PTPLAFL and PTPLAD5 proteins were deduced from transcripts PTPLAfl and PTPLAd5, respectively. Frameshift mutations were found in variants v1, v2, v3 and v4, hence leading to truncated forms of PTPLA. Finally, exons 1 and 6 were in frame in variant v5 sequence that encodes a small truncated protein lacking the PTPL site (Fig. 4C).
The SINE insertion within PTPLA in the PTPLAalf allele strongly affected splicing mechanisms. However, quite surprisingly, an AGGT sequence contained within the 13 bp repeats flanking the insertion (Fig. 3) allowed proper splicing of the SINE thus leading to wt transcripts.
tRNA-like SINE is sufficient by itself to alter splicing
To test whether the SINE is by itself responsible for the characterized splicing defects, we performed ex vivo analyses using exon-trapping experiments. Two PTPLA genomic fragments spanning from intron 1 to intron 4, which differed by the absence or presence of the SINE insertion were cloned into pSPL3 exon-trapping vector (Fig. 5A). Both plasmids, termed pCOS-cnm and pCOS-wt, respectively, were transfected into COS-7 cells and the resulting transcripts were analyzed by RT–PCR. When the pCOS-wt plasmid was used, RT–PCR and nested PCR products corresponded to a properly spliced transcript containing PTPLA exons 2, 3 and 4 flanked by the two 5′- and 3′-β-globin exons from the vector (Fig. 5B). SacI digestion confirmed the presence of PTPLA exon 4 (Fig. 5B). In contrast, only smaller products were observed after RT–PCR and nested PCR when total RNAs from COS-7 cells transfected with the SINE-containing plasmid (pCOS-cnm) were used as template. SacI digestion suggested that PTPLA exon 4 was present in these products (Fig. 5B). This hypothesis was confirmed by sequence analysis. Indeed, the smaller size of the transcripts resulted from the absence of PTPLA exon 2. Additional RT–PCR experiments revealed only a very low level of wt transcript containing PTPLA exons 2, 3 and 4 (data not shown).
Thus, precise splicing of the SINE inserted sequence and skipping of the mutated exon 2 were induced ex vivo using a 1.5 kb mini-PTPLAalf allele, recapitulating most of the splicing defects seen when the complete PTPLAalf allele is transcribed in muscle cells in vivo.
Quantification of PTPLA transcripts in affected dogs' muscles
Seven splicing variants resulting from the transcription of the PTPLAalf allele were detected in cnm−/− muscles, out of which two were wt isoforms, PTPLAfl and PTPLAd5. To determine both the total amount of PTPLA transcripts and the proportion of wt PTPLA transcripts in cnm−/− muscles, we performed quantitative analyses using real-time RT–PCR. All PTPLA transcripts identified in cnm+/− and cnm−/− muscles contained exons 6 and 7 (Fig. 4B). Hence, a specific primer pair encompassing sequences of exons 6 and 7 was designed to quantify the total amount of PTPLA transcripts (Fig. 6A). In contrast, exon 2 is either deleted or partially skipped in all aberrant PTPLA transcripts in cnm−/− (Fig. 4B). Hence, a specific forward primer was designed in the 25 bp exon 2 fragment specific for the wt transcripts and used in combination with a reverse primer located within exon 4 to quantify the amount of wt transcripts (Fig. 6A).
Five healthy cnm+/− controls and five affected cnm−/− dogs were included in the analysis. Amounts of PTPLA transcript were normalized against TATA-box binding protein (TBP) and muscular phosphofructo-kinase (MPFK) housekeeping gene transcripts. Amounts of PTPLA transcripts from cnm+/− and cnm−/− muscles were then compared and histograms summarizing the results are presented in Fig. 6B. PTPLA transcripts detected in muscles of cnm−/− dogs represented ∼35% (34.8±2.4%) of the amount of PTPLA transcripts found in muscles of cnm+/− dogs. Furthermore, PTPLAfl and PTPLAd5 transcripts in cnm−/− muscles represented ∼1% (0.8±0.03%) of PTPLAfl and PTPLAd5 transcripts expressed in cnm+/− muscles. Statistical analysis based on pair-wise fixed reallocation randomization tests (43) confirmed that the observed differences were highly significant (P=0.001 for both analyses).
In conclusion, PTPLAalf allele was transcribed in cnm−/− affected muscles with a rather low efficiency, 35% of total transcripts compared with healthy cnm+/− muscles. Furthermore, this transcription mainly led to abnormal transcripts with either premature stop codons or codons encoding a truncated protein. Finally, transcription of the PTPLAalf allele also led to normal wt transcripts, although at a very low level (<1%).
DISCUSSION
Labrador retriever centronuclear myopathy perfectly mimics clinical and histological features of the human autosomal dominant and recessive forms of centronuclear myopathy. To date, neither genetic linkage nor molecular data are available for these two forms of the human disease. To the best of our knowledge, canine centronuclear myopathy in Labrador retrievers is the only spontaneous animal model for the disease described so far. We have previously mapped the canine cnm locus to a 18.1 cM interval within the centromeric region of CFA2. Here, we report the identification of a canine gene, PTPLA, that likely plays a critical role in the pathogenesis of centronuclear myopathy seen in Labrador retrievers. PTPLA has been shown to be expressed early in myogenic precursors during mouse embryogenesis (32) and afterwards in adult skeletal muscles of mouse (32) and human (33).
In our experiments, PTPLA was expressed in canine skeletal muscles at all tested stages from birth to adulthood (data not shown). This temporal expression pattern is compatible with the clinical and histological evolution of the disease in affected dogs. Furthermore, we observed a homozygous tRNA-like SINE insertion within PTPLA exon 2 of affected dogs. In the human genome, Alu elements are the most abundant SINEs and numerous Alu insertions into genes have been associated with hereditary disorders (36). Not all of these cases have been demonstrated to be directly causative for the disease, but the lack of other detectable mutation and the rarity of such Alu insertions strongly suggest that they are a likely candidate. In the case of the Labrador retriever centronuclear myopathy, two lines of reasoning suggest that the SINE exonic insertion is responsible for the disease. First, the SINE insertion identified in our pedigree of Labrador retriever perfectly segregates with the disease. We have sampled 47 healthy dogs from 15 pure and mixed breeds and failed to detect this insertion. In particular, the SINE insertion was not found in nine healthy French Labrador retrievers unrelated to our experimental pedigree. The presence of the retroposed SINE within PTPLA is therefore highly correlated with the segregation of centronuclear myopathy in French Labrador retrievers. Second, and most compellingly, we show that the repeat element insertion severely affects the splicing of the corresponding SINE-containing PTPLAalf allele. We provide strong evidence that the tRNA-like SINE is by itself responsible for cumulative transcriptional abnormalities based on the four following features: (i) the SINE can be properly spliced, although at a low level (<1%); (ii) partial exonization of the SINE; (iii) exon-skipping involving several exons and (iv) the total amount of PTPLA transcripts was markedly reduced. Taken together, these observations show that the exonic SINE insertion within the PTPLA gene in cnm−/− dogs acts as a hypomorphic mutation. Existing data and ongoing experiments should help us to confirm that the SINE insertion within PTPLA is a causative event leading to canine centronuclear myopathy. First, our experimental pedigree has been established from two affected Labradors that are unrelated for 11 generations (29). We have found that both dogs are homozygotes for the same insertion within PTPLA exon 2. Genotyping additional affected Labradors coming to the Alfort clinics will allow us to reinforce the association between the SINE insertion and cnm. We also believe that close clinical and histopathological relationships exist between the centronuclear myopathy found in French Labrador retrievers and the HMLR identified in dogs from other countries (29). Although not precisely determined yet, the number of generations from common ancestors of affected Labradors from the USA and France is higher than between affected French dogs. Amplifying the SINE insertion in affected Labrador retrievers from the USA or from any other country will reinforce that the SINE insertion is the causative-disease mutation with founder effect. Finally, characterization of mutations in PTPLA with similar phenotype either in transgenic mice or in human patients affected by autosomal recessive cnm would provide strong evidence that PTPLA plays a central role in the etiology of centronuclear myopathy.
In human, the related X-linked myotubular myopathy has been attributed to truncating or missense mutations in the MTM1 gene encoding myotubularin (8,9,44). Classical and tissue-specific gene-targeting inactivation of Mtm1 in the mouse confirmed that the absence of muscular expression of Mtm1 does not affect myogenesis but induces lethal degeneration of skeletal muscles by 4 weeks of age (45). Myotubularin is a phosphatase with dual activity that specifically dephosphorylates subpools of phosphatidylinositols (PIs) (10,11,13,14). PIs are implicated in a number of physiological processes including cell proliferation, death, motility, cytoskeletal regulation and intracellular vesicle trafficking (46). Ptpla was first annotated in mouse and then in human on the basis of a putative protein tyrosine phosphatase-like catalytic site, with a proline replacing arginine in the consensus (HCX26X2R) sequence (32,33). PTPLA protein shares no sequence homology with the MTM1 protein family (data not shown) but, quite surprisingly, canine cnm-associated histopathological features closely resemble those observed in human and murine XLMTM. Hence, PTPLA protein may be involved in the same PI-dependent functional signaling pathway as MTM1, at least in skeletal muscle. This hypothesis should be addressed by testing the putative PTPLA phosphatase activity and/or investigating functional interactions between PTPLA and MTM1 proteins by analyzing, for example, the subcellular localization of MTM1 in the muscles of affected Labrador or the subcellular localization of PTPLA in the muscles of Mtm1−/− mice.
In the work presented here, we show that the SINE insertion detected within PTPLA exon 2 is associated with various transcriptional features. First, the SINE insertion is flanked by a direct duplication of the insertion site, which is characteristic of the retroposition mechanism of SINEs. As this duplicated sequence contains an AGGT tetranucleotide, the SINE acquired an uncommon intron-like status [5′-splicing donor site (GT) and 3′-splicing acceptor site (AG)] allowing its proper excision. Although it generates complete and normal mature PTPLA transcripts, this mechanism seems to be poorly efficient as only 1% of wt PTPLA transcripts were detected in muscles of affected cnm−/− Labradors compared with healthy cnm+/− dogs. Precise elimination of the SINE observed in vivo was reproduced ex vivo. In addition to this exceptionally fine removal of parasitic sequence, the repeat element insertion has other deleterious consequences on the transcription of the SINE-containing PTPLAalf allele. Indeed, we provide evidence that the splicing machinery recognizes cryptic splice sites within both the SINE and exon 2 sequences, which yields a mature elongated transcript containing a partially exonized SINE sequence. Concomitantly, exon-skipping involving either exon 2 alone or in combination with other exons results in several PTPLA truncated transcripts. Again, this PTPLAalf-associated aberrant exon-skipping could be undoubtedly attributed to the SINE sequence by itself as it was reproduced in ex vivo splicing experiments. Alu insertions have been associated with other diseases (36–39,42,47). These insertions were either associated with exonization of repeat element sequences or with specific exon-skipping. Strikingly, the work presented here highlights that these two mutational mechanisms, previously reported as independent defects, are not necessarily mutually exclusive. The systematic dichotomy observed between exonization and exon-skipping mechanisms may depend on splicing regulatory elements, either contained by the inserted Alu sequence or located in the genomic vicinity. Alternatively, technical limitations might have prevented the identification of concomittant subtle transcriptional aberrations in previously reported studies. Owing to the transcriptional defects associated with the SINE insertion in PTPLA, it is tempting to speculate that this repeat element contains splicing regulatory motifs that are strong enough to drive both exonization and exon-skipping. Furthermore, SINE-associated exon-skipping previously reported only refer to skipping of a single exon. For example, Alu insertion in exon 22 of human BCRA2 gene induces alternative skipping of this exon, which may play a role in the development of breast cancer (48). Surprisingly, in Labradors affected by centronuclear myopathy, not only the SINE-targeted exon 2 but also exons 3, 4 and 5 are abnormally skipped in some muscular transcripts. Introns 2 and 3 of canine PTPLA are 183 and 76 bp in length, respectively. We may thus assume that the SINE hampers the spliceosome machinery to recognize normal consensus splicing sites by modifying spatial conformation of the transcripts and/or specific interaction of RNA with regulatory proteins. Abnormal skipping of exons 3 and 4 would thus result from their closeness with the SINE. Finally, all but one PTPLA abnormal transcripts bear a disruption of reading frame. Princeps studies in yeast and human have paved the way for nonsense-mediated mRNA decay (NMD), a surveillance mechanism that eliminates transcripts containing premature termination codons (49,50). The reduced total amount (35%) of PTPLA transcripts detected in muscles from cnm−/− dogs using real-time RT–PCR analyses could therefore result from NMD leading to the degradation of PTPLA abnormal transcripts.
Deleterious effects of repeat elements have already been reported in dogs. For example, an intronic SINE insertion within the Hypocretin-receptor-2 in Doberman pinschers induces exon-skipping, leading to narcolepsy (51). To date, relatively few canine diseases have been characterized at the genomic level. With the examples of narcolepsy, hemophilia and centronuclear myopathy, diseases associated with insertion of repeat elements seem well represented. The low average divergence of SINEC_Cf repeats in dogs and the high percentage of bimorphic loci with respect to the presence or absence of SINE insertion (6.7–11.5% between dogs of different breed) indicate that expansion of SINEs is a recent and still active process (52; Ewen Kirkness, personal communication). The high incidence of disease-associated insertions in dogs may then be explained by the recent expansion of repeat sequences, which represent ∼31% of the canine genome (52). By comparison, finding 406 gross insertion and duplication mutations in the Human Genetic Mutation Database that contains 44 090 characterized human mutations suggests that Alu insertions, which occur at a lower rate, hardly contribute to 0.09% of human genetic diseases. Besides their deleterious consequences on gene expression, insertions of SINEs in dogs and their observed bimorphism may contribute to interbreed or intrabreed modulation in gene expression accompanying physiological phenotypic diversity (52). The functional effects of PTPLA mutation reported in the present paper disclose a rare example of a single SINE insertion with several modifications of RNA processing, which provides a molecular basis for the dynamic evolutionary mechanism of gene control by mobile elements (reviewed in 53).
In conclusion, the work presented here shows that the exonic SINE insertion within the PTPLA gene in cnm−/− dogs acts as a hypomorphic mutation and recapitulates most of the SINE-dependent mutational molecular mechanisms. These results also illustrate the relevance of the canine model in genetic studies. Indeed, major advances in the development of canine genomic and genetic resources have been reported recently or will be soon available: (i) an integrated 4249 marker FISH/RH map (31); (ii) the publication of the 1.5× Poodle genome (52); (iii) the partial annotation of a 7.6× canine genome assembly and the identification of 500 000 SNPs (Kerstin Lindblad-Toh, personal communication) and (iv) a 10 000 gene map of the canine genome that will help in assembly of the sequence and will be useful in comparative mapping (Christophe Hitte, personal communication). These resources will provide long-awaited means for mapping genes underlying the striking diversity among breeds in morphology and behavior as well as canine disease genes.
MATERIALS AND METHODS
mRNA extraction from biceps femoris and cDNA synthesis
Frozen muscular biopsies of 150 mg were grinded into a mortar containing liquid nitrogen. The resulting powder was resuspended into 1 ml of Dynabeads mRNA direct kit lysing/binding buffer (DYNAL) and incubated at 55°C with proteinase K for 2 h. Crude extracts were centrifuged at 5000g during 5 min and poly(A)+ mRNA isolation from the supernatant was performed using the Dynabeads mRNA direct kit (DYNAL). First-strand cDNA was generated from 200 ng of poly(A)+ mRNAs using Superscript II reverse transcriptase (200 U, Invitrogen), anchored oligo(dT)25 (1.6 µM; Q-Biogen) and RNase inhibitor (40 U; RNaseOUT, Invitrogen). Synthesis of cDNA was performed at 55°C for 90 min and was followed by a 20 min RNase H treatment (4 U; Invitrogen).
Mutation screening of the PTPLA locus and segregation analysis in centronuclear myopathic canines
Primers designed at the 5′ and 3′ ends of PTPLA exons 1 and 2 were used to amplify these exons from healthy carriers (cnm+/−) and affected dogs' (cnm−/−) samples. PCR and RT–PCR were performed on genomic DNA and poly(A)+ mRNAs, respectively, using the following primer pairs: exon1, Ex1F1 (5′-GACGAGGACGGCACCAAC-3′)/Ex1R1 (5′-GGTCATGGCATGTTGTAGA-3′) and exon 2, Ex2F1 (5′-GCTATTGCCATGGTACGTTTT-3′)/Ex2R1 (5′-AAGCAAGGCAAATGTTTGGA-3′). Cycling conditions were 94°C (2 min), 35 cycles of 94°C (30 s), 60°C for exon 1 or 50°C for exon 2 (30 s), 72°C (1 min); 72°C (5 min). After the initial detection of SINE insertion, new primers specific for exon 2 (Ex2F2, 5′-GGAAAAAGGAACACACAAAGG-3′) and exon 3 (Ex3R1, 5′-ACCAATTAAACAGTGGACTAT-3′) were designed to detect the SINE insertion following PCR amplification of genomic DNA. PCR conditions were 94°C (3 min), two cycles of 94°C (30 s), 57°C (30 s), 72°C (2 min), two cycles of 94°C (30 s), 55°C (30 s), 72°C (2 min), two cycles of 94°C (30 s), 53°C (30 s), 72°C (2 min), 29 cycles of 94°C (30 s), 50°C (30 s), 72°C (2 min); 72°C (5 min).
Cloning and sequencing of the SINE insertion
PCR products were extracted from 1% Seakem GTG agarose (TEBU) gels using the NucleoSpin Extract II (Macherey-Nagel) and ligated to pCR4-TOPO plasmid (Invitrogen) following manufacturers' recommendations. Inserts were sequenced using T3 and T7 primers.
Splicing pattern analysis of the PTPLAalf allele
PTPLA transcripts were amplified by RT–PCR using muscular poly(A)+ mRNAs samples from cnm+/+, cnm+/− and cnm−/− dogs as template and Ex1F2 (5′-CTACAACATCGCCATGACC-3′) and Ex7R (5′-CACCTCTCCATGAAGCACCT-3′) as primers. PCR conditions were 94°C (2 min), two cycles of 94°C (20 s), 62°C (30 s), 72°C (2 min), three cycles of 94°C (20 s), 60°C (30 s), 72°C (2 min), 35 cycles of 94°C (20 s), 58°C (30 s), 72°C (2 min); 72°C (5 min). RT–PCR products were either visualized using standard electrophoresis techniques and distinct fragments were purified, cloned and sequenced as described earlier.
Ex vivo splicing experiments
Primers containing EcoRI and NotI restriction sites were designed from the 3′ end of PTPLA intron 1 (Int1F, 5′-CGGAATTCCGCAGATGCTCAACCACCCAG-3′) and 5′ end of PTPLA intron 4 (Int4R, 5′-TTGCGGCCGCAACCCTTCCAACTTTTCATTTAGTCTT-3′). They were used to amplify wt and SINE-containing (cnm) genomic inserts. Amplified fragments were checked for size and ligated to pSPL3 vector using standard EcoRI/NotI double digestion and ligation protocol. The resulting pCOS-wt and pCOS-cnm plasmids were transfected into COS-7 cells as follows: cells were grown at 37°C with 8% CO2 in DMEM medium (Gibco) supplemented with 10% fetal calf serum (D. Deutscher), penicilin (100 U/ml, Invitrogen) and streptomycin (100 µg/ml, Invitrogen). At 80% confluence, cells were trypsinated, centrifuged and resuspended at a final concentration of 1×107 cells/ml. Cells were incubated on ice for 10 min with 2 µg of plasmids and electroporation (200 Ω, 25 µF, 1.5 kV) was performed in 0.4 mm cuvettes using Biorad GenePulser. Electroporated cells were grown using standard cell culture conditions for 48 h at 37°C with 8% CO2. Total RNA extraction from transfected COS-7 cells was done using Trizol Reagent (Invitrogen) according to the manufacturer's recommendation and first-strand cDNA synthesis was performed as described earlier. Amplification was performed using SA2 (5′-ATCTCAGTGGTATTTGTGAGC-3′) and SD6 (5′-TCTGAGTCACCTGGACAACC-3′) primers. A 300-fold dilution of this first PCR was used as template for a nested PCR using SA4 (5′-CACCTGAGGAGTGAATTGGTCG-3′) and SD2 (5′-GTGACCTGCACTGTGACAAGC-3′) primer pair. PCR conditions were: 94°C (2 min), 20 cycles of 94°C (1 min), 60°C for first PCR or 63°C for nested PCR (1 min) and 72°C (2 min); 72°C (5 min).
Real-time RT–PCR analyses
Prior to first-strand cDNA synthesis, genomic DNA was removed from 200 ng mRNA samples using the DNA-free kit (Ambion). Three independent cDNA samples from the skeletal muscle of a healthy Golden retriever (cnm+/+) were used as an internal calibrator to normalize data from independent reactions. Primer pair sequences and MgCl2 final concentrations used are presented in Supplementary Material, Table S1. Annealing temperature was 64°C for all samples. Real-time PCR reactions were performed on a LightCycler (ROCHE) using cDNA dilutions, specific primers (0.4 µM) and the LC-FastStart DNA Master SYBR Green I kit according to the manufacturer's recommendations. For each sample evaluated, crossing-points (Cp) were automatically calculated using the ‘second derivative maximum’ method of the LightCycler Software. Standard curves for each primer pair were generated using duplicated 4-fold dilution series of the calibrator sample. The corresponding PCR efficiencies are indicated in Supplementary Material, Table S1.
Five healthy Labradors (cnm+/−) and five affected Labradors (cnm−/−) samples were tested in duplicates. Each experiment included calibrator-based internal controls that were used to normalize the calculated Cp-value for each sample (Supplementary Material, Table S2). Relative expression analyses with efficiency correction were performed using REST software described in Pfaffl et al.(43). As TBP and MPFK expression levels were not significantly different between healthy carriers (cnm+/−) and affected dogs (cnm−/−), they were used as references. Expression ratios were then calculated for each target, using healthy carriers as controls and affected dogs as samples. Two thousands randomization tests were performed to assess statistical significance of the observed differences in transcript amounts between control and cnm affected dog samples.
Bioinformatics
All primers were designed using the Primer3 software (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi). BLAST searches against the human genome and the canine genome assemblies were performed from the NBCI BLAST page (http://www.ncbi.nlm.nih.gov/blast). Sequence alignments were performed using CAP and dialign2 softwares, freely available from the Infobiogen services home page (http://www.infobiogen.fr/services/menuserv.html). Canine TBP and MPFK sequences were retrieved from the SRS database (http://srs.ebi.ac.uk). Quantitative expression analysis REST Software was downloaded from Gene-quantification Academic & Industrial Information Platform for qPCR (http://www.gene-quantification.com).
SUPPLEMENTARY MATERIAL
Supplementary Material is available at HMG Online.
ACKNOWLEDGEMENTS
We thank Denis Houzelstein (UMR 7592, Institut Jacques Monod) for helpful discussions on real-time PCR analysis, Stéphanie Le Bras for technical expertise in exon-trapping experiments, members from the UMR 955 for their assistance in molecular techniques, Kelly Rogers (Pasteur Institute) and Ewen Kirkness (The Institute of Genome Research) for improving the manuscript. We also thank Jean-Laurent Thibaud, Nicolas Granger (UETM) for medical expertise and Xavier Cauchois, Stéphanie Le Mevel, Ingrid Gruyer and Serge Kouame for taking care of the animals. This work received financial support from the Association Française contre les Myopathies (AFM).
Conflict of Interest statement. None declared.
References
Karpati, G., Carpenter, S. and Nelson, R.F. (
Wallgren-Pettersson, C., Clarke, A., Samson, F., Fardeau, M., Dubowitz, V., Moser, H., Grimm, T., Barohn, R.J. and Barth, P.G. (
Jeannet, P.Y., Bassez, G., Eymard, B., Laforet, P., Urtizberea, J.A. Rouche, A., Guicheney, P., Fardeau, M. and Romero, N.B. (
Meyers, K.R., Golomb, H.M., Hansen, J.L. and McKusick, V.A. (
Herman, G.E., Finegold, M., Zhao, W., de Gouyon, B. and Metzenberg, A. (
Wallgren-Pettersson, C., Clarke, A., Samson, F., Fardeau, M., Dubowitz, V., Moser, H., Grimm, T., Barohn, R.J. and Barth, P.G. (
Helliwell, T.R., Ellis, I.H. and Appleton, R.E. (
Hu, L.J., Laporte, J., Kioschis, P., Heyberger, S., Kretz, C., Poustka, A., Mandel, J.L. and Dahi, N. (
Laporte, J., Hu, L.J., Kretz, C., Mandel, J.L., Kioschis, P., Coy, J.F., Klauck, S.M., Poustka, A. and Dahl, N. (
Biancalana, V., Caron, O., Gallati, S., Baas, F., Kress, W., Novelli, G., D'Apice, M.R., Lagier-Tourenne, C., Buj-Bello, A., Romero, N.B. et al. (
Blondeau, F., Laporte, J., Bodin, S., Superti-Furga, G., Payrastre, B. and Mandel, J.L. (
Taylor, G.S., Maehama, T. and Dixon, J.E. (
Laporte, J., Blondeau, F., Buj-Bello, A. and Mandel, J.L. (
Schaletzky, J., Dove, S.K., Short, B., Lorenzo, O., Clague, M.J. and Barr, F.A. (
Tronchere, H., Laporte, J., Pendaries, C., Chaussade, C., Liaubet, L., Pirola, L., Mandel, J.L. and Payrastre, B. (
Bolino, A., Muglia, M., Conforti, F.L., LeGuern, E., Salih, M.A., Georgiou, D.M., Christodoulou, K., Hausmanowa-Petrusewicz, I., Mandich, P., Schenone, A. et al. (
Azzedine, H., Bolino, A., Taieb, T., Birouk, N., Di Duca, M., Bouhouche, A., Benamou, S., Mrabet, A., Hammadouche, T., Chkili, T. et al. (
Senderek, J., Bergmann, C., Weber, S., Ketelsen, U.P., Schorle, H., Rudnik-Schoneborn, S., Buttner, R., Buchheim, E. and Zerres, K. (
Laporte, J., Bedez, F., Bolino, A. and Mandel, J.L. (
Nandurkar, H.H., Caldwell, K.K., Whisstock, J.C., Layton, M.J., Gaudet, E.A., Norris, F.A., Majerus, P.W. and Mitchell, C.A. (
Nandurkar, H.H. and Huysmans, R. (
Kim, S.A., Vacratsis, P.O., Firestein, R., Cleary, M.L. and Dixon, J.E. (
Mochizuki, Y. and Majerus, P.W. (
Nandurkar, H.H., Layton, M., Laporte, J., Selan, C., Corcoran, L., Caldwell, K.K., Mochizuki, Y., Majerus, P.W. and Mitchell, C.A. (
Kramer, J.W., Hegreberg, G.A., Bryan, G.M., Meyers, K. and Ott, R.L. (
McKerrell, R.E. and Braund, K.G. (
Watson, A.D., Farrow, B.R., Middleton, D.J. and Smyth, J.B. (
Gortel, K., Houston, D.M., Kuiken, T., Fries, C.L. and Boisvert, B. (
Bley, T., Gaillard, C., Bilzer, T., Braund, K.G., Faissler, D., Steffen, F., Cizinauskas, S., Neumann, J., Vogtli, T., Equey, R. et al. (
Tiret, L., Blot, S., Kessler, J.L., Gaillot, H., Breen, M. and Panthier, J.J. (
Kramer, J.W., Hegreberg, G.A. and Hamilton, M.J. (
Breen, M., Hitte, C., Lorentzen, T.D., Thomas, R., Cadieu, E., Sabacan, L., Scott, A., Evanno, G., Parker, H.G., Kirkness, E.F. et al. (
Uwanogho, D.A., Hardcastle, Z., Balogh, P., Mirza, G., Thornburg, K.L., Ragoussis, J. and Sharpe, P.T. (
Li, D., Gonzalez, O., Bachinski, L.L. and Roberts, R. (
Minnick, M.F., Stillwell, L.C., Heineman, J.M. and Stiegler, G.L. (
Antonarakis, S.E. (
Deininger, P.L. and Batzer, M.A. (
Tighe, P.J., Stevens, S.E., Dempsey, S., Le Deist, F., Rieux-Laucat, F. and Edgar, J.D. (
Ganguly, A., Dunbar, T., Chen, P., Godmilow, L. and Ganguly, T. (
Ricci, V., Regis, S., Di Duca, M. and Filocamo, M. (
Sorek, R., Ast, G. and Graur, D. (
Mustajoki, S., Ahola, H., Mustajoki, P. and Kauppinen, R. (
Claverie-Martin, F., Gonzalez-Acosta, H., Flores, C., Anton-Gamero, M. and Garcia-Nieto, V. (
Pfaffl, M.W., Horgan, G.W. and Dempfle, L. (
Laporte, J., Biancalana, V., Tanner, S.M., Kress, W., Schneider, V., Wallgren-Pettersson, C., Herger, F., Buj-Bello, A., Blondeau, F., Liechti-Gallati, S. et al. (
Buj-Bello, A., Laugel, V., Messaddeq, N., Zahreddine, H., Laporte, J., Pellissier, J.F. and Mandel, J.L. (
Oldridge, M., Zackai, E.H., McDonald-McGinn, D.M., Iseki, S., Morriss-Kay, G.M., Twigg, S.R., Johnson, D., Wall, S.A., Jiang, W., Theda, C. et al. (
Miki, Y., Katagiri, T., Kasumi, F., Yoshimoto, T. and Nakamura, Y. (
Hentze, M.W. and Kulozik, A.E. (
Cartegni, L., Chew, S.L. and Krainer, A.R. (
Lin, L., Faraco, J., Li, R., Kadotani, H., Rogers, W., Lin, X., Qiu, X., de Jong, P.J., Nishino, S. and Mignot, E. (
Kirkness, E.F., Bafna, V., Halpern, A.L., Levy, S., Remington, K., Rusch, D.B., Delcher, A.L., Pop, M., Wang, W., Fraser, C.M. et al. (