Proximal spinal muscular atrophy (SMA) is a neuromuscular disorder caused by homozygous mutations of the SMN1 gene. SMN1 interacts with multiple proteins with functions in snRNP biogenesis, pre-mRNA splicing and presumably neural transport. SMN2, a nearly identical copy of SMN1, produces predominantly exon 7-skipped transcripts, whereas SMN1 mainly produces full-length transcripts. The SR-like splicing factor Htra2-beta1 facilitates correct splicing of SMN2 exon 7 through direct interaction with an exonic splicing enhancer within exon 7. In rare cases, siblings with identical 5q13-homologues and homozygous absence of SMN1 show variable phenotypes, suggesting that SMA is modified by other factors. By analysing nine SMA discordant families, we demonstrate that in all families unaffected siblings produce significantly higher amounts of SMN, Gemin2, Gemin3, ZPR1 and hnRNP-Q protein in lymphoblastoid cell lines, but not in primary fibroblasts, compared with their affected siblings. Protein p53, an additional SMN-interacting protein, is not subject to an SMN-dependent regulation. Surprisingly, Htra2-beta1 is also regulated by this tissue-specific mechanism. A similar regulation was found in all type I-III SMA patients, although at a different protein level than in discordant families. Thus, our data show that reduced SMN protein levels cause a reduction in the amount of its interacting proteins and of Htra2-beta1 in both discordant and non-discordant SMA families. We provide evidence that an intrinsic SMA modifying factor acts directly on the expression of SMN, thus influencing the SMA phenotype. Further insights into the molecular pathway and the identification of SMA modifying gene(s) may help to find additional targets for a therapy approach.

The survival motor neuron (SMN) protein and the SMN interacting protein 1 (SIP1) are part of a 300 kD protein complex with a crucial role in snRNP biogenesis and pre-mRNA splicing. Both proteins are colocalised in nuclear structures called gems and in the cytoplasm. Approximately 96% of patients with autosomal recessive spinal muscular atrophy (SMA) show mutations in the SMN1 gene, while about 4% fail to show any mutation, despite a typical SMA phenotype. Additionally, sibs with identical 5q13 homologs and homozygous absence of SMN1 can show variable phenotypes which suggest that SMA is modified by other, yet unknown factors. Since both genes, SMN1 and SIP1, belong to the same pathway and are part of the same protein complex, it is obvious to ask whether mutations within SIP1 are responsible for both the phenotypic variability and the appearance of non-SMN mutated SMA patients. First, we identified the chromosomal location of SIP1 and assigned it to chromosomal region 14q13-q21 by fluorescence in situ hybridisation. No SMA related disorder has yet been assigned to this chromosomal region. Next, we determined the exon-intron structure of the SIP1 gene which encompasses 10 exons and identified five transcription isoforms. We sequenced either RT-PCR products or genomic DNA covering the complete coding region from 23 typical SMA patients who had failed to show any SMN1 mutation. No mutation and no polymorphism was found within SIP1. Additionally, we sequenced RT-PCR products or genomic fragments of the entire SIP1 coding region from 26 sibs of 11 SMA families with identical genotypes (∆7SMN/∆7SMN or ∆7SMN/other mutation) but different phenotypes and again no mutation was found. Finally, we performed quantitative analysis of RT-PCR products from the same 26 sibs. No difference in expression level of the five isoforms among phenotypically variable sibs was observed. Based on these data, we suggest that neither the phenotypic variability nor the 5q-unlinked SMA are caused by mutations within SIP1.

Spinal muscular atrophy (SMA) is a neurodegenerative disease caused by homozygous mutations of the survival motor neuron gene 1 (SMN1). In rare cases sibs with identical 5q13-homologs and identical SMN1 mutations can show variable phenotypes from unaffected to affected, suggesting the influence of modifying genes. SMN is part of an 800 kDa macromolecular complex that plays an essential role in snRNP biogenesis and pre-mRNA splicing. Due to a single nucleotide difference within SMN1 exon 7 that disrupts an exonic splicing enhancer (ESE), SMN2, a nearly identical copy of SMN1, predominantly expresses alternatively spliced transcripts lacking exon 7, whereas SMN1 mainly produces full-length transcripts. The SR-like trans-acting splicing factor Htra2-beta1 was shown to interact with this ESE and to restore full-length SMN2 expression in vivo in a concentration-dependent manner. Since Htra2-beta1 prevents skipping of exon 7 it is obvious to ask whether mutations within Htra2-beta1 are responsible for the intrafamilial variability of the SMA phenotype. We sequenced either RT-PCR products or genomic DNA covering the complete coding region of Htra2-beta1 as well as the putative promoter of 36 sibs belonging to 15 SMA families with discordant phenotypes but identical genotypes. Neither a mutation nor a polymorphism was found within Htra2-beta1. Additionally, we performed quantitative analysis of Htra2-beta isoforms from 26 sibs without identifying any significant difference between phenotypically discordant sibs. Based on these data, we suggest that the intrafamilial phenotypic variability in SMA families is not caused by polymorphic variants or transcription differences within Htra2-beta1.