Alternative splicing of a cryptic exon embedded in intron 6 of SMN1 and SMN2

Both survival of motor neuron (SMN) genes are associated with spinal muscular atrophy; mutations in SMN1 cause the disease, and SMN2 modulates its severity. It is established that different alternative splicing of exon 7 occurs for SMN1 and SMN2, and a cryptic exon was recently found in intron 6 of both genes. Here, we characterize this cryptic exon and clarify its alternative splicing pattern in control and spinal muscular atrophy cells.

The survival of motor neuron (SMN) genes, SMN1 and SMN2, are closely related to the development and severity of spinal muscular atrophy (SMA). 1 Mutations in SMN1, but not in SMN2, are recognized as causing SMA because SMN1 is completely deleted in more than 95% of SMA patients and is deleteriously mutated in the remaining patients. 2 SMN2 is now considered to be a modifying factor of the SMA phenotype. 3,4 Without exception, all SMA patients retain SMN2. 2 Different patterns of alternative splicing have been observed for SMN1 and SMN2. 1 SMN1 exclusively produces a full-length (FL) functional SMN1 transcript. In contrast, less than 10% of SMN2 transcripts are FL, and more than 90% lack exon 7 (Δ7). 5 In SMA, the low levels of FL-SMN protein generated from SMN2 may partially compensate for the lack of SMN1. Progress in our understanding of SMN splicing mechanisms has spurred the development of therapeutic compounds that modify the splicing of SMN2 exon 7 to increase levels of FL-SMN protein, and several SMA clinical trials with such therapeutic compounds, including ISIS-SMNRx (nusinersen) and RG7800, are currently in progress. 6 However, new alternatively spliced products with a cryptic exon in intron 6 of both SMN1 and SMN2 have recently been registered in the Ensemble database: ENST00000506239.6 is an SMN1 transcript, and ENST00000506734.5 is an SMN2 transcript (the data for SMN1 are also registered as NCBI Reference Sequence XP_011541898.1). Because the cryptic exon is located upstream of exon 7, we named this exon 'exon 7a'. Although splicing of exon 7 in the SMN genes has been thoroughly studied, the details of exon 7a splicing have not yet been determined. In this study, we characterized the cryptic SMN exon, exon 7a, and clarified its alternative splicing pattern in control and SMA fibroblast cells. Control fibroblasts were grown from the skin of a 30-year-old healthy man who carried two copies of each of SMN1 and SMN2, and SMA fibroblasts were grown from the skin of a 13-year-old SMA female who carried three copies of SMN2 and no copy of SMN1. 7 In addition, control white blood cells were collected from a 21-year-old healthy female who carried two copies of SMN1 and no copy of SMN2. This study was approved by the Ethical Committee of Kobe University, and informed consent was obtained from the patient and her parents as well as the two healthy individuals.
First, to confirm the presence of an alternatively spliced SMN product with exon 7a, we performed reverse transcription-PCR (RT-PCR) analysis using primers to amplify exon 6 and exon 7a (Ex6-F: 5′-CTCCCATATGTCCAGATTCTCTTG-3′ and Ex 7a-R: 5′-CCCAGATCTTTGTGCATTAA-3′; Figure 1a). We obtained an amplified product from the control fibroblast cells ('Mock' in Figure 1c) and SMA fibroblast cells (data not shown), and DNA sequencing revealed that it corresponds to a transcript containing exon 6-exon 7a. This observation confirmed the presence of an alternative transcript with exon 7a in both the control and SMA fibroblasts. However, the exon 7a-containing SMN transcript level was too low to be detectable by RT-PCR using primers for exon 6 and exon 8 (data not shown). Thus, exon 7a-specific primers were necessary to detect the exon 7a-containing SMN transcript by RT-PCR.
Next, to determine the alternative splicing pattern of exon 7a, we performed fluorescent RT-PCR using exon 7a and exon 8 primers (Ex 7a-F: 5′-(FAM)-ACAGGGTTTCACTGTGTTAGCC-3′ and Ex-8R: 5′-ACTGCCTCACCACCGTGCTGG-3′) (Figure 2a). The major splicing product containing exon 7a was 'exon 6-exon 7a-exon 7-exon 8' in both control and SMA fibroblast cells (Figure 2b). However, the transcript containing 'exon 6-exon 7a-exon 8', which is described in the Ensemble database, was only a minor splicing product in our study. Interestingly, the 'exon 6-exon 7a-exon 8' transcript was not detected in white blood cells from the control individual retaining SMN1 but lacking SMN2 (Figure 2b). Although it is not certain if such a phenomenon occurs in every tissue, this observation suggests that the 'exon 6-exon 7a-exon 8' transcript is usually generated by SMN2 but not by SMN1 (Figure 2c).
We also determined the entire nucleotide sequence of exon 7a (Figure 1b). When the nucleotide sequence was examined within the context of the open reading frame of the SMN transcript, two premature termination codons, TGA and TAA, were identified at 58 and 17 bp upstream of exon 7a, respectively. Thus, the exon 7a-containing SMN transcript is a possible target of nonsensemediated decay (NMD) because it follows the '50-55 nt rule' of NMD. 8 Cycloheximide (CHX) is an NMD inhibitor. 9 To examine whether the exon 7a-containing SMN transcript is an NMD target, we compared CHX (100 μg/ml)-treated and non-treated (mock) control fibroblasts. The SMN transcript with exon 6 and 7a and a reference gene transcript (GAPDH) were amplified; the primers used to amplify GAPDH were 5′-GAGTCAACGGATTTGGTCGT-3′ and 5′-GACAAGCTTCCCGTTCTCAG-3′. The relative expression level of the exon 7a-containing SMN transcript in CHX-treated cells was four times higher than that in non-treated cells (Figure 1c). This finding strongly suggests that the exon 7a-containing SMN transcript is a target of NMD.
Further studies are necessary to clarify the biological function of the exon 7a-containing SMN transcript. As alternative splicing is a mechanism for the quantitative control of gene expression, which involves the use of NMD, 10 the exon 7a-containing SMN transcript may have a role in attenuating the function of the FL-SMN transcript or the FL-SMN protein. If this is the case, when SMA patients are treated with a splicing-modulation therapy, determining the expression level of the exon 7a-containing SMN transcript may be essential to avoid unexpected off-target effects of the therapeutic compounds.
In conclusion, we characterized SMN exon 7a and its splicing characteristics. SMN exon 7a has two premature termination codons and is a target of NMD. Further studies are needed to clarify whether alternative splicing involving exon 7a regulates SMN gene expression.

NOTE ADDED IN PROOF
During the submission of our manuscript, a study performed by Seo et al. 11 was published online. The authors report the same transcript in intron 6 and termed this 109 bp sequence 'exon 6B'. They confirmed our findings that exon 7a is derived from both SMN1 and SMN2 genes. In addition, they showed that the exon 6B-containing SMN transcript is the target of NMD, which is consistent with our results for exon 7a.
In our experiment using conventional PCR, the exon 7a-containing SMN2 transcript could not be amplified without the use of specifically targeted primers, suggesting that its expression level was much lower than that of the Δ7and FL-SMN2 transcripts. Using a 32 P-labeled multi-exon skipping detection assay (MESDA), Seo et al. 11 also demonstrated that the expression level of the exon 6B-containing SMN2 transcript was lower than that of the Δ7and FL-SMN2 transcripts.
Seo et al. 11 investigated the stability, function and localization of the SMN 6B protein, which contains exon 6B, through overexpression of SMN cDNA plasmid constructs with a 3XFLAGtag anti-FLAG antibodies. They used anti-FLAG antibodies to detect the 3XFLAG-tagged proteins. Their data indicated (1) that the stability of SMN 6B was intermediate between that of FL-SMN and Δ7-SMN, (2) that SMN 6B retained its ability to interact with an FL-SMN counterpart protein, Gemin 2 and (3) that SMN 6B was less localized in the nucleus compared with the cytoplasm.

HGV DATABASE
The relevant data from this Data Report are hosted at the Human Genome Variation Database at http://dx.doi.org/10.6084/m9.figshare.hgv.900.