A novel human-specific splice isoform alters the critical C-terminus of Survival Motor Neuron protein

Spinal muscular atrophy (SMA), a leading genetic disease of children and infants, is caused by mutations or deletions of Survival Motor Neuron 1 (SMN1) gene. SMN2, a nearly identical copy of SMN1, fails to compensate for the loss of SMN1 due to skipping of exon 7. SMN2 predominantly produces SMNΔ7, an unstable protein. Here we report exon 6B, a novel exon, generated by exonization of an intronic Alu-like sequence of SMN. We validate the expression of exon 6B-containing transcripts SMN6B and SMN6BΔ7 in human tissues and cell lines. We confirm generation of SMN6B transcripts from both SMN1 and SMN2. We detect expression of SMN6B protein using antibodies raised against a unique polypeptide encoded by exon 6B. We analyze RNA-Seq data to show that hnRNP C is a potential regulator of SMN6B expression and demonstrate that SMN6B is a substrate of nonsense-mediated decay. We show interaction of SMN6B with Gemin2, a critical SMN-interacting protein. We demonstrate that SMN6B is more stable than SMNΔ7 and localizes to both the nucleus and the cytoplasm. Our finding expands the diversity of transcripts generated from human SMN genes and reveals a novel protein isoform predicted to be stably expressed during conditions of stress.


RT-PCR
Total RNA was isolated using TRIzol reagent (Life Technologies) following the manufacturer's recommendations, digested using RQ1 RNase-free DNase (Promega), then further purified using a Qiagen RNeasy Mini RNA purification kit or phenol:chloroform (OmniPur) extraction followed by ethanol precipitation. RNA concentration was measured using a BioMate 3 spectrophotometer (Thermo Scientific). RNA isolated from human tissue was purchased from Ambion (FirstChoice Human Total RNA Survey panel); samples from each tissue were pooled from three healthy individuals. cDNA was generated from either 0.5 µg (for human tissues) or 1.6-2.0 µg (for cell lines and mouse tissues) of total RNA in a 10 µl reaction using SuperScript III (Life Technologies) and oligo (dT) [12][13][14][15][16][17][18] , oligo (dT) with adapter sequence, or a gene-specific primer (3ʹE8-Dde). cDNA was amplified using Taq DNA Polymerase (New England Biolabs) in the presence of either a 5ʹ-end-P 32 -labelled primer or a trace amount of [α-32 P] dATP (3,000 Ci/mmole; Perkin-Elmer). MESDA was performed as previously described 10,11 . In all cases PCR products were amplified and resolved on a native polyacrylamide gel. For DdeI and BglII digestion PCR products were purified by phenol:chloroform extraction and ethanol precipitation prior to overnight restriction digestion. Analysis and quantifications of splice products were performed using a FPL-5000 Image Reader and Multi Gauge software (Fuji Photo Film Inc). For quantitative real-time PCR (QPCR), reactions were carried out in 20 µl using 1X FastStart Universal SYBR Green Master Mix (Roche), 300 nM of each primer, and 1.5 µl of a 1:20 dilution of cDNA template in a Stratagene Mx3005P thermocycler. For negative controls, cDNA was synthesized in the absence of RTase or water was used as a template during PCR amplification. List of all primers used in PCR is given in Supplementary Table 1.

Whole-cell lysate preparation
Cells were washed with ice-cold PBS three times and collected by scraping. Cell pellets were either immediately used for making whole-cell lysates or were snap-frozen in liquid nitrogen and The lysates were then centrifuged at 13,000 rpm for 20 min at 4 ˚C to remove cell debris. Protein concentration was measured using the Bradford assay (Bio-Rad protein assay).

Generation of polyclonal antibodies (anti-6B-001) and purification of Immunoglobulin G
A cysteine was added toward the N-terminus of the synthetic peptide for conjugation with keyhole limpet haemocyanin (KLH). The 6B polypeptide was then conjugated to KLH using Imject TM Maleimide-Activated mcKLH Spin Kit (Thermo Scientific). Mice were injected with 25 µg of KLH-linked 6B polypeptide by intraperitoneal (IP) route on days 1, 14 and 28 days followed by a bleed on day 42. At this time the sera was tested for immunogenicity and found to have a low response. A fourth injection was then administered, followed by a second bleed 14 days later. The sera was then retested and found to have a high enough response (1:500). Mice were then primed using Pristane (Sigma), and after 5 days administered an IP injection of SP2/O cells. After 10 days, mice were euthanized and polyclonal fluids were collected.
Immunoglobulin (IgG) was purified from the polyclonal fluids with Protein A column (Pierce). Briefly, 2 ml of the polyclonal fluids was diluted 1:1 with Protein A IgG Binding buffer (Pierce) and the applied to a Protein A column. After adding 1 ml protein A plus Agarose, the samples were incubated with Agarose for 2 h with rotating at room temperature. Following the incubation the Agarose were washed three times by rotating for 10 min at room temperature. Incubation of the cells with the antibodies was carried overnight at 4°C. Next morning cells were extensively washed with 1XPBS followed by mounting on glass slides using Vectashield Mounting medium with DAPI (Vector Laboratories). Subsequently coverslips were permanently sealed around the perimeter with nail polish.

Comparative Modeling
A representative model of SMN6B protein was calculated using the RosettaCM protocol 12 6,19 . For modeling of the remaining regions of SMN6B, 3 and 9 residue fragment libraries were derived by a sequencebased search of the Protein Data Bank using the Robetta server 20 . The Grishin format sequence alignments were used for threading the SMN6B sequence into each structure template to generate partial models. The threaded models and fragment libraries were used to generate 100 hybrid models derived from multiple templates through low-resolution Monte Carlo based sampling and quasi-Newton minimization to optimize loop closure and backbone geometry.
Selected models with intact YG Box helices were further refined using a full-atom realistic energy landscape and the lowest energy model was selected as the representative model of the SMN6B protein.
Models of the YG domain of SMN∆7 and SMN6B isoforms were calculated with C2 symmetry using both the human and yeast YG domain crystal structures as templates 6,19 . The C2 symmetry definition was generated with the make_symmdef_file application using the yeast SMN YG domain structure (PDB code: 4RG5) and the noncrystallographic (point) symmetry option 21 . The sequences for comparative modeling were restricted to the C-terminal residues that correlate to the predicted YG domain including residues 252-282 for the SMN∆7 isoform and residues 252-294 for the SMN6B isoform. After threading these sequences into the YG domain structure templates, the hybridize application was used to calculate 10 models for each construct. The hybrid models were screened for dimer formation and an intact YG Box helix. Selected hybrid models were further refined to generate representative models of the coiled-coil dimers for both the SMN∆7 and SMN6B YG domains.