Survival of Motor Neuron (SMN1 and SMN2)

Survival of Motor Neuron 1 (SMN1) and its nearly identical centromeric paralog Survival of Motor Neuron 2 (SMN2) are two tandemly duplicated genes located within a complex ~500-kb inverted segmental duplication on human chromosome 5q13.2 (9 exons designated Exons 1, 2a, 2b, 3, 4, 5, 6, 7, and 8 spanning ~28 kb) that encode the 294-amino-acid (~38 kDa) Survival of Motor Neuron (SMN) protein. Inside the cytoplasm and nuclear Cajal bodies (gems) of all eukaryotic cells, SMN forms the oligomeric core of the macromolecular SMN complex (SMN + Gemins 2–8 + Unrip), an ATP-dependent molecular chaperone required to assemble the heptameric Sm protein ring (SmB/B′, SmD1, SmD2, SmD3, SmE, SmF, SmG) onto uridine-rich small nuclear RNAs (snRNAs: U1, U2, U4, U5, U11, U12, U4atac) to build the core ribonucleoprotein particles (snRNPs) of both the major (U2-dependent) and minor (U12-dependent) spliceosomes.

In 1995, Judith Melki and colleagues discovered that 95% of cases of Spinal Muscular Atrophy (SMA)—an autosomal recessive neurodegenerative disease of spinal cord α-motor neurons (~1 in 10,000 live births; carrier frequency ~1 in 40 to 1 in 50)—are caused by homozygous deletion or gene conversion of telomeric SMN1 (with the remaining ~5% carrying a deletion on one allele and a subtle intragenic point mutation on the other). Because complete genetic loss of both SMN1 and SMN2 is early-embryonic lethal (as occurs in Smn^−/− knockout mice, since rodents and non-hominid primates possess only a single Smn1 gene and lack SMN2), every human SMA patient survives in utero only because they retain 1 to 6 copies of centromeric SMN2. However, a single translationally silent C→T nucleotide transition at position +6 of Exon 7 in SMN2 causes the spliceosome to skip Exon 7 in ~85–90% of SMN2 pre-mRNA transcripts, establishing SMN2 Exon 7 splicing rescue as one of the landmark triumphs of precision drug discovery.

Five-Nucleotide Divergence Between SMN1 and SMN2

Across their entire ~28-kb genomic sequences, telomeric SMN1 and centromeric SMN2 share >99.9% sequence identity, differing by only five canonical single-nucleotide substitutions near their 3′ ends (one in Intron 6, one in Exon 7, two in Intron 7, and one in the non-coding Exon 8 3′ UTR), none of which alter a single amino acid in the 294-residue protein sequence.

Yet the two paralogs exhibit opposite splicing outcomes at Exon 7 (54 bp, encoding residues 279–294 at the C-terminus of the SMN protein):

• Telomeric SMN1 (c.840C at position +6 of Exon 7): Includes Exon 7 with >95–99% efficiency, generating full-length FL-SMN mRNA and stable oligomeric SMN protein (294 aa, half-life >8–12 hours).
• Centromeric SMN2 (c.840T at position +6 of Exon 7): Skips Exon 7 in ~85–90% of transcripts, yielding Only ~10–15% full-length FL-SMN and ~85–90% SMNΔ7 mRNA.

As demonstrated by Adrian Krainer, Gideon Dreyfuss, and Ravindra Singh, the single synonymous C→T (C>U in pre-mRNA) transition at position +6 of Exon 7 (codon 280: TTC [Phe] in SMN1 → TTT [Phe] in SMN2) triggers Exon 7 skipping through a dual enhancer-loss / silencer-gain mechanism:

1. Disruption of an SRSF1 (SF2/ASF) Exonic Splicing Enhancer (ESE): In SMN1, the heptamer sequence CAGACAA (positions +6 to +12 of Exon 7) serves as a high-affinity binding site for the serine/arginine-rich splicing factor SRSF1, which recruits U2AF to the upstream weak 3′ polypyrimidine tract. The +6 C>U substitution (UAGACAA) abolishes SRSF1 binding.
2. Creation of an hnRNP A1/A2 Exonic Splicing Silencer (ESS): Simultaneously, the +6 UAG sequence creates a composite binding motif for the repressive RNA-binding protein hnRNP A1, which cooperates with a second hnRNP A1-binding site inside Intronic Splicing Silencer N1 (ISS-N1, positions +10 to +24 of Intron 7) and with a weak, non-canonical 5′ splice donor site (GGA/GUAAGU, which contains an inhibitory A at position -1 of Exon 7 and folds into an inhibitory terminal stem-loop 2 TSL2 hairpin) to sterically block U1 snRNP recruitment to the Exon 7–Intron 7 junction.

When Exon 7 (54 bp) is skipped, Exon 6 splices directly to non-coding Exon 8, introducing a premature stop codon just four triplets into Exon 8. This replaces the normal 16-amino-acid Exon 7 C-terminus (GFRQNQKEGRCSHSLN) with a short 4-amino-acid tail (EMLA). Because the YG-box self-oligomerization domain (residues 254–280) spans the Exon 6–Exon 7 boundary, truncated SMNΔ7 (282 aa) cannot self-oligomerize into stable octamers; worse, the YG-box + EMLA junction forms a potent 15-amino-acid C-terminal degron that triggers rapid polyubiquitination and 26S proteasomal destruction (half-life <2 hours).

Structurally, the 294-residue SMN monomer comprises three conserved functional modules: (1) an N-terminal Gemin2-binding domain (Exons 1–2a, residues 1–51) that forms the structural pincer gripping the pentameric/heptameric Sm ring; (2) a central aromatic Tudor domain (Exon 3, residues 91–151) folded into a five-stranded β-barrel with an aromatic cage (Trp102, Tyr109, Tyr127, Tyr130) that specifically recognizes symmetrically dimethylated arginine (sDMA) residues (RG/RGG tails modified by the methylosome PRMT5/MEP50) on SmD1, SmD3, and SmB/B′ (missense mutations such as E134K inside the Tudor domain abolish Sm binding and cause severe SMA); and (3) a C-terminal YG-box dimerization helix (Exons 6–7, residues 254–280).

Why does systemic SMN deficiency selectively destroy spinal α-motor neurons and disrupt the neuromuscular junction while sparing most non-neuronal tissues? First, SMN reduction disproportionately depletes the low-abundance minor (U12-dependent) spliceosome snRNPs (U11, U12, U4atac/U6atac), causing intron retention and mis-splicing of U12-intron genes essential for proprioceptive Ia afferent–motor neuron sensory-motor circuits (such as Stasimon / TMEM41B and voltage-gated calcium channels). Second, SMN functions inside motor axons to traffic mRNP granules (β-actin and Agrn (z+ neural agrin)) to presynaptic nerve terminals.

How does the small-molecule drug risdiplam (Evrysdi) selectively restore SMN2 Exon 7 splicing without disrupting thousands of other human exons?

The 5′ splice donor site of SMN2 Exon 7 (GGA/GUAAGU) is anomalous because its last exonic nucleotide (position −1) is an adenosine (A-1) rather than the consensus guanine (G-1). When U1 snRNA base-pairs with this 5′ splice site, A-1 is forced to bulge out of the RNA–RNA duplex helix, weakening U1 snRNP binding. Risdiplam acts as a sequence-specific "molecular glue" that intercalates into the major groove of the U1 snRNA : SMN2 Exon 7 pre-mRNA duplex right at the bulged A-1 pocket, locking U1 snRNP onto the weak splice donor and Cooperating with Far Upstream Element-Binding Protein 1 (FUBP1)/KHSRP at an exonic enhancer (ESE2)—a bipartite structural signature shared by only a tiny handful of human exons (such as FOXM1 and STRN3).

Why do silent (2+0) SMA carriers evade standard quantitative PCR carrier screening tests?

Standard population carrier screening measures the total genomic copy number of SMN1 Exon 7, assuming that non-carriers have one SMN1 copy on each chromosome (1+1 = 2 copies) and carriers have one deleted chromosome (1+0 = 1 copy). However, ~3–5% of normal chromosomes carry a duplication of SMN1 on a single chromosome (2 copies in cis). An individual with 2 copies of SMN1 on one chromosome and 0 copies (a deletion) on the homologous chromosome (2+0 configuration) has a normal total of 2 SMN1 copies by qPCR or MLPA, yet has a 50% risk of transmitting the 0-copy deleted allele to offspring. Detecting 2+0 silent carriers requires phasing linked Chr5q13.2 single-nucleotide polymorphisms (g.27134T>G and g.27706_27707delAT) or long-read human pangenome sequencing.

Why is presymptomatic newborn screening critical for SMA therapeutic efficacy?

Spinal cord α-motor neurons are post-mitotic cells that undergo rapid degeneration during the late third trimester and the first 3 to 6 months of postnatal life in SMA Type I (>50–70% of the motor unit pool is already irreversibly lost by the time clinical hypotonia and areflexia appear). In the NURTURE (nusinersen) and SPR1NT (onasemnogene abeparvovec) trials, infants identified via universal Newborn Screening (NBS) and treated in the first 2 to 6 weeks of life—before motor neuron loss—achieved 92% to 100% independent sitting and walking on normal developmental timelines.

1. Lefebvre, S., Bürglen, L., Reboullet, S., et al. (1995). "Identification and characterization of a spinal muscular atrophy-determining gene." Cell, 80(1), 155–165. https://doi.org/10.1016/0092-8674(95)90460-390460-3)
2. Lorson, C. L., Hahnen, E., Androphy, E. J., & Wirth, B. (1999). "A single nucleotide in the SMN gene regulates splicing and is responsible for spinal muscular atrophy." Proceedings of the National Academy of Sciences, 96(11), 6307–6311. https://doi.org/10.1073/pnas.96.11.6307
3. Cartegni, L., & Krainer, A. R. (2002). "Disruption of an SF2/ASF-dependent exonic splicing enhancer in SMN2 causes spinal muscular atrophy in the absence of SMN1." Nature Genetics, 30(4), 377–384. https://doi.org/10.1038/ng854
4. Finkel, R. S., Mercuri, E., Darras, B. T., et al. (2017). "Nusinersen versus sham control in infantile-onset spinal muscular atrophy." New England Journal of Medicine, 377(18), 1723–1732. https://doi.org/10.1056/NEJMoa1702752