Introduction

SNAP25 on chromosome 20p12.2 encodes a synaptosomal-associated protein of 25 kDa (SNAP25), which is mainly expressed in neurons and neuroendocrine cells [1, 2]. SNAP25 is a plasma membrane soluble N-ethylmaleimide-sensitive factor attachment receptor (SNARE) protein that forms a specific SNARE complex together with syntaxin and synaptobrevin proteins [3]. In synaptic vesicle exocytosis, the SNARE complex enables neuronal vesicles to release their neurotransmitters by mediating calcium-triggered vesicle fusion with the plasma membrane [4, 5].

SNAP25 transcripts exist as two isoforms: SNAP25a and SNAP25b. Snap25a and Snap25b were differentially expressed during mouse development, and were predominantly localized in the embryonic and adult brains of mice, respectively [6]. Snap25b-deficient mice (with protected Snap25a expression) demonstrated neurological hyperactivity, anxiety, learning deficits, and spontaneous seizures [7]. Therefore, Snap25b might be required for synaptic maturation and neuronal function, though the functional difference between the two isoforms remains elusive [8]. To date, six de novo mutations in SNAP25 have been reported in individuals with different types of seizures [9,10,11,12]. Here, we report on two affected siblings with a novel SNAP25b-specific mutation detected by whole-exome sequencing (WES), and discuss its clinical phenotype together with those of previously reported mutations.

Subjects and genetic analysis

An Israeli family with two affected siblings showing seizures and cerebellar dysfunction was recruited in this study (II-2 and II-3; Fig. 1a). The study protocol was approved by Institutional Review Boards of Yokohama City University School of Medicine. Clinical features are summarized in Table 1 and Supplementary information. Samples were collected from familial members after parental consent was given. WES was performed using DNA extracted from peripheral blood leukocytes from one of the affected individuals (II-3) as described in the Supplementary information. We focused on variants under the autosomal dominant or autosomal recessive inheritance model. Stepwise variant selection is described in Supplementary Table S1. Candidate variants were sequenced in all family members (I-1, I-2, II-1, II-2, and II-3) by Sanger sequencing. As for the mosaic c.176G > C mutation in SNAP25 found in the father (I-1), deep sequencing was performed using Miseq on PCR products amplified from DNA of blood leukocytes, saliva, hair roots, and nails (Supplemental information).

Fig. 1
figure 1

Familial pedigree and SNAP25 mutation. a Familial pedigree. The bold asterisk shows the somatic mosaic c.176G > C mutation in I-1. b The two SNAP25 isoforms are shown with previously reported mutations listed above. The novel p.Arg59Pro mutation is shown in bold. E1–E4 exon 1–exon 4, E5a or E5b Exon 5 of SNAP25a or SNAP25b, E6–E8 exon 6–exon 8. c Evolutionary conservation of p.Arg59 in SNAP25. d Human SNAP25b (NP_570824.1) protein structure. t-SNARE CC1/2 t-SNARE coiled-coil homology 1/2, CR cysteine-rich domain

Full size image
Table 1 Summary of clinical features in individuals with SNAP25 mutations
Full size table

Mutation detection

In WES of individual II-3, 94.7% of RefSeq coding DNA sequence (CDS) was covered by 20 reads or more, and the average read depth for the CDS was 75.1×. After variant selection, only the missense SNAP25 variant c.176G > C (p.Arg59Pro; NM_130811.2) remained (Supplementary Table S1 and Fig. 1b). This variant was likely to be pathogenic because it was predicted to be deleterious by three prediction tools (SIFT = 0.001, Polyphen2 = 0.998, and MutationTaster = disease causing). The variant was absent in our in-house Japanese exome data (n = 575), dbSNP build 138, the NHLBI Exome Sequencing Project (http://evs.gs.washington.edu/EVS/), the Exome Aggregation Consortium (http://exac.broadinstitute.org/), the Human Genetic Variation Database (http://www.hgvd.genome.med.kyoto-u.ac.jp/), and Tohoku Medical Megabank Organization database (https://ijgvd.megabank.tohoku.ac.jp/). The p.Arg59 residue is highly evolutionarily conserved from flies to humans in the t-SNARE coiled-coil homology 1 domain (Fig. 1c, d). Sanger sequencing revealed heterozygous changes in the affected individuals and the absence of this mutation in their unaffected mother and elder sister (Supplementary Figure S1a). The father’s electropherogram showed a lower peak for the mutant allele (Supplementary Figure S1a and S1b). Targeted deep sequencing produced read fractions containing the mutant allele from the father’s DNA extracted from peripheral blood leukocytes, saliva, hair roots and nails of 24.9, 30.8, 52.4, and 26.9%, respectively.

Discussion

The p.Arg59Pro mutation within exon 5 is specific to SNAP25b (Fig. 1b). Two splicing variants of SNAP25 contain alternative exon 5s (each 118 bp in size). SNAP25a is highly homologous to SNAP25b, though nine amino acid residues corresponding to different exon 5 are different (Supplementary Figure S2). Of the six previously reported mutations in SNAP25, four missense mutations (p.Lys40Glu, p.Gly43Arg, p.Val48Phe, and p.Asp166Tyr), and one nonsense mutation (p.Gln174*) were found in exons common to the two isoforms, whereas one missense mutation (p.Ile67Asn) was seen in exon 5 of SNAP25b (Fig. 1b) [9,10,11,12].

Most patients reported to carry SNAP25 mutations presented with various types and severities of seizures, intellectual disability, severe speech delay, neurological features including ataxia and muscle weakness, and abnormal brain electroencephalograms (EEGs) or abnormal magnetic resonance imaging (MRI) (Table 1). Although the current affected siblings had generalized tonic–clonic seizures and cerebellar dysfunction with muscle clumsiness, their intelligence were normal with no facial dysmorphism, and no abnormal MRI findings. Furthermore, one affected individual (II-2) had no abnormality in their interictal EEG. Their symptoms were therefore much milder than those of previously reported patients. It attracts our attention that a patient with the p.Ile67Asn mutation showed intellectual disability, severe speech delay and intractable seizures [10]. Thus the exon 5b specific mutations (p.Arg59Pro and p.Ile67Asn) are not necessarily associated with mild clinical phenotypes. SNAP25 with a high pLI score of 0.96 indicates that haploinsufficiency is the most likely mechanism of pathogenicity. Indeed, Snap25+/− mice (with both isoforms affected) showed cognitive and learning deficits, were susceptible to seizures, and showed an abnormal EEG pattern, which was similar to clinical findings in patients with SNAP25 mutations (Table 1) [13, 14]. Interestingly, p.Ile67Asn was reported to act in a dominant-negative manner [10], possibly explaining the difference of clinical features between p.Arg59Pro and p.Ile67Asn.

In conclusion, we report a novel missense SNAP25 mutation in an Israeli family with affected siblings with seizures. This mutation is the second SNAP25b-specific mutation. Further studies are needed to investigate genotype–phenotype correlations in SNAP25 mutations as well as their pathomechanisms.