Jos Jansen

40 Chapter 2 Mutational Analyses To uncover the genetic defect, we performed exome sequencing of individuals F1-II1 and F1-II2 from index family F1. Eight possible candidates were identified on the basis of having autosomal-recessive inheritance (Table S2). Among these candidates was a homozygous missense variant in CCDC115 (c.92T > C [p.Leu31Ser]) (Table 1). We performed a profile-based method, Position-Specific Iterated (PSI)-BLAST,25 to identify possible homologs of the candidate variants and identifiedVma22p (GenBank:NP_011927.1) as theyeasthomologofCCDC115 (GenBank: NP_115733.2) in the second iteration with an E-value of 2e-14 and a reciprocal E-value of 3e-11. Vma22p is a dedicated ER-localized assembly factor of the V-ATPase.(9,26,27) Importantly, Vma22p and CCDC115 were found as each others’ best hits. This suggests that, apart from being homologs, they are likely orthologs with overlapping functions in humans.28 Based on the link between the V-ATPase and abnormal glycosylation, this variant was considered our most likely candidate.7 This was further supported by homozygosity mapping, indicating a small homozygous region on chromosome 2, in which CCDC115 was located (Figure S1). In silico analysis of the p.Leu31Ser substitution with SIFT, PolyPhen-2, and MutationTaster predicted pathogenicity (Table S3). The ExAC database showed a very low allele frequency of 8.253e-06. Sanger sequencing confirmed homozygosity for the affected individuals, heterozygosity for both parents, and homozygous wild-type sequence for a healthy sibling, confirming complete segregation in the family (Figure 1A). Western blotting of fibroblasts derived from individual F1-II4 demonstrated a protein level similar to that of healthy control individuals (Figure S2). For individuals F2-II1 and F5-II1, exome sequencing revealed multiple genetic variants, among which was the same c.92T > C homozygous missense variant (Table 1). Sanger sequencing for individual F2-II1 confirmed homozygosity and heterozygosity for the parents (Figure 1A). For individual F5-II1, Sanger sequencing of parental DNA revealed a maternal heterozygous missense mutation and a paternal wild-type sequence (Figure 1A). We suspected a paternal deletion as possible explanation, given that haplotype analysis excluded non- paternity. MLPA of DNA from individual F5-II1 displayed a heterozygous deletion for all exons of CCDC115 and on the studied position in upstream PTPN18 (Table S4).We did not observe a deletion for the position we investigated within downstream SMPD4 . Segregation analysis showed that the deletion originated