Distinct mutations with different inheritance mode caused similar retinal dystrophies in one family: a demonstration of the importance of genetic annotations in complicated pedigrees

Our study, conformed to the Declaration of Helsinki, was approved and prospectively reviewed by the local ethics committee of People Hospital of Ningxia Hui Autonomous Region (No. 10 [2017]). Eleven participants from family A (Fig. 1a) and 14 participants from family B (Fig. 1b) were recruited from the People’s Hospital of Ningxia Hui Autonomous Region. Written informed contents were obtained from all participants or their legal guardians before their enrollments. Peripheral blood samples were collected from all 25 participants for genomic DNA extraction. Family history and consanguineous marriages were carefully reviewed. Medical records were obtained from all participants. Each participant received general ophthalmic evaluations, while comprehensive ophthalmic examinations were selectively conducted on the eight included patients. Another 150 Chinese healthy controls free of major ocular problems were recruited with their blood samples donated.

To reveal the disease causing mutation in the two families, we selectively performed whole exome sequencing (WES) on three participants in family A (A-IV:3, A-VI:2 and A-VI:3) and two patients in family B (B-III:4 and B-IV:1). WES was conducted with the 44.1 megabases SeqCap EZ Human Exome Library v2.0 (Roche NimbleGen, Madison, WI) for enrichment of 23588 genes on patients from family A [6], and with SureSelect Human All Exon V6 60 Mb Kit (Agilent Technologies, Santa Clara, CA) on patients from family B [7]. Briefly, qualified genomic DNA samples were randomly sheared by Covaris into 200–250 base pair (bp) fragments. Fragments were then ligated with adapters to both ends, amplified by ligation-mediated polymerase chain reaction (LM-PCR), purified, and hybridized. Non-hybridized fragments were then washed out. Quantitative PCR was further applied to estimate the magnitude of enrichment of both non-captured and captured LM-PCR products. Each post-capture library was then loaded on an Illumina Hiseq 2000 platform for high-throughput sequencing.

Raw data were initially processed by CASAVA Software 1.7 (Illumina) for image analysis and base calling. Sequences were generated as 90 bp pair-end reads. Reads were aligned to human h19 genome using SOAPaligner (http://​www.​soap.​genomics.​org.​cn) and Burrows-Wheeler Aligner (BWA; http://​www.​bio-bwa.​sourceforge.​net/​). Only mapped reads were included for subsequent analysis. Coverage and depth were determined based on all mapped reads and the exome region. Atlas-SNP2 and Atlas-Indel2 were applied for variant calling [8]. Variant frequency data were obtained from the following six single nucleotide polymorphism databases, including dbSNP144 (http://​www.​hgdownload.​cse.​ucsc.​edu/​goldenPath/​hg19/​database/​snp135.​txt.​gz.), HapMap Project (ftp://​ftp.​ncbi.​nlm.​nih.​gov/​hapmap), 1000 Genome Project (ftp://​ftp.​1000genomes.​ebi.​ac.​uk/​vol1/​ftp), YH database (http://​yh.​genomics.​org.​cn/​), Exome Variant Server (http://​www.​evs.​gs.​washington.​edu/​EVS/​), and Exome Aggregation Consortium (http://​exac.​broadinstitute.​org/​). Variants with a minor allele frequency of over 1% in any of the above databases were discarded. Sanger sequencing was employed for mutation validation and prevalence test in 150 additional controls using a previously defined protocol [9]. Primer information is detailed in Additional file 1: Table S1 and Additional file 2: Table S2.

We applied vector NTI Advance™ 2011 software (Invitrogen, Carlsbad, CA) to analyze the conservation of the mutated reside by aligning protein sequence of human TULP1 (ENSP00000229771) with sequences of the following orthologues proteins: P. troglodytes (ENSPTRP00000030898), C. lupus (ENSCAFP00000001922), B. taurus (ENSBTAP00000055698), M. musculus (ENSMUSP00000049070), G. gallus (ENSGALP00000010281), D. rerio (ENSDARP00000099556), D. melanogaster (FBpp0088961), and C. elegans (F10B5.4). Crystal structural modeling of the wild type and mutant TULP1 proteins were constructed with SWISS-MODEL online server [10, 11], and displayed with PyMol software.

Two patients from family A, A-VI:2 and A-VI:3, and six patients from family B, B-II:4, B-III:3, B-III:5, B-IV:1, B-IV:2 and B-IV:4, were included in the present study with their clinical details summarized in Table 1. Ophthalmic features of patient A-V:2 were obtained according to his medical records, and were presented in Table 1. All patients from the two families were clinically diagnosed with RP. In family A, all three patients had early onset nyctalopia and rapid disease progress. Best corrected visual acuity was light perception for both patients A-VI:2 and A-VI:3 at their last visit to our hospital at the ages of 25 and 24 respectively. Typical RP presentations and macular degeneration were detected upon their ophthalmic evaluations (Fig. 2A–G and Table 1). In family B, RP onset ages ranged from early childhood to 50 years old (Table 1). RP progression also varied among the 6 patients. Patients B-IV:1 and B-IV:2 reported to have nyctalopia since early childhood, while the other four patients showed RP symptoms elder than 30-year-old. On examination, typical RP presentations were detected for all 6 patients, while patient B-II:4 also had chronic angle closure glaucoma in her right eye (Fig. 2H–S). Noteworthy, all 6 patients presented mild to severe cataracts (Table 1). Patient B-III:3 received bilateral cataract surgeries 2 years ago. No systemic defect was noticed in any of the included patients.

To identify the pathogenic mutations, WES with high quality was selectively performed on individuals A-IV:3, A-VI:2, and A-VI:3 from family A (mean coverage: 98.16%; mean depth: 70.89×), and patients B-III:5 and B-IV:1 from family B (mean coverage: 98.32%; mean depth: 104.66×). NGS data were summarized in Additional file 3: Table S3. Exon-specific coverage report of all known RP genes was presented in Additional file 4: Table S4. For family A, patients A-VI:2 and A-VI:3 were born to parents with consanguineous marriage, supporting potential autosomal recessive inheritance. WES identified 10 homozygous variants and 6 compound heterozygous variants shared by patients A-VI:2 and A-VI:3 (Additional file 1: Table S1). However, Sanger sequencing revealed no variant co-segregated with the disease phenotype. We thus hypothesized that the two patients may have distinct RP causing mutations. Based on WES data, patient A-VI:2 carried a recurrent homozygous C8ORF37 mutation c.555G>A (p.W185*; Fig. 1d and Table 2), while patient A-VI:3 had a novel hemizygous OFD1 mutation c.358A>G (p.T120A; Fig. 1c and Table 2).

As to family B, WES revealed one homozygous variant and 18 compound heterozygous variants shared by patients B-III:4 and IV:2 (Additional file 2: Table S2), while no variant was validated co-segregated with the disease phenotype. According to the family pedigree, patients B-IV:1 and B-IV:2 were born to unaffected parents with consanguineous marriage, indicating a potential autosomal recessive inheritance pattern. However, the RP phenotypes of patients B-III:3 and B-III:4 were likely inherited from the affected mother B-II:4, suggesting a dominant inheritance mode. Upon this hypothesis, a novel homozygous TULP1 mutation c.1255C>T (p.R419W; Fig. 1e and Table 2) was revealed as RP causative for patients B-IV:1 and B-IV:2, and a recurrent heterozygous RP1 mutation c.2285_2289delTAAAT (p.L762Yfs*17; Fig. 1f; Table 2) was found in patients B-II:4, B-III:3 and B-III:4. The mutated residue R419 in TULP1 was highly conserved among all tested species (Fig. 1g). Crystal structures of the wild type and mutant TULP1 proteins were obtained based on human TUB protein (Protein Data Bank ID: 1S31) with a sequence identify of 75.19 and a sequence similarity of 0.54. Our data suggested that the substitution from arginine to tryptophan at residue 419 would lead to the elimination of two hydrogen bonds between residue 419 and residues V488 and S534 (Fig. 1h, i), further supporting that this mutation would disturb the tertiary structure of TULP1 and interrupt its function. Residues R419, N463, V488 and S534 were conserved between TULP1 and TUB proteins (Fig. 1j). All four mutations identified in the two families segregated with the disease phenotype (Fig. 1a, b), and were confirmed absent in 150 Chinese controls free of major ocular problems.