Homozygous mutation of VPS16 gene is responsible for an autosomal recessive adolescent-onset primary dystonia

Dystonia is a neurological movement disorder that is clinically and genetically heterogeneous. Herein, we report the identification a novel homozygous missense mutation, c.156 C > A in VPS16, co-segregating with disease status in a Chinese consanguineous family with adolescent-onset primary dystonia by whole exome sequencing and homozygosity mapping. To assess the biological role of c.156 C > A homozygous mutation of VPS16, we generated mice with targeted mutation site of Vps16 through CRISPR-Cas9 genome-editing approach. Vps16 c.156 C > A homozygous mutant mice exhibited significantly impaired motor function, suggesting that VPS16 is a new causative gene for adolescent-onset primary dystonia.

Genetic screening of known dystonia-related genes. To uncover the underlying genetic defect in the patients, we initially screened 24 known dystonia-related genetic loci using targeted exome-sequencing approach. Neither of them was detected in these patients, suggesting a novel genetic defect is responsible for the dystonia symptom in the family.

Whole Exome-sequencing detects a missense variant in VPS16.
To uncover the underlying genetic defect in those patients, we sequenced the whole exomes of four affected individuals and two unaffected parents. Variants annotation and filtering analysis combined with homozygous mapping and IBD analysis from WES were used for screening candidate dystonia-causative variant in the family (Fig. 2). The study was approved by the Ethics Review Committee (ERC) of Shenzhen University 1 st Affiliated Hospital and BGI. WES produced approximately 100 million paired reads per sample, more than 99% of which were mapped to target region (supplementary Table S1). The average sequencing depth was 74.98X, with more than 85% reads covered > = 20X. Based on the assumption that variants that are common in the population are not likely to be the genetic cause of rare Mendelian diseases, such variants that were reported in the 1000 Genomes, dbSNP, HapMap, and YH databases with high frequency (> 0.5%) are initially filtered out from our analysis (supplementary Table S2). Those rare variants including 9 SNPs and 7 Indels that were shared in all cases and absent in controls were retained as prioritized candidate dystonia-causative variant (supplementary Table S3).
Since the proband (IV6) and two of her affected brothers (IV10, IV18) were offspring of first-cousin marriage, as well as one of her affected nephew (V5), suggesting an autosomal recessive inheritance of the disease in this family. Therefore, the dystonia-causative variant must be located in a case-shared homozygous IBD region [homozygosity-by-descent (HBD)] and having no overlap with the controls. Based on these criteria, three case-shared HBD regions identified from homozygosity mapping and IBD analysis were retained as candidate disease-causing genetic loci (supplementary Table S4): Chr1: 178006795-179040929 bp, Chr2: 42722370-42809031 bp, and Chr20: 1126746-4842635 bp (hg19; UCSC genome browser). Combined with variants filtering analysis we did in the first step, only two candidate recessive variants were found to be located in one of the HBD region on chromosome 20 ( Fig. 3  Both nucleotide substitutions in VPS16 and SIGLEC1 were confirmed in patient V5 by Sanger sequencing (Fig. 4a). His mother (IV4) and his daughter (VI4) with normal phenotype both carried heterozygous mutations in VPS16 and SIGLEC1, whereas V5 was homozygous for both mutations. And the familial segregation of both variants was consistent with disease status, showing an autosomal recessive inheritance in the family (Fig. 4a).
VPS16, encodes a core subunit of VPS-C complex that is required for vesicle transport and fusion process in the late endosomes/lysosomes pathway 20,21 . Importantly, VPS-C complex was found highly enriched in the brain and nervous system 22 , and interacted with syntaxin 1A and neuron specific SNARE proteins regulating pre-synaptic release 23 . The replacement of an Asn residue in the β -propeller domain of VPS16 with a Lys residue is highly conserved, and the homozygous missense mutation, c.156 C > A in VPS16, is predicted to have damaging  impact on protein function through MutationTaster prediction 24 , supporting the pathological role of amino acid substitution (Fig. 4b, supplementary Table S3). Therefore, we highly speculated a causative role of VPS16 gene mutation in the familial adolescent-onset primary dystonia. The other candidate gene, SIGLEC1 encodes an immunoglobulin superfamily protein with proinflammatory functions in macrophages 25 . We didn't pursue it further in our study, because wide-type Asp458 of SIGLEC1 is not evolutionarily conserved in vertebrate ( Fig. 4b and supplementary Table S3) and SIGLEC1-deficient mice only showed a minimal phenotype of immune system 26 . Therefore, based on its expression pattern and reported biological function, we considered that c.1372C > T substitution in SIGLEC1 is an irrelevant polymorphism.
Mutational screening of VPS16 in 200 normal controls. Given that dystonia only affects small numbers of people, it is unlikely the genetic defect would exist in the normal people. Therefore, we screened the c.156 C > A variant of VPS16 in 200 controls from individuals with the same geographic ancestry as the consanguineous family. No homozygous mutation in VPS16 was detected in 200 controls (Fig. 4a), indicating it is a rare polymorphism in normal cohort.
Mutational screening of VPS16 in 14 sporadic dystonia patients. To examine whether VPS16 was associated with dystonia in additional patients, we performed Sanger sequencing of the entire VPS16 exons in 14 unrelated sporadic cases with adolescent-onset dystonia compatible with autosomal recessive inheritance. However, novel or known rare (frequency < 1%) homozygous or compound heterozygous variants of VPS16 were not detected in any of these patients, as were protein-disrupting heterozygous variants, suggesting refinement of this phenotype requires further genetic screening in additional familial and sporadic dystonia cohorts. Generation of Vps16 c.156 C > A mutation mice through CRISPR-Cas9 genome-editing approach. To further investigate the pathological role of VPS16, we generated mice with targeted c.156 C > A mutation of Vps16 through CRISPR-Cas9 genome-editing approach. The mouse Vps16 (NM_030559.3) shows 97% identity with the human sequence (NM_022575.3). All animal experiments were approved by the life ethics and biological safety review committee of BGI (Shenzhen, China). A single-guild RNA (sgRNA) targeting the sequence to be mutated (Fig. 5a) was coinjected along with Cas9 mRNA and a single stranded oligonucleotide (Fig. 5a) containing the c.156 C > A mutation of Vps16 into one-cell stage C57BL/6J embryos. Following selective breeding of the Vps16 mutant mice, we successfully generated three mice that carried homozygous c.156 C > A mutations of Vps16. Successful HDR-mediated mutation was confirmed by Sanger sequencing (Fig. 5b).
Phenotype Analysis of Vps16 c.156 C > A mutant mice. Vps16 c.156 C > A mutant mice were indistinguishable from wild-type (WT) at birth, showing no obvious abnormality in the body size and weight. However, homozygous c.156 C > A mutation of Vps16 is leading to a reduction of VPS16 protein expression in mice (Fig. 5c). Moreover, when hung by their tails at 6 month-old, Vps16 mutant mice stayed almost still with their unbalanced hind limbs overstretched, showing obvious abnormality in the behavior compared to WT (supplementary video 1). To assess the motor balance and coordination, Vps16 mutant and WT mice were tested by the accelerating Rotarod analysis. Three Vps16 mutant mice all showed a significantly reduced latency to fall from the Rotarod compared to WT mice at 6 months of age (p < 0.0001), suggesting the impaired motor function (Fig. 5d, and supplementary video 2).

Discussion
In the past decade, the application of WES has facilitated the discovery of many causative genes of mendelian disorders in pedigree-based genetic mapping. Herein, our study successfully demonstrates the identification of a novel disease-causing gene VPS16 in a Chinese consanguineous family with adolescent-onset primary dystonia based on WES approach. The proband (IV6) and two of her affected brothers (IV10, IV18) are the offspring of first-cousin consanguineous marriage, as well as one of her affected nephew (V5), suggesting an autosomal recessive inheritance of the disease in this pedigree (Fig. 1a).
Combined with variants filtering analysis, homozygous mapping and IBD analysis from WES, a homozygous missense mutation, c.156 C > A of VPS16, was identified as the candidate dystonia-causative variant. Sanger sequencing confirms that the familial segregation of VPS16 variant was consistent with disease status, showing an autosomal recessive inheritance in the family (Fig. 4a).
The frequency of Vps16 c.156 C > A allele mutation has been reported to be 0.0029 in the East Asian population from the ExAc database (http://exac.broadinstitute.org/variant/20-2840713-C-A), whereas the homozygotes has not yet been reported. Moreover, among the 200 normal controls we screened, neither heterozygous nor homozygous mutation of VPS16 is detected (Fig. 4a), indicating it is a rare polymorphism in the normal cohort. So far, we have not identified mutations in VPS16 in additional unrelated patients, possibly due to limited patients we investigated. Therefore, refinement of this phenotype requires further genetic screening in additional familial and sporadic dystonia cohorts.
VPS16 has been reported to be enriched in the brain and nervous system and co-localize with syntaxin-1 in the neuronal processes and axonal outgrowths 22 . However, it has not yet been reported to be related to any diseases. Interestingly, loss of one of its major interactors, VPS18, led to widespread neurogeneration through obstructing endosomal maturation, lysosomal trafficking and autophagosome clearance 27 . Therefore, we highly speculated a causative role of VPS16 gene mutation in the familial adolescent-onset primary dystonia. The replacement of an Asn residue in the β -propeller domain of VPS16 with a Lys residue is predicted to have damaging impact on protein function through MutationTaster prediction 24 , supporting the pathological role of amino acid substitution (Fig. 4b, supplementary Table S3).
It would be interesting to mutate Vps16 to assess its biological function. With the efficiency and ease of the CRISPR-Cas9 genome editing approach 28,29 , we have demonstrated how it can be applied to investing disease-causing gene in dystonia (Fig. 5a,b). Targeted c.156 C > A mutation of Vps16 led to downregulation of VPS16 expression in the mice (Fig. 5c). Moreover, Vps16 mutant mice exhibited obvious abnormality in the behavior and significantly impaired motor function (Fig. 5d, supplementary video 1 and 2), indicating that VPS16 is the genetic cause of familial adolescent-onset primary dystonia.
To our knowledge, this is the first demonstration that the mutation of VPS16 was identified in cases of adolescent-onset primary dystonia. The functional characterization of targeted Vps16 c.156 C > A homozygous mutation in mice also supports its disease-causing role in dystonia, thus providing new insights into the pathogenesis of this common movement disorder.

Methods
Clinical evaluations. The family with adolescent-onset primary dystonia was ascertained from Hunan province. Human studies were conducted in accordance with the Declaration of Helsinki, with formal approval from the Ethics Reviewing Committee of Shenzhen University 1 st Affiliated Hospital and BGI. All related experiments were carried out only after written informed consent was obtained from each individual and/or parents of the children. The medical history was obtained by use of a questionnaire regarding the following aspects: age at onset, set of onset and evolution. The Burke-Fahn-Marsden dystonia rating scale (BFMDRS) was applied for motor assessment. The motor score consists of a scale assess movement impairment of the speech, eyes, trunk and extremities 30 . The movement assessment was performed by a specialized neurologist. MRI images were acquired from patient IV6 on a 1.5 Tesla Siemens scanner to display the brain image (SIEMENS AG, Munich, Germany).
Exome capture and sequencing. Genomic DNA was extracted from peripheral blood samples using DNeasy Blood and Tissue kit (Qiagen, Hilden, Germany). Exome capture and sequencing were performed in BGI according to the manufacturer's protocol. Briefly, genomic DNA was randomly sonicated to generate short fragments between 250 bp and 300 bp. Then adaptors were subsequently added to ligate each ends of the fragments. After amplified by ligation-mediated PCR, the pooled DNA fragments were hybridized to NimbleGen SeqCap EZ Exome (64 M) capture array (Roche NimbleGen Inc, Wisconsin, USA) for target enrichment, followed by washing and amplification. The captured DNA libraries were then sequenced on a Hiseq 2000 platform (Illumina, San Diego, USA).

Read mapping and variant analysis.
The sequencing reads were analyzed as previously described 31 . The raw image files were initially processed using the Illumina pipeline software (version 1.7) for base calling, and generated 90-bp paired-end reads. Clean reads were then mapped to the reference human genome (GRCh37/ hg19) from UCSC genome browser (http://genome.ucsc.edu/) with SOAPalinger software (http://soap.genomics. org.cn/index.html). SNPs were then detected by using SOAPsnp software (http://soap.genomics.org.cn/index. html). Indels were aligned to the UCSC reference human genome using BWA software (http://bio-bwa.sourceforge.net/) and further processed with the Genome Analysis Toolkit (GATK v1.6) for recalling. SNP calls and Indels were filtered to coordinate with the following criteria: 1) consensus quality score ≥ 20; 2) sequencing depth ≥ 4 and ≤ 500; 3) copy number ≤ 2; 4) distance between two adjacent SNPs ≥ 5 bp.Variants in non-coding region and synonymous mutations were removed from our analysis. Non-synonymous mutations, frame-shifting Indels, splice acceptor and donor site mutations that could potentially affect normal protein functions were retained. Common SNPs and Indels presented in dbSNP (http://www.ncbi.nlm.nih.gov/projects/SNP/, build 137), 1000 Genomes (ftp://www.1000genome.org), HapMap and YH database with high frequency (> 0.5%) were also filtered out. Novel variants that were presented as homozygous call in at least one case were considered as prioritized candidate variants. Given that all the affected individuals were related, only the recessive alleles that presented in the homozygous region were retained as disease-causing candidate variant for Sanger validation. Candidate variants were analyzed using SIFT 32 , Polyphen-2 33 and Mutation Taster 24 software to determine possible changes in protein structure that could affect phenotype.
Exome Homozygosity mapping. The data from WES were used for examining large stretches of homozygous region as previously reported 34,35 . All the autosomal dbSNP sites and novel SNPs that had ≥ 20-fold coverage of the exome target regions were selected as markers to create a genetic map. Variants displaying an identical SNP allele with ≥ 95% of all reads and covering at least 5-fold were considered to be homozygous markers. SNPs with 30-70% variation reads and covered at least 10-fold were considered to be heterozygous markers. And SNPs with < 30% or with 70-95% variation reads were considered ambiguous. Statistics on the distribution of map markers along the genome were analyzed with Perl scripts. We adopted a window of 500 markers, containing maximum Scientific RepoRts | 6:25834 | DOI: 10.1038/srep25834 two heterozygous markers and allowing the gap between two adjacent markers was maximum 500 Kb. A genomic region was identified as a homozygous stretch as coalescence of all qualified windows with minimum of 1 Mb in length 36 . The two complementary oligos were denatured at 95 °C for 5 minutes, ramp cooled to 25 °C over a period of 45 min to allow annealing, and finally ligated with the linearized pX458. A single stranded oligonucleotide containing the desired mutation in the targeted region of Vps16 was designed and ordered. The oligonucleotide sequence was as follows:
Microinjection of zygotes. C57BL/6J mice were purchased from the Animal Center of Guangdong Medical Laboratory (Guangzhou, China). Eight to ten week-old female mice were superovulated by injection with 10 IU of equine chorionic gonadotrpin and HCG (Ningbo second hormone factory, Ningbo, China) and mated with 8-10 week-old males. One-cell stage embryos were collected for microinjection of Cas9 mRNAs (100 ng/μ l), sgRNA (50 ng/μ l) and oligonucleotide (100 ng/μ L), using a standard microinjection system (Eppendorf, Hamburg, Germany). The embryos were then cultured in M2 medium (Omega, USA), and immediately after turning into the blastocyst stage, the embryos were transferred into pesudopregnant female mice.
Sanger sequencing of Vps16 mutant mice. Genomic DNA was prepared from the tails of three-week-old mice and CRISPR/Cas9-induced Vps16 c.156 C > A mutations were identified through Sanger sequencing. The region including CRISPR/Cas9 target site was PCR-amplified by using primer as follow: Western Immunoblot analysis. Retro-orbital blood samples were collected from the right retro-orbital plexus of anesthetized mice as described 38 . Anesthesia was induced by placing each mouse in an inhalation chamber with 4% isoflurane (RWD Life Science, San Diego, USA) regulated with a calibrated vaporizer. For each mouse approximately 0.5 ml of heparin-anticoagulated blood was collected. Whole blood samples were incubated for 10 min with erythrocyte lysis buffer and centrifuged at 1500 rpm for 10 min. For western blotting, harvested cells were lysed in RIPA buffer. After determining the protein content of the cell lysates, the protein extracts were separated by 10% SDS-PAGE, transferred to a PVDF membrane and incubated with primary antibody (VPS16, β -actin antibodies were purchased from Santa Cruz Biotechnology, Inc., Texas, USA). The signal was detected by ECL detection system (GE Healthcare, USA).

Rotarod test.
The rotarod equipment (YLS-4C, Jinan Yiyan Scientific Research Company, Shandong, China) was used to examine the motor balance, coordination and strength of mice. Six-month old mice were trained three times for three consecutive days at 30 rpm per day. On the fourth day, WT mice and Vps16-mutant mice underwent testing. The latency to fall from the Rotarod apparatus was recorded automatically and analyzed later. The experiment was repeated at least three times.
Data analysis. All data are the mean ± SE for at least 3 experiments. Data were analyzed statistically using the unpaired Student's t test. The criterion for a significant difference was ***p < 0.0001.