Associations Among Genotype, Clinical Phenotype, and Intracellular Localization of Trafficking Proteins in ARC Syndrome

Arthrogryposis–renal dysfunction–cholestasis (ARC) syndrome is a rare autosomal recessive multisystem disorder caused by mutations in vacuolar protein sorting 33 homologue B (VPS33B) and VPS33B interacting protein, apical–basolateral polarity regulator (VIPAR). Cardinal features of ARC include congenital joint contractures, renal tubular dysfunction, cholestasis, severe failure to thrive, ichthyosis, and a defect in platelet alpha-granule biogenesis. Most patients with ARC do not survive past the first year of life. We report two patients presenting with a mild ARC phenotype, now 5.5 and 3.5 years old. Both patients were compound heterozygotes with the novel VPS33B donor splice-site mutation c.1225+5G>C in common. Immunoblotting and complementary DNA analysis suggest expression of a shorter VPS33B transcript, and cell-based assays show that c.1225+5G>C VPS33B mutant retains some ability to interact with VIPAR (and thus partial wild-type function). This study provides the first evidence of genotype–phenotype correlation in ARC and suggests that VPS33B c.1225+5G>C mutation predicts a mild ARC phenotype. We have established an interactive online database for ARC (https://grenada.lumc.nl/LOVD2/ARC) comprising all known variants in VPS33B and VIPAR. Also included in the database are 15 novel pathogenic variants in VPS33B and five in VIPAR. Hum Mutat 33:1656–1664, 2012. © 2012 Wiley Periodicals, Inc.

and 'ARC syndrome' into PubMed (NCBI) and screened abstracts and full-text articles for variants. The database also contains 188 unique variants taken from dbSNP (http://www.ncbi.nlm.nih.gov/projects/SNP) (Sherry et al., 2001) and the most up to date information from the 1000 Genomes Project (http://www.1000genomes.org/), including information on frequency. Patients with classical ARC phenotype were referred for clinical advice and also for molecular diagnosis to the University of Birmingham and West Midlands Regional Genetics laboratories where all coding exons and intron-exon boundaries were screened for mutations in VPS33B and VIPAR. One patient was screened for mutations in VPS33B by Prevention Genetics (http://www.preventiongenetics.com). Here we report 19 previously unpublished pathogenic variants.
Variant data for VPS33B and VIPAR were initially stored in separate offline Microsoft Excel worksheets (Microsoft, Redmont, WA) and were subsequently imported in tab-delimited format into the ARC database within the LOVD platform (Fokkema et al., 2011). Variants were named according to HGVS nomenclature guidelines (http://www.HGVS.org) and numbered using the VPS33B reference sequence (NG_012162.1, NM_018668.3) and the VIPAR reference sequence (NG_023421.1, NM_022067.3). Previously published mutations were renamed accordingly and all annotations were checked using Web-based Mutalyzer software linked to LOVD (Wildeman et al., 2008).
VPS33B and VIPAR each have a gene homepage. This provides general information about the gene, as well as a guide to the variant tables in the database and an explanation of the searches that can be implemented. Additional information about the gene is also accessible through links to other resources provided on the homepage. In the "variants" section are variant-specific fields for each gene (exon, pathogenicity, template/technique used for detection, DNA/RNA/protein change, frequency, number of times reported, database ID, phenotype, and reference). Records describing variants per individual patient can be accessed, along with information on patient phenotype and ethnic/geographic origin. The database also contains a "documentation" section, which includes a guide for submitters. After initial registration, new variants can be submitted directly to the database, where they will be made public following curation.

Characterization of new variants in VPS33B and VIPAR
Of the disease-causing variant types, most common are substitutions (n=11: 9 splice-site and 2 nonsense). The effect of the splicing mutations was predicted by bioinformatic analysis using the BDGP Splice Site Prediction software NNSPLICE (http://www.fruitfly.org/seq_tools/splice.html). We identified 3 deletions which both are predicted to result in frameshift and premature termination of transcription, as well as one whole-gene deletion. A variant of particular interest is the splicing mutation c.1225+5G>C, as it is associated with an attenuated ARC phenotype.
We also report 5 novel mutations in VIPAR (Table 1), including substitutions (n=3: 2 missense and 1 nonsense), a deletion predicted to result in frameshift and premature termination of transcription, and a splicing mutation. All are classed as 'pathogenic' or 'probably pathogenic'.
qRT-PCR to determine exon copy number qRT-PCR was performed using 25ng genomic DNA to identify deletions of VPS33B exon 4 using the method described (Hoebeeck et al. 2004) Normalization was carried out using 2 reference genes, GAPDH and actin with normal copy number. qRT-PCR was performed with Power SYBR Green MasterMix (Applied Biosystems, Warrington, UK) and an ABI 7500 thermal cycler (Applied Biosystems, Warrington, UK).

Co-Immunoprecipitation
For co-immunoprecipitation, 20 μg of anti-HA monoclonal antibodies were covalently conjugated to 100 μl of Dynabeads Protein G (Invitrogen, Paisley, UK) using dimethyl pimelimidate and triethanolamine according to the manufacturer's instructions. HEK293 cells growing on 6-well plates and transfected with a total of 4 μg of plasmid DNA constructs. These were allowed to recover for 48h before the protein was extracted as above. Extracted proteins (250μg) were mixed with 20 μL of antibody-conjugated Dynabeads and incubated on a blood rotor with end-over-end mixing (at 4°C for 3 h). The complexes were then washed 3 times using cell lysis buffer supplemented with 150 mM NaCl, after which proteins were eluted by boiling in 2x SDS loading buffer (10 min °C) and supernatants were removed for immunoblotting.

Immunoblotting
The absence or presence of VPS33B was determined by separating 25μg of extracted protein on a 12% SDS-PAGE gel. Proteins were transferred to transblot polyvinylidene difluoride membranes (Hybond-P; Amersham Biosciences, Little Chalfont, UK).
Immunoblotting was carried out according to standard protocols (Harlow and Lane, 1998) using a 1/500 dilution of the primary antibody rabbit anti-human VPS33B (Proteintech, Manchester, UK) and a 1/1000 dilution of goat anti-rabbit HRP conjugate (Dako, Cambridge, UK).

RNA Extraction and cDNA Synthesis
Fibroblast cells were grown to confluence in 75 cm 2 flasks and HEK293 cells were grown in 6well plates before RNA was extracted using RNAzol B reagent (Campro Scientific, UK) according to manufacturer's instructions. For cDNA synthesis, 1μg of total RNA was reverse transcribed using ImProm-II Reverse Transcription Systems and oligo dT primers (Promega, Southampton, UK) according to manufacturer's protocol. 3'-RACE was carried out using the SMART RACE cDNA Amplification kit (Clontech, UK) according to manufacturer's conditions.

Supp. References
Hoebeeck (C) PCR products using primers flanking the deletion indicate that maternal DNA contains the deletion but DNA from father and control do not. qRT-PCR results using genomic DNA (D) confirm this finding. Exon 4 is normalized against GAPDH and actin. Samples containing 2 copies of the exon will give the result 1, and those containing one copy due to a heterozygous deletion will give the result 0.5. Samples from the mother and Patient AB contain only one copy of exon 4.
Supp. Figure S2. Supp. Figure S3. Co-immunoprecipitation of VIPAR with HOPS complex proteins. (A) HEK293 cells were co-transfected with HA-empty or HA-VIPAR and myc tagged VPS11, VPS18, VPS39 or VPS41. Co-immunoprecipitation experiments revealed interaction between over-expressed VIPAR and HOPS complex proteins. More VPS18 than VPS11, VPS39 and VPS41 was pulled down. Although these proteins were present in the lysates, none were pulled down together with HA-empty. (B) HEK293 cells were co-transfected with HA-empty or HA-VIPAR, YFP-empty or YFP-VPS33B and myc-VPS18. A co-immunoprecipitation experiment showed that VIPAR and VPS18 interacted when the over-expressed VPS33B was absent. In presence of over-expressed VPS33B substantially less VPS18 was pulled down.