Whole-Exome Sequencing of Congenital Glaucoma Patients Reveals Hypermorphic Variants in GPATCH3, a New Gene Involved in Ocular and Craniofacial Development

Congenital glaucoma (CG) is a heterogeneous, inherited and severe optical neuropathy that originates from maldevelopment of the anterior segment of the eye. To identify new disease genes, we performed whole-exome sequencing of 26 unrelated CG patients. In one patient we identified two rare, recessive and hypermorphic coding variants in GPATCH3, a gene of unidentified function, and 5% of a second group of 170 unrelated CG patients carried rare variants in this gene. The recombinant GPATCH3 protein activated in vitro the proximal promoter of CXCR4, a gene involved in embryo neural crest cell migration. The GPATCH3 protein was detected in human tissues relevant to glaucoma (e.g., ciliary body). This gene was expressed in the dermis, skeletal muscles, periocular mesenchymal-like cells and corneal endothelium of early zebrafish embryos. Morpholino-mediated knockdown and transient overexpression of gpatch3 led to varying degrees of goniodysgenesis and ocular and craniofacial abnormalities, recapitulating some of the features of zebrafish embryos deficient in the glaucoma-related genes pitx2 and foxc1. In conclusion, our data suggest the existence of high genetic heterogeneity in CG and provide evidence for the role of GPATCH3 in this disease. We also show that GPATCH3 is a new gene involved in ocular and craniofacial development.


Supplementary Methods
Exome capture and sequencing. For whole exome capture the SeqCap EZ Exome Enrichment Kit v3.0 (Roche NimbleGen), based on the human genome build GRCh37/hg19 with a 64.19 Mb sequence capture design, was used following a modified manufacturer's protocol with pre-capture multiplexing adapted to the Sciclone Liquid Handling System. In short, after Covaris E210 fragmentation of genomic DNA (1 μg) the Illumina TruSeq DNA libraries were prepared on the robotic system using Gold Standard DNASeq protocol. Adapter ligated DNA fragments were enriched by eight cycles of pre-capture PCR. Samples were manually pre-capture pooled with a combined mass of 750 ng for the baits hybridization step (47°C for 71 hours). After washes, the captured Illumina paired-end sequencing. In brief, genomic DNA (3.0 μg) was sheared on a Covaris E210 instrument followed by Agilent adaptor ligation. The adaptor-modified DNA fragments (150-300 bp) were PCR amplified (six cycles using Herculase II fusion DNA polymerase) with SureSelect Primer and SureSelect Pre-Capture Reverse PCR primers (Agilent technologies) and hybridized to SureSelect baits (Agilent Technologies) for 24 hours at 65°C, using an Applied Biosystems 2720 Thermal Cycler. After the washes the eluate was PCR amplified (16 cycles) in order to add the index tags using Sure SelectXT Indexes for Illumina (Agilent Technologies). The final library size and concentration and sequencing were carried out as described previously with a paired end run of 2x101 bp to reach mean coverage of 7596X of the approximately 51.2 Mb target region.
Image analysis, base calling and quality scoring of all the sequencing runs were processed using the manufacturer's software Real Time Analysis (RTA 1.13.48) and followed by generation of FASTQ sequence files by CASAVA.
Exome data analysis. Sequencing reads were trimmed from the 3' end up to the first base with a Phred quality >9 and were mapped to the Human reference genome (GRCh37) using GEM 1 to generate BAM files containing only properly paired and uniquely mapped reads (see coverage on the target regions per sample in Supplementary   Table 1; available at www.aaojournal.org). The files were further processed with Picard tools (http://picard.sourceforge.net/) to remove duplicates, and local realignment was performed with the Genome Analysis Tool Kit (GATK). SAMtools 2 was used on the processed BAM files to call single nucleotide variants (SNVs) and small insertion deletions (InDels). Functional annotations from Ensembl (http://www.ensembl.org/index.html) were added to the resulting VCF using snpEff 3 . snpSift 4 was used to add information from dbSNP 5 , the 1000 Genomes Project 6 , the NHLBI Exome Sequencing Project and a variety of conservation and deleteriousness predictions included in dbNSFP 7 .
Site-directed mutagenesis and cloning of GPATCH3 mutations. The wild-type GPATCH3 coding sequence obtained from a commercial cDNA (Origene, NM_022078, SC122923) was subcloned into the XbaI and BamHI restriction sites of the modified mammalian cell expression vector pcDNA3.1(-) (Cterminal Myc-Tagged) 8 by directional PCR using the following primers: 5'AATTTCTAGAATGGCGGTGCCCGGCGAG3' (P1) and 5'AAGGATCCAGGTCAGGCAATGAGGGGCT3' (P2) (XbaI and BamHI sequences are indicated in bold, respectively). This cDNA construct was used as a template to obtain the GPATCH3 variants (p.Asn234Ser and p.Gly475Glu) by site-directed mutagenesis using the QuickChange Lightning SiteDirected Mutagenesis Kit (Agilent Technologies), with the primers and PCR conditions indicated in the Supplementary Table 3, according to the manufacturer's instructions. All site-directed mutagenesis products were sequenced to confirm the presence of the mutation and were subcloned into either the XbaI and BamHI restriction sites of pcDNA3.1(-).

Transactivation activity assays.
A 600-bp fragment of the human CXCR4 gene promoter, which contains one Fox Binding Element, was cloned into the NheI/NcoI restriction sites of the pGL3-basic vector (Promega) by directional PCR using normal human genomic DNA as previously described. 9 HEK293T cells in 24-well tissue culture plates (2x10 5 cells/well) were transfected with 500 ng of either the different GPATCH3 variants expression vectors, the FOXC2 expression vector [pcDNA3.1(-) myc-his] as a positive control or the MYOC expression vector [pcDNA3.1(-) myc-his] as a negative control, along with 25 ng of the recombinant pGL3-basic-CXCR4 luciferase reporter and 50 ng of the CYP1B1 expression vector [pcDNA3.1(-) myc-his] as a transfection control. The transactivation assays were performed 24 hours after the transfection. The cells were harvested and assayed for firefly luciferase activity using the Luciferase Assay System according to the manufacturer's instructions. GPATCH3, FOXC2 (positive control), myocilin-HA (MYOC, negative control) and CYP1B1-myc (transfection control) recombinant proteins and endogenous lactate dehydrogenase (LDH, loading control) were detected by Western blot.
Immunocytochemistry. The HEK-293T cells were seeded on coverslips placed into 24-well plates were transiently transfected with DNA constructs encoding FOXC2 and different GPATCH3 variants. All of the transfections were performed using the SuperFect Transfection Reagent. After the transfection, the cells were washed once with DPBS (Dulbecco's Phosphate-Buffered Saline) and cultured for 24 hours. The cells were then fixed with 4% paraformaldehyde for 10 min at room temperature, followed by incubation with phosphate buffered solution containing 0.2% Triton X-100, 10% fetal bovine serum (FBS), and 5% bovine serum albumin for 30 min at room temperature. The recombinant proteins were detected using an anti-myc antibody (Santa Cruz Biotechnology) (at a 1:500 dilution) at 4°C overnight followed by a Cy2-conjugated anti-mouse IgG (Jackson ImmunoResearch Labs,1:1000 dilution) for 2 hours at room temperature. Finally, the coverslips were mounted on glass slides using polyvinyl alcohol mounting medium with DABCO containing DAPI (40,60-diamidino-2phenylindole), and the cells were viewed under an LSM 710 Zeiss confocal microscope Bioinformatic analyses. The deleterious effect of mutations was predicted on line with the Sorting Intolerant From Tolerant (SIFT), 10 and PolyPhen programs. The UniProt database was searched to identify protein domains.
The Consite software was used to identify putative transcription factor target sequences 11 . The MicroSNiPer software was used to identify putative miRNA target sequences 12 . The effect of variants on splicing was predicted using the Human Splicing Finder software 13 . Amino acid sequence alignments were carried out with Clustalw 14 .
Preparation of tissue extracts for Western blot. Human ocular tissues were obtained from enucleated eyes of dead donors through the National Disease Research Interchange (Philadelphia, PA, USA). Dissection was carried out, and tissues were collected in Eppendorf tubes. Proteins were extracted by mechanical pulverization of tissues in liquid N2 as previously described 8 . Non ocular tissue extracts were obtained from BioChain Institute, Inc. and Santa Cruz Biotechnology, Inc. Zebrafish embryos (96 hpf) were dechorionated by pronase treatment (1 mg/ml) for 5 min at room temperature 15 , and yolk sacks were manually removed under a light stereomicroscope SMZ18 (Nikon) using microdissection needles. Pools of 50 zebrafish embryos were homogenized in extraction buffer (Tris HCl pH 7.4 50 mM, IGEPAL 1%, NaCl 150 mM, EDTA 1 mM, PMFS 1 mM, Leupeptin 1 µg/µl, and NaF 1 mM). The extract was centrifuged, and the supernatant was frozen at -80°C. The protein content of each sample was quantitated using the Bicinchoninic Acid Protein Assay Kit (Thermo Scientific), following the manufacturer's recommendations.
Alcian blue cartilage staining. 96 hours post fecundation (hpf) MO injected and wild type larvae were fixed in 4% paraformaldehyde overnight. Larvae were washed with PBS-Tween (PBST) to remove paraformaldehyde, bleached in 3% hydrogen peroxide and 1% KOH solution to remove pigment, and then were washed again three times in PBST. Larvae were incubated overnight in an Alcian blue solution (0.1% Alcian blue, 5% HCl and 70% etanol), rehydrated in a staged ethanol series and treated with 50 mg/ml trypsin for 2 hours at 37ºC to dissolve brain tissue and allow visualization of pharyngeal arch cartilages. Bright-field images were taken using a Nikon SMZ18 microscope equipped with a Nikon DSRi2 digital camera.

Confocal imaging of Tg(sox10:eGFP) reporter fishes. Gpatch3 knockdown of embryos obtained from
Tg(sox10:eGFP) reporter fishes (kindly provided by Dra. Cavodeassi) was performed using the gpatch3 ATG morpholino oligonucleotide (MO) purchased from Gene Tools, LLC. Embryos were injected with 2.0 ng of MO at the 1-2 cell stage. Embryos (96 hpf) were anesthetized with 0,02% Tricaine (MS222) (Sigma), mounted in 2% methylcellulose (Sigma) and visualized using an LSM710 Zeiss confocal microscope. Fluorescence emitted by the GFP reporter protein was registered at 494-552 nm. Z-Stack maximum intensity projections of embryos were obtained with ZEN software (Zeiss). 83.64 PCG-208 95.38 a c10: percentage of the mappable exome covered by at least 10 uniquely mapped and properly paired (according to orientation and distance) read pairs. Note that six samples  were captured with Agilent SureSelect XT Human All Exon v4 (54Mb) and the rest with Nimblegen SeqCap EZ Exome Enrichment Kit v3.0 (64Mb).