Bardet Biedl syndrome (BBS) is a multisystem genetically heterogeneous ciliopathy that most commonly leads to obesity, photoreceptor degeneration, digit anomalies, genito-urinary abnormalities, as well as cognitive impairment with autism, among other features. Sequencing of a DNA sample from a 17-year-old female affected with BBS did not identify any mutation in the known BBS genes. Whole-genome sequencing identified a novel loss-of-function disease-causing homozygous mutation (K102*) in C8ORF37 , a gene coding for a cilia protein. The proband was overweight (body mass index 29.1) with a slowly progressive rod-cone dystrophy, a mild learning difficulty, high myopia, three limb post-axial polydactyly, horseshoe kidney, abnormally positioned uterus and elevated liver enzymes. Mutations in C8ORF37 were previously associated with severe autosomal recessive retinal dystrophies (retinitis pigmentosa RP64 and cone-rod dystrophy CORD16) but not BBS. To elucidate the functional role of C8ORF37 in a vertebrate system, we performed gene knockdown in Danio rerio and assessed the cardinal features of BBS and visual function. Knockdown of c8orf37 resulted in impaired visual behavior and BBS-related phenotypes, specifically, defects in the formation of Kupffer’s vesicle and delays in retrograde transport. Specificity of these phenotypes to BBS knockdown was shown with rescue experiments. Over-expression of human missense mutations in zebrafish also resulted in impaired visual behavior and BBS-related phenotypes. This is the first functional validation and association of C8ORF37 mutations with the BBS phenotype, which identifies BBS21. The zebrafish studies hereby show that C8ORF37 variants underlie clinically diagnosed BBS-related phenotypes as well as isolated retinal degeneration.
Bardet Biedl syndrome (BBS, MIM209900) is a genetically heterogeneous pleiotropic disorder of autosomal recessive inheritance and is primarily characterized by photoreceptor degeneration, obesity, digit anomalies, genito-urinary anomalies and cognitive impairment ( 1 , 2 ). BBS is considered a ciliopathy, as genes associated with the disease are associated with primary cilia function ( 3–22 ). BBS has been an important disease model to highlight the clinical and molecular overlaps that now characterize ciliopathies ( 23 , 24 ). Mutations in BBS genes may also cause other syndromic conditions that mimic BBS or other ciliopathies and result in non-syndromic (isolated) retinal dystrophies such as cone-rod dystrophy (CRD, MIM 120970) and retinitis pigmentosa (RP, MIM 268000) ( 25 , 26 ). The genetic heterogeneity of non-syndromic inherited retinal dystrophies involves over 250 causally associated genetic loci (RETNET, http://www.RETNET.org , last accessed April 19, 2016), a third of which are genes expressed in primary cilia ( 27 ).
RP and CRD are inherited photoreceptor degeneration diseases that differ by the predominant photoreceptor cell type involved. The prevalence of RP is 1 in 4000 individuals and 1 in 40 000 individuals for CRD. In both conditions, the vision loss is usually progressive and severe for which there is no treatment at this time. Mutations in C8ORF37 were recently identified in individuals with autosomal recessive RP type 64 (RP64) and CRD type 16 (CORD16) ( 27–32 ). Nine different C8ORF37 mutations have been reported ( Fig. 1 A), none of which were associated with BBS ( 27 , 32 ).
The C8ORF37 (NP_808880.1) protein is encoded by six exons ( Fig. 1 A) and is 207 amino acids in length (NM_177965.3). The overall function and functional domains of C8ORF37 are unknown. C8ORF37 is expressed ubiquitously in human tissues, with robust expression in the brain, heart and retina ( 27 ). C8ORF37 is localized to the base of the photoreceptor connecting cilium ( 27 ). As such, it is likely that mutations in this gene disturb intracellular trafficking and can contribute to the pathophysiology of RP, CRD as well as BBS.
In this study, we report that a novel homozygous C8ORF37 mutation causes BBS. Using gene knockdown in zebrafish, we observed visual impairment and BBS-specific phenotypes, specifically Kupffer’s vesicle (KV) formation defects and melanosome transport delay. We also tested the pathogenicity of two previously reported missense mutations associated with retinal dystrophies ( 32 ) by using established BBS-related phenotypes in zebrafish and visual function assessment ( 33 ). In this study, we took advantage of rapid disease modeling in zebrafish to elucidate the role of the Danio rerio c8orf37 in BBS-related functions and vision thus providing phenotype–genotype correlation for BBS21.
A 12-year-old Caucasian female, born to consanguineous parents ( Fig. 1 B), was initially referred for progressive myopia (2012) with no significant family or personal history of vision loss or health issues and underwent a thorough assessment.
Ocular phenotype: high myopia with slowly progressive rod-cone degeneration
Correction of the myopia of −9.0 + 2.50 axis 110° (oculus dexter [OD]) and −10.0 + 2.25 axis 80° (oculus sinister [OS]) led to a best-corrected visual acuity of 20/40 in either eye which remained unchanged over 3 years (ages 12–15 years). Retinal atrophy was seen around the optic nerve (peripapillary) and inferiorly, which findings were in keeping with myopic degenerative changes ( Fig. 2 A). However, mild vessel attenuation and a flat looking retina were observed and suggestive of RP ( Fig. 2 A). Fundus autofluorescence showed a paramacular ring of hyper-AF; increased AF was also noted at the fovea, whereas inferior retina revealed a speckled pattern of AF, suggesting diffuse retinal pigment epithelium dysfunction ( Fig. 2 B). Spectral-domain optical coherence tomography of the central retina showed areas of disruption of the photoreceptor outer segment (OS) - inner segment (IS) junction the external limiting membrane, a network-like structure that is situated at the bases of the rods and cones, appeared to be reasonably preserved at the macula ( Fig. 2 C). The central retinal thickness, on the lower end of normal, did not change between 2011 (OD: 202 µm; OS: 196 µm) and 2014 (OD: 199 µm; OS: 200 µm). A grade 1 foveal hypoplasia ( 34 ) was coincidentally observed with no retinal re-organization and a well-preserved retinal lamination ( Fig. 2 C).
The full-field electroretinogram (ERG) recordings were severely decreased suggesting a progressive rod-cone dystrophy and recording was still measurable at 16 years of age. More specifically, the dim light scotopic ERG (rods) was non-detectable. The bright flash ERG (7.6 cds.m-2) showed reduced but detectable a-wave (40 microvolts and 33 microvolts in the RE and LE, respectively); the b-waves were smaller than the a-wave. The cone ERGs were also severely attenuated in both eyes. Although her central visual acuity remained stable over 3 years, the fields of vision constricted by 50%, to less than 20 degrees ( Fig. 2 D and E), meeting the criteria of legal blindness. Color vision was not measurable using the HRR (Hardy Rand Ritler test) plates by the age of 14 years and her contrast vision was reduced.
Systemic findings meet the diagnostic criteria of BBS
The patient was overweight (with a body mass index of 29.1 at 15 years) and had three limbs post-axial polydactyly, a horseshoe kidney (with normal function), mild learning difficulty and elevated liver enzymes (alanine transaminase: 81 units/l). Liver ultrasound was normal but she had an abnormally positioned uterus. Hearing, behavior and brain magnetic resonance imaging were normal.
Patient genotyping identified a novel homozygous K102* mutation in C8ORF37
Mutational analysis of known BBS disease-causing genes using a Clinical Laboratory Improvement Amendments-approved laboratory first showed a heterozygous variant in BBS4 (K46R) that was predicted to be benign, and no deletion was identified using microarray analysis. Whole-genome sequencing covered 97% of the whole-genome sequence and 98.8% of the known exons. The genome-wide SNP-heterozygous/homozygous ratio (Ti/Tv) was less than would be expected for a proband of a non-consanguineous family (∼2), at 1.527 ( 35 ). From the whole-genome filtering ( Supplementary Material, Fig. S1 ), 24 variants were selected according to our criteria of conservation and pathogenicity prediction ( Supplementary Material, Table S1 ). The only variant involving a gene associated with retinal degeneration ( 27 , 31 ) was a novel homozygous nonsense variant (NM_177965.3: c.304A > T/p.K102*) in C8ORF37 , which was highly conserved ( Fig. 3 C) and segregated with disease status in the family ( Fig. 1 B). The mutation was within a large homozygous region (32 Mb) on chromosome 8 ( Supplementary Material, Fig. S2 ) and was not reported in any control databases (1000 genomes, NHLBI, cg and dbSNP). Only one heterozygous K102* allele was observed in ExAc from 121 116 individuals (frequency 8.257 × 10 −6 ). K102* was predicted to be disease causing by Mutation Taster with probability of 0.999 and deleterious according to Combined Annotation Dependent Depletion (CADD)_Phred. In addition, c.304A > T met the criteria of pathogenicity as defined by the American College of Medical Genetics and Genomics standards and guidelines ( 36 ).
K102* leads to nonsense-mediated decay
Nonsense-mediated decay (NMD), a process by which cells recognize and degrade mRNA containing premature stop mutations to prevent toxic effects of truncated peptides, was activated by the nonsense mutation c.304A > T ( Fig. 1 C). The C8ORF37 reverse transcriptase-polymerase chain reaction (RT-PCR) product from the cDNA matched with the control transcript with higher expression for the gene, NM_177965.3. The RT-PCR showed a difference in transcript level between the mutant samples treated with cycloheximide, a translation inhibitor, versus the samples treated with dimethyl sulfoxide (DMSO). In comparison, there were no differences in transcript levels with either treatment in the wild-type (WT) samples.
Mutations in C8ORF37 are a rare cause of BBS
Direct sequencing of the C8ORF37 coding sequence was performed in 53 unrelated clinically diagnosed BBS patients without mutations in known BBS genes based on molecular testing. No additional disease-causing mutations were identified in C8ORF37 .
c8orf37 is expressed during zebrafish embryonic development and in adult tissues
The D. rerio chromosome 16 open reading frame of human C8ORF37 (c16h8orf37) protein (XP_003200284.2) is encoded by six exons and is 202 amino acids in length (XM_003200236.3) ( Fig. 3 A and B). The predicted zebrafish ortholog shows 44% amino acid identity (61% homology) to human C8ORF37 (Clustal2.1) ( Fig. 3 B). Four of the C8ORF37 mutations sites in exon 6 (L166, R177, Q182 and W185) reported to cause retinal degeneration ( 27 ) and the site of the novel nonsense mutation (K102*) are conserved in zebrafish ( Fig. 3 B and C) as well as in rat, mouse, chicken and frog ( Fig. 3 C). The implication of mutating R177 and Q182 were hereby specifically explored. c8orf37 is ubiquitously expressed during early zebrafish development and in adult tissues including the eye, brain, heart and kidney ( Fig. 3 D and E). The high conservation in the pathogenic amino acid sequences of zebrafish C8orf37 compared with that of human C8ORF37 and the comparable expression of c8orf37 and C8ORF37 through embryogenesis to adulthood supports the role of D. rerio as a good model system to study the potential pathogenicity of human mutations in this gene.
Knockdown of c8orf37 in zebrafish results in BBS-related phenotypes
To establish the functional role of c8orf37 in syndromic as well as non-syndromic retinal degeneration, we performed gene knockdown experiments in the zebrafish model. This approach was previously successful in evaluating hypomorphic mutations associated with RP in genes that also caused different syndromic disorders ( 37 ). For gene knockdown, antisense morpholino oligonucleotides (MO) targeting the c8orf37 translation start target (ATG MO) and the splice junction of intron1 and exon2 (Splice MO) were injected into 1–2 cell stage embryos at different doses. RT-PCR of splice morphants verified that knockdown persisted out through 5 dpf (data not shown). The gross morphology of both the ATG and splice morphants appeared normal without visible defects such as bent body axis and renal cysts. To verify our hypothesis that C8ORF37 causes BBS, we evaluated the c8orf37 morphants for BBS-related phenotypes such as defects in KV formation and retrograde melanosome transport, in addition to visual behavior.
KV formation is impaired in c8orf37 morphants
Kuppfer’s vesicle (KV) is a ciliated organ that develops from the dorsal forerunner cells and is found transiently in the posterior tailbud of developing zebrafish embryos from 10 to 30 h post-fertilization (hpf). The KV is analogous to the mammalian node and plays a role in left–right axis patterning. During the somite stage, the segmentation period, the ciliated KV normally forms in the tailbud of the zebrafish and can be readily inspected under a microscope. KV morphology was scored between the 10 and 13 somite stages, and reduction of the KV size to less than notochord width or its absence was considered defective ( Fig. 4 A). Similar to what has been seen in BBS ( 24 , 33 ), knockdown of c8orf37 using either the ATG MO or splice MO resulted in statistically significant KV defects in a dose-dependent manner. At the highest dose (8 ng splice MO), approximately 53% (57/107) of embryos displayed KV defects, while the lower doses of 5 ng ATG MO or 4 ng Splice MO generated 20% and 29% defects, respectively ( Fig. 4 C). Analysis of ATG morphants demonstrates a general decrease in vesicle size and cilia content ( Fig. 4 B).
Delayed retrograde melanosome transport in c8orf37 morphants
Pigment distribution in zebrafish serves as a model for cellular transport and specific motors lead to disperse melanosome (anterograde transport) or the perinuclear aggregation of melanosomes (retrograde transport) in response to environmental or chemical cues. A cardinal feature of knocking down BBS function is a delay in the rate of retrograde intracellular transport within the melanophore ( 9 , 25–27 ). We determine the rate of melanosome transport to the perinuclear regions in c8orf37 morphants. Day 5 larvae reared in a dark background had melanophores with dispersed melanosomes ( Fig. 4 D; before). Trafficking of melanosomes from the periphery to the perinuclear region is stimulated by addition of epinephrine ( Fig. 4 D; after). The time to reach the perinuclear region was measured in individual larva and the average time of retrograde melanosome trafficking for each group was plotted ( Fig. 4 E). WT and control MO injected larvae responded within 1.60 ± 0.05 min and 1.76 ± 0.01 min, respectively. ATG morphants (2.5 and 5.0 ng) and splice morphants (8 ng) showed statistically significant delays in melanosome transport ( P < 0.05) ( Fig. 4 E). These data show that knockdown of c8orf37 function generates a KV defect and delays in melanosome transport, known cardinal features of zebrafish knockdown of BBS gene expression.
c8orf37 morphants have impaired visual behavior
Because mutations in C8ORF37 were also associated with severe non-syndromic retinal dystrophies in patients, we tested the visual function of morphants. Zebrafish larvae have an adaptive escape response of change in swimming direction when exposed to sudden changes in light ( Supplementary Material, Fig. S3A ). We have used this vision startle assay in analysis of several genes associated with human visual disorders ( 37 , 38–41 ). The numbers of larval responses to five visual startle stimuli were recorded. Knockdown of the Cone-Rod Homeobox ( crx ) was used as a control of visually non-responsive morphants. WT embryos and control morphants typically respond >4 times and crx morphants average ∼2 times ( Fig. 5 ). Both ATG and splice morphants showed a statistically significant decrease in visual startle responses with splice morphants compared to controls demonstrating greater impairment in a MO-dose-dependent manner: 2.5 ng (3.58 ± 0.4, t -test P < 0.05), 5 ng (3.46 ± 0.32, t -test P < 0.05) and 8 ng (1.98 ± 0.22, analysis of variance [ANOVA] with Tukey P < 0.01). Indeed, the 8-ng splice morphant response showed no significant difference from crx morphants ( Fig. 5 ).
Specificity of c8orf37 knockdown and gene product rescue
To validate the specificity of MO knockdown and to functionally validate the non-syndromic retinal dystrophy-associated alleles, we tested the extent to which full-length mRNAs of c8orf37 rescued c8orf37 morphant defects. We selected splice MO (8 ng) as it produced robust knockdown phenotypes. We performed sequential injections introducing c8orf37 RNA (∼200 pg) followed by 8 ng Splice MO to directly assess the impact of the RNA compared to MO only and RNA only siblings. Embryos injected with WT c8orf37 RNA appeared morphologically normal. c8orf37 RNA injected in the splice morphants rescued MO-induced KV defects ( Fig. 6 A), melanosome transport delay ( Fig. 6 B), and the visual defect ( Fig. 6 C).
Functional analysis of human missense mutations
We have established that knockdown of c8orf37 generated BBS-like phenotypes and visual impairment and demonstrated that wt- c8orf37 can rescue knockdown defects. We next evaluated the functional properties of human missense mutations and the relationship of the non-syndromic c8orf37 mutations with BBS-related phenotypes. Our previous work evaluating the extent to which an allele can suppress the knockdown defects has revealed both syndromic and non-syndromic aspects to BBS3 mutations ( 38 ).
To evaluate the level of protein expression in the injected embryos, we placed N-terminal myc-tags on WT and missense c8orf37 constructs. Zebrafish embryos were injected with RNA of the WT or missense forms. Protein was isolated from RNA injected embryos at day 5 and the abundance of myc-tagged product was determined by western blot. WT, R177W and Q182R c8orf37 products were expressed ( Supplementary Material, Fig. S5 ). Because of the activation of NMD by K102* shown in Human, this allele was not characterized. We next determined if myc-tagged wt- c8orf37 suppressed the MO knockdown, similar to untagged form. Sequential injection of RNAs and c8orf37 splice MO demonstrated that both the untagged and myc-tagged form suppress the MO-induced KV defect. c8orf37 morphants showed 31% KV defects, while the untagged and tagged RNA plus MO were at 13.8% (Fisher’s exact, P < 0.01) and 12.5% (Fisher’s exact, P < 0.05), respectively ( Fig. 6 A).
Melanosome transport rates were evaluated at 5 dpf in the sequentially injected embryos. Both tagged and untagged forms rescued the MO defect ( Fig. 6 B). The melanosome transport rate in splice morphants (2.65 ± 0.17 min) was reduced to 1.76 ± 0.09 min ( P < 0.05) in untagged and 1.80 ± 0.11 min ( P < 0.05), in the myc-tagged RNA plus MO groups. Finally, both tagged and untagged WT c8orf37 rescued the MO-induced vision defect with morphants responding 2.29 ± 0.28 times, and untagged and tagged RNA plus MO responding 3.68 ± 0.27 times and 3.41 ± 0.30 times, respectively ( Fig. 6 C). We showed that the N-terminal myc tag did not interfere with C8orf37 protein function.
The pathogenicity of two of the reported C8ORF37 missense mutations associated with retinal degeneration was determined using overexpression of the mutant RNAs (R177W-RNA and Q182R-RNA) in WT zebrafish. Overall, embryos injected individually with R177W-RNA or Q182R-RNA showed normal morphology. However, expression of R177W and Q182R generated somewhat different phenotypes relative to KV formation, melanosome transport and vision. Injection of R177W-RNA into WT embryos induced statistically significant KV formation defects ( Fig. 6 A), melanosome transport delays ( Fig. 6 B) and visual impairment ( Fig. 6 C), while Q182R-RNA injected embryos displayed statistically significant KV formation defects ( Fig. 6 A). Overexpression of Q182R did not disrupt melanosome transport ( Fig. 6 B). Statistically significant visual impairment was observed when data were analyzed with one-tailed Student’s t -test ( P = 0.03) ( Fig. 6 C). These data support the pathogenicity of both mutations.
We next evaluated the extent to which R177W and Q182R expression rescued knockdown phenotypes with sequential injection of Splice MO and RNA. Sequential injection of R177W-RNA with the Splice MO failed to rescue the KV defect (23.5% of embryos with abnormal KV formation compared to 31% in Splice MO only) ( Fig. 6 A) and melanosome transport delay ( Fig. 6 B). While R177W-RNA plus Splice MO appeared to have decreased visual response rates compared to Splice MO only, it was not statistically significant. In summary, for all phenotypes examined (KV defects, melanosome transport and vision), sequential injection of Splice MO with either R177W or Q182R RNA failed to significantly rescue the phenotype compared to Splice MO only. Relative to the KV defect, Q182R-RNA alone was not significantly different than Splice MO-alone, and although Q182R-RNA in morphants showed a 19% KV defect (compared to 30% MO), it was not significantly different than Splice MO only ( Fig. 6 A). Q182R-RNA alone did not significantly alter melanosome transport nor did it suppress Splice MO-induced melanosome transport delay ( Fig. 6 B). Moreover, Q182R-RNA in morphants showed the same visual defect as MO-alone ( Fig. 6 C). The failure of either mutation to rescue knockdown with MO supports the pathogenicity of R177W and Q182R.
The novel C8ORF37 p.K102* variant reported here is the first association of BBS with C8ORF37 ( BBS21 ). C8ORF37 is not predicted to be a component of the BBSome nor based on homology would it be expected to be a component of the BBS gene network. However, our zebrafish model system showed that in vivo c8orf37 knockdown led to the zebrafish “BBS phenotypes,” which supports that the homozygous p.K102* variant found in a single BBS patient is likely the cause of the signs observed. Specifically, we determined that MO knockdown of c8orf37 results in delayed retrograde melanosome transport, abnormalities in KV morphogenesis, in addition to an abnormal visual function. In addition, we validated the pathogenicity of C8ORF37 mutations and showed that the K102*-containing transcript undergoes NMD. These findings support an interaction of C8ORF37 with the BBS gene network, yet to be characterized.
C8ORF37 (MIM614500) is a ciliary cytoplasmic protein that has no significant sequence homology to any other human protein. It has been given its own protein class: the retinal maintenance protein superfamily and has no known function. C8ORF37 is ubiquitously expressed ( http://www.gtexportal.org/home/gene/C8orf37 , last accessed April 19, 2016) and co-localizes with polyglutamylated tubulin at the base of the primary cilium in human retinal pigment epithelium cells and at the base of the photoreceptor connecting cilium ( 27 ). It is proposed to have interactions with NRF1 (alpha palindromic-binding protein), which function links the transcriptional modulation of key metabolic genes to cellular growth and development. The ubiquitous expression patterns throughout development are typical of BBS genes and further support C8ORF37 being involved with primary cilia function and implicated in the development of BBS [( http://bgee.unil.ch/bgee/bgee , last accessed April 19, 2016), ( http://www.ncbi.nlm.nih.gov/unigene , last accessed April 19, 2016), ( http://www.proteinatlas.org/ , last accessed April 19, 2016) and ( http://www.ebi.ac.uk , last accessed April 19, 2016)]. While C8ORF37 has not been previously reported to be involved in BBS, two siblings with retinal disease associated with a C8ORF37 mutation ( C8ORF37 c.156-2A->G) also had polydactyly on the right foot or hand ( 7 , 8 ). Whether this finding was truly isolated is unclear as signs of BBS may be mild and overlooked if not functionally significant ( 42 ).
In humans, allelic heterogeneity of C8ORF37 leads to different retinal degeneration phenotypes ( 27–32 ) and now includes that of BBS (BBS21). As study of other ciliary proteins has shown, mutations can lead to important phenotype heterogeneity ( 42 , 43 ). The effect of an early null mutation (p.K102*) may explain the full blown BBS phenotype. The potential modifying role of the likely benign heterozygous BBS4 (p.K46R) on C8ORF37 p.K102* expression is yet to be determined. The recent report of a known C8ORF37 mutation (R177W) associated with a BBS-like phenotype ( 44 ) also suggests that early disruption of C8ORF37 may lead to a more syndromic phenotype with extra-ocular features. Our work supports that C8ORF37 mutations associate with syndromic and non-syndromic retinal degeneration.
Zebrafish have been a useful model in allowing validation of a range of conserved novel variants. In this study, we find that WT c8orf37 can rescue knockdown defects related to BBS and to vision, unlike the missense mutations p.Q182R and p.R177W that do not globally suppress the knockdown phenotypes. Based on the zebrafish work, it could be that missense mutations result in a less deleterious effect than an early null mutation, as indicated by the human data.
Although C8ORF37 mutations are a rare cause of BBS (BBS21), such an association is important and suggests that patients with C8ORF37 mutations should have a medical workup to assess the potential presence of BBS features including obesity, renal abnormalities, diabetes and hypertension. Furthermore, this expands the components of the BBS-related network. How C8ORF37 interacts with the BBS protein network including the BBSome remains to be determined.
Materials and Methods
This work was approved by the Hospital for Sick Children (Toronto, Canada) Ethics Review Board respecting the Tenets of the Declaration of Helsinki and the University Animal Care and Use Committee at the University of Iowa (IA). All oligonucleotides used are listed in Supplementary Material, Table S1 .
Clinical assessment details
The patient was recruited through the Ocular Genetics Clinic of the Hospital for Sick Children (Toronto) and met the clinical diagnostic criteria of BBS ( 1 , 45 ). Phenotype information collected included a comprehensive eye exam and kinetic visual fields (Goldmann perimeter). The spectral-domain optical coherence tomography (Cirrus, Carl Zeiss Meditec) was also used to provide high-resolution (axial resolution: 5 µm) cross-sectional images of the central retinal layers. Central retinal thickness referred to the 1 mm diameter circle centered on the fovea ( 46 ). The recording of the full-field ERG met the International Society for Clinical Electrophysiology of Vision standards ( 47 ).
Molecular assessment of the patient
Genotyping of patient
Genotyping on blood derived DNA first excluded the known BBS genes using a Clinical Laboratory Improvement Amendments-approved laboratory. We followed this testing with standard whole-genome re-sequencing using mated gapped reads was performed by Complete Genomics ( http://www.completegenomics.com/ , last accessed April 19, 2016) ( 48 ). Mapping to genome assembly GRCh37/hg19 and annotation was done in our in-house facility; TCAG ( http://www.tcag.ca/facilities/dnaSequencingSynthesis.html , last accessed April 19, 2016). Details of the filtering steps are available in Supplementary Material, Fig. S1 . Potential disease-causing variants were validated according to internal protocols ( Supplementary Material, Fig. S1 ). Novel mutations were assessed using information from the literature and various mutations databases. The potential effect of novel variants was predicted using publicly available software: Polyphen-2 ( 49 ) and SIFT ( 50 ) followed by segregation analysis and determination of allele frequency in the normal population.
Testing of NMD on lymphoblast cell lines
To determine whether the variant c.304A > T (p.K102*) causes NMD ( 51–54 ), total RNA was extracted from the lymphoblast cell lines (patient and control) with RNeasy (Qiagen, Hilden, Germany) according to manufacturer’s instructions. Subsamples of the cells (106) were treated with 1.5 mg of cyclohexamide (stock 100 mg/ml in DMSO, 4 hours incubation at 37°C), which inhibits translation (NMD) ( 55 ) and an equivalent subsample with only DMSO (15 µl). RT-PCR used primers flanking C8ORF37 exon 3 ( C8ORF37 -mRNA, Supplementary Material, Table S1 ) where the nonsense variant is located and co-amplified with those of the Cyclophilin A (PPIA, Supplementary Material, Table S1 ) a housekeeping gene to control for experimental variation. The C8ORF37 cDNA RT-PCR products from mutant and control were validated by Sanger sequencing.
c8orf37 Zebrafish Model
Zebrafish were maintained in compliance with the University Animal Care and Use Committee at the University of Iowa. Embryos were collected from natural spawning and staged as described previously ( 56 , 57 ).
Total RNA was extracted from pools of WT embryos using the Total RNA Mini Kit (Tissue) (IBI Scientific® Peosta, IA) at the following stages: maternal (2–4 cells), 10–13 somites stage, 1 day post-fertilization (dpf), 2 dpf and 5 dpf as well as from the adult brain, eye, intestine, heart, kidney, liver, testis and ovary (oocytes). cDNA was synthesized utilizing oligo DT primers, the expression of c8orf37 was evaluated using primers spanning exon1 and exon5 ( c8orf37 -f and -r, Supplementary Material, Table S1 ) and β -actin expression served as a control ( Supplementary Material, Table S1 ).
Knockdown with antisense oligonucleotide morpholinos (MO)
In this article, the predicted zebrafish ortholog of human C8ORF37 is referred to as c8orf37 . We used two different antisense MO to knock down c8orf37 : antisense c8orf37 MOs targeting the translation start site (ATG MO, Supplementary Material, Table S1 ) or intron1-exon2 junction (Splice MO, Supplementary Material, Table S1 ) were purchased from Gene Tools (LLC, Philomath, OR). As the phenotype associated with previously reported mutations, we validated the pathogenicity of two previous reported variants. We characterized the phenotype of two missense mutations of C8ORF37 reported in patients with non-syndromic retinal degeneration: p.R177W and p.Q182R. We assessed the effects of the different MOs on visual behavior, KV formation and retrograde intracellular trafficking in vivo (melanosome transport assay), which are phenotypes known to be associated with BBS. MOs were air-pressure-injected into one- to two-cell staged embryos at concentrations ranging from 2 to 10 ng. The efficacy of splice MO on transcript was assayed by RT-PCR using cDNA generated from pools of MO-injected embryos (morphants) at 10–13 somite and 5 dpf stages using primers c8orf37 -rt-f and c8orf37 -rt-r ( Supplementary Material, Table S1 ). Cone-rod homeobox (crx) MO injection served as a control for the vision startle assay ( 58 , 59 ) and standard control MO injection served as an injection control ( Supplementary Material, Table S1 ).
Zebrafish BBS-related phenotypes assessment
Morphological analysis and immunohistochemistry of KV
The KV is a transient ciliated organ formed in the tailbud of somite-stage zebrafish ( 60–62 ). Morphological analysis of KVs in live 10–13 stage embryos (∼14 hpf) was performed using a stereoscope and size was scored relative to the width of the notochord ( 33 , 63 ). Representative live embryos were photographed with a Zeiss Axiocam camera. To assess KV, 10–13 somite stage embryos were fixed, stained against acetylated tubulin (Sigma- Aldrich (T7451)), goat anti mouse Alexa Fluor® 488 (Invotrogen (A-21121)) and photographed with an Olympus Confocal microscope (Fluoview FV10i, Olympus Corporation, Tokyo, Japan).
Retrograde melanosome transport assay
BBS knockdown was shown to delay the rate of melanosome retrograde transport ( 33 , 40 , 63 ). To disperse the melanosomes, 5-day-old larvae were incubated overnight in the dark and then treated with epinephrine (final concentration 500 µg/ml, Sigma E4375). The time for melanosomes to move to the perinuclear area was documented under a microscope (Olympus SZX7, objective lens 2X).
Vision startle response assay
A behavioral assay based on a startle response by a light evoked stimulus was performed as previously described ( 38 , 40 , 63–65 ). In short, 5-day-old larvae were light adapted for at least 1 h in advance and then evoked by 1 s block in bright light intensity for five times spaced 30 s apart. Swimming patterns were observed under a microscope. At the end of the assay, larvae were probed with a blunt needle. Those that failed to respond to mechanical stimuli were excluded from the data analysis.
Analysis of retinal morphology
After the vision startle response assay, larvae were fixed with 4% paraformaldehyde in phosphate-buffered saline, infiltrated with 15% sucrose, 30% sucrose followed by 100% optimal cutting temperature compound (OCT, Sakura, Tokyo, Japan). Larvae were cryo-sectioned at 10–12 µm thicknesses and stained according to a standard H&E staining protocol. Images were taken with an Olympus IX71 inverted microscope. The length of the individual photoreceptors was measured using Image J program. Coronal sections of the photoreceptors of the central retina of the brain side from three of WT and six c8orf37 morphants were studied.
Cloning zebrafish c8orf37 and generation of two human missense mutations
Full-length c8orf37 was amplified using cDNA synthesized from 5 dpf embryos with the primers ( c8orf37-EcoRI- f and XhoI-r , Supplementary Material, Table S1 ), cloned utilizing the StrataClone Blunt PCR Cloning Kit (Agilent Technologies, Santa Clara, CA) and subsequently subcloned into pCS2+. The cloned product was mutagenized using primers c8orf37 R177W f and r, Q182R f and r in Supplementary Material, Tables S1 and PfuUltraII fusion HS DNA polymerase. PCR products were digested with EcoRI and XhoI and then subcloned into a pCS2 + MT tagging myc N-terminally. The sequence of all products was verified. The cloned product was mutagenized using primers c8orf37 R177W f and r, Q182R f and r ( Supplementary Material, Table S1 ) and PfuUltraII fusion HS DNA polymerase. PCR products were digested with EcoRI and XhoI and then subcloned into a pCS2 + MT tagging myc N-terminally. Bold letters were mutated C-> T , T-> G and A-> G in order. Products were sequenced using SP6 and T7 primers ( Supplementary Material, Table S1 ).
RNA synthesis for rescue studies
Full-length c8orf37 RNA, N-terminally myc tagged RNA (myc-RNA), myc-R177W RNA and myc-Q182R RNA were in vitro transcribed utilizing the mMessagemMachine transcription kit (SP6) (Ambion. Austin, TX). For rescue experiments, synthesized RNA products (100 ng/µl RNAs) were injected into one-cell stage embryos followed by the injection of the splice MO.
Supplementary Material is available at HMG online.
We thank Autumn Marsden and Charles. C. Searby (University of Iowa, IA) for technical support, and Daniele Merico and Christian Marshall from The Centre of Applied Genomics (The Sick Children Hospital, Toronto, Canada) for WGS filtering pipeline. We also thank Hélène Dollfus, MD, and Corinne Stoetzel who sequenced 44 families to look for C8ORF37 variants (from the Laboratoire de Génétique Médicale INSERM U1112 and Centre de référence pour les affections rares en génétique ophtalmologique (CARGO) (Strasbourg, FRANCE).
Conflict of Interest statement . None declared.
The Mira Godard Research Fund (to E.H.); The McLaughlin Foundation Accelerator grant (to E.H.); the Ophthalmology Research Fund (to E.H.); US National Institutes of Health grants (R01EY011298 and R01EY017168 to V.C.S.); Carver Charitable Trust (V.C.S.); Howard Hughes Medical Institution (V.C.S.) and Cardiovascular Interdisciplinary Research Fellowship (Institutional NRSA HL007121 to J.E.G.).