Novel OPN1LW/OPN1MW Exon 3 Haplotype-Associated Splicing Defect in Patients with X-Linked Cone Dysfunction

Certain combinations of common variants in exon 3 of OPN1LW and OPN1MW, the genes encoding the apo-protein of the long- and middle-wavelength sensitive cone photoreceptor visual pigments in humans, induce splicing defects and have been associated with dyschromatopsia and cone dysfunction syndromes. Here we report the identification of a novel exon 3 haplotype, G-C-G-A-T-T-G-G (referring to nucleotide variants at cDNA positions c.453, c.457, c.465, c.511, c.513, c.521, c.532, and c.538) deduced to encode a pigment with the amino acid residues L-I-V-V-A at positions p.153, p.171, p.174, p.178, and p.180, in OPN1LW or OPN1MW or both in a series of seven patients from four families with cone dysfunction. Applying minigene assays for all observed exon 3 haplotypes in the patients, we demonstrated that the novel exon 3 haplotype L-I-V-V-A induces a strong but incomplete splicing defect with 3–5% of residual correctly spliced transcripts. Minigene splicing outcomes were similar in HEK293 cells and the human retinoblastoma cell line WERI-Rb1, the latter retaining a cone photoreceptor expression profile including endogenous OPN1LW and OPN1MW gene expression. Patients carrying the novel L-I-V-V-A haplotype presented with a mild form of Blue Cone Monochromacy or Bornholm Eye Disease-like phenotype with reduced visual acuity, reduced cone electroretinography responses, red-green color vision defects, and frequently with severe myopia.


Introduction
OPN1LW and OPN1MW encode the apo-proteins of the long-wavelength-sensitive (LWS) and middle-wavelength-sensitive (MWS) cone photopigments, respectively. OPN1LW and OPN1MW share high sequence homology and are arranged as a tandem repeat gene cluster on Xq28 with a single OPN1LW gene followed by one or multiple OPN1MW gene copies [1]. Their expression is governed by the locus control region (LCR), an upstream Int. J. Mol. Sci. 2022, 23, 6868 2 of 11 enhancer element that selectively contacts the proximal promoter of OPN1LW or OPN1MW to drive mutual exclusive expression of the one or the other gene in individual cone photoreceptors [2].
In 2012, Ueyama and colleagues showed that a small fraction of protanopic and protanomalous color vision defects are linked to rare combinations of exon 3 single nucleotide polymorphisms in OPN1LW which induces mis-splicing and skipping of exon 3 [7]. Ever since, we and others have corroborated and extended exon 3 haplotype-induced splicing defects of OPN1LW and OPN1MW genes as a frequent cause of cone dysfunction disorders in patients with BCM, Xl-CD and BED [8][9][10][11][12].
So far, only a few distinct exon 3 haplotypes have been observed recurrently in patients and validated by in vitro splicing assays including G-C-G- Here we present clinical and functional analysis data which demonstrate that G-C-G-A-T-T-G-G ('L-I-V-V-A') is another deleterious exon 3 haplotype that induces exon 3 skipping and is associated with cone dysfunction disorders in seven male patients from four families.

Identification of Patients with the Novel LIVVA Exon 3 Haplotype in the OPN1LW and/or the OPN1MW Gene and Functional Minigene Splicing Assays
The common denominator of the probands in this study is the presence of a novel rare exon 3 haplotype G-C-G-A-T-T-G-G ('L-I-V-V-A') in one or all copies of the X-linked cone opsin genes (Table 1, Figure 1). All probands had a structurally normal OPN1LW/OPN1MW gene cluster with a single OPN1LW and a single OPN1MW gene copy.
The novel G-C-G-A-T-T-G-G ('L-I-V-V-A') haplotype was present either in both OPN1LW and OPN1MW (BCM363_19041), only in the OPN1LW gene (ZD569 and BCM375 family, two affected brothers each) or only in the OPN1MW gene (both brothers of the BCM353 family). Both brother pairs in the ZD569 and the BCM375 families carried the A-A-C-G-G-T-G-G ('M-V-V-V-A') exon 3 haplotype in their OPN1MW gene copy, and both brothers of the BCM353 family carried the G-C-G-A-T-C-G-T ('L-I-A-V-S') exon 3 haplotype in their OPN1LW gene (Table 1, Figure 1).
) exon 3 haplotype in their OPN1MW gene copy, and both brothers of the BCM353 family carried the G-C-G-A-T-C-G-T ('L-I-A-V-S') exon 3 haplotype in their OPN1LW gene (Table 1, Figure 1). We used an established minigene assay to analyze the effect of the novel G-C-G-A-T-T-G-G ('L-I-V-V-A') exon 3 haplotype in HEK293 cells. Assays were also performed in parallel for the two other exon 3 haplotypes observed in the probands as well as two control haplotypes, G-C-G-A-T-C-G-G ('LIAVA') and A-A-C-G-G-C-A-T ('MVAIS'). In HEK293 cells, we observed a strong, near-complete splicing defect of the novel G-C-G-A-T-T-G-G ('L-I-V-V-A') haplotype with only 5.28 ± 0.09% of correctly spliced transcripts (Table 2, Figure 2). The fractions of correctly spliced transcripts for the other haplotypes were: 28.36 ± 0.5% for G-C-G-A-T-C-G-T ('L-I-A-V-S'), 74.41 ± 0.5% for A-A-C-G-G-T-G-G ('M-V-V-V-A'), and for the controls: non-detectable levels of correctly spliced transcripts for G-C-G-A-T-C-G-G ('LIAVA') and 98.12 ± 0.07% for A-A-C-G-G-C-A-T ('MVAIS'), values which are similar to previous reports [7][8][9][10]. The results of the minigene assay in HEK293 cells suggest a strong deleterious effect on exon 3 splicing for the novel haplotype We further investigated the reliability of the minigene assay in HEK293 cells in comparison with WERI-Rb1, a retinoblastoma cell line with cone-like character and expression profile [13,14]. In fact, we found that WERI-Rb1 cells well express OPN1LW (exon 3 haplotype: A-A-C-G-G-C-A-T ['MVAIS']) and OPN1MW (exon 3 haplotype: A-A-C-G-G-C-A-G ['MVAIA']) endogenously with no evidence of aberrantly spliced transcripts ( Figure  2). We found that the splicing outcome of minigene constructs expressed in WERI-Rb1 cells was highly similar to that observed in HEK293 cells ( Table 2), confirming that HEK293 cells are well suited for the analysis of OPN1LW and OPN1MW exon 3 splice defects. We used an established minigene assay to analyze the effect of the novel G-C-G-A-T-T-G-G ('L-I-V-V-A') exon 3 haplotype in HEK293 cells. Assays were also performed in parallel for the two other exon 3 haplotypes observed in the probands as well as two control haplotypes, G-C-G-A-T-C-G-G ('LIAVA') and A-A-C-G-G-C-A-T ('MVAIS'). In HEK293 cells, we observed a strong, near-complete splicing defect of the novel G-C-G-A-T-T-G-G ('L-I-V-V-A') haplotype with only 5.28 ± 0.09% of correctly spliced transcripts (Table 2, Figure 2). The fractions of correctly spliced transcripts for the other haplotypes were: 28.36 ± 0.5% , and for the controls: non-detectable levels of correctly spliced transcripts for G-C-G-A-T-C-G-G ('LIAVA') and 98.12 ± 0.07% for A-A-C-G-G-C-A-T ('MVAIS'), values which are similar to previous reports [7][8][9][10]. The results of the minigene assay in HEK293 cells suggest a strong deleterious effect on exon 3 splicing for the novel haplotype G- Table 2. Splicing of OPN1LW and OPN1MW exon 3 haplotype minigene constructs in HEK293 and WERI-Rb1 cells.

Figure 2.
Bioanalyzer separation of RT-PCRs from HEK293 and WERI-Rb1 cells. RT-PCR products obtained with RNA from HEK293 and WERI-Rb1 cells transfected with different exon 3 minigene constructs were separated on a Bioanalyzer DNA-1000 chip. The exonic composition of the products is indicated on the right. RT-PCR products with RNA from non-transfected WERI-Rb1 cells showing expression and correct splicing of endogenous OPN1LW and OPN1MW transcripts were separated in comparison (lanes WERI-Rb1 (LW) and WERI-Rb1 (MW)). Note that different reverse primers were used for amplification of cDNA from minigene transfected cells (i.e., construct-specific reverse primers) and non-transfected WERI-Rb1 cells, which also results in fragment size differences. STD-Size Standard.

Clinical Findings
Seven male patients were examined in this project, and all but one (BCM363_19041) had at least one follow-up examination ( Table 3). The youngest patient was 4 years old, and the oldest at follow-up was 15 years old. The age of onset of the first visual symptoms varied between 3 years and 10 years. In two patients, the age of onset was unknown as they did not perceive any subjective visual difficulties prior to eye examinations. For the rest of the patients, decreased BCVA was the leading first symptom. One patient (BCM363_19041) reported dyschromatopsia in ambient bright light conditions.
The BCVA of all patients was usually decreased to the middle range, but varied between 20/125 and 20/20. There was no clear difference between the BCVA values of patients with the different OPN1LW and OPN1MW exon 3 haplotype combinations; however, full visual acuity was only reached by ZD569_26825, an almost asymptomatic patient carrying the novel LIVVA exon 3 haplotype in OPN1LW and the common MVVVA exon 3 haplotype in OPN1MW. All but two patients (BCM363_19041 who was emmetropic and BCM353_31197 who was hyperopic at the first visit at the age of 4 years and emmetropic at the follow-up visit at the age of 10 years) had high myopia up to −20 spherical diopters with astigmatism. Five patients reported photophobia, only two (BCM353_31197 and _31198) did not. If color vision tests were performed, the majority of the patients had mixed confusions along the red-green axes, but in two of the patients carrying the LIVVA exon 3 haplotype in OPN1MW pure green-color deficits were recorded. The borders of visual fields were normal in all subjects. RT-PCR products obtained with RNA from HEK293 and WERI-Rb1 cells transfected with different exon 3 minigene constructs were separated on a Bioanalyzer DNA-1000 chip. The exonic composition of the products is indicated on the right. RT-PCR products with RNA from non-transfected WERI-Rb1 cells showing expression and correct splicing of endogenous OPN1LW and OPN1MW transcripts were separated in comparison (lanes WERI-Rb1 (LW) and WERI-Rb1 (MW)). Note that different reverse primers were used for amplification of cDNA from minigene transfected cells (i.e., construct-specific reverse primers) and non-transfected WERI-Rb1 cells, which also results in fragment size differences. STD-Size Standard.
We further investigated the reliability of the minigene assay in HEK293 cells in comparison with WERI-Rb1, a retinoblastoma cell line with cone-like character and expression profile [13,14]. In fact, we found that WERI-Rb1 cells well express OPN1LW (exon 3 haplotype: endogenously with no evidence of aberrantly spliced transcripts (Figure 2). We found that the splicing outcome of minigene constructs expressed in WERI-Rb1 cells was highly similar to that observed in HEK293 cells ( Table 2), confirming that HEK293 cells are well suited for the analysis of OPN1LW and OPN1MW exon 3 splice defects.

Clinical Findings
Seven male patients were examined in this project, and all but one (BCM363_19041) had at least one follow-up examination ( Table 3). The youngest patient was 4 years old, and the oldest at follow-up was 15 years old. The age of onset of the first visual symptoms varied between 3 years and 10 years. In two patients, the age of onset was unknown as they did not perceive any subjective visual difficulties prior to eye examinations. For the rest of the patients, decreased BCVA was the leading first symptom. One patient (BCM363_19041) reported dyschromatopsia in ambient bright light conditions. The BCVA of all patients was usually decreased to the middle range, but varied between 20/125 and 20/20. There was no clear difference between the BCVA values of patients with the different OPN1LW and OPN1MW exon 3 haplotype combinations; however, full visual acuity was only reached by ZD569_26825, an almost asymptomatic patient carrying the novel LIVVA exon 3 haplotype in OPN1LW and the common MVVVA exon 3 haplotype in OPN1MW. All but two patients (BCM363_19041 who was emmetropic and BCM353_31197 who was hyperopic at the first visit at the age of 4 years and emmetropic at the follow-up visit at the age of 10 years) had high myopia up to −20 spherical diopters with astigmatism. Five patients reported photophobia, only two (BCM353_31197 and _31198) did not. If color vision tests were performed, the majority of the patients had mixed confusions along the red-green axes, but in two of the patients carrying the LIVVA exon 3 haplotype in OPN1MW pure green-color deficits were recorded. The borders of visual fields were normal in all subjects.
In full-field electroretinography (ERG), all patients showed normal scotopic recordings confirming well-functioning rods. Photopic recordings, however, showed various degrees of reduced cone functionality with reduced amplitudes and some prolonged implicit times, without clear differences in cone-driven recordings between patients with the LIVVA exon 3 haplotype in OPN1MW or OPN1LW.
Retinal imaging was almost normal, but all patients had a rather hypopigmented fundus appearance with myopic peripapillary atrophy in particular in the myopic patients, as well as some slight temporal paleness of the optic nerve head (Figure 3). Although OCTs showed normal layering of the central retina, all patients with the LIVVA exon 3 haplotype in OPN1MW had slight signs of diffuse disruption of the central outer segments. Additionally, patients BCM375_31869 and BCM375_31870 showed a very mild form of foveal hypoplasia in FAF. Decent foveal and perifoveal changes were observed in the patient cohort in FAF, which were characterized by a hypopigmented fovea with usually a slight hyperpigmentation perifoveal ring. An overview of the clinical findings is given in the Figure 3 and Table 3.

Discussion
Exon 3 haplotype-induced splicing deficiency comprises a rather new category of molecular pathology in OPN1LW/OPN1MW opsin gene cluster linked disease. So far, only a small number of unequivocally deleterious exon 3 haplotypes have been recurrently reported and their splicing deficiency functionally confirmed by minigene assays ( [7][8][9][10]. Herein, we report a novel splicing-deficient exon 3 haplotype, G-C-G-A-T-T-G-G ('L-I-V-V-A'), observed in a number of male probands with cone dysfunction disorders. The minigene assay confirmed a severe splicing deficiency with only about 5.4% and 3.9% of correctly spliced transcripts in HEK293 and WERI-Rb1 cells, respectively. This fraction is higher than in the prototypical splicing-deficient 'L-I-A-V-A' haplotype but lower than for the 'L-I-A-V-S' haplotype first described by Mizrahi-Meissonier and colleagues [5] and also found in the OPN1LW gene copy in the two brothers of the BCM353 family. To the best of our knowledge, our study for the first time investigated the exon 3 haplotypelinked splicing defect in WERI-Rb1, a retinoblastoma cell line with cone-like character and expression profile. In fact, we demonstrated endogenous OPN1LW and OPN1MW gene expression and its proper splicing in WERI-Rb1 cells (Figure 2). The results obtained with the WERI-Rb1 cell line closely resemble and support the reliability-also in quantitative terms-of the HEK293 cell system for the assessment of OPN1LW and OPN1MW exon 3 linked splicing defects.
The clinical phenotype of the patients reported here can be described as BCM or BED. BED has been described as represented by X-linked myopia with astigmatism, deuteranopia and reduced cone responses with hypoplasia of the optic nerve head [15][16][17][18]. BCM is a congenital X-chromosomal functional disorder of both red and green cones [19,20].
Although the phenotypes of BED and BCM overlap, optic nerve head atrophy, high myopia and better preserved red color perception are rather characteristic of BED. The cohort described here had a fundus appearance compatible with BED with slight temporal paleness of the optic nerve head and hypopigmented fundi. Retinal imaging with OCT showed a slight diffuse disruption of the central photoreceptor outer segments in patients carrying the LIVVA exon 3 haplotype in OPN1MW, but these changes were rather discreet. The remaining patients had a rather normal foveal morphology (ZD569_26825 and _26827) or a small level of foveal hypoplasia (BCM535_31197 and _31198).
The functional aspects of the vision in our cohort did not allow for a precise differentiation into the categories of BED or BCM based on exon 3 haplotypes. Still, a pure green-color defect was observed only in two patients carrying the LIVVA haplotype in OPN1MW (BCM363_19041 and BCM353_31198). Although this would be characteristic for BED, exactly these two patients were emmetropic, thus not fulfilling the high myopia criterion of BED. Additionally, in hue tests, the green and red color axes are similar to each other, thus a pure green-color defect detected by hue tests might be to some part also mixed with red-color vision problems. All other patients had a color vision defect of both red (L) and green (M) cones. Rod functionality, measured with full-field ERG, seemed normal in the whole cohort. This is in contrast with the previously published phenotype of exon 3 opsin phenotypes described as X-linked myopia with normal visual acuity, color perception and rod affection [21].
In our cohort, cones were affected to different extents; however, color disturbance or some degree of BCVA reduction and reduced photopic ERG were present in all patients. However, there were two patients who did not report subjective symptoms before ophthalmologic examination. Otherwise, the cone functional loss in the sense of decreased visual acuity at preschool age or early school age was the leading symptom. These findings are in concordance with BED [22] but also BCM.
Because our cohort included only minors, we cannot report clinical progression of the cones' functional loss into adulthood. Still, the follow-ups in these patients covering preschool age and teenage confirmed rather stable findings consistent with BCM in general and corresponded with previously reported findings [23].
Morphological changes of the retinas in these patients were rather decent. Close to normal retinal layering was observed in OCT images and retinal appearance showed some degree of hypopigmentation and optic nerve head paleness. Because many of the patients had myopia, myopic changes such as peripapillary atrophy were present, too. Findings of macular atrophy, such as reported in affected adults by Orosz and colleagues [21], were not observed in our cohort. There were only small changes in the perifoveal autofluorescence in our patients, reflecting probably a small degree of pigment epithelium changes in the area of the highest red (L) and green (M) cone density.
In summary, the phenotype of this cohort represents a BCM phenotype with various degrees of color perception problems, mostly with myopia, astigmatism and slight optic nerve head atrophy, sharing many features with the BED phenotype [6] although none of our patients fulfilled all typical BED criteria, i.e., a deutan color vision defect, high myopia and optic nerve head atrophy. A clear phenotypic distinction between subjects with the L-I-V-V-A exon 3 haplotype either in OPN1LW or in OPN1MW was not observed in our small patient cohort.

Patient Recruitment and Clinical Evaluation
The study was conducted in accordance with the tenets of the World Medical Association Declaration of Helsinki. Patients were recruited as part of local clinical studies or ad hoc at different centers specialized in inherited retinal diseases during routine clinical diagnostics. Blood samples for genetic analysis were obtained from all patients upon written informed consent. The study was approved by the institutional review board of the Ethics Committee of the University Hospital of Tübingen (project No. 116/2015BO2, last update and review 9 February 2021).
Patients underwent multimodal clinical examinations including functional tests of the retina as well as retinal imaging. All patients were asked for detailed ophthalmological history with age of first symptom onset, presence of night blindness and photophobia. Examinations included best corrected visual acuity (BCVA), color vision testing using mostly the Farnsworth 15 hue test, full-field electroretinography and kinetic perimetry or gross confrontational perimetry in children. Retinal imaging with fundus photography, spectral domain optical coherence tomography (Spectralis OCT) and fundus autofluorescence (FAF, both with Heidelberg Engineering, Heidelberg, Germany) were also performed.

Genotyping of the OPN1LW/OPN1MW Gene Cluster
The integrity and composition of the OPN1LW/OPN1MW gene cluster was analyzed by a custom screening protocol as described previously [24]. Long-distance PCR was used to specifically amplify parts of the OPN1LW gene and OPN1MW gene, respectively, and to determine the exon 3 sequences by Sanger sequencing.
Multiplex ligation-dependent probe amplification (MLPA) was used to determine the copy number at the OPN1LW/OPN1MW gene cluster. For MLPA we used a premarketing release version of the SALSA X080-B1 Opsin probe-mix (MRC Holland, Amsterdam, The Netherlands) targeting sequences of the LCR and individual exons of OPN1LW and OPN1MW specifically. MLPA was performed according to the manufacturer's recommendations and amplification products were separated by capillary electrophoresis on an ABI 3130XL (Applied Biosystems/ThermoFisher Scientific, Weiterstadt, Germany) instrument. Data analysis was performed using the Coffalyser.Net software (MRC Holland, Amsterdam, The Netherlands) which calculates ratios in comparison to reference samples co-processed with the test samples.

Genotyping of the OPN1LW/OPN1MW Gene Cluster
The prototype OPN1LW opsin minigene construct was kindly provided by Dr. Ueyama (Shiga University, Otsu, Japan) and used to generate minigene constructs with OPN1LW or OPN1MW gene variant combinations in exon 3, flanked by its full-length upstream and downstream introns and the remaining human OPN1LW cDNA [9]. In brief, a fragment encompassing exon 3 and parts of the flanking introns were PCR-amplified from genomic DNA from probands and inserted into the minigene vector backbone by conventional cloning procedures. The inserted sequences were verified by Sanger sequencing.
Data Availability Statement: All analyzed data are included in the manuscript. Raw data may be obtained from the authors upon written request.