Retrospective analysis in oculocutaneous albinism patients for the 2.7 kb deletion in the OCA2 gene revealed a co-segregation of the controversial variant, p.R305W

Background Oculocutaneous albinism (OCA) is an autosomal recessive disorder. A significant portion of OCA patients has been found with a single pathogenic variant either in the TYR or the OCA2 gene. Diagnostic sequencing of the TYR and OCA2 genes is routinely used for molecular diagnosis of OCA subtypes. To study the possibility that genomic abnormalities with single or multiple exon involvement may account for a portion of the potential missing pathogenic variants (the second), we retrospectively analyzed the TYR gene by long range PCR and analyzed the target 2.7 kb deletion in the OCA2 gene spanning exon 7 in OCA patients with a single pathogenic variant in the target genes. Results In the 108 patients analyzed, we found that one patient was heterozygous for the 2.7 kb OCA2 gene deletion and this patient was positive with one pathogenic variant and one possibly pathogenic variant [c.1103C>T (p.Ala368Val) + c.913C>T (p.R305W)]. Further analysis of maternal DNA, and two additional OCA DNA homozygous for the 2.7 kb deletion, revealed that the phenotypically normal mother is heterozygous of the 2.7 kb deletion and homozygous of the p.R305W. The two previously reported patients with homozygous of the 2.7 kb deletion are also homozygous of p.R305W. Conclusions Among the reported pathogenic variants, the pathogenicity of the p.R305W has been discussed intensively in literature. Our results indicate that p.R305W is unlikely a pathogenic variant. The possibility of linkage disequilibrium between p.R305W with the 2.7 kb deletion in OCA2 gene is also suggested. Electronic supplementary material The online version of this article (doi:10.1186/s13578-017-0149-3) contains supplementary material, which is available to authorized users.


Background
Albinism is a disease caused by a lack of or reduction in the production of melanin pigment. There are many genes involved in the complex production of melanin, known as melanogenesis. Albinism affects about 1 in 17,000 individuals worldwide, with varied subtypes and prevalence among different ethnic populations [1]. Oculocutaneous albinism (OCA) is associated with a hypopigmented state of melanin in skin, hair, and eyes. In addition, OCA may manifest with eye symptoms such as photophobia, strabismus, nystagmus, and reduced visual acuity. Nonsyndromic OCA describes the disorders in which the manifestations are apparently limited to skin and eye functions. Nonsyndromic OCA is an autosomal recessive condition with genetic heterogeneity. Six genes and one additional locus have been identified as responsible for seven nonsyndromic OCA subtypes [2]. They are OCA1 (TYR gene), OCA2 (OCA2 gene), OCA3 (TYRP1 gene), OCA4 (SLC45A2 gene), OCA6 (SLC24A5 gene), and OCA7 (C10orf11 gene). OCA5 was identified by linkage analysis to a genomic locus at 4q24 [3]. OCA1 is the most common subtype found in Caucasians and accounts for approximately 50% of cases worldwide [4,5]. OCA-2 accounts for 30% of cases worldwide and is most prevalent among African and African-American populations. OCA2 is estimated to affect one in 10,000 individuals of African background, while it affects 1 in 36,000 individuals of Caucasian background [6].
The OCA2 gene (also known as Pink-Eyed Dilution, P gene) codes for a membrane protein localized at melanosomes in the melanocytes. Its function may involve transport of tyrosine and some other small molecules [7]. The OCA2 protein may also play a role in stabilizing some melanosomal proteins such as tyrosinase and initiating melanin synthesis in melanocytes. Mutations in the OCA2 gene (cause the OCA2 phenotype (MIM 203200). The OCA2 gene consists of 24 exons (23 coding), spanning almost 345 kb of genomic DNA in the region of 15q11.2-q12. The OCA2 genomic location is also highly polymorphic.
In this study, we evaluated 108 clinically diagnosed OCA patients by sequencing their TYR and OCA2 genes. Then, we retrospectively analyzed the previously reported 2.7 kb OCA2 gene deletion using long-range PCR. We also confirmed the co-segregation of the deletion and the known controversial variant, p.R305W. We suggested that the p.R305W was the more ancestral allele and the 2.7 kb deletion was introduced into this allele more recently. This association will have a clinical impact on the pathogenicity assessment of the p.R305W variant and genetic consultation.

Results
In our past clinical genetic tests, OCA patients were sequenced for pathogenic variants in the TYR gene at the Molecular Diagnostics Laboratory in the University of Minnesota Medical Center, Fairview (Table 1). In the 68 OCA patients tested for the TYR gene mutations, single pathogenic variant was found in 23 patients and 2 pathogenic variants were found in 45 patients. Long range PCR was performed on these patients with a series of overlapped primer sets and with OCA patients with two pathogenic variants including one equivocal mutation of the single mutation samples as PCR controls (Additional file 1: Table S1). No abnormal PCR fragment was identified (data not shown). OCA2 gene sequencing test was developed later and used as a reflex clinical test for the patients without two pathogenic variants in the TYR gene sequencing at the National Eye Institute (Table 2). For the patients without pathogenic variant or with single pathogenic variant or variants with unknown significance in either TYR gene or OCA2 gene, we analyzed the OCA2 gene 2.7 kb exon 7 deletion. We found that none of these patients had the common exon 7 deletion. We then extended our exon 7 deletion analysis to the rest of the samples in the Tables 1 and 2. We discovered that one patient (NE27) was indeed heterozygous for the exon 7 deletion (Fig. 2). Interestingly, this patient NE27 was also positive for one known OCA2 gene pathogenic variant, c.1103C>T, p.Ala368Val [8], and one variant with uncertain significance but likely pathogenic, c.913C>T, p.R305W, as we interpreted based on our previous analysis and literature search. Moreover, the patient NE27's mother was phenotypically normal. We found that this mother was homozygous for the p.R305W variant (data not shown) and heterozygous for the exon 7 deletion (Fig. 2), suggesting the variant of p.R305W is likely to be a benign variant. In addition, the ExAC reported 383 homozygous of p.R305W in 119616 alleles worldwide (ExAC, 3/12/2017). This is strongly support the variant is polymorphism not a pathogenic mutation. We sequenced the 2.7 kb deletion junction fragment isolated from patient NE27 and determined the exact genomic abnormality for this deletion (Fig. 1) as reported previously [7]. Our study identified the deletion of 2958 bases at g.28,017,720_28,020,677 (GRCh38) and followed by replacement with triple A (c.646+1823_c.807+678delinsAAA, p.?), which deleted the exon 7 of OCA2 gene, leading to an exon skip and potential frameshift (because the exon 7 contains 161 bp which is not in-frame if deleted). This deletion generated an 802 bp PCR fragment by the MHR51F and MHR72R primers. We further analyzed two previously reported homozygotes of the 2.7 kb deletion for the p.R305W variant [7] and found that they were all homozygous for p.R305W too. One additional patient who was heterozygous for the deletion was also heterozygous for the p.R305W variant. It clearly indicated the phase that the 2.7 kb deletion spanning the exon 7 of OCA2 gene was indeed in cis with the p.R305W variant in the exon 9 of OCA2 gene.

Discussion
In our study, we sequenced DNA samples from 108 OCA patients for TYR or OCA2 gene pathogenic variants. Our results consistently supported the published allele distributions [4]. We also performed long range PCR to search for potential large genomic rearrangements such as deletion or duplication in the TYR locus and for the 2.7 kB deletion in OCA2 gene. We did not identify any abnormality at the TYR intronic regions using the long range PCR methodology, suggesting that other mechanisms, such as splicing error variants, might be involved. We cannot rule out that this approach was not sensitive enough for smaller sizes of deletions/insertion or even larger massive genomic rearrangements [10]. In addition, we also noticed high frequent of the Alu repetitive sequences in the TYR intronic regions.
The three primer PCR assay system that we adapted from Durham et al. [9] enabled us to determine the genotype of the individuals in a single multiplex PCR with an internal wild type intron sequence control. Observation of the 802 bp deletion-specific junction fragment, the 3.5 kb fragment for the intact genomic region, and the 237 bp fragment spanning one of the breaking points in OCA2 heterozygotes provided a convenient carrier status confirmation analysis for routine clinical diagnostics (Fig. 2).
In a comparison of gender, ethnic background, genotypes from sequencing results and the deletion status in five samples that were positive for the p.R305W variant (Table 3), besides the patient NE27, we found that the OCA2 gene in patient NE12 contains two heterozygous pathogenic variants: an A-G change at nucleotide 1465 (c.1465A>G) in exon 14, and a T-G change at nucleotide 2310 (c.2310T>G) in exon 22. These pathogenic variants change the amino acid Asparagine (N) to aspartic acid (D) at residue 489 (p.N489D) and amino acid Tyrosine Asterisk represents termination codon a The pathogenic alleles in these patients have been reported previously [7] (Y) to a termination codon (*) at residue 770 (p.Y770*), respectively, in the P protein. The pathogenic variant p.N489D has been identified in unrelated patients diagnosed with Oculocutaneous Albinism Type 2. The pathogenic variant p.Y770* has not been identified in patients diagnosed with Oculocutaneous Albinism Type 2 to our knowledge. However, this nonsense variant is predicted to cause a premature termination in translation and lead to protein product truncation and expected to be pathogenic. The p.R305W was also present in this patient with unknown phase. In three other patients (NE21, NE22, NE30), it was interesting to note that they were all positive for a single pathogenic variant in the TYR gene. It was more likely that there would be an unknown secondary pathogenic variant at their TYR locus on the other chromosome.
The OCA2 (P) gene consists of 24 exons (23 coding), spanning almost 345 kb of genomic DNA in the region of 15q11.2-q12. The 2.7 kb deletion was expected to result in a frameshift in the first luminal loop, predicting a truncation of the polypeptide and a non-functional gene product. Regarding the protein structure and predicted function, the p.R305 locates in the first luminal loop. Kamaraj and Purohit provided a computational model to depict the functional consequence of p.R305W change [2]. It was predicted that the p.R305W leads to more flexible in conformation, which might alter the catalytic function in melanin biosynthesis. We propose that such functional alteration is not sufficient to lead   [7]. Lane 2 was NE27. Lane 3 was the mother of NE27. Lane 4 was patient NE28. Lane 5 was no DNA control to OCA. Therefore, homozygous genotype itself would not present OCA and it is demonstrated by unaffected individuals in the population. But in a subpopulation of p.R305W, co-segregated 2.7 kb exon 7 deletion was the real disease-causing mutation. Our previously identified exon 7 deletion positive samples in homozygous and heterozygous status were consistently associated with the p.R305W variant. It suggests that the p.R305W variant was more likely an associated variant as a more ancient allele in linkage disequilibrium. The pathogenicity of the p.R305W has been discussed intensively in literature [2,3,11,12]. Until this study, we were still interpreting that the p.R305W was a VUS with likely pathogenic. Based on the analysis on allele origination, our results are not consistent with the p.R305W variant functioning as a pathogenic variant in an autosomal recessive inheritance pattern. We speculate that the dispute for the pathogenicity of p.R305W was due to lack of analysis on this exon 7 related deletion in the past, which presented as disease associated marker. It should be brought to an attention for our clinical molecular diagnosis on OCA2. More attention should be given to the patient ethnic background and the possibility of a 2.7 kb exon 7 deletion associated with the p.R305W variant in sequencing tests. The p.R305W variant itself is likely benign variant in cis with the exon 7 deletion and likely misleading for OCA accurate molecular diagnosis.

Conclusion
In summary, we herein evaluated 108 clinically diagnosed OCA patients by sequencing their TYR and OCA2 genes, and alkalized the 2.7 kb OCA2 gene deletion using longrange PCR. We confirmed the co-segregation of the deletion and the known controversial variant, p.R305W. Our results suggest that the p.R305W is the more ancestral allele and the 2.7 kb deletion is introduced into this allele more recently. These results provided evidence that do not support a pathogenic classification for the p.R305W variant. This association will have a clinical impact genetic consultation for OCA patients. in Bethesda, Maryland. This study was reviewed and approved by the University of Minnesota University IRB and the Combined Neuroscience Institutional Review Board of the National Institutes of Health. Informed consent was obtained from each participant as adhering to tenets of the Declarations of Helsinki.

Patients
Genomic DNA was extracted using the Puregene genomic DNA extraction kit (Qiagen), and quantified with the ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE). Genomic DNA samples from patients were sequenced using protocol described previously [13].
Long range PCR was performed using the LA PCR kit 2.1 from Takara (Clontech Laboratories, Mountain View, CA). PCR primers for the TYR gene intron regions were designed using GeneRunner 3.01 software [13] and PCR condition were optimized by annealing temperatures, magnesium concentrations, and DMSO concentrations (Additional file 1: Table S1). PCR primers for the OCA2 2.7 kb deletion and PCR conditions were adapted from