We studied 39 individuals with ocular albinism and members of their families. We found mutations in 12 individuals out of 39 (31%) with a clinical diagnosis of ocular albinism. We found two previously reported mutations and eight novel sequence changes in the OA1 gene. Among the previously reported mutations, the L134P change was found in what seemed to be three unrelated French families. These families were initially examined in two different countries, by three different ophthalmologists and three different medical geneticists. Details about the extended family were either overlooked or not known. The same nucleotide mutation (401T>C) was not found in 100 ethnically matched control patients. From a meticulous study by interview of the patients' pedigree, it was found that these families are three branches of the same family (Figure ). The unmarried lone daughter (II3) of one family and the mothers of the other two families (II2 and II4) are in fact three sisters who had not readily disclosed their biological familial links. Sequence alignment analysis showed that L134 is conserved among all tetrapod OA1 proteins. This result suggests that this amino acid may play a key role in OA1 protein function. Therefore a mutation of this leucine residue, located in the third transmembrane domain, may result in a non-functional protein and explains the severe visual phenotype observed in the affected patients.
The mutation R285X resulted in the OA1 protein being truncated in its third luminal loop.
The G81V mutation affects a glycine residue conserved among the vertebrate orthologs of OA1. This valine residue, located in the second transmembrane domain, is larger than the glycine residue and may prevent the protein from folding correctly.
The C116W mutation affects a cysteine residue that is highly conserved between all species and among all GPCRs. This cysteine is thought to form a disulfide bond with C184, allowing the correct folding of the receptor. The mutation modifies the protein in the very end of the first luminal loop. Thus, the protein cannot fold properly and is therefore completely non-functional [17
]. Two other mutations have been reported to affect this amino acid: C116S and C116R [9
Alignments of all the vertebrate OA1 proteins show that a hydrophobic amino acid is found in most vertebrate OA1 proteins at the position corresponding to the human OA1 amino acid residue 166, which is normally a threonine (figure ). Although threonine and asparagine are both polar amino acids, threonine is much less hydrophilic than asparagine which is the mutated amino acid. Moreover, substituting the threonine at codon 166 with an asparagine substantially changes the residue side chain at a critical position of the fourth transmembrane domain, in the OA1 protein. Therefore, this amino acid substitution very likely causes structural and functional alterations to OA1 and is thus very likely to be highly pathogenic.
The c.163_170dupGCGGGCCC duplication is located in the region encoding the first intracellular loop and induces a frameshift and a premature stop codon (p.G58fsX29). This truncates the protein in its second transmembrane domain, completely preventing any function of the OA1 protein.
Both c.504_505delCT and c.601_602insT mutations induce a frameshift and a premature stop codon, truncating the protein in its third intracellular loop.
We used NNSPLICE in order to analyze the mutated mRNAs resulting from the splice site mutations, and we found that they had displaced splice sites, with neighboring cryptic splice sites being used instead of the missing normal splice site. This resulted in short erratic sequences ending with a stop codon after the normal sequence. Mechanistically, this observation is somewhat reminiscent of that previously reported by our lab in a case of pyruvate dehydrogenase deficiency [19
]. Thus, if the abnormal mRNA are actually translated and escape degradation, the encoded proteins are truncated and non-functional.
Recent findings have shown that new splice sites can appear within introns, leading to aberrant mRNA [19
]. These splice sites cannot be found at the genomic DNA level by using screening techniques limited to exons and canonical splice sites at the exon boundaries [19
]. We observed 21 patients who had apparently no OA1
gene mutations or deletions despite presenting an apparently X-linked form of ocular albinism. These patients may have this type of intron mutation or may have mutations that alter the regulatory regions of the gene, such as the promoter region, silencers or enhancers.
Normally, with the techniques currently used for the molecular genetics diagnosis of X-linked ocular albinism, one third of affected individuals is not found to carry such mutations [10
]. Other patients have a completely or partially deleted gene.
Apparently sporadic cases may be due to several reasons. In some instances, we could simply not be able to get all the DNA samples of the families. Sometimes, de novo
germinal alterations could have arisen, the mechanisms of which are diverse, including DNA lesions and repair or recombinations during meiosis [21
]. The latter mechanism looks more plausible, as there may be many repeated sequences within introns or around the gene, leading to misalignments of X chromosomes during meiosis and unequal crossing-over(s). This may explain the high deletion rate in OA1
gene. Germinal mosaicism may also be a cause of sporadic deletion in OA1
Another possibility is that some of these patients may have been considered erroneously as affected by X-linked ocular albinism while they are in fact affected by very mild forms of oculocutaneous albinism. The limited size of the pedigrees analyzed in this study prevents us from stating with certainty the mode of inheritance of all the clinically diagnosed ocular albinisms included in our cohort. We cannot rule out that ocular albinism may constitute a heterogenous genetic ensemble despite being apparently clinically homogenous. However, in light of all the studies published so far, this hypothesis is very unlikely and all clinical forms of ocular albinism appear with a very high degree of probability linked to genetic alterations of the OA1 gene, despite the variability of symptoms observed in different patients [22
Studying the size and the sequences of the mRNA from genes encoding proteins involved in melanogenesis is hindered by the extreme difficulty in their amplification by RT-PCR of total RNA extracted from lymphoblastoid or fibroblast cell lines from affected patients. The illegitimate transcription of such cell lines to obtain cDNA corresponding to the transcripts encoding melanogenic proteins has never been shown to be particularly efficient. Indeed, these genes either are not transcribed or are only weakly transcribed in these cell lines. Therefore, we would need to obtain skin biopsies from affected patients, purify melanocytes from these samples, culture them and then extract total RNA from these cultures before carrying out RT-PCR.
Our mutational screening would certainly benefit from quantitative genomic PCR [23
] in order to increase the probability of detecting other OA1
mutations that may have gone unnoticed by the sequencing procedure used for small, amplified exon fragments studied in this report.