A Comprehensive Genetic Study of Autosomal Recessive Ocular Albinism in Caucasian Patients

Abstract

purpose. Autosomal recessive ocular albinism (AROA) is a group of genetic disorders in which reduced pigmentation of the eye is associated with decreased visual acuity, nystagmus, strabismus, and photophobia, although pigmentation of skin and hair is relatively normal. Previous studies have shown that AROA in some cases constitutes a clinically mild presentation of oculocutaneous albinism (OCA), due to mutations in either the TYR (OCA1) or OCA2 (P) genes. The purpose of this study was to characterize the relative prevalence of different genetic forms of AROA, and to characterize a sample repertoire of gene mutations in a large series of Caucasian patients with AROA.

methods. Thirty-six unrelated Caucasian patients carrying the clinical diagnosis of AROA were studied by DNA sequence analysis of the four classic OCA genes: TYR, OCA2 (P), TYRP1, and SLC45A2 (MATP), as appropriate. In all patients with no apparent pathologic mutations in these genes, DNA sequence analysis was performed of a candidate OCA gene, SILV, and the two genes most often involved in Hermansky-Pudlak syndrome, HPS1 and HPS4, the most frequent syndromic form of OCA.

results. TYR gene mutations were identified in 20 (56%) patients, OCA2 mutations in 3 (8%), mutations in both TYR and OCA2 in 2 (6%), and possible TYRP1 mutations in 2 (6%). In at least nine patients, no mutations were found in any of the genes studied. Almost all patients with OCA1-related AROA were compound heterozygous for severe OCA1 mutant alleles and the common R402Q variant.

conclusions. Most patients with AROA represent phenotypically mild variants of OCA, well over half of which is OCA1.

Ocular albinism (OA) is a group of genetic disorders in which reduced pigmentation of the eyes is associated with decreased visual acuity, nystagmus, strabismus, and photophobia. Pigmentation of the skin and hair is normal, by definition, although patients often appear relatively lightly complected compared with their unaffected relatives. 

At least two general forms of OA are distinguished: the X-linked recessive Nettleship-Falls form (OA1; OMIM 300500; Online Mendelian Inheritance in Man; http://www.ncbi.nlm.nih.gov/Omim/ provided in the public domain by the National Center for Biotechnology Information, Bethesda, MD) and so-called autosomal recessive ocular albinism (AROA ¹ ). OA1 results from mutations in the OA1 locus. ² AROA, in at least some cases, results from mutations in TYR ³ or OCA2, ⁴ two of the 18 genes involved in the more severe oculocutaneous albinism (reviewed in Ref. ⁵ ). AROA in these patients thus represents clinically mild variants of oculocutaneous albinism type 1 (OCA1; OMIM 203100) and type 2 (OCA2; OMIM 203200). One patient with AROA has been reported to have de novo deletion of region q13-q15 of chromosome 6 ⁶ (OMIM 203310), although no pigmentary genes in this genomic region have yet been identified. 

We have shown that patients with TYR-related AROA may be compound heterozygous for various severe OCA1 mutations on one allele, the other allele carrying a polymorphic variant of the TYR gene, rs1126809, that is quite common among Caucasians, with allele frequency ∼0.278 (Ref. ⁷ and dbSNP; http://www.ncbi.nlm.nih.gov/SNP/snp_ref.cgi?rs=1126809). The rs1126809 variant encodes a tyrosinase enzyme with an arginine-to-glutamine substitution at codon 402 (R402Q ³ ), resulting in a tyrosinase polypeptide that is thermolabile⁷ and subject to retention in the endoplasmic reticulum, ⁸ yielding only 25% of the catalytic activity of wild-type enzyme at 37°C. ⁷ Patients with TYR-related AROA thus have total tyrosinase catalytic activity ∼12% to 13% of normal. One unique family segregates AROA along with Waardenburg syndrome type 2, the result of digenic inheritance of the TYR402Q allele and a frameshift mutation of MITF, resulting in simultaneous downregulation of TYR transcription and the thermolabile R402Q tyrosinase. ⁹ OCA2-related AROA ^{4 10} appears to be less frequent, but these patients likewise are compound heterozygous for one severe OCA2 allele and the other allele having a mutation with only modest negative effect on function. ¹¹  

All published studies to date have described molecular analyses of patients with AROA on a gene-by-gene basis, precluding assessment of either the relative frequencies of the different types of AROA or of different gene mutations. Furthermore, it is possible that in some patients AROA results from mutations of other genes associated with other types of albinism, such as OCA3, OCA4, and perhaps the various types of Hermansky-Pudlak syndrome (HPS) or other syndromic forms of OCA. ⁵ In this article, we describe comprehensive molecular genetic analysis of 36 unrelated Caucasian patients with the clinical diagnosis of AROA, yielding insight into the relative prevalence of different types of AROA and a repertoire of pathologic gene mutations in this group. 

Methods

Subjects and Diagnostic Criteria

A total of 36 unrelated Caucasian patients were studied. All were referred for molecular analysis to the investigator by their respective ophthalmologists, each carrying the presumptive clinical diagnosis of autosomal recessive ocular albinism. Six of the patients were also examined clinically by the investigator. Patients’ clinical data are presented in Supplementary Table S1. All patients had various ophthalmic features typical of AROA, including reduced visual acuity, iris transillumination, photophobia, nystagmus, strabismus, hypopigmented retinas, and abnormal fovea and macula. All patients had normal pigmentation of the skin and hair, though most, but not all, were relatively lightly complected in comparison with their families. Patients with mutations in the OA1 gene were excluded. This study conformed to the principles expressed in the Declaration of Helsinki and was approved as a no-consent study by the Combined Institutional Review Board (COMIRB) of the University of Colorado at Denver and Health Sciences Center on the grounds that only archived samples were used for the original purpose for which the samples were obtained. 

Molecular Genetic Analyses

DNA was prepared from peripheral blood leukocytes and quantified (ND-1000 Spectrophotometer; NanoDrop Technologies, Wilmington, DE). For patients with less than 30 ng/μL DNA, whole-genome amplification was performed (REPLI-g Midi Kit; Qiagen, Valencia, CA) and the products quantified (Quant-iT PicoGreen dsDNA Quantification Kit; Invitrogen, Carlsbad, CA). 

For each patient, amplicons containing each exon and adjacent flanking regions of the TYR, OCA2 (P), TYRP1, and SLC45A2 (MATP) genes, the 5′ promoter regions of TYR (1555 nucleotides [nt], excluding a simple sequence repeat from nt 88549828-88550258) and OCA2, and a conserved 647-bp segment located 8989 bp upstream from the TYR major mRNA 5′ terminus that may represent a locus control region, ¹² were amplified by touchdown PCR for DNA sequencing. For patients with no apparent pathologic mutations in any of these genes, amplicons containing each exon of the HPS1, HPS4, and SILV genes were then amplified for sequencing (in many patients DNA sequencing of multiple genes was performed in parallel). PCR primers are listed in Supplementary Table S2. PCRs were performed in 25-μL volumes containing 30 ng DNA, 5 pmols of each primer, 2.5 μL of 10× PCR buffer, 1.5 mM MgCl₂, 1.25 M betaine, 0.2 mM dNTPs (GeneAmp dNTP Blend; Applied Biosystems, Inc. [ABI], Foster City, CA), and 2.0 U Taq DNA polymerase (Platinum Taq; Invitrogen). For most amplicons, DNA was denatured at 94°C for 10 minutes followed by 15 cycles of denaturation at 94°C for 30 seconds, annealing from 63°C to 56°C for 45 seconds decreasing 0.5°C each cycle, and elongation at 72°C for 1 minute, followed by an additional 25 cycles of denaturation at 94°C for 30 seconds, annealing at 56°C for 45 seconds, and elongation at 72°C for 1 minute, followed by a final extension at 72°C for 10 minutes in a thermocycler (model 9600 or 9700; ABI). For TYR exon 4 the annealing range was further decreased to 54°C by adding four more cycles at the initial stage and by decreasing the annealing temperature of the following 25 cycles to 54°C. 

DNA Sequencing

PCR products were purified either by using a kit (QIAquick PCR Purification Kit; Qiagen) or by using a modified shrimp alkaline phosphatase (SAP)/exonuclease I method, in which for every 5 μL of PCR product, we added 2 μL SAP, mixed for 1 minute; added 1 μL of exonuclease I, mixed again for 1 minute; and incubated samples at 37°C for 15 minutes and then at 80°C for 15 minutes. 

DNA (100 ng) of each PCR product was sent to the University of Colorado Cancer Center DNA Sequencing and Analysis Core and sequenced (3730 DNA Analyzer; ABI). Analyses of DNA sequences were performed on computer (Sequencher; Gene Codes Corp, Ann Arbor, MI) software. Codon and nucleotide enumeration is referent to TYR transcript ENSG00000077498, OCA2 transcript ENST00000354638, TYPR1 transcript ENST00000381142 (the longest transcript, which includes all others), SLC45A2 (OCA4) transcript ENST00000382102, HPS1 transcript ENST00000359632, HPS4 transcript ENST00000336873, and SILV transcript ENST00000358822. Evolutionary conservation of variant amino acid residues was evaluated by alignment of orthologous human, chimpanzee (Pan troglodytes), macaque (Macaca mulatta), dog (Canis familiaris), mouse (Mus musculus), rat (Rattus norvegicus), and chicken (Gallus gallus) protein sequences, obtained from the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/ Bethesda, MD) or Ensembl (http://www.ensembl.org/). ¹³ Any variants listed as polymorphisms in dbSNP or other public variation databases and novel variants observed in patients with homozygous pathologic mutations in another gene were considered to be nonpathologic variants. Mutation nomenclature conforms to standard convention. ¹⁴  

Results

We studied 36 unrelated Caucasian patients referred with the clinical diagnosis of AROA. As shown in Supplementary Table S1, specific clinical details were often incomplete, and in some cases were lacking altogether. Nevertheless, findings of reduced visual acuity, nystagmus, strabismus, iris transillumination and photophobia, retinal hypopigmentation, macular or foveal abnormalities, and light complexion of the skin and hair were typical, though none were universal, even when specifically commented on. Although no specific genotype–phenotype correlation was evident, visual acuity tended to be better among older patients, and in several patients, progressive improvement of visual acuity and reduction of nystagmus was well documented over time, in some cases becoming apparently normal. Likewise, in many patients, apparent skin and hair pigmentation increased with age; several patients had had a diagnosis of possible OCA as newborns. 

To establish molecular diagnoses, we performed DNA sequence analysis of the four genes associated with classic OCA, TYR (OCA1), OCA2, TYRP1 (OCA3), and SLC45A2 (MATP; OCA4). In all patients in whom no pathologic mutations were found, we sequenced a candidate OCA gene, SILV, as well as HPS1 and HPS4, which are associated with the two most frequent types of Hermansky-Pudlak syndrome, which is the most frequent syndromic OCA disorder. Two other syndromic OCA disorders, Chediak-Higashi syndrome and Griscelli syndrome, are readily distinguished from other forms of albinism on clinical grounds; therefore, these corresponding genes were not sequenced. As shown in Table 1 , altogether 20 (56%) patients had pathologic mutations in TYR, 3 (8%) had pathologic mutations in OCA2, and 2 (6%) had novel variants in TYRP1 (OCA3), although these last were not definitively pathologic. Two (6%) patients had apparent pathologic mutations in both TYR and OCA2 and thus presented difficulty in interpretation. Nine (25%) patients had no apparent pathologic mutations in any of the genes studied. 

TYR (OCA1)

We sequenced all five TYR exons, the adjacent intron and flanking sequences, and 5′ promoter region ¹⁵ in all 36 patients. In addition, in all patients who carried the R402Q (c.1205G>A) variant previously associated with risk of AROA, ³ we also sequenced a conserved 647-bp DNA segment, located 9 kb upstream of the major TYR mRNA 5′ start site, that regulates transcription of TYR mRNA and may represent a locus control region, ¹² though we found no mutations in that segment. Overall, we identified apparent pathologic mutations in 20 (56%) patients. All of these patients were compound heterozygous for the R402Q variant on one allele and for various severe OCA1 mutations on the other allele. In all patients in whom family studies were performed, these mutations were found to be in trans. As shown in Table 1 , besides R402Q, we identified 13 different OCA1 mutations: one nonsense mutation (R402X [c.1204C>T]; patient 1), two frameshifts (47delC [c.47delC], patient 2; 1467insT [c.1467-1468insT], patient 3), one splice junction mutation (IVS2-1G>A [c.1037-1G>A]; patient 4), one splice consensus mutation (IVS2-7T>A [c.1037-7T>A), patients 5 and 6), one missense substitution involving the translation initiation codon (M1V [c.1A>G], patient 7), and seven other missense substitutions (P45T [c.133C>A], patient 8; C55Y [c.164G>A], patient 9; P81L [c.242C>T], patients 10 and 11; R217Q [c.650G>A], patient 12; T373K [c.1118C>A], patients 13 to 16; D383N [c.1147G>A], patients 17 to 19; and R422Q [c.1265G>A], patient 20]. All but one of these are known pathologic OCA1 mutations; only the 47delC frameshift is novel. 

OCA2

We sequenced 24 exons of the OCA2 gene (the first of which is noncoding), the adjacent intron and flanking sequences, and the 5′ promoter region ¹⁶ in all patients who lacked pathologic mutations of TYR. We did not sequence exon 19, an alternative exon that contains an in-frame terminator and thus does not encode functional OCA2 mRNA. ¹⁶ We found known pathologic OCA2 mutations in three patients. Patient 21 was compound heterozygous for two missense substitutions: V443I (c.1327G>A) and M446V (c.1336A>G). Two other patients were heterozygous for pathologic missense substations: in one case (patient 22) V443I and in the other (patient 23) N489D (c.1465A>G). In both of these patients no other apparent pathologic OCA2 mutations were found. The patterns of OCA2 polymorphisms indicated that they did not share a second OCA2 allele in common and furthermore indicated that neither was heterozygous for occult large gene deletions. Of interest, both of these patients were also heterozygous for the TYR R402Q variant. 

TYRP1 (OCA3)

We sequenced the eight exons of the TYRP1 gene and the adjacent intron and flanking sequences ¹⁷ in all patients who lacked pathologic mutations of TYR and OCA2. Two patients were found to be heterozygous for novel missense substitutions: T253M (c.758C>T; patient 24) and M451V (c.1351A>G; patient 25); in neither patient was a second mutation found. Neither of these substitutions has been identified as polymorphic SNPs, and both amino acid residues have been highly conserved through mammalian evolution; nevertheless, it is difficult to be sure these are pathologic mutations. In addition, the first patient was homozygous for the TYR R402Q variant and both had various polymorphic variants in the OCA2, SLC45A2 (MATP), and HPS4 genes. 

SLC45A2 (MATP; OCA4), HPS1, HPS4, and SILV

We sequenced the 7 exons of the SLC45A2 gene, ¹⁸ the 12 exons of the SILV gene, ¹⁹ the 20 exons of the HPS1 gene, ²⁰ the 13 exons of the HPS4 gene, ²¹ and the adjacent intron and flanking sequences in all patients who lacked pathologic mutations in TYR and OCA2. No patients had apparent pathologic abnormalities in any of these genes, although several had various known nonpathologic polymorphisms (Table 1) . 

Patients with Pathologic Mutations in Both TYR and OCA2

As shown in Table 1 , two patients had pathologic mutations in both TYR and OCA2, and thus present difficulties in interpretation. Patient 26 was compound heterozygous for a TYR frameshift (c.1467insT) and the common TYR R402Q substitution. This genotype offers a sufficient explanation of AROA; however, this patient was additionally heterozygous for the OCA2 V443I mutation. Patient 27 was homozygous for a novel missense substitution in OCA2, G27R (c.79G>A), confirmed by restriction enzyme cleavage analysis of the patient and her parents, but furthermore was heterozygous for the known TYR IVS2-7T>A mutation. OCA2 residue G27 has not been highly conserved through mammalian evolution, but the occurrence of this mutation among OCA patients with other OCA2 mutations (Hutton S, Spritz R, unpublished data, 2008) suggests that it is pathologic. 

Patients with No Apparent Pathologic Gene Mutations

Nine subjects (patients 28–36) had no apparent pathologic mutations in any of the genes studied. However, patient 29 was homozygous for all observed OCA2 polymorphisms, and patients 30 to 32 were homozygous for polymorphisms through long segments of the gene. This finding suggests that patients 29 to 32 may be hemizygous due to large deletions spanning all or part of OCA2. Nevertheless, no other OCA2 mutations were found in these patients, and so even this would not explain AROA. Six of these nine patients were female and so do not represent unrecognized cases of X-linked recessive OA1. 

Discussion

In general, AROA often appears to represent clinically mild variants of the classic OCA disorders. To date, AROA has only been associated with mutations in TYR (OCA1) and OCA2, but it seems likely that some patients with AROA may harbor mutations in TYRP1 (OCA3), SLC45A2 (OCA4), and perhaps other genes. Among the 36 patients with AROA studied, we identified clearly pathologic mutations in at least 25 (69%), including 20 (56%) with pathologic mutations in TYR, 3 (8%) with mutations in OCA2, and 2 (6%) with mutations in both TYR and OCA2. Two (5%) additional patients were heterozygous for novel variants of TYRP1, although it is not clear that these are pathologic. Nine (25%) patients had no apparent pathologic mutations in any of the genes analyzed and either harbor occult gene mutations or deletions not detected by our PCR-based DNA sequencing or perhaps have mutations in other genes yet to be discovered. 

Of the total 22 patients with mutations in TYR, 21 (95%) were compound heterozygous for various severe OCA1 mutant alleles and the R402Q variant that encodes a temperature-sensitive tyrosinase polypeptide with reduced catalytic activity. ⁷ These patients with TYR-related AROA thus express reduced overall tyrosinase catalytic activity, with most severe pigmentary deficiency in tissues near to core body temperature, such as the eye and optic neural tracts. 

Interpretation of variation in OCA2 is more problematic. The OCA2 gene is very large, >655 kb, and has several nonpathologic polymorphisms. One patient with OCA2-related AROA was compound heterozygous for two different pathologic gene mutations. Two other patients were heterozygous for known pathologic OCA2 mutations, and presumably carried occult mutations on their other alleles. Four additional patients, in whom no apparent mutations were detected in any gene, were homozygous for polymorphisms distributed across most or all of the OCA2 locus, suggesting that they might be hemizygous, with a large deletion of one allele and occult mutations of the other allele. 

Two patients had pathologic mutations of both TYR and OCA2 and thus presented unique interpretational difficulties. One was compound heterozygous for a TYR frameshift and the common TYR R402Q variant, sufficient to account for the AROA phenotype, but also was also heterozygous for a known severe OCA2 mutation. The second was homozygous for a novel OCA2 missense substitution that appears to be a common cause of OCA2 (Hutton S, Spritz R, unpublished data, 2008), and also was heterozygous for a known pathologic TYR mutation that is associated with mild OCA1. These may be merely chance concurrences or, alternatively, the heterozygous mutations may somehow exacerbate the phenotype resulting from the compound heterozygous or homozygous genotypes. 

Two patients were heterozygous for novel missense substitutions of TYRP1. Although the involved amino acid residues T253 and M451 have both been completely conserved across mammalian evolution, no other TYRP1 mutation was found in either patient, and it thus remains uncertain whether these TYRP1 variations are pathologic. 

Taken together, our results indicate that most patients with AROA represent phenotypically mild cases of OCA. In good agreement with our findings among Caucasian patients with typical OCA (Hutton S, Spritz R, unpublished data, 2008), well over half of patients with AROA have mutations in TYR. Almost all patients with TYR-related AROA are compound heterozygous for severe OCA1 mutations and the common R402Q variant, a genotypic combination that should occur in approximately 1 per 280 Caucasian individuals. However, the prevalence of AROA, while unknown, certainly is far lower than that, indicating that the penetrance of the AROA phenotype must be very low, given a susceptible genotype. About one tenth of patients with AROA have pathologic mutations in OCA2, although this may be an underestimate due to occult OCA2 gene deletions. A small percentage of patients with AROA have variants of TYRP1, although it is unclear whether these are truly pathologic. One fourth of patients with AROA have no apparent pathologic mutations in any of the genes we analyzed. Most of these patients are female and do not represent unrecognized cases of X-linked recessive OA1. Thus, it may be that some cases of AROA result from abnormalities in genes yet to be discovered. 

 Supported by Grant EY015626 from the National Institutes of Health.

 Submitted for publication June 28, 2007; revised October 24, 2007; accepted January 16, 2008.

 Disclosure: S.M. Hutton, None; R.A. Spritz, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Gene Mutations in 36 Caucasian Patients with AROA

Gene	Patient	Sex	Mutation 1	Mutation 2	Other Polymorphisms
TYR	1	M	c.1204C>T (R402X)	c.1205G>A (R402Q)
	2	F	c.47delC (47delC)	c.1205G>A (R402Q)
	3	M	c.1467-68insT (1467insT)	c.1205G>A (R402Q)
	4	M	c.1037-1G>A (IVS2-1G>A)	c.1205G>A (R402Q)
	5	F	c.1037-7T>A (IVS2-7T>A)	c.1205G>A (R402Q)
	6	M	c.1037-7T>A (IVS2-7T>A)	c.1205G>A (R402Q)	SLC45A2: R298C (H)
	7	F	c.1A>G (M1V)	c.1205G>A (R402Q)
	8	M	c.133C>A (P45T)	c.1205G>A (R402Q)	HPS4: E229G (h), L443V (H), V552M (h), H606Y (h), Q625H (h)
	9	M	c.164G>A (C55Y)	c.1205G>A (R402Q)
	10	F	c.242C>T (P81L)	c.1205G>A (R402Q)
	11	M	c.242C>T (P81L)	c.1205G>A (R402Q)
	12	M	c.650G>A (R217Q)	c.1205G>A (R402Q)
	13	M	c.1118C>A (T373K)	c.1205G>A (R402Q)
	14	F	c.1118C>A (T373K)	c.1205G>A (R402Q)
	15	F	c.1118C>A (T373K)	c.1205G>A (R402Q)
	16	F	c.1118C>A (T373K)	c.1205G>A (R402Q)	TYR: S192Y (h); OCA2: IVS5-19A>G (H), IVS21 + 18A>G (H): SLC45A2: L374F (h)
	17	M	c.1147G>A (D383N)	c.1205G>A (R402Q)
	18	M	c.1147G>A (D383N)	c.1205G>A (R402Q)
	19	F	c.1147G>A (D383N)	c.1205G>A (R402Q)
	20	M	c.1265G>A (R422Q)	c.1205G>A (R402Q)
OCA2	21	F	c.1327G>A (V443I)	c.1336A>G (M446V)
	22	M	c.1327G>A (V443I)	—	TYR: R402Q (H); OCA2: R305W (H); SLC45A2: L374F (h)
	23	F	c.1465A>G (N489D)	—	TYR: R402Q (H); OCA2: R305W (G); HPS4: E229G (h), V552M (h), H606Y (h), Q625H (h); SILV: G325V (H)
TYRP1	24	F	c.758C>T (T253M)	—	TYR: R402Q (h); OCA2: R419Q (H); SLC45A2: L374F (h); HPS4: E229G (H), L443V (H), V552M (H), H606Y (h), Q625H (h)
	25	F	c.1351A>G (M451V)	—	TYR: S192Y (H); OCA2: R419Q (H); SLC45A2: L374F (h); HPS4: E229G (h), L443V (h), V552M (H), H606Y (h), Q625H (h)
TYR and OCA2	26	M	TYR: c.1467-68insT (1467insT) OCA2: c.1327G>A (V443I)	TYR: c.12G>A (R402Q) —	TYR: S192Y (H); OCA2: R419Q (H); SLC45A2: L374F (h)
	27	F	TYR: c.1037-7T>A (IVS2-7T>A)	—	TYR: S192Y (H); OCA2: L440F (h)
			OCA2: c.79G>A (G27R)	OCA2: c.79G>A (G27R)
No mutations	28	F	—	—	TYR: S192Y (h); HPS4: E229G (H), V522M (H), H606Y (H), Q625H (H)
	29	M	—	—	TYR: S192Y (h); OCA2: R305W (h); SLC45A2: L374F (h); HPS4: E229G (h), V552M (h), H606Y (h), Q625H (H)
	30	F		—	SLC45A2: L374F (h); HPSI: R158H (H); HPS4: E229G (h), L443V (h), V552M (h), H606Y (h), Q625H (h)
	31	M	—	—	TYR: S192Y (h), R402Q (H); HPS4: E229G (h), L443V (H), V552M (h), H606Y (h), Q625H (h)
	32	F	—	—	TYR: S192Y (H), R402Q (H); OCA2: R305W (H); SLC45A2: L374F (h); HPS1: G283W (H); HPS4: E229G (h), L443V (h), V552M (h), H606Y (h), Q625H (h)
	33	F	—	—	HPS4: E229G (h), L443V (H), V552M (h), H606Y (h), Q625H (h)
	34	F	—	—	HPS4: E229G (h), V552M (h), H606Y (h), Q625H (h)
	35	F	—	—	TYR: S192Y (h); OCA2: R305W (H); SLC45A2: L374F (h); HPS4: E229G (h), L443V (h), V552M (h), H606Y (h), Q625H (h)
	36	M	—	—	TYR: S192Y (H), R402Q (H); OCA2: R305W (H), R419Q (H); SLC45A2: L374F (h); HPS4: E229G (h), V552M (h), H606Y (h), Q625H (h)

 H, heterozygous; h, homozygous.

Supplementary Materials

Supplementary Table S1 - (.doc) 

Supplementary Table S2 - (.doc)