Novel mutations in BBS5 highlight the importance of this gene in non-Caucasian Bardet–Biedl syndrome patients

Tina Duelund Hjortshøj,* Karen Grønskov, Alisdair R. Philp, Darryl Y. Nishimura, Adebowale Adeyemo, Charles N. Rotimi, Val C. Sheffield, Thomas Rosenberg, and Karen Brøndum-Nielsen Kennedy Center, Medical Genetics Laboratory, Glostrup, Denmark Department of Paediatrics, Howard Hughes Medical Institute, University of Iowa, Iowa City, Iowa Department of Ophthalmology, Howard Hughes Medical Institute, University of Iowa, Iowa City, Iowa National Human Genome Center, Howard University, Washington, District of Columbia Kennedy Center, National Eye Clinic, Hellerup, Denmark

Bardet-Biedl syndrome (BBS) is a rare genetically heterogeneous disorder presenting with retinal dystrophy, postaxial polydactyly, obesity, renal malformations, learning disabilities, and male hypogenitalism [Beales et al., 1999]. To date, 12 genes have proven to be implicated in the disease, accounting for the mutational load in 70-90% of the patients [Blacque and Leroux, 2006;Stoetzel et al., 2007]. BBS is inherited as an autosomal recessive disease though in some instances, digenic triallelic inheritance has been suggested [Katsanis et al., 2001]. Although mutations in many different BBS genes have been described in Caucasians, BBS1 and BBS10 are the two major genes accounting for greater than 50% of the BBS patients [Mykytyn et al., 2002;Stoetzel et al., 2006a]. Mutational findings in the remaining known genes have predominately been reported in non-Caucasian patients [Ansley et al., 2003;Chiang et al., 2004;Stoetzel et al., 2006b]. BBS5 is a minor contributor to BBS as only 2% of families from various ethnic backgrounds harbor BBS5 mutations [Li et al., 2004].
We studied five BBS patients from two nonconsanguineous families residing in Denmark: A Somali family (Family 1) with five siblings of whom four were affected, and an affected boy from Sri Lanka (Family 2). The Sri Lankan patient was a single adopted child and no further information of his biological parents was available. The patients were identified from the files of the Retinitis Pigmentosa Registry at the National Eye Clinic, Hellerup, Denmark [Haim, 2002]. Diagnosis was based on an ERG-verified panretinal photoreceptor dystrophy in association with three or more systemic manifestations; that is, postaxial polydactyly, obesity, cognitive impairment, renal signs, and male hypogenitalism [Beales et al., 1999]. DNA from the patients was collected for mutation analysis. The control group consisted of 43 East Africans (Kenya) and 58 West Africans (Nigeria) for family 1 [Rotimi et al., 2001], and 54 Indian individuals for family 2. Appropriate informed consent was obtained from the patients and their families.
As part of a larger study screening of BBS1, BBS2, BBS4, MKKS, and BBS10 was done by denaturing high performance liquid chromatography (DHPLC, Varian, Inc., Palo Alto, CA) followed by DNA sequencing of aberrant products on an ABI3100 automated capillary sequencer using Big Dye Terminator v.3.1 (Applied Biosystems, Foster City, CA; authors' unpublished data, manuscript submitted). BBS5 mutational analysis was likewise performed by DHPLC. Primer sequences and PCR protocols for BBS5 are available upon request.
Genotyping of the two families was done separately. For Family 1 we used the Affymetrix Gen-eChip Mapping 10K 2.0 Array to search for regions of shared genotypes among the affected siblings. Sample processing for this part of the study was carried out at the Microarray Facility in Tübingen, Germany. For Family 2, SNP genotyping was performed on Affymetrix GeneChip Human Mapping 50K Hind 240 SNP microarrays (Affymetrix, Santa Clara, CA). We allowed SNPs that could not be scored to be included in the regions of interest. An average call rate >95% was obtained.
Screening of BBS1, BBS2, BBS4, MKKS, and BBS10 by DHPLC did not reveal causative mutations in the five patients. We therefore performed SNP genotyping in order to identify other BBS loci. Since we had no information of consanguinity in Family 1, we searched for regions of shared genotypes covering more than 5 Mb among the four affected sibs and different from the unaffected sib. We identified three major regions: Two regions at 6q15 (23 Mb) and 12q32 (34 Mb) did not contain any known BBS gene while a region spanning 11 Mb at 2q31 contained the BBS5 locus (see the online Table I at http:// www.interscience.wiley.com/jpages/1552-4825/ suppmat/index.html). Mutational analysis of BBS5 revealed a homozygous nucleotide change, c.214G > A (p.Gly72Ser) in exon 4, identified in all four affected siblings in Family 1 but not in the unaffected sib. Both parents were carriers. The mutation was absent in 202 ethnically matched control chromosomes. Sequence alignment showed the change to be localized within a conserved region (Fig. 1).
SNP genotyping of the patient in Family 2 lead to the identification of nine homozygous regions including two regions containing a known BBS gene. One locus spanning 19.6 Mb at 1p32 contained TRIM32 (BBS11) while the other locus spanning 16.6 Mb at 2p31 contained BBS5 (see the online Table I at http://www.interscience. wiley.com/jpages/1552-4825/suppmat/index.html). Sequence analysis of TRIM32 did not reveal any causative nucleotide changes in the patient. However, direct screening of the entire BBS5 gene identified a novel single base pair change, c.547A > G, in exon 7 in the homozygous state predicted to result in a non-conserved amino acid change, p.Thr183Ala (Fig. 1). The mutation was absent in 108 ethnically matched control chromosomes. Furthermore, it was not detected among 60 BBS patients primarily of Northern European origin (authors' unpublished data, manuscript submitted). No family members were available for testing.
We report here on two novel missense mutations in BBS5. Both mutations are localized within conserved regions of the gene, are present in the homozygous state in the patients, and are absent in control chromosomes. In silico analysis predicts the mutations to affect protein function and in Family 1 the mutation segregates with the phenotype. Most of the known mutations in BBS5 are localized within either of two putative domains, called DM16 (Fig. 2); as for the mutations reported here, p.Gly72Ser is localized in the first domain while p.Thr183Ala is localized in the second DM16 domain of BBS5. DM16 is a domain of unknown function and evolutionary conserved among many species [Li et al., 2004]. The fact that the mutations reported here are located in the DM16 domains supports their pathogenecity. We cannot rule out though, the existence of another mutation in linkage disequilibrium with these missense mutations. However, all coding regions plus 20 base pairs in the flanking regions were sequenced. Our findings represent the first missense mutations detected in the homozygous state in BBS5 patients. Thus, these BBS5 missense mutations might be useful for functional studies (Table II).
Only five other point mutations, two indels and one large deletion have been previously reported in BBS5 (Table II) [Li et al., 2004;Nishimura et al., 2005;Smaoui et al., 2006]. Most of these were detected in the homozygous state with the exception of the two previously reported missense variants (p.Asn184Ser and p.Arg207His). The previously published mutations are identified in patients from Africa and the Middle East, and in a single patient from New Foundland [Li et al., 2004;Smaoui et al., 2006].
In conclusion, we report two novel missense mutations in BBS5. Several pieces of information support that the mutations are pathogenic. Both patients are non-European. Screening of 60 patients from Northern Europe revealed no mutations in BBS5. These data might have implications for the mutational screening strategy of BBS5.