Homozygosity mapping with SNP arrays identifies TRIM32, an E3 ubiquitin ligase, as a Bardet-Biedl syndrome gene (BBS11)

Abstract

The identification of mutations in genes that cause human diseases has largely been accomplished through the use of positional cloning, which relies on linkage mapping. In studies of rare diseases, the resolution of linkage mapping is limited by the number of available meioses and informative marker density. One recent advance is the development of high-density SNP microarrays for genotyping. The SNP arrays overcome low marker informativity by using a large number of markers to achieve greater coverage at finer resolution. We used SNP microarray genotyping for homozygosity mapping in a small consanguineous Israeli Bedouin family with autosomal recessive Bardet-Biedl syndrome (BBS; obesity, pigmentary retinopathy, polydactyly, hypogonadism, renal and cardiac abnormalities, and cognitive impairment) in which previous linkage studies using short tandem repeat polymorphisms failed to identify a disease locus. SNP genotyping revealed a homozygous candidate region. Mutation analysis in the region of homozygosity identified a conserved homozygous missense mutation in the TRIM32 gene, a gene coding for an E3 ubiquitin ligase. Functional analysis of this gene in zebrafish and expression correlation analyses among other BBS genes in an expression quantitative trait loci data set demonstrate that TRIM32 is a BBS gene. This study shows the value of high-density SNP genotyping for homozygosity mapping and the use of expression correlation data for evaluation of candidate genes and identifies the proteasome degradation pathway as a pathway involved in BBS.

Fig. 1.

BBS11 pedigree and shared haplotype. (A) Pedigree of BBS family. Filled symbols indicate affected individuals. (B) Segregation of STRP haplotype in parents (IV-1 and IV-2) and offspring. The disease haplotype is indicated by the boxed genotypes. Recombinant events observed in affected individuals V-5 and V-9 define the interval.

Fig. 2.

Representative TRIM32 sequence. (A) Normal proline homozygote at position 130 (CCT). (B) Heterozygous sequence. (C) Mutant serine homozygote (TCT).

Fig. 3.

Schematic diagram of TRIM32 (653 residues). N-terminal tripartite motif (zinc RING finger, zinc B-box, and coiled-coil domains) and five NHL repeats (solid boxes) are shown.

Fig. 4.

Representative KV phenotypes and summary of zebrafish trim32 knockdown. (A–D) Photographs of live zebrafish embryos at the 10- to 13-somite stage. (A) KV (dashed box) located in the posterior tailbud in a representative control-injected embryo. (B) Control KV (arrowhead). (C) trim32 MO-injected embryo with a reduced KV (arrowhead). (D) trim32 MO-injected embryo with no morphologically visible KV (arrowhead). (Magnifications: A, ×5; B–D, ×10.) (E) Percentage of zebrafish with altered KV (reduced or absent). MO refers to zebrafish trim32 antisense MO-injected embryos. In rescue experiments, WT, P130S, or D487N containing full-length trim32 mRNA was coinjected with the trim32 MO. Controls were injected with an MO containing mismatched bases to the trim32 sequence. Thirty-six percent of trim32 MO-injected embryos displayed KV defects, whereas only 2% of control-injected embryos exhibited KV defects (P < 0.0001). Both WT human TRIM32 (4%) and the D487N allele (11%) rescued the KV phenotype (not significantly different from controls); however, the P130S allele (30%) failed to rescue the KV phenotype (P < 0.0001 compared with controls).

Fig. 5.

Summary of the melanosome transport assay in 5-day zebrafish embryos injected with trim32 MO with and without mRNA rescue. Control MO- and trim32 MO-injected embryos were observed for melanosome transport response time after epinephrine treatment. Embryos treated with trim32 MO alone showed an average response time of 178 s compared with an average 94-s response time for embryos treated with the control MO (P < 0.0001). Both WT human TRIM32 (103 s) and the D487N allele mRNA (103 s) rescued the melanosome transport defect (not significantly different from controls). The P130S allele (158 s) failed to rescue the transport defect (P < 0.0001 compared with controls).