Detailed characterisation of the Opitz syndrome protein MIDl, its homologues and molecular interactions

Abstract

Opitz G/BBB syndrome (OS) is a genetically heterogeneous congenital disorder with both X-linked and autosomal forms. Although there are established diagnostic criteria, the phenotypic variability among patients can be considerable. Most commonly, OS patients present with craniofacial, laryngotracheoesophageal, cardiac, and genital anomalies. OS has been linked to Xp22.3 and 22q 11.2; and additional patients with unique chromosomal anomalies have allowed the tentative assignment of three other autosomal loci at 14ql 1.2, 13q32.3ter and 5p12-13. Yet despite differing genetic bases, the clinical presentation in either f01m is essentially indistinguishable, despite marked variability in phenotypic presentation even among related family members. Our laboratory has previously shown that approximately 50% of patients with a diagnosis of OS harbour mutations in the X-linked MIDI gene. MIDI encodes a 667 amino acid RING finger, B-box, coiled-coil (RBCC) microtubule binding protein that is expressed widely throughout embryonic development and congruent with tissues affected in OS. A highly related protein, MID2, is the closest homologue to MID 1 and has been found to similarly associate with the microtubule network either as homomultimers, or as heteromultimers with MID 1. In this thesis, several approaches were undertaken in order to better understand the sub-cellular and molecular aspects of OS. Firstly, a search for MIDlinteracting proteins was undertaken, and secondly an analysis of homologous proteins was completed with a view to providing insight based on the function of similar proteins. To identify protein partners of MID 1 that might be components of the large MID complexes, a yeast two-hybrid screen was performed. Using this approach, Alpha 4, a regulatory subunit of PP2-type phosphatases was identified as a strong interactor of MID 1 and additionally through separate testing, MID2. Cell localisation studies showed that both MID 1 and MID2 tethered Alpha 4 to the microtubules. Co-expression of domain-deleted and mutant forms of MID 1 with Alpha 4 localised the interaction domain of MID 1 to the B-boxes, which was confirmed by yeast two-hybrid assays. In addition to MID 1 and MID2 interacting with Alpha 4, it was also found that MID 1 and MID2 interacted with oneanother through their coiled-coil domains. This further implicated MID2 as a potential modifier of the OS phenotype. Therefore, in order to better understand the basic functional aspects of MID 1 and MID2, a detailed search for other homologues was canied out. The RBCC tripartite (TRIM) motif skewed protein BLAST searches for N-terminal matches for MID 1/MID2 homologues, so combinations of N-terminal and C-te1minal domains were also used to identify structurally related proteins. BLAST and PSIBLAST searches identified the homologous proteins, TRIM9 and TRIM36 as well as partial transcripts of two new proteins TRIFIC and TNL (IRIM.2.-like). Subsequently, full sequences were determined for both of these genes. This group of six related RBCC proteins were designated as the RBCC C-I subfamily. To further identify sequence identities within the C-1 subfamily and the rest of the RBCC protein superfamily, a relatively novel technique using HMMER was employed. This showed that the most highly scoring region across the C-I subfamily was not a known domain but a region between the coiled-coil and Fibronectin type-III domains. This new region, called the COS box, was shown to be restricted to a select group of RBCC and non-RBCC proteins, all which associate with microtubules. Mutation of paiiicular amino acids in the COS box completely abolished microtubule association, and addition of a MIDI C-terminal fragment containing the COS box to a non-microtubule associating RBCC protein directed the hybrid protein to microtubules. These findings completely change the previous understanding that the coiled-coil or C-terminus of MID 1 and other RBCC proteins were responsible for microtubule association. The work presented in this thesis shows that normal association of MID 1 to microtubules is dependent on the COS box. In addition, the B-box domain is necessary for tethering Alpha 4, a protein phosphatase 2A regulatory protein, to the microtubules. The homology of MID 1 and MID2, which extends functionally to homo and heterodimerisation, microtubule association, and interaction with Alpha 4, has led to the suggestion of potential heterogeneity m Opitz syndrome. In this light, Alpha 4 was subsequently shown to be mutated m a phenotypically similar syndrome, further supporting the importance of the molecular interaction between these proteins. Potential molecular pathways in which these proteins act are discussed, as are the effects of disruption of these pathways, and the impact such changes may have on the development of key tissues affected in the diseases.