Search

A network of interactions between transcription factors and cis-regulatory sequences controls the expression of developmental genes. Changes in either the cis- or trans acting components of a developmental interaction can have consequences for both the output of the interaction and for the greater network of interactions. Thus, the evolution of regulation is considered to be a major force in the evolution of morphological diversity. To investigate the evolution of an interaction, this thesis has compared the Bicoid-dependent regulation of hunchback and orthodenticle expression in the Dipterans, Drosophila melanogaster, Musca domestica, Calliphora vicina, Lucilia sericata and Megaselia abdita. hb genes were isolated from Calliphora and Lucilia and these encode a number of domains that are conserved in hb from other species such as Drosophila, Musca and Tribolium. In contrast to the coding sequences, the hb promoters from Calliphora and Lucilia differ from each other and from the Drosophila and Musca hb promoters in terms of the number, sequence, orientation and spacing of Bed-binding sites that they contain. Analysis of intra-specific variation in the M. domestica hb gene demonstrated that both coding and non-coding sequences are subject to slippage generated turnover of simple motifs and that the extent of this turnover is dependent upon region specific constraints. This suggests that mechanisms of turnover are responsible for the different hb promoter configurations observed in the Dipterans. To investigate any functional consequences of the differences in both bed and hb between Drosophila, Musca and Megaselia, transcription assays were carried out using homogeneous and heterogeneous combinations of these two components in yeast. The results of these assays suggest that differences in Bed and the hb promoters between these species may have co-evolved to maintain the interaction. Therefore, to investigate how other Bed-regulated genes have evolved in Musca, the expression patterns and coding sequences of the otd gene were characterized in Musca and compared to those of Drosophila otd.

A characteristic of viruses in the family Bunvaviridae is that the viral genome consists of three RNA segments of single stranded RNA, designated large (L), middle (M) and small (S). The Bunvaviridae family is sub-divided into five genera, but in all members of the family the S RNA segment encodes the RNA- binding nucleocapsid (N) protein. This study is restricted to viruses in the Bunvavirus genus and concerns the expression strategy of S segment encoded proteins and the nature of the interaction of N protein with viral RNA. For viruses of the Bunvavirus genus the S segment also encodes a non-structural protein, termed NSs, in an overlapping reading frame downstream of the start of the N open reading frame (ORF). The NSs ORF initiates from a tandem AUG in most bunyaviruses, such as Bunyamwera the prototype virus. The Maguari (MAG) virus belongs to the same serogroup as Bunyamwera virus (BUN), with which it shares high S segment homology, but the MAG S segment in addition to encoding N and NSs proteins, possess a third ORF, potentially encoding a protein of Mr 9.3K, downstream of the N and NSs ORFs. The MAG S segment is 945 bases in length and most of the MAG S segment (bases 22-937) was available on a single cDNA clone, pMAG60. In order to investigate the expression strategy of the MAG S segment fragments of the cDNA, containing either the complete S segment coding region or parts of it, were subcloned downstream of a T7 RNA polymerase promoter sequence in the vector pTZlSU. A series of recombinant pTZ plasmids were used as templates to generate message-sense RNA in T7 RNA polymerase directed run-off transcription reactions. Aliquots of the RNA were translated in a cell free translation system (rabbit reticulocyte lysate and wheat germ extract) and the synthesised proteins, which were labelled with 35S methionine, were analysed by polyacrylamide electrophoresis. Translation of RNA derived from templates containing the entire MAGS coding region produced three major products with electrophoretic mobilities equivalent to the predicted Mr of the encoded products, N, NSs and ORF3 proteins. Translation of transcripts derived from plasmids containing NSs and 9.3K ORFs or 9.3K ORF only produced two major products or one major product repectively, corresponding to NSs and/or the 9.3K protein. These experiments indicated that the proposed sites of initiation of translation for N, NSs and 9.3K proteins were correct but in addition it demonstrated that all three encoded products were capable of being translated from a single mRNA species o The NSs tandem initiation codons for NSs in the MAG S cDNA were mutated either singly or together to leave either the first or second or no potential codons at the start of the NSs ORF. Translation of transcripts derived from these templates indicated that NSs translation was initiated at the tandem AUG and that both the first and the second AUG were equally capable of initiating NSs translation. In addition NSs translation was found to be unaffected by the loss of the upstream N initiation codon or coding sequences. However, synthesis of the 9.3K protein was reduced when the upstream N initiation codon was inactivated. Although the 9.3K protein product could be translated in vitro, its presence could not be demonstrated in radiolabelled MAG virus infected cells due to a high background of labelled host proteins. In an attempt to make a specific antiserum to this protein the 9.3K protein was expressed as a beta-galactosidase (beta-gal) fusion protein in bacterial cells. A protein band corresponding to the fusion protein was isolated from preparative polyacrylamide gels and used to inoculate a rabbit at 3-4 week intervals. Test bleeds were used in immunoprecipitation assays of appropriately programmed rabbit reticulocyte lysates, but the serum failed to immunoprecipitate the 9.3k protein. However, the serum was able to recognise the beta-gal carrier protein in a seperate immunoblot assay of beta-gal protein. It was concluded that the 9.3K protein was poorly immunogenic and no further attempt at raising an antiserum to this protein was made. The N protein interacts with each RNA genome segment to form internal nucleocapids within the Bunvaviridae particle. Each nucleocapsid consists of a single species of RNA genome segment (L, M or S), multiple copies of N protein and minor amounts of L protein. In virus infected cells N protein has high specificity for full length negative-sense genomic or positive-sense antigenomic RNA, but not viral mRNAs, which have different 5' and 3' ends, and host cell RNA. The nature of this specificity is unknown and was investigated by studying in vitro the RNA binding activity of recombinant expressed N protein with various S segment derived RNA species.

The interaction between the cytoskeletal proteins α-actinin and vinculin is considered to be fundamental to the link between actin microfilaments and the cytoplasmic face of integral membrane proteins that support cell adhesion. Evidence is provided here for an interaction between α-actinin and vinculin. Electron microscopy of negatively stained samples was used to directly visualise α-actinin/vinculin complexes. The micrographs showed that an α-actinin dimer could bind two vinculin molecules, one at each end of the α-actinin rod. Low speed equilibrium centrifugation demonstrated that the two proteins interacted with a Kd of approximately 1.6X10-5M. An interaction between α-actinin and vinculin was also demonstrated using a 125I-vinculin overlay technique. Appropriate concentrations of unlabelled vinculin or α-actinin inhibited the binding of 125I-vinculin to α-actinin electroblotted onto nitrocellulose. The vinculin binding site of α-actinin has been localised to a 37 amino acid sequence (residues 713 to 749). Thermolysin generated fragments of α-actinin and glutathione-S-transferase linked α-actinin fusion proteins were used to define the binding site. Their ability to bind vinculin was assayed using 125I-vinculin overlays, by their ability to compete in 125I-vinculin overlays, and in a solid phase binding assay. Determination of the three dimensional structure of α-actinin may prove useful in elucidating a number of α-actinin/protein interactions. As yet, intact α-actinin has not been crystallised. The actin binding domain of α-actinin (residues 1 to 268) was expressed in E. coli. A purification protocol that produces milligram quantities for crystallisation studies was devised. No crystals have as yet been produced. The repeats of α-actinin (residues 240 to 749) were expressed. However, a satisfactory purification protocol could not be devised. The possible homology between the actin binding domains of α-actinin and filamin was investigated. An immunological approach, in conjunction with peptide mapping, was used to identify the actin binding domain of filamin for N-terminal sequencing.