Periplakin and envoplakin are components of the epithelial cornified cell envelope (CE). The CE is a crosslinked protein structure that forms beneath the plasma membrane of differentiating keratinocytes and is essential for epidermal barrier function. Periplakin and envoplakin belong to the plakin family of cytolinker proteins. Their N termini are thought to mediate their interaction with the plasma membrane. I found that the N terminus of envoplakin does not efficiently localise to desmosomes or microvilli, suggesting a dependence on periplakin for membrane localisation. I found that the first 133 amino acids of periplakin targets the protein to the apical plasma membrane. In a yeast two-hybrid screen, using this region as 'bait', a novel protein was identified and named kazrin. There are four alternatively spliced transcripts, encoding three proteins with different N termini. The interaction between periplakin and all three kazrin isoforms was confirmed by pull-down reactions. Kazrin was expressed in all layers of stratified squamous epithelia and was incorporated into the CEs of cultured keratinocytes. Kazrin localised to the cell periphery of differentiated cells, however it was more diffusely localised in cells of the basal layer. In stratified keratinocyte cultures kazrin partially colocalised with desmoplakin and periplakin at the desmosomes and periplakin at the interdesmosomal plasma membrane. In transient transfections, kazrin proteins localised predominantly to the apical plasma membrane of primary keratinocytes. The exogenous expression of kazrin proteins caused changes in keratinocyte cell shape and the actin cytoskeleton. I carried out in vivo analysis of the function of periplakin and envoplakin using mice lacking periplakin, envoplakin and involucrin expression. Triple knockout mice produced slightly larger CEs and had a slight delay in epidermal barrier formation relative to control mice. In future it would be interesting to establish the consequence of deleting the kazrin gene in mice.

Chromodomains are a superfamily of protein domains that are implicated in the modulation of chromatin structures. The chromodomain superfamily can be subdivided into two subfamilies, chromodomains and chromo shadow domains. Chromodomains function to localise proteins to specific sites on chromatin. An NMR structure of a chromodomain has been previously solved. Chromo shadow domains are protein-protein interaction domains. These recruit other chromatin associated proteins to their sites of action. The two domains have sequence similarities and are likely to have similar structures. The basis for their functional divergence is not known. In the present study, an X-ray structure of the chromo shadow domain of fission yeast Swi6 protein was solved to 1.9A. The structure reveals that the chromo shadow domain is a dimer with monomers closely resembling the chromodomain structure. Dimerisation of the chromo shadow domains creates a cleft that is commensurate with peptide binding. Binding studies were carried out to further investigate chromo shadow domain function. Recently, proteins of the <i>Drosophila </i>dosage compensation complex were found to contain divergent chromodomain-like regions. These domains have been noted to have RNA binding properties. Preliminary structural studies were carried out to determine if these domains shared a common fold with chromodomains and chromo shadow domains.

This thesis is concerned with the loading of transferrin with manganese and some of its chemical and biological properties. Manganese is bound to transferrin as Mn(III) with a characteristic ligand (Tyr) to metal (Mn(III) charge-transfer band at a wavelength of 430 nm. Caeruloplasmin is shown to enhance the uptake of manganese from MnC1₂ by apo-hTf. However binding is often incomplete and slow. A novel method of loading hTf with Mn using KMnO₄ is reported. This method leads to rapid uptake and inductively coupled plasma atomic emission spectroscopy (ICP-AES) determinants confirmed the binding of at least two Mn per hTf molecule. The possible oxidising effects of MnO₄^- on protein amino acid side chains was considered. In model systems MnO₄^- oxidises methionine to methionine sulfoxide and methionine sulfone. Evidence of structural changes in apo-hTf induced by Mn(III) binding was obtained by studies using [e-¹³C]Met-hTf. Preliminary work suggests that Mn(III), like several other metals studied, preferentially binds to the C-lobe first, although this may result in an open domain conformation. Fe(III) as Fe(NTA)₂ was found to displace Mn(III) from hhTf but displacement was slower when hTf had been loaded using KMnO₄ rather than MnCl₂. KMnO₄ was not able to displace Fe(III) from Fe₂-hTf. Attempts to crystallise Mn-hTf to characterise these structural changes proved difficult. Crystals grew but were of poor quality and did not diffract. Many large crystals were obtained from solutions of Fe₂-hTf. The crystals were red/orange and ellipsoidal in shape. Of the Fe₂-hTf crystals grown, one diffracted to 3.3 Å with the data being complete to 90%, but not enough information was gained for adequate molecular replacement and structural solution.

Specific enzyme-linked immunosorbent assays (ELISA) for rat mast cell proteinase I and II (RMCP I and II), intestinal mast cell proteinase (IMCP) and sheep mast cell proteinase (SMCP) were developed. Sheep serum or lymph contained potent inhibitory factors which interfered with the ELISA for SMCP, whereas little or no effect was demonstrated in the rodent ELISA by homologous serum. Secretion of SMCP into gastric lymph was noted in immune sheep following oral challenge with <i>Ostertagia circumcincta</i>. Development of immunity to <i>Haemonchus contortus</i> in sheep, expressed as the ability to exclude larvae from the mucosal surface, was characterised by an increase in mucosal mast cells (MMC) and increased abomasal tissue concentrations of SMCP. Increased concentrations of SMCP were demonstrated in serum and mucus in immune, but not in naive, sheep following direct abomasal challenge with 1 x 10<SUP>6</SUP> L<SUB>3</SUB> <i>Haemonchus</i> larvae. The abrogation of immune exclusion by treatment with corticosteroids was associated with a significant reduction in the number of MMC and concentrations of SMCP in abomasal tissue. Immune exclusion persisted for 6 weeks but had declined by 12 weeks following the cessation of larval challenge. This decline was associated with a significant reduction in the number of MMC and concentrations of SMCP in abomasal tissue. Ovine mast cells derived from <i>in vitro</i> culture of bone marrow cells (BMMC) were compared with MMC. BMMC contained similar constituents to MMC including SMCP, histamine, dopamine and arylsulphatase. BMMC contained an additional <SUP>3</SUP>[H]-DFP reactive protein not demonstrated in MMC. In the mouse, the major source of IMCP was the gastrointestinal tract. During infection with <i>Trichinella spiralis</i>, secretion of IMCP was demonstrated. Murine mast cells of the gastrointestinal tract exhibited heterogeneity in their proteinase phenotype, with many cells apparently containing a mixture of proteinases. Heterogeneity in the proteinase expression of mast cells in the rat was also demonstrated, with mast cells containing either RMCP I, RMCP II or RMCP I and II. Cells expressing dual proteinase phenotype were observed in some non-mucosal locations. During a primary infection of <i>Nippostrongylus brasiliensis</i>, changes in the RMCP I and II concentrations occurred in almost all tissues of the rat. These included significant increases in RMCP II concentrations in mesenteric lymph node, lung and intestinal tract 12 days after infection. Other changes, including those of RMCP I concentrations in bone marrow, are described.

No description available.

This thesis describes a novel DNA microarray method for determining the sequence specificity of single-stranded nucleic acid binding proteins (SNABPs). In <i>Salmonella typhimurium </i>six homologous CSPs (<i>St</i>CspA, <i>St</i>CspB, <i>St</i>CspC, <i>St</i>CspD, <i>St</i>CspE, <i>St</i>CspH) have been identified, although their functions are yet to be clearly elucidated. The novel microarray assay revealed two different types of ss DNA binding preferences for either purine or pyrimidine rich sequences. CspD bound purine (guanine) rich sequences, with the consensus binding sequence, 5’-ACGGgg-3’. CspA, CspB, CspC, and CspE bound pyrimidine (thymine) rich sequences, with an identical consensus core binding sequence. 5’-TCTTT-3’. The kinetics and thermodynamics of CSP/ss DNA interactions were examined for <i>St</i>CspE and <i>St</i>CspD, using the surface plasmon resonance method and isothermal titration calorimetry, which complemented the initial results determined by the novel microarray method. In addition, the X-ray crystal structure of <i>St</i>CspE was determined at 1.1 Å resolution and refined to R = 0.203. A computer generated model of <i>St</i>CspD was also created. The consensus ss DNA binding sequences for <i>St</i>CspE and <i>St</i>CspD (5’GTCTTTT-3’ and 5’-ACGGGG-3’, respectively), were docked onto the structures to reveal key molecular interactions, which accounted for the observed ss DNA sequence specificities. This work reveals key differences in selective ss DNA binding, existing within a small highly conserved family of CSPs, thus reflecting potential differences in function. Classification of SNABPs in this manner may provide a means of elucidating their cellular function and identifying gene networks regulated by specific SNABPs.

Recent developments of Metabolic Control Analysis include the derivation of a general method (the ‘universal method’) which aims to define the conditions under which a pre-determined change in the flux to a specific end-point metabolite may be achieved without changing other fluxes or metabolic concentrations. This thesis describes the application of this method in the yeast, <i>Saccharomyces cerevisiae</i>, by over-expressing genes of the shikimate and tryptophan pathways and, by analysing the outcomes, attempts an evaluation of the method’s validity. A single-copy plasmid, pPH28, bearing the 4 genes encoding enzymes of the shikimate pathway was constructed and introduced into a strain along with a multi-copy plasmid, pME554, bearing the 5 genes of the tryptophan pathway. In this way the activity of the shikimate enzymes and tryptophan enzymes were over-expressed by factors of about 2-5 and 20-60, respectively. The stability of the plasmids, both separately and together, under selective and non-selective conditions in different media was studied, along with enzyme activities. Results showed higher losses of plasmids when grown without selection, greater loss of the 2-micron plasmid than for the centromeric plasmid, and, in general, corresponding rates of loss of enzyme activities. The effects of the plasmids upon the production of tryptophan, phenylalanine and tyrosine were studied. pME554 and, subsequently, pPH28 had no observable, effect upon the total flux to tryptophan, phenylalanine and tyrosine. However, fluxes to tryptophan, phenylalanine and tyrosine pools were altered. pME554 increases intracellular concentrations of tryptophan and decreases those of phenylalanine and tyrosine, all by small amounts. When pPH28 is also included, intracellular tryptophan, phenylalanine and tyrosine concentrations increase by factors of 2-3. There are also corresponding changes in the amounts excreted into the medium.

Nuclear pre-mRNA splicing occurs within the spliceosome. The spliceosome is a dynamic structure in which multiple interactions among splicing factors catalyze the removal of introns from pre-mRNAmolecules. The <i>PRP8</i> gene of <i>Saccharomyces cerevisiae</i> encodes a 280 kD protein that is essential for splicing (Jackson <i>et al</i>., 1988). Previously, antibodies were raised against PRP8 protein, and an immunological approach was used to identify PRP8 protein as a stable component of both the U5 snRNP and the U4/U5/U6 multi-snRNP complex (Lossky <i>et al</i>., 1987). In this thesis, these immunological studies are extended, and the possible role of PRP8 protein as a spliceosomal pre-mRNA splicing factor is investigated. The association of PRP8 protein with yeast splicing complexes was demonstrated by the ability of PRP8-specific antibodies to precipitate precursor RNA, splicing reaction intermediates and lariat intron product from <i>in vitro</i> splicing reactions. PRP8 protein maintains its association with spliceosomal complexes throughout both steps of the splicing reaction and remains associated with the excised intron in a post-splicing complex that does not contain the spliced exons. The association of PRP8 protein with pre-mRNA molecules is splicing-dependent, for PRP8-specific antibodies did not precipitate mutant precursor RNAs nor did they precipitate wild-type pre-mRNA in the absence of ATP. A novel method for the immuno-affinity purification of yeast spliceosomes was developed. <i>In vitro</i> splicing reactions were assembled with polyadenylated pre-mRNA substrates. Poly(A)-binding protein (PAB), endogenous to yeast splicing extracts, associated with the poly(A) tail of the pre-mRNA, and spliceosomes were precipitated with antibodies raised against yeast PAB. Analysis of the affinity-selected spliceosomal proteins revealed the presence of PRP8 protein; thus is was established that PRP8 protein is a component of yeast spliceosomes. The possible function of PRP8 protein within spliceosomal complexes was explored by a combination of immunological and UV-crosslinking techniques. Short-wave UV light forms covalent crosslinks between adjacent protein and RNA molecules. PRP8-specific antibodies precipitated PRP8 protein-pre-mRNA adducts from UV-irradiated <i>in vitro</i> splicing reactions. This demonstrated that the two molecules were in direct contact at the time of exposure to UV light. The interaction between PRP8 protein and substrate RNA is a splicing-specific event. The protein-RNA interaction is dependent on the presence of ATP, and there was no crosslinking observed between PRP8 protein and RNA substrates that were not spliced, even if PRP8 protein assembled into complexes with these substrates. It was established that PRP8 protein would only interact with pre-mRNA molecules capable of participating in at least the first step of the splicing reaction, thereby implicating additional <i>cis</i>- or <i>trans</i>-acting factors in the PRP8 protein-pre-mRNA contact event. Finally, Ribonuclease T1 protection mapping studies were conducted to locate PRP8 protein's binding site(s) on pre-mRNA and pre-mRNA-derived molecules. Preliminary results are presented which suggest that PRP8 protein may interact with both the 5' splice site region of pre-mRNA molecules and the branch structure of the lariat intermediate and/or lariat intron. The implications of these results are explored.

This work presents the crystal structure of the C-terminal 10 kDa- sub-domain from <i>C. elegans</i> Hsp70. Despite a high degree of sequence identity, the <i>C. elegans</i> domain is shown to adopt a conformation distinct from the rat crystal structure, consistent with the more distantly related bacterial homologous. Comparison with the rat structure reveals an intriguing putative domain-swap dimerisation mechanism though the isolated <i>C. elegans</i> domain was found to exist exclusively as a monomer in solution. A previous study identified two TPR domain containing <i>C. elegans</i> putative proteins predicted to interact with Hsp90. These proteins were identified as the <i>C. elegans</i> homologues for small glutamine-rich TPR containing protein (SGT) and Hsp70/Hsp90 organising protein (HOP). These proteins have been successfully cloned, expressed and purified. SGT forms homo-dimers in solution. Its hydrodynamic dimensions in relation to its molecular weight suggest a protein with a low level of compactness and an extended conformation. SGT interacts with the C-terminal peptides from both Hsp70 and Hsp90 with equal affinities. Studies on <i>C. elegans</i> HOP suggested it might exist as a dimer in solution. In addition, a tight binding interaction was demonstrated with human and <i>C. elegans</i> Hsp90 homologues. A thorough search of the complete <i>C. elegans</i> proteome and genome was performed to identify the complete repertoire of TPR domain containing proteins likely to interact with Hsp70 or Hsp90. A profile HMM based search of the published <i>C. elegans</i> protein and DNA databases identified 12 proteins; nine of which are homologues of proteins known to interact with Hsp70 or Hsp90. The remaining three are uncharacterised putative proteins and represent targets for further study.

No description available.