The work in this thesis was originally directed towards a comparison of mitochondrial electron transport systems rendered artificially deficient in ubiquinone with those systems which did not possess ubiquinone, described by other workers. The results summarized in Chapter 4 however, show that contrary to these claims the tissues of lower vertebrates such as Elasmobranchs do in fact contain ubiquinone at concentrations comparable with those of other respiratory components. Furthermore, work on systems rendered deficient in ubiquinone by various solvent extraction techniques led to doubts as to the validity and usefulness of studies on such systems (Chapter 3). It was decided therefore to start work on another problem, the electron transport system of Acetobacter suboxydans, to form part of a detailed and systematic investigation of bacterial electron transport systems already being undertaken in this laboratory. This work is described in Part II.

No description available.

Three independent mutants induced with ethylmethane- sulphonate and selected for their inability to grow on acetate, were shown by enzymic assays in comparison with their wild-type parents and revertants to be devoid of isocitrate lyase. Genetic mapping by interrupted conjugation, recombination analysis and phage transduction showed each of these lesions to lie within the same gene locus situated at 78.3 min on the E. coli linkage map in the order met A, icl, pgi. Furthermore, the occurrence of Acetate recombinants from intragenic crosses between these mutants demonstrated their heteroallelism and located their sites within the icl locus in the order metA, icl-1, icl-2,icl-3. Mutants lacking a second anaplerotic enzyme, PEP-synthase were also obtained through E.M.S. mutagenesis but were selected for their inability to grow on lactate, pyruvate or alanine. Using the previously described mapping techniques the structural gene specifying this enzyme was shown to lie in close proximity to the aroD marker at 32.5 min on the E. coli linkage map. These latter mutants were then employed in the isolation of spontaneously occurring id regulatory mutants (iclR), which, by their constitutive production of isocitrate lyase, permitted the growth of these strains on lactate but neither on pyruvate nor acetate plus pyruvate. By interrupted mating and three-point transduction, the iclR lesions within two such independent mutants were mapped and shown to lie adjacent to the icl structural gene in the order metA, icl, iclR. Furthermore, "cis-trans" tests employing a merodiploid of this region demonstrated the transdominance of iclR~.' over iclR and showed this lesion to mark the site of an icl regulator locus. These results together with the location of the glyoxylate cycle malate synthase (masA) in the position masA, icl iclR (Vanderwinkel & de Vlieghere, I968) and the domonstration of the co-ordinate formation (Kornberg, I96I, I966) and de-repression (Vanderwinkel, Liard, Ramos & Wiame, I963) of these enzymes provide both structural and biochemical evidence for the occurrence of a glyoxylate cycle operon.

The following D. discoideum glycosidases were studied: beta-hexosaminidase (EC 3.2.1.30 and EC 3.2.1.17), alpha-mannosidase (EC 3.2.1.24), beta-glucosidase (EC 3.2.1.21), beta-galactosidase (EC 3.2.1.23) and alpha-glucosidase (EC 3.2.1.20) or glucoamylase (EC 3.2.1.3). beta-hexosaminidase and beta-glucosidase were purified to greater than 90% homogeneity, alpha-mannosidase was purified to two protein components both with enzymic activity; whereas alpha-glucosidase and beta-galactosidase could not be separated and the purified preparation contained several components. The biochemical and biophysical properties of the purified enzymes are described. Antibodies against these enzymes were prepared and the immunological properties of purified and crude samples of glycosidases from various stages of the slime mould life-cycle were examined. These antibody preparations were used as tools to isolate isotopically labelled enzymes from crude extracts in experiments which show that beta-hexosaminidase and glucosidase are synthesized de novo and are stable during growth and the first few hours of development. From these studies and others it is concluded that the rates of accumulation of assayable enzyme activity of beta-hexosaminidase and glucosidase reflects precisely the rates of de novo enzyme synthesis; the effect of enzyme degradation, inhibition or activation being negligible. It is concluded that the marked Increases in the levels of beta-hexosaminidase, glucosidase and galactosidase activities per cell during the life-cycle of D. discoideum are determined by changes in the rate of general protein synthesis and degradation and changes in the rate of cell division rather than changes in the rate of enzmye synthesis. For alpha-mannosidase changes in the rate of enzyme synthesis may also be a factor. Studies on the physiological role of the glycosidases indicate that they have an intracellular digestive function during various stages of the life-cycle; during growth some of the enzymes act on the exogenous food source within food vacuoles, and during development some of the enzymes act on endogenous material, probably' within autophagic vacuoles.

The aim of this work was to develop a more efficient and usable basic form of cloning vehicle for genetic engineering than those currently in use. Subsequent work would then enable a new generation of cloning vectors to be constructed. The characteristics that such a vehicle would require were determined by considering the basic processes involved in DNA cloning and the kinds of facilities that vector replicas can provide. Against this background the development of currently used phage and plasmid vectors was considered and the properties they offer examined. Both growth forms have different kinds of advantages and disadvantages in use. It is postulated that a plasmid/phage growth cycle of the vector replicon would harness the advantages of both growth phases and enable their disadvantages to be avoided. Consequently, a plasmid-phase growth cycle of bacteriophage lambda has been developed to fulfil this requirement. Immediately after infection of a bacterial cell,the phage lambda genome exists transiently as a plasmid replicon before growth diverges to the lysogenic or lytic pathways. A potential therefore exists for the development of an alternative plasmid growth pathway. Evidence had suggested that an N mutation arrests development at this stage (Signer, 1969;Lieb, 1970;Kleckner and Signer, 1977). However,this investigation found that such plasmid formation is abortive. A series of further experimental investigations is reported that have enabled the elucidation of the genetic elements controlling the plasmid pathway. The plasmid mode of growth of lambda genomes is subject to a novel kind of fine genetical control, quite different to that of the lytic or lysogenic modes. Furthermore, lambda genomes growing as plasmids can be recovered as phages, if subjected to a precise molecular switch:-this facilitates the PHASMID growth cycle. Application of lambda phasmids for DNA cloning is considered. Precise genetical control and characterization,coupled with possible alternate growth phases, make lambda phasmids potentially extremely useful for the propagation and manipulation of cloned genes.

The work presented in this thesis is an investigation of factors involved in 3' splice site selection. To try to determine the intrinsic strengths of 3' splice site sequences, a cis-competition assay system was used. This system was based on the large intervening sequence (IVS-2) of the rabbit beta-globin gene. Synthetic 3' splice site sequence oligonucleotides were inserted into the EcoRl restriction site, forty nine nucleotides downstream of the authentic rabbit beta-globin IVS-2 3' splice site. The oligonucleotides conformed to the established 3' splice site consensus sequence and allowed for variations within this sequence. The authentic site served as a constant reference site against which the strengths of the synthetic sites could be measured. When spliced in HeLa cells in vivo, all constructs tested were seen to choose the authentic 3' splice site over the synthetic 3' splice site under test. A series of mutageneses was carried out to try to decrease the intrinsic strength of the authentic site and/or improve the environment of the synthetic site such that the overall strengths of the two sites might be balanced. An AG→CG mutation at the authentic 3' splice site caused the synthetic 3' splice site to be activated as a cryptic site in vivo and in vitro. In this case lariat formation was mapped to an artificially created branch point within exon 3. Splicing component binding to both 3' splice sites was investigated by looking at protection of the RNA from oligonucleotide directed cleavage by RNase H. Initial protection of both 3' splice sites was independent of the final choice of site. However, branch point protection was dependent on the 3' splice site chosen. Components bound to the authentic 3' splice site could be immunoprecipitated whether that site was chosen or not. The synthetic 3' splice site was poorly precipitated even when it was chosen. This data tends to suggest that the synthetic 3' splice site directs inefficient complex assembly, and that at least partial complex assembly occurs at a 3' splice site which has been inactivated by an AG→CG mutation. Preliminary work was carried out to develop a method for the analysis of splicing component binding to either or both 3' splice sites of material within fully and partially assembled splicing complexes (spliceosomes) isolated by sucrose gradient sedimentation.

A quantitative RT-PCR method for analysis of alternative isoforms of RNA, and a method for the transient expression of mini-genes in differentiated muscle cells have been developed. This has enabled the analysis of endogenous RNA transcribed from rodent N-CAM and from human tropomyosin mini-genes which were expressed in COS cells and mouse C-2 myoblasts and myotubes. The sequences of the previously unreported mouse homologues of human exons MSD1b and MSD1c of N-CAM have been determined, and deposited in the EMBL data base. The tissue- and stage-specific alternative splicing patterns of exons within the muscle-specific domain (MSD) of N-CAM have also been established; the exons were normally incorporated as a unit in muscle cells, but were not included in transcripts derived from non-muscle myoblasts and neural cells. The triplet AAG exon was also included in a stage- and tissue-specific manner, but independently of inclusion of other exons of the MSD. Transfection of C-2 myoblasts with mutant mini-gene constructs of human tropomyosin determined the cis-acting elements which regulate the mutually exclusive alternative splicing of the central exons (NM and SK) in both non-muscle and muscle cells. In non-muscle, these were found to be due either to cis-acting repressor sequences within the SK exon or cis-acting activator sequences within the NM exon. In differentiated cells, exclusion of the NM exon is not via cis-acting repressor sequences within the NM exon, but because the upstream (NM) exon site is dormant and is therefore skipped by the splicing machinery. The evolution of alternative pre-mRNA splicing has also been discussed, and on this basis and from analysis of the data presented here, I conclude that regulation of alternative pre-mRNA splicing of transcripts from different genes may be founded upon a common mechanism which is largely dependent upon the presence of sub- optimal splice-signals and the potential for variation in the relative concentrations of certain splicing factors.

A Rhizobium genomic library was constructed in Escherichia coli NM522 and the library was screened for the ability to grow on the halogenated compound 2-chloropropionate (2CP). Two positive clones were identified and the plasmids designated pSC1 and pSC530. Both clones allowed E. coli to grow on 2CP at growth rates slower than that seen for the Rhizobium. Investigation of the dehalogenase activities of the clones showed that extracts from cells containing pSC1 were active against D-, L- and D/L-2CP and extracts from cells containing pSC530 were also active against dichloropropionate. Restriction enzyme mapping of the clones indicated that they were unique regions of the genome that conferred the new growth ability on E. coli. Southern analysis of the Rhizobium genomic DNA using insert DNA as probes confirmed that both of the clones were from the Rhizobium and that the clones were not contiguous on the genome. Subcloning of pSC1 resulted showed that the insert encoded two stereospecific dehalogenase genes and two plasmids were constructed by subcloning that were each able to express each of the two dehalogenases. The stereospecific dehalogenases (HadD and HadL) were purified and the N-terminal amino acid sequences were determined. HadL was also purified from the Rhizobium and the N-terminal sequence was shown to be the same as the cloned HadL. The complete nucleotide sequences of the hadD and hadL genes were determined and showed little similarity to each other or to other known dehalogenases. HadD and HadL were used, both in vivo and in vitro, to resolve a racemic solution of 2CP, an analagous situation to possible industrial applications of dehalogenases.

The role of vinculin in cell adhesion was investigated by isolating NIH3T3 and Balb/c 3T3 cells containing a plasmid vector expressing a vinculin antisense RNA under the control of the MMTV promoter. Stable cell lines of NIH3T3 cells were cultured with dexamethasone but none showed any reduction in vinculin protein levels. Two Balb/c 3T3 cell lines, when cultured with dexamethasone displayed a marked reduction in vinculin levels and displayed an altered cell shape from a spread to a more spindle-shaped morphology. These phenotypic changes were reversible upon removal of the dexamethasone. The reduced adhesion corresponded with a reduction in the growth rate of the clones. The cells were also unable to spread on fibronectin and the ususal increase in the phosphotyrosine content of two signalling proteins, paxillin and pp125FAK, normally associated with cell adhesion, was not observed. The results establish that vinculin is a key component of integrin-mediated cell adhesion. To explore the structure-functional relationship of talin, six anti-talin monoclonal antibodies were microinjected into human fibroblasts and the effect upon the actin cytoskeleton was assessed. Two of the antibodies, TA205 and TD77 resulted in the disintegration of actin stress fibres, and migration of CEF was also severely impaired following microinjection of either antibody. The epitopes recognised by TA205 and TD77 had been mapped to talin residues 102-497 and 2269-2541 respectively. Microinjection of CEF with a polypeptide containing residues 102-497 demonstrated that they were mainly associated with the detergent-soluble cytoplasmic portion of the cell and were also able to disrupt stress fibre and focal adhesion integrity. Residues 2269-2541 were associated with the detergent-insoluble cytoskeletal fraction of the cell but did not affect stress fibre or focal adhesion integrity. Attempts to identify proteins that interacted with residues 102-497 of talin were unsuccessful but an actin-binding site was identified within residues 2269-2541. The results indicate the presence of domains within the N- and C-terminus of talin that are essential to the involvement of talin in the formation of focal adhesions.

Recombinant serine acetyltransferase was overexpressed in Escherichia coli and purified 13-fold to homogeneity. The specific activity of the purified enzyme was 719 units/mg. Hydrodynamic and quasi-elastic light scattering studies indicated that SAT has an open hexameric quaternary structure, furthermore, chemical cross-linking studies indicated that the fundamental building block is a trimer. Steady-state and stopped-flow kinetic studies indicated that the forward reaction proceeds via a ternary complex, and dead-end inhibition analyses indicated that the order of substrate addition was random. Substrate inhibition by serine indicated that the breakdown of the SAT-CoA complex is partially rate-determining, whereas the linearity of primary double-reciprocal plots and inhibition replots suggested that the interconversion of ternary complexes is not significantly faster than kcat. Hence the kinetic data is consistent with a steady-state random-order reaction mechanism in which the interconversion of ternary complexes and the dissociation of CoA may be both partially rate-determining. Steady-state kinetic studies with the alternative acyl acceptor threonine were consistent with such a mechanism, whereas, those with propionyl CoA satisfied the requirements of a rapid-equilibrium random-order reaction mechanism. Steady-state kinetic studies of the reverse reaction were consistent with a steady-state random-order reaction mechanism for SAT in which the breakdown of the enzyme-serine complex is partially rate-determining. Calorimetric titration data indicated that cysteine does not bind at the coenzyme binding site and may not bind at the serine binding site; hence it may bind at an allosteric site. The pH dependence of kcat/Km suggested that a base, with a pK in the region of 6, might have a role in catalysis. Chemical modification of all three cysteines in SAT by [14C]- iodoacetamide demonstrated that the enzyme exists in the fully reduced state. Furthermore, SAT was irreversibly inhibited by iodoacetamide and substrates conferred protection against such inhibition. Silica thin layer eletrophoresis demonstrated the labelling of both histidine and cysteine in native and denatured preparations of [14C]-iodoacetamide-modified SAT. Peptide sequencing was employed to further characterise the residues labelled in the native state.