Search

Porphobilinogen deaminase has been purified and crystallized from an overproducing recombinant strain of Escherichia coli harbouring a hemC containing plasmid which has permitted the isolation of milligram quantities of the enzyme. Determination of the molecular weight of the enzyme by SDS gel electrophoresis (35,000) and gel filtration (32,500) agrees with the gene derived molecular weight of 33,857. The enzyme has a of 19fJ.M, an isoelectric point of 4.5 and an N terminal sequence NH2- MLDNVLRIAT. The enzyme is shown to contain a novel dipyrromethane cofactor which is linked covalently to the protein. The structure of the cofactor is proposed on the basis of its reaction with Ehrlich's reagent and from its chemical properties. Specific labelling of the dipyrromethane cofactor by growth of the E. coli in the presence of 5-amino[5- 14Cjlaevulinic acid demonstrated that the cofactor is not subject to catalytic turnover. The structure of the cofactor was further confirmed as a dipyrromethane made up of two linked pyrrole rings by [13C] n.m.r. studies after the deaminase was specifically labelled with [13C] by growth of the bacteria on 5-amino[5-13C]laevulinic acid. The chemical shift data indicate that one of the pyrrole rings of the cofactor is covalently attached to the deaminase enzyme through a cysteine residue. Evidence from protein chemistry studies confirm that cysteine-242 is the covalent binding site for the cofactor. The formation of the dipyrromethane cofactor of E. coli was shown to depend on the presence of 5-aminolaevulinic acid. A hemA" mutant formed inactive porphobilinogen deaminase when grown in the absence of 5-aminolaevulinic acid since this strain was unable to biosynthesise the dipyrromethane cofactor. The mutant formed normal levels of deaminase, however, when grown in the presence of 5 aminolaevulinic acid. Porphobilinogen, the substrate, interacts with the free a-position of the dipyrromethane cofactor to give three stable enzyme-intermediate complexes, ES, ES2 and ES3. Experiments with regiospecifically labelled intermediate complexes have shown that, in the absence of further substrate molecules, the complexes are interconvertible by the exchange of the terminal pyrrole ring of the complex. The formation of enzyme-intermediate complexes is accompanied by the exposure of a cysteine residue suggesting that substantial conformational changes occur on binding substrate. Experiments with the a-substituted substrate analogue, abromoporhobilinogen, have provided further evidence that the cofactor is responsible for the covalent binding of the substrate at the catalytic site. Based on these findings it has been possible, for the first time, to construct a mechanistic scheme for the deaminase reaction involving a single active site which is able to catalyse the addition or removal of either NH3 or H2O. The role of the cofactor as both a primer and as a means forregulating the number of substrates bound in each catalytic cycle is discussed.

Membrane-bound structures such as intracellular compartments and extracellular vesicles appear to play important roles in bacterial biochemistry and survival. Whilst bacterial extracellular vesicles (BEVs) are naturally secreted, intracellular vesicle formation in bacteria has only been observed following the overproduction of some membrane proteins. The proteins of the LemA family have previously been shown to have membrane restructuring capabilities when recombinantly expressed in Escherichia coli. Consequently, this project set out to investigate the potential of utilising LemA proteins together with protein engineering approaches to produce intracellular and extracellular vesicles in E. coli. Initial studies focussed on two distinct LemA proteins from Pseudomonas aeruginosa (Pa) LemA1 and LemA2, which were predicted to be targeted to the inner and the outer membranes, respectively. Transmission electron microscopy (TEM) analysis revealed the formation of intracellular membrane vesicles in both samples, while scanning electron microscopy (SEM) studies showed increased outer membrane vesiculation in the cells overexpressing PalemA2; findings that were consistent with the predicted localisation of these proteins. Moreover, the purification and structural studies of PaLemA1 provided a solid basis for future structural work on this protein. The construction and overproduction of the transmembrane and soluble domains of MamQ, LemA.153, LemA.159, LemA.501 and PaLemA1 proteins individually resulted in inclusion body formation in the majority of samples. However, the expression of the transmembrane domain (TMD) of PalemA1 alone was sufficient to induce intracellular vesicle production. A 'mix-and-match' approach of the different TMDs and soluble domains further yielded a wide range of novel membranous phenotypes in E. coli. Together, these results provide good evidence for the recombinant production of membranous compartments following LemA protein overproduction in E. coli. Furthermore, such membranous structures hold a lot of potential for targeted protein engineering approaches in the generation of a novel vaccine delivery platform.

Approximately half of all algal species are dependent on an external supply of the corrinoid cobalamin (vitamin B12) for growth. Algal phyla differ in the proportion of species dependent on B12. Extremes of this are the Chlorophyta (30% B12 dependent) and the Dinophyta (91% B12 dependent). B12 is only synthesised by prokaryotic organisms, therefore, algae must rely on an external source of this nutrient. Little is known about how B12 is taken up in algae at the molecular level. One protein that has been characterised is Cobalamin Acquisition Protein 1 (CBA1) - identified in the diatoms Phaeodactylum tricornutum and Thalassiosira pseudonana - which was found to increase the B12 uptake rate when overexpressed. However, no proteins required for B12 uptake have been identified in green algae. In this study, the model green alga Chlamydomonas reinhardtii was used to investigate B12 uptake. Candidate genes related to B12 transport and binding were identified in C. reinhardtii from sequence similarity to genes known to bind or transport B12 in other organisms. Six genes were tested in vivo using knockdown and knockout strains. Lines with disruptions in Cre02.g081050, which encodes the C. reinhardtii CBA1 homologue (CrCBA1), were unable to take up B12, and complementation lines restored the ability to take up B12. To assess the subcellular location of CrCBA1, a line was made with a fluorescent tag at the C-terminus of CrCBA1. Confocal microscopy was used to image this line and showed its likely membrane association. To exert control over B12 uptake, a thiamine-responsive riboswitch was used to control CrCBA1 expression. In the absence of thiamine, the line could take up B12 to a similar extent as the wild type strain, but addition of thiamine impaired the ability of this line to take up B12. The presence of CBA1 was assessed in algae more generally, and sequences similar to CBA1 were identified in many algal species belonging to different phyla. This indicated that CBA1 is much more prevalent in nature than previously thought. The diversity of CBA1 sequences allowed highly conserved residues to be identified, which provides a promising route for future investigation. A forward genetic screen was undertaken to search for other proteins important in B12 uptake, using insertional mutagenesis. A reporter line was made with a paromomycin-resistance marker under control of a B12-repressible promoter, so that it showed repression of growth in paromomycin under high media vitamin B12 concentrations (+P+B12). As this line was thought to respond to intracellular B12 only, the screen was expected to detect mutants deficient in taking up B12 by selecting for those that grew in +P+B12. Insertional mutagenesis of the reporter strain was performed, and 8 insertional mutant lines were identified. One of these lines (1.G2) was found to be unable to take up B12. This was confirmed by incubating a fluorescent B12 analogue (B12-BODIPY) with algal cultures: 1.G2 did not show a B12-BODIPY signal that co-localised with the cell, whereas in lines able to take up B12, a B12-BODIPY signal with multiple circular foci within the cell was observed. The site of the transgene insertion was mapped using DNA sequencing to Cre12.g508644, however, knockout strains of this locus were able to take up B12, indicating that this gene was not responsible for B12 uptake. CRISPR/Cpf1 lines targeting Cre12.g508644 were made, however, no lines with successful genome editing of Cre12.g508644 were identified. Higher coverage DNA sequencing of 1.G2 was performed, which revealed that in addition to Cre12.g508644, there was also disruption of CrCBA1, by a putative transposon that was previously uncharacterised in C. reinhardtii. B12 transport was restored by transformation with CrCBA1, demonstrating the causality of this insertion. To identify genes that are differentially expressed in the presence of B12, RNA sequencing was performed. Seven genes were identified, 5 of which had not been associated with B12 before. RNA sequencing of 1.G2 found it had extensive transcriptomic differences compared to the background line. This provided new candidate genes to assess for association with CrCBA1 and B12 transport that could be investigated in future studies. Collectively, this work has contributed to a greater understanding of B12 acquisition in C. reinhardtii and algae more generally, and has provided algal lines, insight and techniques for future research in this area.

Bacterial microcompartments (BMCs) are proteinaceous organelles that encapsulate specific metabolic pathways to enhance key catalytic processes and protect the cell from toxic intermediates. Carboxysomes and the 1,2-propanediol utilising and ethanolamine utilising metabolosomes have been studied extensively. However, bioinformatic analyses have revealed a host of BMCs that are yet to be experimentally investigated. This study examined the production by E. coli of previously uncharacterised shell proteins from two organisms: Rhodospirillum rubrum and Clostridium autoethanogenum. The flexible nature of microcompartment formation was revealed by the different structures formed by the R. rubrum shell proteins both in vivo in E. coli and following their isolation. In vivo, shell proteins formed swirled sheets that were unable to form compartments. However, upon isolation the components of the protein structures were able to re-assemble to form apparently closed "empty" compartments. This highlighted the complex nature of the protein interactions involved in microcompartment assembly and the effect of a changing environment upon these interactions. The production of C. autoethanogenum shell proteins in E. coli revealed a previously unobserved interaction of protein sheets with ribosomes within the cytoplasm. This phenotype was shown to involve a C-terminally extended hexameric shell protein, Caethg_3286. The function of C-terminally extended shell proteins, which are encoded in many BMC operons, is unknown at present although the structure of one, EutK, was resolved and shows a high degree of similarity to nucleic acid binding domains. The crystal structure of the BMC domain of Caethg_3286 revealed that the C-terminus is on the concave surface of the hexamer allowing the C-terminal extension to extend into the cytoplasm and interact with ribosomes. The ribosomal interaction revealed in this study and the potential nucleic acid binding capacity of another C-terminal extension may indicate a role of the C-terminal extension in the regulation of BMC formation at the level of translation.

Cobalamin, more commonly known as vitamin B12, is a highly complex molecule whose biosynthesis in bacteria and archaea involves around 30 enzyme-mediated steps. The biosynthesis of vitamin B12 occurs by one of two distinct pathways, which are referred to as the aerobic and the anaerobic routes. The work described in this thesis relates to the aerobic pathway, and is mainly focussed on two enzymes responsible for the amidation of the corrin ring component of cobalamin, CobB which is the enzyme responsible for the amidation of HBA to its product HBAD and CobQ which amidates cobyrinic acid a, c-diamide to cobyrinic acid. These enzymes amidate the peripheral carboxylic acid sidechains of the corrin moiety, and produce hydrogenobyrinic acid (HBA) and adenosylcobyrinic acid a, c-diamide, respectively. CobB, amidates the a and c side chains attached to C2 and C7 of the corrin ring, of HBA. The product of this multi-functional enzyme is initially hydrogenobyrinic acid c- monoamide (HBAM) followed by hydrogenobyrinic acid a, c-diamide (HBAD). The gene encoding CobB was cloned from Brucella melitensis, overexpressed and its encoded protein purified. After crystallisation of the B. melitensis CobB, the first X- ray crystal structure of this enzyme was solved using selenomethionine labelled protein to 1.6 Å resolution. Further high resolution structures were obtained containing substrate and co-factors. To complement the structural information, kinetic data was acquired for the binding of CobB with its substrates, using stopped flow analysis. This kinetic approach helped identify a two-phase binding process, indicating that the initial HBA molecule binds quickly to the dimeric enzyme, whilst the second HBA molecule binds at a much slower rate. Collectively, the binding data suggest that CobB undergoes a conformational change upon binding the first HBA substrate that alters the binding affinity of the second HBA-binding site. 3 CobH, the enzyme prior to CobB in the cobalamin pathway, is known to bind its product (HBA) very tightly, which poses the question as to how CobB interacts with CobH in order to access its substrate. To investigate if CobB and CobH interact a model was constructed to determine if the two enzymes could form a complex that facilitates direct metabolite channelling. Evidence for direct metabolite transfer was obtained from fluorescence spectroscopy, where fluorescence emission was used to monitor changes associated with substrate transfer of HBA from CobH to CobB Co- crystallisation trials to obtain a complexed crystal structure of CobH/HBA/CobB was attempted, but this approach resulted only in the generation of CobH crystals. CobQ, the second enzyme in the amidation process, was also produced recombinantly and purified. This enzyme amidates positions b, d, e and f on the corrin ring. Crystallisation trials were attempted, and a low resolution data set was collected (4.5 Å). However, the structure was not determined due to the poor quality of the data. CobB and CobQ are key enzymes involved in the amidation process in the vitamin B12 pathway. To date there is no structural information in the literature as to how these enzymes function. Obtaining a crystal structure of either CobB or CobQ would enhance our knowledge on these multi-functional enzymes, enabling a clearer understanding on how they bind their substrates, and how these enzymes pass their product to the next enzyme in the vitamin B12 pathway.

Bacterial microcompartments (BMCs) are cytoplasmic polyhedral organelles that are composed of a semi-permeable proteinaceous shell made from a number of different proteins, which encapsulate a specific metabolic process. Recent studies have revealed that the basic functions of these compartments are not only to increase the metabolic efficiency of the enclosed enzymes but also to protect the cell from toxic intermediates. BMCs are found in about 20 % of all bacteria and spread widely across the kingdom. One of the bacteria that contains operons encoding for such protein compartments is Clostridium phytofermentans, a bacterium that was originally isolated from forest soil and can directly convert plant biomass into biofuels. The bacterial genome of C. phytofermentans has three BMC-encoding loci, one of which is a 13-gene operon that houses the genetic information for a predicted fucose/rhamnose utilisation system. Within this operon, six genes encode for shell proteins, which form the BMCs outer structure, and the other seven genes encode for metabolic enzymes that are responsible for converting fucose/rhamnose into ethanol, propanol, and propionate. Based on this genome analyses, research has been undertaken to clone the various BMC-associated genes of the C. phytofermentans fucose/rhamnose system in order to generate a recombinant BMC in E. coli. This was attempted firstly by the individual cloning and recombinant expression in E. coli of the genes that encode the shell proteins. This approach also allowed the characterisation of the shell proteins and determine their ability to form functional BMCs. Secondly, engineer empty BMCs from the six shell proteins and investigate the minimum number of genes required for BMC formation. One of the shell proteins characterised in this study, Cphy_1186 (PduT-like), was crystallised and had its structured determined by X-ray crystallography. The protein was found to have a trimeric arrangement of subunits, each of which contains a double-BMC-domain consistent with it belonging to the BMC-T class. The trimer forms a flat approximately hexagonally shaped disc with a central pore that is suitable for binding a 4Fe-4S cluster. The pore is lined with six cysteine residues, more than enough for Fe-S centre formation The structure of the shell protein Cphy_1182 (PduA-like) was also solved. The protein contains a single-BMC-domain within each subunit and forms a hexamers, thereby belonging to the BMC-H class. Overproduction of Cphy_1182 by itself results in the formation of filaments or nanotubes in the bacterial cell. The first attempt at engineering empty metabolosomes from the six shell proteins that are involved in the formation of the C. phytofermentans metabolosome is presented. Although ultimately this was unsuccessful the associated research has given greater insight into the structure/function relationship between the individual components.

Cobalamin or vitamin B12 is one of the most complex small molecules produced by prokaryotic organisms. B12 is a member of the metabolically diverse group of compounds called tetrapyrroles, whose members include haem, chlorophyll, sirohaem and co-enzyme F430. The metabolic activity of B12 is dependent upon the cobalt ion located at the centre of a contracted corrin macrocycle. Replacement of this central cobalt ion with an alternative metal ion results in the formation of a compound that structurally resembles B12 but shares none of its vitamin function. Zincobalamin (Zbl), the zinc analogue of B12, was synthesised from an intermediate of B12 biosynthesis, hydrogenobyrinic acid a, c diamide (HBAd). This intermediate is the last metal-free intermediate of B12 biosynthesis. The synthesis of Zbl from HBAd involves a combination of chemical synthesis and biochemical modification of the starting material. Regiospecific amidation of 4 out the 5 carboxylic acid sidechains decorating the macrocycle of HBAd is required in order to produce Zbl. During B12 biosynthesis these amidations are catalysed by a single enzyme CobQ. Previously characterised CobQs are highly specific for a cobalt containing intermediate of B12 biosynthesis. A CobQ capable of acting "out of turn" with respect to the normal order of B12 biosynthesis was identified within the genome of Allochromatium vinosum. This enzyme was used to good effect in the conversion of HBAd into hexa-amidated compound called hydrogenobyric acid (Hby), the metal free analogue of a late intermediate of B12 biosynthesis cobyric acid. A structural reason as to why A. vinousm CobQ is able to recognise HBAd as a substrate was investigated through sequence alignment and structural prediction. The remaining steps towards Zbl synthesis were completed chemically through the abiotic insertion of zinc into the macrocycle of Hby and the attachment of the pre-fabricated lower nucleotide loop. As both Hby and its zinc containing counterpart have not been previously described in the literature all three compounds including Zbl were subject to biochemical characterisation by UV-visible, NMR and mass spectroscopy. The influence of these analogues on the growth of a strain of Salmonella enterica dependent on exogenous B12 for growth, was investigated.

Enhancing the catalytic efficiency of enzymes has been an aspiration for many years and protein engineering has been quite successful in this area. However, researchers are now turning their attention to the organisation of enzymes within cells to enhance whole pathway productivity. As such, bacterial microcompartments (BMCs) have attracted significant interest due to their ability to not only concentrate enzymes and metabolites but also prevent the release of toxic intermediates into the cell owing to their semi-permeable protein shell. The aim of the research described in this thesis was to develop a number of mechanisms to enhance the spatial organisation of proteins within E. coli and thereby increase metabolic productivity. The ability of BMC encapsulation peptides to target foreign proteins to a BMC was initially investigated. However, we found that the fusion of BMC encapsulation peptides to the enzymes of a 1,2-propanediol (1,2-PD) synthesis pathway resulted in the aggregation of the tagged enzymes rather than their localisation to a BMC. Nonetheless, the tagged enzymes were found to produce significantly more 1,2-PD than strains producing untagged enzymes, irrespective of the presence of BMCs, suggesting that it is the enzyme aggregate that results in enhanced substrate channelling. Furthermore, we showed that a highly conserved hydrophobic motif is conserved across a wide range of BMC targeting peptides and suggest that it is the amphipathic nature of these peptides that results in aggregation of tagged components. The ability of PduA, a component of the BMC shell, to form filamentous structures was investigated for its potential as a cytoplasmic scaffold. Targeting proteins to the cytoscaffold was explored through the use of de novo designed coiled-coil peptides. Localising the enzymes pyruvate decarboxylase and alcohol dehydrogenase on this scaffold by tagging them with the cognate coiled-coil peptide resulted in a significant enhancement in ethanol production in comparison to a strain lacking the scaffold. Additionally, we showed that it was possible to locate this intracellular scaffold to the inner membrane of the cell, further demonstrating the flexibility of this system. Finally, we utilised the same coiled-coil peptide technology to target fluorescent proteins to BMCs, overcoming the aggregation behaviour of the native targeting peptides. Through the re-design of a BMC shell protein we showed targeting occurs to both the external and luminal faces of these BMCs. Taken together, the evidence presented in this thesis provides a number of alternative strategies to not only enhance spatial organisation within a cell but also to increase productivity of engineered metabolic pathways.

Bacterial microcompartments are isohedral structures composed entirely of protein and contain a 2-4 nm thick proteinaceous shell. This shell encases enzymes which act in a sequential manner to carry out a specific metabolic reaction. There are seven essential shell proteins encoded on the propanediol utilization microcompartment (Pdu) operon, Pdu -A -B -B' -J -K -N and - U, each having a critical role in the formation of the compartment. Other genes in the Pdu operon interact to form enzymes in the metabolic pathway with some having unknown functions such as PduV which has been shown to form filament-like structures. These filaments only appear in the presence of microcompartments and often co-localise with them and it is thought that PduV plays a role in the spatial distribution of microcompartments. This study has explored the interaction between PduV and the microcompartment shell. The shell protein involved in this interaction is thought to be PduK, and to characterise the mechanism of this interaction PduK truncations were generated and novel microcompartments containing these truncations were engineered. The ability of these microcompartments to form was examined, as well as the ineraction between PduV and these truncated forms of PduK. Filament formation of the small GTPase, PduV, was studied by site directed mutagenesis of the GTP binding site, which was then examined by microscopic and biochemical approaches. In addition, to further characterise how microcompartments form each individual shell protein was tagged with mCherry. Microscopy techniques such as live cell imaging and transfer electron microscopy were used in order to determine if these novel compartments were able to form correctly, if PduV filament formation was affected in any way and whether this subsequently had an effect on the distribution of microcompartments within the bacterial cell. The work presented has given insight into the formation of the microcompartment shell, higlighting the sensitivity of individual shell proteins to minor modifications such as fluorescent tagging and the effect this can have on the formation of the shell. It has also revealed possible previously undiscovered regulatory mechanisms of shell proteins on PduV filament formation and length, with the modification of shell proteins affecting PduV filaments. Finally, it has given greater evidence to previous suggestions that PduV may be a GTPase.

Cobalamin, also known as vitamin B12, is an essential nutrient for many different organisms including mammals, fish, birds, nematodes, and a variety of bacteria. However, cobalamin is only synthesised by a few bacteria and archaea. Organisms that cannot synthesise cobalamin de novo must obtain it from their diet. In humans, the cobalamin uptake mechanism has been studied in detail, but in many organisms, such as Caenorhabditis elegans, no method of transport has been defined, and their need for cobalamin is recognised by a cobalamin deficiency phenotype. Corrin ring modified fluorescent analogues of cobyric acid and ribose conjugated fluorescent analogues of cobalamin were synthesised in order to follow the uptake and localisation of these corrinoids in a variety of organisms. Both the C5 corrin-ring modified and the ribose conjugated analogues were absorbed by Salmonella enterica, using the B12 uptake system (Btu) and could be converted into active coenzyme forms. The imaging of these fluorescent analogues enabled the identification of the coelomocytes in C. elegans as a possible storage cell for cobalamin. However, the C5 cobyric acid analogue was not recognised which suggests that the C. elegans cobalamin transport mechanism is specific for complete corrinoid molecules. Lepidium sativum, garden cress, was shown to take up both cobalamin analogues from the roots and store it in the vacuoles of the cotyledons in seedlings, even though plants have no cobalamin requirement. In contrast, Arabidopsis thaliana did not transport any of the cobalamin analogues. Cobalamin deficiency has been implicated in impeding disease progression in a number of diseases, such as tuberculosis. The Mycobacterium tuberculosis cobalamin uptake protein, BacA, has only recently been identified, and there is still much to learn about the relationship between M. tuberculosis and cobalamin. Incubations of a cobalamin dependent strain of M. tuberculosis, ΔmetE, with a selection of cobalamin biosynthesis intermediates showed that cobyric acid is the earliest intermediate to be taken up and converted into the cofactor form. The C5 corrin ring modified cobyric acid fluorescent analogue is also capable of rescuing this ΔmetE strain, and is taken up faster than the ribose conjugated cobalamin analogue. Overall, the research outlined in this thesis demonstrates that fluorescent corrinoid analogues can be used to follow the journey of cobalamin in a broad range of different organisms and systems.

Bacterial microcompartments (BMCs) likely represent the largest protein complex found in bacterial cells (estimated 18,000 subunits). BMCs are associated with specific metabolic processes such as carbon fixation in the case of carboxysomes or carbon utilisation in the case of the 1,2-propanediol utilisation metabolosome. With all BMCs there is still much to learn about the structure and function of these supramolecular assemblies. In this project the structure and function of the Pdu BMC has been studied through a combination of recombinant DNA technology, proteomics, fluorescence microscopy, transmission electron microscopy (TEM) and atomic force microscopy (AFM). Initially, wild type Citrobacter freundii (C. freundii) Pdu BMCs were compared to recombinant BMCs that had been produced in Escherichia coli (E. coli). These were found to have very similar profiles in terms of BMC shell composition, size (120-130 nm on average) and shape. However, recombinant empty BMCs (eBMCs), generated by the coexpression of the genes for just the shell proteins, were found to be distinctly different in that they incorporated more of the first shell protein of the operon, PduA, and were also significantly smaller (55-80 nm on average) than the complete BMCs. This work provides the first recorded application of AFM for the study of BMCs and revealed that the eBMCs were stiffer and less flexible than complete BMCs and it is unclear if differences in shell protein composition or the lack of the internal scaffold is causing the reported biophysical differences. One key property of the BMCs is their ability to absorb substrate and cofactor molecules to allow the internalised metabolic pathway to operate. Adenosylcobalamin (Ado-B12) is required as a coenzyme for the diol dehydratase conversion of 1,2-propanediol (1,2-PD), but the uptake of cobalamin into the BMC was not shown. By feeding cells exogenous cobalamin, it was possible to demonstrate that cobalamin is specifically partitioned from the cytoplasm into the BMC. Moreover the synthesis of fluorescently labelled versions of cobalamin allowed this partitioning to be followed ex-vivo with purified BMCs.

Magnetosomes are membranous organelles found in magnetotactic bacteria (MTB). The organelle consist of ferromagnetic crystals housed within a lipid bilayer chained together by an actin-like filament and allows MTB to orient within magnetic fields. The genetic information required to produce these organelles has been linked to four different operons, encoding for 30 genes. These membranous organelles and the magnetic minerals housed within have various biotechnological applications, therefore enhanced recombinant production of such structures in a model organism holds significant potential. The research described in this thesis is focuses on the production of recombinant magnetosomes in the model organism Escherichia coli. Cloning the genes involved in the generation of the organelle individually or in various combinations resulted in the construction of over 100 different plasmids, compatible with the model organism. SDS-PAGE and electron microscopy analysis was used to characterise E. coli cells harbouring these constructs. The observation of electron dense particles, arranged in a chain structure, show that magnetosome generation in the model organism is possible, but is highly dependent on the growth conditions used. The need for specific growth conditions is later backed up by the analysis of the maturation of the cytochrome c proteins involved in magnetosome biomineralisation, which can only be correctly processed under certain conditions. Individual production of two different magnetosome proteins, MamQ or MamY, allowed the generation of various membranous structures in E. coli observed in 48.9% and 56.2% of the whole population of cells respectively. Combinations of these with MamI, MamL or MamB in a variety of combinations led to a variation in the phenotype observed. Bioinformatics analysis of MamQ led to the discovery of a novel membrane restructuring protein family, the LemA protein family, present in a broad range of bacteria. Four different LemA proteins from Bacillus megaterium, Clostridium kluyveri, Brucella melitensis or Pseudomonas aeruginosa were then produced in E. coli and the analysis of the resulting strains revealed the presence of novel intracellular membranous structures which vary in size, form and localisation. Furthermore, when attempts were made to target these proteins for the modification of the outer membrane, a mechanism for increased outer membrane vesicle generation was serendipitously discovered and different effects of these proteins were once again observed. Together, the results described shows good evidence for recombinant magnetosome production in E. coli and opens a new avenue of membrane engineering in this commonly used organism. Such membranous structures have various biotechnological applications, such as enhanced metabolic engineering potential or specialised lipid vesicle production.

Haem, a cyclic tetrapyrrole, is found in organisms from all three domains of life. Haem is a prosthetic group for many proteins involved in essential biological processes such as respiration and oxygen transport. Synthesis of haem in eukaryotes and most bacteria follows a well defined route with highly conserved intermediates. However, an alternative haem biosynthesis pathway in Archaea and some bacteria was recently elucidated. This newly discovered pathway utilises sirohaem as a metabolic intermediate rather than a prosthetic group. The alternative haem biosynthesis pathway is catalysed by the Ahb enzymes A, B, C and D. Initial decarboxylation of sirohaem occurs at the two acetic acid side-chains attached to carbons C12 and C18 to give didecarboxysirohaem, a process catalysed by AhbA and AhbB. Subsequently, the radical SAM enzyme AhbC converts didecarboxysirohaem to Fe-coproporphyrin III. Finally, AhbD catalyses the conversion of Fe-coproporphyrin III into haem in another SAM dependent reaction. This study focused on the AhbA and AhbB proteins from three sources, Desulfovibrio desulfuricans, Desulfovibrio vulgaris and Methanosarcina barkeri. Purifications of individual recombinantly produced proteins revealed that both AhbA and AhbB are highly unstable. However, low concentrations of D. desulfuricans and D. vulgaris AhbA and AhbB proteins were isolated and were discovered to have decarboxylase activity. Simultaneous overproduction of AhbA and AhbB proteins facilitated the co-purification of stable heteromeric AhbA/B complexes from all three organisms. However, despite their sequence similarities, distinctly different properties were observed between the homologues including differences in oliogmeric state and haem/product binding capabilities. The D. desulfuricans enzyme has been crystallised and its structure has been elucidated in both apo- and product-bound forms. Mutagenesis of the D. desulfuricans complex has provided further information about active site residues and a mechanism of action has been proposed. Finally, novel functional chimeric complexes were produced using the Desulfovibrio proteins.