Thesis (Ph.D.) -- University of Texas at Arlington, 2008.

Thesis (Ph. D.)--Massachusetts Institute of Technology and Woods Hole Oceanographic Institution, 2001. / "February, 2001." "Funding was provided by a grant from the National Science Foundation (OCE-9618729), the National Science Foundation Coastal Traineeship (DGE-9454129), and by an Environmental Protection Agency STAR Graduate Fellowship (U-914966-01-0)." Includes bibliographical references.

Bdellovibrio bacteriovorus are small, predatory bacteria that grow within the periplasmic space of a host bacterium. Bdellovibrio has a biphasic life-cycle switching from a uni-nucleoid, growth-senescent ‘attack-phase’ to a novel, multi-nucleoid filamentous ‘growth-phase’, which elongates and divides, growing saprophytically within the periplasmic space of their prey. Little is known to date about Bdellovibrio developmental processes and cell division within this periplasmic niche. Recent publications have demonstrated that bacterial cytoplasms house highly organised matrices of protein structures, called the bacterial cytoskeleton. The Bdellovibrio processes of prey-cell entry, filamentous cell growth and division coordination brings cellular morphological changes and challenges that could be coordinated by cytoskeletal elements. Green Fluorescent protein (GFP)-tagging and gene knock out approaches were used to gain insights into the function of these elements including: an Intermediate filament like protein Ccrp, which has a role in the maintenance of cell morphology; two actin homologues, which appear to function at different points in the predatory cycle, MreB1 and MreB2; and a new type of cytoskeletal element designated ‘bactofilin’, which may have a role in cell division control. Recent advances in GFP technologies have led to the development of optimised GFP variants, such as mTFP1 and mCherry. These have been used to reveal previously unseen detail of Bdellovibrio development within prey. Bdellovibrio do not follow the familiar pattern of bacterial cell division by binary fission, instead divide synchronously at multiple sites along their length, once prey resources are depleted. This yields both odd and even numbers of progeny Bdellovibrio.

Planktonic algae are useful as indicators of water quality because their composition and distribution reflects environmental condition in lakes. Therefore, understanding their dynamics can aid certain water quality management goals. Lake Diefenbaker is a large mesotrophic reservoir in the Canadian Prairies. Approximately 98 % of its inflow is from the South Saskatchewan River. The composition and ecology of the phytoplankton community has not been reported comprehensively since the 1980s. This is a potential problem for a reservoir with multiple end users. Therefore, I collected epilimnetic whole water samples along its length from June to October in 2011 and in 2012. I examined the phytoplankton community and related their distribution to environmental factors. A total of 72 phytoplankton genera were observed with the chlorophytes having the highest number of genera (33). The increased nutrient load and non-algal turbidity associated with high inflow from the South Saskatchewan River may be related to the dominance of the cryptophytes and bacillariophytes (together constituting ~89 % of the total phytoplankton biomass). The cryptophytes were abundant during periods of high flow rates and thermal stratification whereas the bacillariophytes were abundant during cool, isothermal conditions. Lake Diefenbaker is characterized by numerous embayments. Some of these embayments are exposed to human activities including development (housing, golf courses, marinas) and livestock operations (e.g., cattle watering). These localized activities could increase the frequency or size of algal blooms that will adversely affect the water quality. Therefore, I compared the phytoplankton community composition from eight exposed embayments, four unexposed embayments and six main channel sites. Phytoplankton community compositions were not significantly different in exposed, unexposed embayments and main channel sites (P > 0.05). High flows may have overridden localized influence from embayments. Hence, similar environmental conditions were present in the embayments and main channel. Blooms of cyanobacteria are of concern because of the potential of some genera to produce cyanotoxins. I examined cyanobacteria in Lake Diefenbaker. Cyanobacterial biomass was low in Lake Diefenbaker (< 5 %). However, I observed some potential toxin and bloom-forming genera that may threaten the water quality under different environmental conditions in the future.

Cyanobacteria are prokaryotes that can grow photoautotrophically using oxygenic photosynthesis. Some filamentous cyanobacteria in media with insufficient fixed nitrogen develop a regular pattern of heterocyst cells that fix nitrogen for the remaining vegetative cells. We have built an integrated computational model of fixed nitrogen transport and cell growth for filamentous cyanobacteria. With our model, two qualitatively different experimentally observed nitrogen distributions between a pair of heterocysts are reconciled. By adding dynamic heterocyst placement into our model, we can optimize heterocyst frequency with respect to growth. Further introduction of modest leakage leads to distinct growth rates between different heterocyst placement strategies. A local placement strategy yields maximal growth and steady state heterocyst spacings similar to those observed experimentally. Adding more realistic fixed nitrogen storage based heterocyst commitment together with lateral inhibition to the model allows us to address initial heterocyst commitment and qualitatively reproduces many aspects of heterocyst differentiation. We also investigate patterns of starving cells and correlations of fixed nitrogen in filaments without heterocysts. We find percolation transitions in both spatial one dimensional patterns and space-time two dimensional patterns.

To investigate the effectiveness of physical forces in destroying microorganisms a heat resistant (D7Soc=20 min) Gram-positive vegetative bacterium, Microbacterium lacticum, a Gram-negative vegetative bacterium, Escherichia coli, and vegetative cells and spores of Gram-positive bacterium, Bacillus subtilis, were subjected to high mechanical energies using gelatin and maize grits as carriers. A twin screw extruder, a piston capillary rheometer and a rotational rheometer were used. When extruded with gelatin there was a strong correlation between the destruction of M. lacticum and the die wall shear stress and specific mechanical energy (SME). Within the limit of detection, no surviving M. lacticum could be detected in gelatin at the highest die wall shear stress of 409 kPa and a SME of 390 kJ -kg" , giving at least 5.3 decimal reductions. There was no surviving M. lacticum in maize grits at 289 kPa die wall shear stress and 294 kl-kg" SME, giving a 4.6 decimal reduction. The temperature at the extruder die remained below 61°C for all the extrusion experiments indicating that the bacterial destruction was due to combined shear and thermal forces rather than thermal forces alone. It was suggested that the thermal energy supplied during extrusion weakened the bacterial cell wall, making the cells susceptible to shear forces. A maximum 3.2 decimal reduction in the number of spores of B. subtilis in maize grits was obtained at 595 kPa die wall shear stress and 844 kJ.kg-1 SME, below 43°C extruder die temperature. There was no statistically significant difference in the survival of B. subtilis PS346 and B. subtilis PS361, which is a heat sensitive strain due to lack of (l- and P-SASP proteins, under the same extrusion conditions, suggesting that the main destruction mechanism was not heat. High reduction in the number of the viable spores suggested a possible "mechanical germination" inside the dynamic environment of the extruder. A 4.2 logarithmic reduction in the number of M lacticum in 30% (wwb) moisture content gelatin was observed in an unsheared sample in the piston capillary rheometer at 192 MPa and 60°C, showing that pressure could cause major destruction at high temperatures (60-75°C). No survival of M. lacticum was detected beyond 695 kPa shear stress and 64 MPa at 60°C, suggesting that an optimum combination of shear, thermal and pressure forces can cause an important reduction in the numbers of vegetative cells. Shearing of the microorganisms in the rotational rheometer in gelatin showed that the shear resistance of the microorganisms were different. Although only three species of bacteria were tested, it appeared that Gram-positive bacteria were more resistant to shear forces than Gram-negative bacteria. The results suggested that the destruction of the microorganisms at low shear forces (-3 kPa shear stress) was due to weakening of the bacterial cell wall at temperatures above 60°C. A maximum 1.4 logarithmic reduction in the number of M. lacticum was achieved after 4 min of shearing at 804 S-1 shear rate and 75°C. Based on the heat resistance data, thermal forces were not enough to cause significant destruction in the numbers of the microorganism, however, the temperature played a significant role by weakening the bacterial cell wall making it susceptible to shear forces. In this context, it is possible that there was a synergistic relationship between the shear and thermal forces. A shear D-value concept was introduced which was used to evaluate the shear resistance of microorganisms at different temperatures. Starch conversion (determined by differential scanning calorimeter) due to low temperature extrusion of maize grit inoculated with M. lacticum and spores of B.subtilis showed that there is a positive correlation between the bacterial destruction and the starch conversion. Up to 94% starch conversion was obtained during low temperature extrusion of maize grit where the estimated degree of starch conversion due to heat alone was 3.8%. The results suggested that if the shear forces can be optimally combined with thermal forces, an acceptable sterility can be achieved at significantly lower temperatures which would help to keep the quality of food products high.

Bacteria have been isolated from virtually every environment on Earth. The Antarctic continent is no exception. In this extremely cold and dry environment bacteria have colonised various refugia and have evolved a number of strategies for coping with the extreme physico-chemical fluctuations they are exposed to within the environment. These psychrophilic adaptations include cold adapted proteins and lipids which are interest for biotechnology in areas such as frozen foods, agriculture and cryogenic storage. One type of cold adapted protein of particular interest is the antifreeze protein (AFP) for its recrystalisation inhibition and thermal hysteresis activity. It was first isolated from Antarctic fish in the 1970, but has since been found in plants, fungi, insects and bacteria. Over 800 bacterial isolates were cultured from lakes of the, Vestfold Hills, Larsemann Hills and MacRobertson Land, Antarctica. Approximately 87% were Gram negative rods. A novel AFP assay designed for high-throughput analysis in Antarctica, demonstrated putative activity in 187 isolates. Subsequent SPLAT analysis (qualification assessment of recrystalisation inhibition activity) of the putative positive isolates showed 19 isolates with significant recrystalisation inhibition activity. These 19 isolates were cultured from five separate lakes with substantial physico-chemical differences. The 19 AFP active isolates were characterised, using amplified ribosomal DNA restriction analysis (ARDRA) and 16S rDNA sequencing, as predominantly belonging to genera from the a- and y-Proteobacteria, although they were more prominent in the gamma subdivision. One of these isolates (213, Halomonas sp.) was shown as dominant within its community by DGGE analysis, indicating a possible selective advantage for AFP active bacteria. This is the first report of the phylogenetic distribution of AFP activity within bacteria, thus providing information which could enable future bacterial AFP assessments to be aimed at specific taxonomic groups.

Ten overlapping fragments covering the entire Clostridium difficile toxin A gene were cloned and expressed in Escherichia coli. Eight fragments (a', a2, b, c, d, e, f and g) represented the first 5.55kb of the gene whereas two fragments (hl and h2) each spanned the entire C-terminal repeat region of the molecule. All activities relating to binding to carbohydrates (i. e. cold haemagglutination of rabbit erythrocytes), binding to bovine thyroglobulin and non-specific binding to a murine monoclonal antibody were restricted solely to peptides H1 (amino acids [aa] 1834-2683) and H2 (aa 1832-2683). Peptide H2 alone also displayed the ability to bind to cells and to be internalised into endosome-like compartments within the cells. Taken together with the observation that peptide H2 caused a cytopathic effect on Vero cells which was atypical of the holotoxin, these results may indicate that the repeat region of toxin A stimulates intracellular signalling pathways prior to Rho glucosylation. Peptide A2 (aa 1-536) glucosylated recombinant RhoA (rRhoA) in vitro, whereas peptides A'(aa 1-205), B (aa 542-859), C (aa 114-859), D (aa 869-1330), E (aa 542- 1161), G (aa 869-1830) and H2 (aa 1832-2683) did not. The results obtained for peptides A', A2 and C indicate that the first 536 as encompass the catalytic domain for this activity, that more than the first 205 as alone are needed for expression of enzymic activity, and that for a peptide to be active it must not lack the first 113 aa. The first 113 as of the holotoxin are probably essential for the correct folding of the catalytic domain and expression of its activity. These studies were also the first to locate the toxin A ATP binding site to a peptide spanning as 542-859 (peptide B) of the holotoxin. Antibody reaction profiles of antiserum to holotoxin A against toxin A peptides and of antiserum to the peptides against holotoxin A indicate that this region is unexposed in the native state. Also of interest was the observation that the only peptides, which contained the nucleotidebinding site (B and E), lacked the ability to glucosylate rRhoA. Further peptide A2, which possessed glucosyltransferase activity, lacked the nucleotide-binding site. These studies therefore, suggest that a nucleotide-binding site is not required for in vitro glucosylation of rRhoA by toxin A, and fail to identify a role for the toxin A nucleotide binding site. An engineered truncated form of toxin A, consisting of the first 539 as of the holotoxin (encompassing glucosyltransferase activity) fused to the 852 as C-terminal peptide H2 (repeat end binding portion) caused a conventional cytopathic effect (CPE), but was 1,400 fold less cytotoxic to Vero cells than the holotoxin. Peptide A2 (aa 1-536) alone had no effect on Vero cells or in rabbit ileal loops suggesting that peptide H2 aided delivery of the glucosyltransferase molecule into cells leading to a CPE. The truncated toxin lacked the nucleotide binding site and the putative membrane-translocating domain (internal hydrophobic region). The reduced activity of the truncated toxin suggests that although not essential for cytotoxic activity, the nucleotide-binding site and the internal hydrophobic region are important for stability and/or efficient translocation of the holotoxin into the cytosol.

In this study several components responsible for the formation and function of the unidirectional flagellum of R. sphaeroides WS8 were identified via the characterisation of motility impaired TnphoA mutants. The role of the alternative sigma factor sigma 54 in flagellar gene regulation was also examined. Mutant M18 was defective in a fliI homologue, characterisation of this mutant revealed that FliI is not essential for flagellar formation in R. sphaeroides. This differs from that reported in the literature for S. typhimurium and so highlights the importance of studying R. sphaeroides as a model for flagellar motility. Analysis of another mutant Nm7 revealed that it was defective in FliF, a rotor component around which other flagellar components assemble. Overexpression of a FliF fusion protein allowed the production of anti FliF antiserum. DNA sequencing upstream and downstream of the fliF gene, revealed several other genes encoding flagellar components and a potential flagellar gene regulator (Torf). fliE, encoding a component of the basal body of unknown function, was identified upstream of fliF, an interposon mutant was created and was unable to be complemented by the wild type gene in trans suggesting a dominant effect. This is the first dominant mutation to be isolated in any fliE . The gene encoding the motor component FliG was also identified downstream of fliF and its C-terminal motility domain was found to contain regions that are conserved between FliG proteins from unidirectional and bidirectional motors, these may play a role in motor rotation and not switching. An overexpressed poly histidine FliG fusion protein was found to form a complex with the FliF-GST fusion protein ill vitro. The torf gene encodes a protein with homology to sigma 54 enhancer binding proteins. The Torf protein lacks any obvious DNA binding motif and may represent a novel member of the sigma 54 enhancer binding protein family.

The prevalence of the cyanobacterial metabolites: MIB, geosmin and microcystin in drinking water is a major concern to the water industry as these metabolites can compromise the quality of drinking water. Consequently, effective removal of these metabolites from drinking water is paramount. The combination of ozone (O3) and granular activated carbon (GAC) has been shown to be effective for the removal of these metabolites from drinking water. In this study, the ozonation of MIB and geosmin was affected by the character of natural organic material (NOM). In particular, NOM containing compounds of high UV absorbing properties and high molecular weight (MW) resulted in greater destruction of MIB and geosmin due to the formation of hydroxyl (OH) radicals. In addition, alkalinity also affected the ozonation process, with waters containing higher alkalinity resulting in decreased destruction of MIB and geosmin. Laboratory scale minicolumn experiments, coupled with the homogenous surface diffusion model (HSDM), were found to be ineffective in predicting the GAC breakthrough behaviour of MIB and microcystin at two different pilot plants. This can be attributed to the biological degradation of the metabolites at the pilot plants which cannot be modelled by the HSDM. In addition, the volume of GAC used in the minicolumn experiments may not have been appropriate for the predictions, rather, larger laboratory scale columns were found to be more applicable in mimicking pilot plant results. Microcystins were shown to be readily degraded by the bacteria attached to the GAC. Furthermore, the lag period prior to the onset of degradation, which is indicative of most biological degradation studies, was effectively eliminated and in one instance abated. This finding suggests that biological filtration of microcystin is practically feasible especially since the occurrence of microcystins in water supplies is seasonal. This study expands on previous research in the area of O3 and GAC for the treatment of MIB, geosmin and microcystin. With the imminent increase of the use of O3 and GAC in Australian water treatment plants (WTPs), this study provides valuable information for the use of these processes both alone and in combination, particularly since minimal research in this area has been conducted in Australia. / thesis (PhDAppliedScience)--University of South Australia, 2004.