<p> Sigma S (RpoS), encoded by rpoS, is a subunit of RNA polymerase holoenzyme that controls the expression of many genes in stationary phase of various gram negative bacteria. Escherichia coli expresses these genes to withstand environmental stress and nutrient starvation. Several naturally occurring mutant alleles of the gene have been reported and indicate key differences from laboratory strains. We sought to explore the role of natural alleles of the rpoS gene (from non- K12 strains) and thus the sigma subunit relative to the K12 allele. To study the effect of the rpoS polymorphism on gene expression ofRpoS regulon members, rpoS alleles from ECOR- 21, ECOR-28, ECOR-37 and ECOR-40 as well as MG1655 were cloned into the same background, MG1655 ΔrpoS:cat osmY-lacZ. Sequence analysis showed rpoS alleles from all the natural strains tested were different from MG1655 and each other. The strain with rpoS allele from ECOR-28 had increased expression of osmY and katE similar to MG1655. In contrast, rpoS allele from ECOR- 37 showed low expression of osmYbut not as low as ECOR-21 and ECOR-40 which had expression similar to the rpoS mutant. Not surprisingly, recombinant strains with rpoS alleles from ECOR-21, ECOR-37 and ECOR-40 showed no expression of katE (HPII). These suggest that RpoS in ECOR-28 has high activity similar to wildtype K12 strain while RpoS in ECOR-21, ECOR-37 and ECOR-40 has very low or no activity. We conclude that natural E. coli strains have polymorphism in their rpoS ORF which cause variation in the regulatory activities of RpoS on its regulon. </p> / Thesis / Master of Science (MSc)

The Gram-negative bacterium 𝘌𝘴𝘤𝘩𝘦𝘳𝘪𝘤𝘩𝘪𝘢 𝘤𝘰𝘭𝘪 can live on long chain fatty acids as its sole carbon source. However, the rate of fatty acid uptake was reduced after starvation (Mangroo, 1992). In this study, it was found that this starvation effect of reducing oleate uptake rate could only be observed in the initial minute post starvation. The starvation effect was not due to the depletion of its substrates ATP or reduced coenzyme A. Neither was the effect caused by a lowered acyl-CoA synthetase level. This reduction in oleate uptake rate can be reversed by incubation in oleate. It has been reported that lactate activated oleate uptake in 𝘌𝘴𝘤𝘩𝘦𝘳𝘪𝘤𝘩𝘪𝘢 𝘤𝘰𝘭𝘪 (Mangroo and Gerber, 1993). In the present study, it was discovered that when [9, 10-³H]oleate was used as the radioactive tracer in uptake assay, the majority of the radioactivity effluxed from the cells. It was found that lactate did not have an effect on the overall oleate uptake. Its apparent activation on oleate uptake was due to the reduction in the efflux rate of radioactive probes, instead of increasing the uptake rate of oleate. Acyl-CoA synthetase in 𝘌𝘴𝘤𝘩𝘦𝘳𝘪𝘤𝘩𝘪𝘢 𝘤𝘰𝘭𝘪 has always been described as a protein of around 45 kDa (Kameda and Nunn, 1981; Kameda 𝘦𝘵 𝘢𝘭., 1986). However, the 𝘧𝘢𝘥𝘋 gene that encodes the 𝘌𝘴𝘤𝘩𝘦𝘳𝘪𝘤𝘩𝘪𝘢 𝘤𝘰𝘭𝘪 acyl-GoA synthetase predicts a 62 kDa protein (Black 𝘦𝘵 𝘢𝘭., 1992; Fulda 𝘦𝘵 𝘢𝘭., 1994). This new form of acyl-CoA synthetase has been partially purified. This form, which has a size of around 62 kDa, can be converted to its 45 kDa counterpart in a Triton X-1 00 and temperature dependent manner. This induced processing of the 62 kDa enzyme was found to be inhibited by oleate, a natural substrate of acyl-CoA synthetase. The two forms of acyl-CoA synthetase have also been shown to be immunologically related. A much simpler purification scheme was developed for the 45 kDa acyl-CoA synthetase using the membrane bound form of acyl-CoA synthetase. The selective extraction of the 62 kDa acyl-CoA synthetase from the membrane, together with the induced processing into the 45 kDa form has provided the basis for this comparatively simple process. The availability of the expression system made the low yield associated with this simple process acceptable. Using oleate as substrate, the Kₘ and Vₘₐₓ for the 45 kDa acyl-CoA synthetase were determined to be 85 ± 18μM and 1550 ± 165 nmole/min/mg protein respectively, whereas the Kₘ and Vₘₐₓ for the 62 kDa enzyme were determined to be 38 ± 12μM and 633 ± 79 nmole/min/mg protein respectively. This suggested that the 62 kDa enzyme has higher affinity towards oleate than its 45 kDa counterpart under standard assay conditions while the ratio of Vₘₐₓ /Kₘ remained relatively constant. Evidence suggesting Triton X-1 00 activates acyl-CoA synthetase by providing a surface for catalysis was presented. This interaction between acyl-CoA synthetase and the Triton X-1 00 I oleate mixed micelle could be affected by the surface charge of the mixed micelle. / Thesis / Master of Science (MS)

This thesis is missing page 53, no other copies have this page. -Digitization Centre / Bacterial adaptation to changing conditions and to the host environment requires coordinated changes in gene expression that permit more efficient utilization of metabolites and increased survival. An important form of gene control is through the use of alternative sigma factors that direct RNA polymerase to recognize a distinct group of genes. One such sigma factor is RpoS, which is widely present in Proteobacteria including many serious human pathogens. As a key stress response regulator, RpoS plays an important role in adaptation, but its effect on virulence varies in different species. RpoS contributes to virulence through either enhancing survival against host defense systems or directly regulating expression of virulence factors in some pathogens, while RpoS is dispensable, or even inhibitory, to virulence in others. The primary objective of this study is to understand the mechanism of RpoS control in gene expression and pathogenesis of Escherichia coli. This thesis first describes the characterization of RpoS regulon in laboratory and pathogenic E. coli strains by transcriptome profiling analysis. Comparison of RpoS regulons identifies a core set of RpoS-controlled genes as well as strain-specific groups of genes, including many implicated in virulence. The contribution of RpoS to enteropathogenesis in vivo was tested using a Citrobacter rodentium (CR)mouse infection model that is commonly used to simulate E. coli infection in human intestine. Mutations in rpoS result in reduced colonization and delay in mortality, indicating RpoS is important for full virulence. Clinical and natural E. coli isolates exhibit variable abilities in stress resistance and virulence, which is partly attributable to attenuating polymorphisms of rpoS commonly found in E. coli populations. A possible mechanism responsible for the occurrence of rpoS polymorphisms in pathogenic E. coli is addressed. Using a group of representative enterohemorrhagic E. coli strains, we report that growth-enhanced mutants can be selected during growth on succinate and other poor carbon sources under both aerobic and anaerobic conditions. The majority of these mutants carry nonsense or missense mutations in rpoS. Phenotypic microarray analysis reveals that rpoS mutations result in increased utilization of 92 nitrogen and 8 carbon sources. Therefore, the occurrence of rpoS polymorphisms may increase the fitness of the population as a whole for better nutrient scavenging. In conclusion, RpoS may be viewed as a transient regulator that orchestrates the temporal expression of a large regulon for better adaptation under specific conditions including natural and host environments. Under conditions not requiring RpoS, its functions can be turned off through decreasing expression, rapid proteolysis, inhibition of RpoS activity, or selection of attenuating mutations. The final part of this thesis reviews the distinct and niche-dependent involvement of RpoS in virulence of many rpoS-bearing pathogens. - / Thesis / Doctor of Philosophy (PhD)

An increase in antimicrobial resistance bacteria has endangered our ability to treat infectious diseases. Lack of good in-vivo model has made it difficult to study antimicrobial resistance. In this study, we have used an inexpensive and short life span in-vivo model namely, Caenorhabditis elegans (C. elegans) to study antimicrobial treatment using pathogenic Escherichia coli O157:H7, a multidrug resistance bacterium that causes life threatening infection in humans. We have investigated the influence of live vs. heat killed non-pathogenic E. coli OP50 (OP50) as a food source on the growth and survival of infected C. elegans mutant AU37 with E. coli O157:H7 in the presence and absence of antibiotics. This is analyzed using a liquid-based C. elegans-E. coli O157:H7 infection assay. C. elegans was synchronized and grown on a lawn of live OP50 till they reached L4-young adult stage. L4-young adults were transferred to liquid medium where the C. elegans was infected with live E. coli O157:H7 or live non-pathogenic OP50 for 24 hours. After infection, C. elegans were fed live or heat killed OP50 depending on the experiment, and the life span and levels of E. coli O157:H7 were monitored, with and without ampicillin treatment in a 96 well transwell plate. Our results indicate that live OP50 is an ideal food source for C. elegans growth and survival to study antimicrobial treatment. C. elegans growth rate and survival decreased in presence of heat killed OP50, which makes heat killed OP50 as a non-ideal food source for antimicrobial assay. Moreover, using live OP50 we have discovered that the ampicillin dose 8mg/ml, 16mg/ml, and 32mg/ml are effective in increasing the survival of C. elegans infected with E. coli O157:H7. However, treatment on C. elegans infected with acid stressed E. coli O157:H7 is controversial.

Pathogenic E. coli infections cause approximately one quarter of disease losses in commercial turkey flocks. A small subgroup of E. coli causes most infections. Epidemiologic studies of this disease have been hindered by a lack of reliable markers to discriminate between pathogenic and fecal E. coli and by the diversity of poultry strains. Reliance on antimicrobials to control E. coli infections has caused widespread antimicrobial resistance. One hundred five clinical E. coli were obtained, and 1104 isolates were collected from fecal specimens of 20 flocks in an integrated turkey operation. Biochemical fingerprinting and antimicrobial susceptibility tests were performed on all isolates, and somatic antigen serologic testing and PCR for potential virulence genes were conducted on 299 strains including all clinical isolates and fecal isolates that had similar traits to clinical isolates. Most avian E. coli infections were caused by a few clonal strains that were uncommon in normal fecal flora. The potential virulence genes iss, K1 and tsh were detected more frequently among clinical than fecal isolates; however, the pattern of occurrence did not suggest that these genes were useful markers for identifying pathogenic strains. Syndromes consistent with colibacillosis were the most commonly reported illness and principal rationale for antimicrobial therapy in sampled flocks. Most clinical E. coli isolates were resistant to gentamicin, sulfamethoxazole and tetracycline. Although resistance to fluoroquinolones and β-lactam antibiotics occurred less frequently, the potential for resistance to emerge to these antimicrobials was evident. A Bayesian model to estimate sample size confirmed the diversity of avian fecal E. coli strains. Studies are needed to define risk factors for infection with and identify markers for avian pathogenic E. coli strains. These research priorities are complementary and may lead to the identification of new interventions to prevent this important infectious disease of poultry. / Ph. D.

Ph. D.

The glpD gene, which encodes aerobic sn-glycerol 3-phosphate dehydrogenase, and the glpR gene, which encodes a repressor that negatively regulates the expression of the g/p regulon, map near minute 75 on the linkage map of Escherichia coli K-12. In the present study, the nucleotide sequence of the 2895 base pair of DNA containing the glpD control region and the glpE, glpG, glpR genes was determined. The translation initiation codons with adjacent ribosome-binding sites were found for these four genes. The transcription start site of the glpD gene was identified 42 base pairs upstream from the proposed methionine start codon, preceded by a region containing typical -10 and -35 sequences found in bacterial promoters. A binding site for the cyclic AMP-cAMP receptor protein complex was located just upstream from the -35 sequence, centered at position -63. The interaction site for the glp repressor was identified by using DNase I footprinting. This region contained two tandemly repeated sequences which started at the -10 sequence and continued to position +38. The glp repressor contained 252 amino acid residues and had a molecular weight of 28,046 which was deduced from the nucleotide sequence. The position of the initiation codon was verified by determination of the amino acid sequence of the N-terminus of the purified gip repressor. The presumptive helix-turn-helix region of the repressor was located near the N-terminus (amino acids 22 to 41) at a poSition analogous to that found for the operator binding domain of other repressors such as the deo and Jac repressors. The recognition helix of the glp repressor and the nucleotide sequence of the glp operator were very similar to those of the deo system. The presumptive glpR recognition helix was changed to the deoR recognition helix and the sixth amino acid arginine of the recognition helix was changed to alanine by site-directed mutagenesis. The mutant forms of the repressor had a greatly reduced affinity for the glpD operators in vivo, determined by measuring β-galactosidase activity in a strain carrying a glpD-lacZ fusion. / Ph. D.

The glpF gene, which encodes a cytoplasmic membrane protein that facilitates the diffusion of glycerol into the cell, and the glpK gene, which encodes glycerol kinase, map near minute 88 on the linkage map of Escherichia coli K-12. In the present work, the nucleotide sequence of the 843 base Pair glpF gene, 430 base pairs of the glipK gene, and the intervening sequence between the two genes were determined. The control region for the glpFK operon was identified and sequenced. The gipK gene product was purified to near homogeneity by streptomycin sulfate and ammonium sulfate fractionation with subsequent DEAE Sephadex chromatography. N-terminal amino acid analysis identified the startpoint of translation for the gipK gene. The transcription start site was identified 71 base pairs upstream from the proposed translation start codon for glpF. Preceding the transcription start site were -10 and -35 sequences Similar to the consensus sequences for gram bacterial promoter elements. DNase I footprinting was used to identify two binding sites for the CAMP-cAMP receptor protein (CRP) complex upstream from and overlapping the putative -35 sequence. Four binding sites for the glp repressor were located sequentially along the DNA extending from -89 (relative to the start point of transcription) to within the -10 region. Two additional repressor binding sites were identified within the glpK coding region. Interaction of these operator sites with those in the control region was identified. The affinity of the gilp repressor for the control regions of the glpD, glpACB-gipTQ, and glpFK operons was compared by titration studies using a strain harboring a glpT:lacZ fusion and a glpRⁿ mutation. / Ph. D.

A rapid, chemical method for the detection and quantitation of Iipoic acid has been developed. Lipoic acid produces a yellow color when reacted with PdCl₄⁻² in 1 N HCI. The colored complex formed is extractable into methylene chloride, which can be readily concentrated, increasing the color intensity. The limit of sensitivity of this assay for lipoic acid detection in supernatants from Escherichia coli K-12 cell cultures is 2.5 x 10⁻⁵ M, giving an absorbance of 0.10 at 408 nm. Using this assay, it would be possible to screen for a mutant of Escherichia coil K-12 which excretes 100-fold more lipoic acid than the parent strain. Cellular lipoic acid in an anaerobic Escherichia coli K-12 culture remained constant (5 to 6 µg/g dry weight) during growth in a. minimal salts medium containing 1% glucose. However, cellular lipoic acid in an aerobic Escherichia coli K-12 culture increased from 15 µg/g dry weight to 20 to 25 µg/g dry weight during cell growth in the same medium. Aeration (2 L air/L medium/ minute) of a mid-log phase anaerobic Escherichia coli K-12 culture resulted in a doubling of cellular lipoic acid levels within the first thirty minutes and a four fold increase over the next five hours of aerated cell growth. Chinese hamster ovary cells were found not to incorporate the two known bacterial precursors, [²H₃]-acetate and [U-²H₁₅]-octanoate, into lipoic acid. Further studies with both Chinese hamster ovary and mouse fibroblast cells which were designed to demonstrate that these two transformed eel 1 1 ines require lipoic acid for maximal growth were inconclusive. / Master of Science

Ornithine decarboxylase (ODC) catalyzes the decarboxylation of ornithine to produce the diamine, putrescine, in the bacterium Escherichia coli. The speC gene encoding ODC has been shown to be subject to transcriptional repression by either putrescine or the cAMP- cAMP receptor protein (CRP) complex. To determine whether these regulatory modes are independent, the expression of ODC was determined by measuring the specific activity of ODC in crude extracts prepared from exponentially grown cultures of wild type ~ coli K-12 as well as in strains unable to synthesize cAMP (cya) and/or CRP (crp). 1-5 mM cAMP repressed ODe activity 22-38% in wild type, 57-66% in the cya strain, and only 7-18% in the ~ strains. 2-10 mM putrescine repressed ODe activity 30-32% in wild type, 48-49% in the cya strain, and 37-38% in the ~ strain. As putrescine repressed ODe activity in the absence of eRP protein (i.e. in a crp strain), putrescine-mediated repression of ODe appears to be independent of the repression of ODC by the cAMP-eRP complex. This conclusion was verified by demonstrating th t .oDC repression by putrescine and cAMP together was additive. / Master of Science