Thesis advisor: Mary F. Roberts / Thesis advisor: Jianmin Gao / The research reported in this thesis focused on three proteinaceous virulence factors of the intracellular bacterial pathogen Listeria monocytogenes: listeriolysin O (LLO), broad-range phospholipase C (PC-PLC), and phosphatidylinositol-specific phospholipase C (PI-PLC). Based on sequence homology of LLO with other cholesterol-dependent cytolysins (CDC), the protein has four domains of which domain 4 is thought to anchor the protein to cholesterol-containing surfaces while domain 3 mediates protein-protein binding on the membrane and contributes α-helices that convert to two β-strands that form the large β-barrel pore. It was previously assumed that the sequential and cooperative behaviors of domain 3 in each LLO monomer required D4 to bind to cholesterol-enriched membranes. By cloning and expressing a separate protein containing domains 1, 2, and 3 (D123) and the isolated domain 4 (D4) of LLO, I could uncouple some of the events in its membrane binding and pore-formation. Flow cytometry, used to investigate protein binding to vesicles and to red blood cells, showed that D123 had no membrane affinity on its own, but became membrane-bound when sub-lytic amounts of LLO were added. D123, not membrane-lytic by itself, became hemolytic when trace amounts of LLO were present to provide a membrane anchor for D123 proteins. FRET and fluorescence correlation spectroscopy were used to show that D123 and LLO formed oligomers at nanomolar concentration and could also associate with one another in the solution. These results suggest that D4 provides an initial membrane attachment but need not be present on all monomers to trigger the cooperative conformational change that leads to membrane insertion and pore formation. The gene for L. monocytogenes PC-PLC was obtained, expressed in E. coli and the product protein purified and characterized. The zinc content of this metalloenzyme was analyzed with ICP-MS. The dissociation constants of the three zinc ions proposed as necessary for PC-PLC activity ranged from 0.05 to 60 μM. Enzymatic activities of PC-PLC were analyzed for various substrates, include long-chain phospholipid in vesicles (LUVs, SUVs) and micelles (Triton X-100), and short-chain lipids (diC4PC, diC6PC, diC7PC) mono-dispersed in solutions. Key results include the following: (1) the L. monocytogenes PC-PLC has an acidic pH optimum (in contrast to other bacterial PC-PLC enzymes) consistent with its role in vacuole lysis upon acidification; (2) the preference of PC-PLC for longer chain monomeric substrates is not because of a higher kcat but a reduced Km suggesting some amount of hydrophobicity is important for substrate binding in the active site; (3) the apparent Kd of PC-PLC for Zn2+ derived from kinetics at pH 6.0 (1.94 ± 0.22 μM) is lower that that from ICP-MC; and (4) PC-PLC enzymatic activity is not enhanced by added LLO that generates pores in vesicles (likewise, PC-PLC does not affect the membrane lytic activity of LLO) indicating no synergism between the two virulence factors. These results should aid in understanding the function of PC-PLC in L. monocytogenes pathogenicity. The L. monocytogenes PI-PLC and a variant with reduced catalytic activity were expressed and are currently used in a collaborative project with the Portnoy laboratory at the University of California at Berkeley. / Thesis (PhD) — Boston College, 2014. / Submitted to: Boston College. Graduate School of Arts and Sciences. / Discipline: Chemistry.

Staphylococcus aureus causes significant illnesses throughout the world, including toxic shock syndrome (TSS), pneumonia, and infective endocarditis. Major contributors to S. aureus illnesses are secreted virulence factors it produces, including superantigens and cytolysins. Rabbit cardiac physiology is considered similar to humans, and rabbits exhibit susceptibility to S. aureus superantigens and cytolysins. As such, rabbits are an excellent model for studying pneumonia, infective endocarditis, and sepsis, We examined the ability of USA200, USA300 and USA400 strains to cause vegetations and lethal sepsis in rabbits. USA200, TSST-1+ strains that produce only low amounts of Α-toxin, exhibited modest LD50 in sepsis (1x108-5x108) colony-forming units (CFUs), and 3/4 caused significant IE. USA200 strain MNPE, which produces high levels of Α-toxin, was both highly lethal (LD50 5x106 CFUs) and effective in causing IE. In contrast, USA300 strains were highly effective in causing lethal sepsis (LD50s 1 x 106 and 5 x 107 CFUs) but were minimally capable of causing IE. USA400 strains were both highly lethal (LD50s of 1 x 107 and 5 x 107 CFUs) and highly effective causes of IE. Additional studies investigated the role of phenol soluble modulins in infection. We showed that PSMs are important for the ability of S. aureus to cause sepsis but not infective endocarditis. Additionally, immunization against PSMs did not protect rabbits from lethal infection. Our studies show that clonal groups of S. aureus differ in abilities to cause infective endocarditis and lethal sepsis and suggest that secreted virulence factors, including superantigens and cytolysins, account for some of these differences. This thesis also investigates the use of superantigens and cytolysins as staphylococcal vaccine candidates. We generated three TSST-1 mutants; G31S/S32P, H135A, and Q136A. All rabbits administered these TSST-1 toxoids generated strong antibody responses (titers>10,000) that neutralized native TSST-1 in TSS models, both in vitro and in vivo. These TSST-1 mutants lacked detectable residual toxicity. Additionally, the TSST-1 mutants exhibited intrinsic adjuvant activity, increasing antibody responses to a second staphylococcal antigen (Β-toxin). This effect may be due to TSST-1 mutants binding to the immune co-stimulatory molecule CD40. The superantigens TSST-1 and SEC and the cytolysin Α-toxin are known to contribute to staphylococcal pneumonia. Immunization of rabbits against these secreted toxins provided complete protection from highly lethal challenge with a USA200 S. aureus strain producing all three exotoxins; USA200 strains are common causes of staphylococcal infections. The same three exotoxins plus the cytolysins Β-toxin and Γ-toxin contribute to infective endocarditis and sepsis caused by USA200 strains. Immunization against these five exotoxins protected rabbits from infective endocarditis and lethal sepsis. Additionally, a heptavalent vaccine composed of the pentavalent units plus SEB and SE-l X protected rabbits from lethal pneumonia caused by USA100 strain 209. Passive immunization using pooled sera protects previously non-immunized rabbits from lethal pneumonia due to MNPE. These data suggest that immunization against toxoid proteins of S. aureus exotoxins protects from serious illnesses, and concurrently superantigen toxoid mutants provide endogenous adjuvant activity.

Cytolysin A (ClyA) is a member of a class of proteins called pore-forming toxins (PFTs). ClyA is secreted by Gram-negative bacteria, and it attacks a number of mammalian cells by inserting into and forming channels within the cell membrane (Oscarsson J et al., 1999). It has been suggested that ClyA binds to cholesterol (Oscarsson J et al., 1999) and thus can insert into the membranes of many different cell types of eukaryotic origin. In our studies we propose to engineer a ClyA protein that can only attack a small subset of cell types. We propose to engineer ClyA that can be only activated when exposed to specific cell-surface proteases produced by a specific cell type. We ultimately want to target breast cancer cells that differentially secrete or express proteases such as matrix-metalloproteases (Stautz D et al., 2012; Zhang, M et al. 2013). To engineer this protein we took advantage of the N-terminus of ClyA. The N-terminus of ClyA, which is highly hydrophobic (Oscarsson J et al), undergoes a conformational change to insert into the target cell membrane (Oscarsson J et al). This conformational change allows ClyA to penetrate the target membrane to form a transmembrane domain of ClyA. The hydrophobic nature of lipid membranes makes it highly unfavorable for any charged residues to cross the membrane (Hunt J 1997). With this in mind, we hypothesize that negative charges inserted into the N-terminus of ClyA will inhibit it from inserting into the membrane. Thus, we mutated the N-terminus of the ClyA protein by inserting an inactivation site composed of negatively charged amino acids that we hypothesize would prevent insertion into the plasma membrane of the target cell. Once we confirmed that this construct was an inactive ClyA mutant, we inserted a thrombin cleavage site right after the inserted negative charges. This site should allow us to remove the negative charges once the protein is exposed to thrombin. Once the negative charges are removed, the protein should recover its activity. This approach will allow us to create a version of ClyA that is protease-switchable.

Pore forming toxins are known for their ability to efficiently form transmembrane pores which eventually leads to cell lysis. PFTs have potential applications in devel-oping novel drug and gene delivery strategies. Although structural aspects of many pore forming toxins have been studied, very little is known about the dynamics and subsequent rupture mechanisms. In the first part of the thesis, a combined experimental and modeling study to understand the lytic action of Cytolysin A (ClyA) toxins on red blood cells (RBCs) has been presented. Lysis experiments are carried out on a 1% suspension of RBCs for different initial toxin concentrations ranging from 100 – 500 ng/ml and the extent of lysis is monitored spectrophotometrically. Using a mean field approach, we propose a non – equilibrium adsorption-reaction model to quantify the rate of pore formation on the cell surface. By analysing the model in a pre-lysis regime, the number of pores per RBC to initiate rupture was found to lie between 400 and 800. The time constants for pore formation are estimated to lie between 1-25 s and monomer conformation time scales were found to be 2-4 times greater than the oligomerization times. Using this model, we are able to predict the extent of cell lysis as a function of the initial toxin concentration. Various kinetic models for oligomerization mechanism have been explored. Irreversible sequential kinetic model has the best agreement with the available experimental data. Subsequent to the mean field approach, a population balance model was also formulated. The mechanics of cell rupture due to pore formation is poorly understood. Efforts to address this issue are concerned with understanding the changes in the membrane mechanical properties such as the modulus and tension in the presence of pores. The second part of the thesis is concerned with using atomic force microscopy to measure the mechanical properties of cells. We explore the possibility of employing tapping mode AFM (TM-AFM) to obtain the elastic modulus of soft samples. The dynamics of TM-AFM is modelled to predict the elastic modulus of soft samples, and predict optimal cantilever stiffness for soft biological samples. From experiments using TM-AFM on Nylon-6,6 the elastic modulus is predicted to lie between 2 and 5 GPa. For materials having elastic moduli in the range of 1– 20 GPa, the cantilever stiffness from simulations is found to lie in the range of 1 – 50 N/m. For soft biological samples, whose elastic moduli are in the range of 10-1000 kPa, a narrower range of cantilever stiffness (0.1 – 0.6 N/m), should be used.

Historically, infectious bovine keratoconjunctivitis (IBK) was believed to be under the exclusive competence of Moraxella bovis. However, the roles of other species of Genus Moraxella are also being considered in the pathogenesis of IBK, such as Moraxella ovis and particularly Moraxella bovoculi. This thesis describes phylogenetic and genotypic analysis based on the genes encoding type IV pili Q- and I-type (TfpQ/I) of M. bovis and cytotoxin of M. bovis (MbxA), M. bovoculi (MbvA) and M. ovis (MovA). The distinction between M. bovis, M. bovoculi and M. ovis was previously performed by PCR (16S-23S intergenic region) according to the protocol established in the literature. Then, there is described a molecular analysis based on the 3' region of genes tfpQ/I (including α1-C N-terminal subdomain and C-terminal domain) of 16 field strains and five vaccine strains of M. bovis from South America. All 47 sequences of tfp Q- and I-type genes analyzed resulted in 31 alleles designated 1 to 31. The phylogenetic reconstruction resulted in a distinction of 31 alleles in eleven groups (designated A through J and Epp). The analysis of the deduced amino acid sequence (aa) of the C-terminal region showed similarity levels between 67 and 100% within the groups, while the analysis of the D region (C-terminal subunit) resulted in levels of similarity between 60 and 100%. In addition, a phylogenetic analysis based on the 3' region of the cytotoxin gene was performed to investigate the genetic relationship among M. bovis (n = 17), M. bovoculi (n = 11) and M. ovis (n = 7) strains and reference strains. Phylogenetic reconstruction allowed the differentiation among species, and the older M. bovoculi strains remained in branch closer to M. bovis strains. The amino acid similarity level among the MbxA sequences stood at an average of 99.9%, while among the MbvA and MovA sequences the similarity was respectively 98.8% and 99.3%. The similarity between MbvA and MovA was 96.6%, while MbxA for MbvA and MovA was 77.6%. Thus, it is possible to conclude that the tfp gene may be inferred suitable for differentiate among M. bovis strains, while the cytotoxin-encoding gene is suitable for phylogenetic classification of M. bovis, M. bovoculi and M. ovis strains and perhaps for understanding the evolutionary relationships among three species. / Historicamente, acreditava-se que a ceratoconjuntivite infecciosa bovina (CIB) estáva sob competência exclusiva de Moraxella bovis. Contudo, outras espécies do Gênero Moraxella também vêm sendo estudadas quanto à participação na patogenia da CIB, como Moraxella ovis e principalmente Moraxella bovoculi. Esta tese descreve análises filogenéticas e de dados genotípicos com base nos genes codificadores dos pili tipo IV dos tipos Q e I (TfpQ/I) de M. bovis e da citotoxina de M. bovis (MbxA), M. bovoculi (MbvA) e M. ovis (MovA). A diferenciação entre M. bovis, M. bovoculi e M. ovis foi previamente realizada por PCR (região intergenica 16S-23S) conforme protocolo estabelecido na literatura. Após, é descrita uma análise molecular com base na região 3' dos genes tfpQ/I (compreendendo subdomínio α1-C N-terminal e o domínio C-terminal) de 16 isolados de campo e cinco cepas vacinais de M. bovis, provenientes da América do Sul. Todas as 47 sequencias do gene tfp tipo Q e tipo I analisadas resultaram em 31 alelos designados de 1 até 31. A reconstrução filogenética resultou em uma distinção dos 31 alelos em onze grupos (designados de A até J e Epp). A análise da sequencia de aminoácidos (aa) deduzidos da região C-terminal mostrou níveis de similaridade entre 67 e 100% dentro dos grupos, enquanto a análise da região D (subunidade C-terminal) resultou níveis de similaridade entre 60 e 100%. Além disso, um estudo filogenético com base na região 3' dos genes da citotoxina foi realizado para investigar a relação genética entre os isolados de M. bovis (n = 17), M. bovoculi (n = 11) e M. ovis (n = 7) e cepas de referência. A reconstrução filogenética permitiu a diferenciação entre as espécies, sendo que os isolados mais antigos de M. bovoculi permaneceram em ramo mais próximos aos isolados de M. bovis. O nível de similaridade de aminoácidos entre as sequencias de MbxA ficou em 99.9% de média, enquanto entre as sequencias de MbvA e MovA foi respectivamente de 98.8% e 99.3%. A similaridade entre MbvA e MovA foi de 96.6%, enquanto MbxA em relação a MbvA e MovA foi de 77.6%. Assim, é possível concluir que o gene tfp pode ser adequado para inferir distinção entre os isolados de M. bovis, enquanto o gene codificador da citotoxina é adequado para classificação filogenéticas dos isolados de M. bovis, M. bovoculi e M. ovis, e, talvez para a compreensão das relações evolutivas.

The production of membranous vesicles is observed to occur among organisms from all domains of the tree of life spanning prokaryotes (bacteria, archaea) and eukaryotes (plants, animals and fungi). Bacterial release of membrane-derived vesicles (MVs) has been studied most extensively in cases of Gram-negative species and implicating their outer membrane in formation of extracellular MVs. However, recent studies focusing on Gram-positive bacteria have established that they also undergo MV formation. Membrane vesicles are released during normal bacterial growth, they are derived from the bacterial membrane(s) and may function as transporters of different proteins, DNA and RNA to the neighbouring bacteria or to the cells of a mammalian host. The transport of virulence factors in a condensed manner via MVs to the host cells presumably protects these proteins from degradation and, thereby, targets the host cells in a specific manner. The aim of my thesis is to investigate secretion of MV-associated virulence factors and to study interactions of MVs produced by two selected Gram-negative and Gram-positive bacteria, i.e. Vibrio cholerae and Listeria monocytogenes, with eukaryotic host cells. Depending on whether the bacterium acts as an extracellular or intracellular pathogen, MVs may be considered to have specific functions, which may lead to the different outcomes of MV-host interactions. V. cholerae transport systems for virulence factors include the Type VI secretion system and MVs (also referred to as the “Type 0” secretion system). We have identified that the biologically active form of PrtV protease in different V. cholerae serogroups is transported via MVs. PrtV protease is essential for V. cholerae environmental survival and protection from natural predator grazing. We demonstrated that PrtV is primarily translocated via the inner membrane to the periplasmic space, where it undergoes autoproteolysis, and the truncated version of PrtV protein is packaged inside the MVs and released from the surface of bacteria. MV-associated PrtV protease showed a contribution to bacterial resistance towards the antimicrobial peptide LL-37, thereby, enhancing bacterial survival by avoiding this innate immune defense of the host. We also studied another virulence factor of V. cholerae, the pore-forming toxin VCC, which was found to be transported by MVs. MV-associated VCC is biologically active and triggers an autophagic response in the target cells. We suggested that autophagy serves as a cellular defense mechanism against the MV-associated bacterial virulence factor of V. cholerae. Listeria monocytogenes is a Gram-positive intracellular and facultative anaerobic food-borne pathogen causing listeriosis. It causes only sporadic outbreaks in healthy individuals, however, it is dangerous for a fetus or newborn child, and for pregnant and immunocompromised people, leading to a deadly infection in one third of the cases. We have analyzed MVs produced by L. monocytogenes and their interaction with eukaryotic cells. Confocal microscopy analysis showed that MVs are internalized into HeLa and HEK293 cells and are accumulated in lysosomes. Moreover, L. monocytogenes produces MVs inside the host cells and even inside the phagosomes. We found that the major virulence factor of L. monocytogenes, the cholesterol-dependent pore-forming protein listeriolysin O (LLO), is entrapped inside the MVs and resides there in an oxidized inactive state. LLO is known to induce autophagy by making pores in the phagosomal membrane of targeted eukaryotic cells. In our studies, we have shown that MVs effectively abrogated autophagy induced by Torin1, by purified LLO or by another pore-forming toxin from V. cholerae. We also found that MVs promote bacterial intracellular survival inside mouse embryonic fibroblasts. In addition, MVs have been shown to have a strong protective activity against host cell necrosis initiated by pore-forming toxin. Taken together, these findings suggested that in vivo MVs production from L. monocytogenes might be a relevant strategy of bacteria to manipulate host responses and to promote bacterial survival inside the host cells.

Cell cycle regulation is an important part for the detection and destruction of mutated and dysregulated cells as it is a natural protection against degenerative diseases and cancer. The ability of the body to detect and destroy these cells is a vital part in maintaining homeostasis in the body. Once cells have circumvented this line of defence, dismantling these cells would become very difficult.  Research into new ways to target and destroy mutated cells are constantly evolving in hopes of being able to control and direct lysis of target cells using therapeutic drugs. One of the possibilities for such a method are Cholesterol Dependent Cytolysins (CDC), specific proteins found in bacteria. These proteins are dependent on the ability to bind to cholesterol in cell-walls to form pores that lyse and effectively destroy cells.   This project aims to study the structure and mechanistic details of pore formation by CDC mitilysin using cryogenic electron microscopy (cryo-EM). Mitilysin was purified by affinity chromatography and its pore formation ability was confirmed by calcein release assay and hemolysis assay. The pore structures of mitilysin were observed by transmission electron microscopy (TEM) using liposomes composed of both 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) and cholesterol as model membranes. Detergent screening directed separation of pores from liposomes; so that they could be visualized by cryo-EM. While these steps were optimized and proven successful, they were time-consuming. An initial 3D-model of pore-structures was rendered, but no molecular characteristics could be determined at the end of the allotted time. The study does lay the ground steps for obtaining the complete structure of mitilysin pores.