Mechanisms of Staphylococcus Aureus Persistence And

MECHANISMS OF STAPHYLOCOCCUS AUREUS PERSISTENCE AND

ERADICATION OF CHRONIC STAPHYLOCOCCAL INFECTIONS

by Rebecca Yee

A dissertation submitted to the Johns Hopkins University in conformity with the requirements for the degree of Doctor of Philosophy

Baltimore, Maryland December, 2018 ABSTRACT

Bacteria can exist in different phenotypic states depending on environmental conditions.

Under stressed conditions, such as antibiotic exposure, bacteria can develop into persister cells that allow them to stay dormant until the stress is removed, when they can revert back to a growing state. The interconversion of non-growing persister cells and actively growing cells is the underlying basis of relapsing and chronic persistent infections. Eradication and better treatment of chronic, persistent infections caused by Staphylococcus aureus requires a multi- faceted approach, including a deeper understanding of how the bacteria persist under stressed conditions, regulate its cell death pathways, and development of novel drug therapies.

Persisters were first discovered in the 1940s in a staphylococcal culture in which penicillin failed to kill a small subpopulation of the cells. Despite the discovery many decades ago, the specific mechanisms of Staphylococcus aureus persistence is largely unknown. Recently renewed interest has emerged due to the rise of chronic infections caused by pathogens such as

M. tuberculosis, B. burgdorferi, S. aureus, P. aeruginosa, and E. coli. Treatments for chronic infections are lacking and antibiotic resistance is becoming a bigger issue. The goal of our research is to define the mechanisms involved in S. aureus persistence and cell death to improve our knowledge of genes and molecular pathways that can be used as targets for drugs to eradicate chronic infections. Using several high-throughput genetic screens, we identified and confirmed several core regulators of bacterial persistence and cell death upon exposure to bactericidal antibiotics and environmental stresses such as heat and acid stress.

Currently, drug combinations approved for clinical usage do not target persister cells.

One exception is the treatment regimen for tuberculosis that includes pyrazinamide, an anti- persister drug that targets unconventional drug targets such as proteins involved in energy

ii metabolism and trans-translation. Upon the addition of pyrazinamide, which kill persister bacteria to the combination of rifampin and isoniazid which kill growing bacteria, the duration of tuberculosis treatment was shortened from 9-12 months to 6 months. Here, using the treatment for tuberculosis as an example that demonstrates the powerful activity of an anti-persister drug such as pyrazinamide, we screened for drugs with high activity against growing and non-growing forms of S. aureus to formulate drug combinations that can effectively kill the heterogeneous population of bacteria in biofilms. To test this approach in a clinically-relevant animal model, we established a chronic, skin infection mouse model of S. aureus to validate the improved efficacy of drug combinations in clearing persistent skin infections than currently-approved regimens used in the clinics. Our findings reveal that drug combinations consisting of drugs targeting both actively growing bacteria and persister cells, such as the combination of meropenem + daptomycin + clinafloxacin, can eradicate biofilms more effectively than doxycycline + rifampin, a combination used currently on patients. Our in vivo mouse model studies further validate the efficacy of such drug combinations in clearing the bacteria in the lesions, reducing the pathology, and completely healing the chronic lesions formed on the skin. These findings have important implications for treating other persistent bacterial infections.

Advisor: Dr. Ying Zhang

Thesis Committee: Dr. Jie Xiao Dr. J. Marie Hardwick Dr. Meghan Davis Dr. Nicole Parrish

Alternates: Dr. Valeria Culotta Dr. Gary Ketner

iii

ACKNOWLEDGEMENTS

I would first like to express my gratitude to my advisor, Dr. Ying Zhang, for his support and encouragement throughout the years. I gained independence, a wide range of experience in bacteriology research, and confidence as a scientist. Thank you for listening to me when I expressed interest to pursue research on S. aureus and allowing me to conduct research not limited to the genetic basis of S. aureus persistence but also treatment and even diagnostics. I am extremely grateful for your support of my ultimate career goal in becoming a clinical microbiologist and for your creative research ideas that were applicable to the clinical field.

Thank you for giving me the freedom to explore and grow. It has been an amazing time working with the past and present members of the Zhang lab. Thank you all for keeping the lab exciting and lively at all times. I especially have to thank Dr. Wanliang Shi for training me during my rotation in the lab and providing a positive experience that drew me back to joining the lab.

I would like to thank my thesis committee members (both past and present), Dr. Jie Xiao,

Dr. J. Marie Hardwick, Dr. Meghan Davis, Dr. Nicole Parrish, and Dr. Arturo Casadevall for providing constructive feedback and also having conversations about my career goals along the way. In particular, I would like to thank both Dr. Hardwick and Dr. Parrish who have been with me since the beginning of my PhD career, from my preliminary oral exam to my final oral exam.

You both watched me grow and I appreciated all our conversations ranging from personal to professional career advice. The members of the Hardwick lab (Zach, Jason, Heather, and

Madhura) were also instrumental in helping me execute some of my studies. I also need to thank my past exam committee members, Dr. David Sullivan, Dr. Winnie Tang, and Dr. Randy Bryant, who still periodically check up on me about the progress that I have made. I thank you all for the continuous support throughout the years.

Special thanks go to my classmates and colleagues at Johns Hopkins University. To my collaborators, Naina and Andreina, we came to know each other very well after some unsuccessful but then successful experiments together. I am happy to be able to learn a lot from you both but more importantly, we are friends and not simply colleagues. From MMI, a special big thanks to Snow, Phil, Jane, Yuting, Jasmine, and Gail. Thank you all for your friendship, encouragement, and our random conversations after Forum or Seminar. It's great to have a life outside of lab! A bigger thanks to my roommate, Jessie, for being a great roommate, scientist, friend and confidant as we both navigated our PhD journey together. And, the biggest thanks go to my partner, Eric. Thanks for being a great listener and constant supply of humor, optimism, laughter, food, and bubble tea. I always knew you all would have my back and made sure I was not falling through anywhere. I will miss you all but I am confident that we will all do great things in the future and we will cross paths once again.

Lastly, I express much thanks to my family members. Thank you for letting me pursue my dreams and despite the long duration spent in education, you only encouraged me to push further with your unconditional love and support. I thank my parents, especially my father who was a teacher in his past career, for instilling the importance of education and discipline at a young age. By forcing me to memorize my multiplication tables and perfect my cursive before any of my peers did taught me how a strong work ethic during times of struggle will lead to many successes, such as a PhD. Thank you to my sister Jessica for being around to pick up the slack when I could not be home. I'm glad we have each other during the good and bad times.

TABLE OF CONTENTS

Abstract ...... ii Acknowledgements ...... iv Table of Contents ...... vi List of Figures ...... viii List of Tables ...... x Chapter 1: Introduction ...... 1 Staphylococcus aureus ...... 2 Antibiotic Resistant S. aureus ...... 3 Mechanisms of Antibiotic Resistance ...... 4 Drug Resistance Mechanisms in S. aureus ...... 5 Detection of Antibiotic Resistance in Bacteria ...... 6 The Phenomenon of Antibiotic Persistence ...... 7 Persisters and Persistent Infections ...... 8 Biofilms ...... 9 The Yin-Yang Model of Persistence ...... 10 Mechanisms of Persisters ...... 10 Methods of Isolating Persisters ...... 12 Cell Death ...... 13 Hallmarks of Cell Death in Eukaryotes ...... 14 Cell Death in Unicellular Eukaryotes ...... 15 Cell Death in Bacteria ...... 16 Inducers of Bacterial Cell Death ...... 17 Mechanisms of Bacterial Cell Death after Cidal Drug Treatment ...... 18 Methods of Studying Cell Death in Unicellular Organisms ...... 21 Treatment of Persistent S. aureus Infections ...... 23 Approaches in Eradicating Persisters ...... 24 Conclusions ...... 26 Chapter 2: Genetic Screen Reveals The Role of Purine Metabolism in Staphylococcus aureus Persistence to Rifampicin ...... 27 Chapter 3: Identification of a Novel Gene ArgJ Involved in Arginine Biosynthesis Critical for Persister Formation in Staphylococcus aureus ...... 49

Chapter 4: Identification of Genes Regulating Cell Death in Staphylococcus aureus ...... 72

Chapter 5: Drug Combinations Targeting Growing and Persister Cells Eradicate Chronic Staphylococcus aureus Infection ...... 91

Chapter 6: Conclusions and Future Directions ...... 119

Appendix ...... 132

References ...... 145

Curriculum Vitae ...... 158

vii

LIST OF FIGURES

Genetic Screen Reveals The Role of Purine Metabolism in Staphylococcus aureus Persistence to Rifampicin

Fig 2.1: Work flow for isolation of mutants with defect in persistence ...... 41

Fig 2.2: Characterization of genes involved rifampicin persistence ...... 43

Fig 2.3: Confirmation of defective persistence in Pur mutants ...... 45 Fig 2.4: Restoration of persistence phenotype in complemented mutants...... 46 Identification of a Novel Gene ArgJ Involved in Arginine Biosynthesis Critical for Persister Formation in Staphylococcus aureus Fig 3.1: Summary of the genetic screen...... 63 Fig 3.2: Comparison of persistence phenotypes of the ArgJ mutant and parental strain ...... 65 Fig 3.3: Persistence to antibiotics can be restored in the ArgJ mutant by complementation ...... 67 Fig 3.4: Arginine biosynthesis through the Arg pathway plays a role in persistence...... 69 Fig 3.5: The ArgJ mutant has attenuated virulence in vivo ...... 70 Identification of Genes Regulating Cell Death in Staphylococcus aureus Fig 4.1: Identification of genes involved in causing bacterial cell death-resistance and cell death- sensitivity...... 85 Fig 4.2: Strains that are more death-resistant caused more virulent infections in C. elegans...... 87 Fig 4.3: Strains that are more death-sensitive also show decreased persistence to bactericidal antibiotics...... 88 Drug Combinations Targeting Growing and Persister Cells Eradicate Chronic Staphylococcus aureus Infection Fig 5.1: Clinically recommended treatments for chronic S. aureus infections partially eradicate biofilms in vitro...... 110 Fig 5.2: Identification of drug combinations that kill MRSA biofilms ...... 111 Fig 5.3: Evaluation of oritavancin in killing biofilms as a single drug or in combination...... 112 Fig 5.4: Establishing a chronic S. aureus skin infection mouse model...... 113 Fig 5.5: Validation of drug combinations in chronic skin infection model ...... 114 Fig 5.6: Meropenem + Daptomycin + Clinafloxacin reduced immune response of infected skin tissues...... 115

viii

Fig 5.7: Chemical structure of the quinolones tested ...... 116 Conclusions and Future Directions Fig 6.1: A model of potential cell persistence and cell death pathways based on our genetic screen results...... 129 Fig 6.2: Overlapping pathways in S. aureus persistence and cell death found in our genetic screens ...... 130

LIST OF TABLES

Introduction

Table 1.1: Diseases caused by S. aureus ...... 4

Table 1.2: Recommended treatment for persistent infections caused by S. aureus ...... 24

Genetic Screen Reveals The Role of Purine Metabolism in Staphylococcus aureus Persistence to Rifampicin Table 2.1: Primers used in this study ...... 48

Identification of a Novel Gene ArgJ Involved in Arginine Biosynthesis Critical for Persister Formation in Staphylococcus aureus

Table 3.1: Oligonucleotide primers used for qRT-PCR ...... 71

Identification of Genes Regulating Cell Death in Staphylococcus aureus

Table 4.1: Top 4 genes whose mutations result in resistance to cell death in both heat-ramp and acetic acid stress ...... 89

Table 4.2: Top 4 genes whose mutations result in sensitivity to cell death in both heat-ramp and acetic acid stress ...... 90

Drug Combinations Targeting Growing and Persister Cells Eradicate Chronic Staphylococcus aureus Infection

Table 5.1: Drug dosage, scheduling, and administration ...... 117

Table 5.2: Ranking of fluoroquinolones based on their activity in killing biofilms ...... 118

Conclusions and Future Directions

Table 6.1: Overlapping genes found in our persistence screens ...... 131

CHAPTER 1

INTRODUCTION

Staphylococcus aureus

Staphylococcus aureus is a major opportunistic, Gram-positive bacterial pathogen that causes a range of clinical infections. Studies reveal that in some clinical cohorts, S. aureus causes over 80% of the skin and soft tissue infections making it a leading cause of skin and soft tissue infections (SSTI). It is the most common pathogen that can be isolated from surgical site infections, subcutaneous abscesses, purulent cellulitis, and chronic ulcers and wounds [1,

2]. Upon entering the bloodstream, S. aureus can cause invasive infections such as osteomyelitis, osteoarticular infections, endocarditis, meningitis, pneumonia, sepsis, and device-related infections [3] (Table- 1). In the United States, skin infections caused by S. aureus account for over 11 million outpatient visits and 500,000 hospital admissions per year [1]. Infections associated with S. aureus result in a five-fold increase of hospital mortality among patients [4].

Despite the pathogenic nature of the bacteria, one-third of the human population is colonized with S. aureus.

Transmission of S. aureus is mainly due to contact with infected individuals or contaminated abiotic materials and surfaces, allowing this bacterium to be easily transmitted at clinical settings. Thus, people who work in healthcare settings or visit clinical settings at a regular basis are at higher risk of infection. Other risk factors include age (young and elderly), usage of prosthetic devices, HIV infection, organ transplantation, and diabetes. S. aureus bacteremia is usually due to common primary clinical sources of infections such as catheter- related infections, skin and soft-tissue infections, endocarditis and bone infections [5]. In the past two decades, infections caused by S. aureus are a major public health concern given the increasing number of nosocomial infections caused by S. aureus and the widespread emergence of antibiotic resistance as seen in methicillin-resistant S. aureus (MRSA) that is driving an

epidemic of community-associated skin and soft tissue infection. From a MRSA surveillance study, representing geographic areas across the United States and comprised of over 14 million people, the percentage of nosocomial infections caused by MRSA is 76% [6].

Antibiotic Resistant S. aureus

MRSA was initially found to be common in children, athletes, military members, inmates, and men who have sex with men [5]. Among patients with SSTIs, several studies have reported a range of 50-90% of the cases was found to be caused by MRSA [7]. The prevalence of MRSA in causing difficult-to-treat infections such as cardiac implant infections, osteomyelitis, and prosthetic joint infections are up to 51%, 22%, and 12%, respectively [5]. While the overall rates of invasive infections caused by MRSA are declining, MRSA is still responsible for causing over

80,000 invasive infections and 11,000 related deaths per year [8]. Approximately 1-2% of the carriers of S. aureus are colonizers of MRSA.

Healthcare-associated MRSA (HA-MRSA) is used to define a strain isolated from a patient with cultures that were obtained on or after the fourth calendar day of admission and/or a strain that is isolated from patients with residence in a long term care facility [6]. Community- associated MRSA (CA-MRSA) is defined onset of disease that is not related to hospital admittance based on previous classifications above [6]. In the past two decades, there is an increasing prevalence of SSTIs caused by CA-MRSA. A CDC surveillance study consisting of over 14 million people from all geographic areas of the United States, including states such as

California, Connecticut, Minnesota, New York and Tennessee, showed that CA-MRSA was isolated more compared to HA-MRSA in diseases such as bloodstream infections, osteomyelitis, endocarditis, cellulitis, and skin abscesses [6]. During the first outbreak of CA-MRSA in 1990s, a dominant circulating strain was USA400 [5]. Since 2000, USA300 has been the dominating

strain in the United States. In several large cohort studies, studies revealed that up to 85% of

SSTIs and 65% of all CA-MRSA were caused by USA300, further confirming the stable transmission of this strain. Hospital-acquired MRSA (HA-MRSA) strain USA100 and other

CA-MRSA strains such as USA500 are found in low prevalence in the US [6]. Despite the decline in the number of invasive MRSA infections from 2005 to 2011, the CDC and WHO still listed MRSA as a "serious threat" due to S. aureus being a leading cause of hospital-associated infections [5, 8].

Table 1.1. Diseases caused by S. aureus

Percentage caused by S. aureus Disease (or MRSA) Endocarditis 23-33% SSTI: Impetigo 60-80% (4-13%) SSTI: Uncomplicated 31-88% (8-90%) SSTI: Complicated 20-71% (10-70%) Osteomyelitis 32-55% (22%) Native Joint Septic Arthritis 39-79% (2-8%) Prosthetic Joint Infection 18-73% (8-12%) Cardiac Implant Infections 23-46% (51%) Intravascular Catheter Infections 14-41% Pneumonia 2.5-40%

Mechanisms of Antibiotic Resistance

Antibiotic resistance, as exhibited by MRSA and vancomycin-resistant S. aureus (VRSA) strains, is the primary reason why many bacterial infections are increasing globally and now deemed as untreatable. The CDC has listed that antibiotic resistance is one of the greatest threats to human health and in the United States, antibiotic resistant bacteria in general account for

infections of 2 million people and 23,000 deaths annually [9]. Mechanisms of resistance tend to be acquired through horizontal and vertical transfer of resistance genes carried on plasmids or transposable elements between bacteria. Point mutations in the bacterial genome can also introduce for new resistant traits; for example, in Mycobacterium tuberculosis, resistance is mainly generated through genomic mutations.

In general, resistance mechanisms fall in several main categories [10]: 1) decreased penetration of the drug into the bacterial cell either due to enhanced efflux or decreased influx via activity of efflux pumps or porin expression, 2) modification of the drug target, 3) inactivation of the antibiotic by bacterial enzymes causing hydrolysis of the drug or direct modification of chemical groups on the drugs, and lastly, 4) inactivation of a bacterial protein that allows for drug activation (e.g. isoniazid or pyrazinamide used in tuberculosis treatment need KatG and PncA, respectively, for the drugs to be converted to their active products). Overexpression of efflux pumps can confer high levels of resistance and due to transport of a wide range of substrates, multidrug resistance may arise.

Drug Resistance Mechanisms in S. aureus

Due to S. aureus’s resistance nature and high prevalence of MRSA in many clinical infections, it should briefly be noted that methicillin-resistance is conferred by the acquisition of the mecA gene that is located on the staphylococcal cassette chromosome mec (SCCmec) element which allows for modification of the target for methicillin (and related classes of drug).

The mecA gene encodes for a variant of the penicillin binding protein (called PBP2a) that has lower binding affinity against β-lactam antibiotics and as a result, PBP2a can continue to participate in cell wall synthesis even in the presence of β-lactams [11].

On the other hand, vancomycin is a glycopeptide that targets the D-alanyl-D-alanine residue on the building blocks of peptidoglycan and prevents transpeptidase from cross-linking the peptidoglycan layer and forming the cell wall [12]. Intermediate-resistant S. aureus (VISA) strains mainly harbor mutated genes such as walkR, vraSR, and the two-component sensory system yvqF/vraSR that are involved with cell wall metabolism and biosynthesis [13, 14].

Vancomycin-resistant S. aureus (VRSA) strains were shown to harbor a transposon insertion element Tn1546 which altered vancomycin’s target of D-ala-D-ala to d-alanyl-d-lactate (D-ala-

D-lac), a dipeptide residue with much lower affinity for the antibiotic [15].

Additionally, multi-drug resistance can also be found in S. aureus which is attributed to activity of the efflux pumps encoded by the lmrs gene [16]. The chloramphenicol-florfenicol resistance methyltransferase protein, encoded by the cfr gene, methylates the A2503 site in the

23s rRNA causes resistance to many drugs such as phenicols, streptogramins, lincosamides and oxazolidinones as their drug targets have been modified [17].

Detection of Antibiotic Resistance in Bacteria

To determine the drug susceptibility profile of a particular bacterial strain, the bacterial samples are subjected to antibiotic susceptibility tests in clinical laboratories. A classical method to perform antimicrobial susceptibility testing is the macrobroth dilution method [18].

Antibiotic-containing tubes (usually prepared in two-fold dilutions of antibiotics) are inoculated with bacterial suspension and the tubes are examined for visible bacterial growth usually after

18-20 hours. The minimal inhibitory concentration (MIC) of the tested drug is determined as the lowest concentration of antibiotic that can prevent visible growth. Another popular method is the disk diffusion susceptibility method which is performed by applying paper antibiotic disks onto a lawn of bacterial on an agar plate [19]. Upon incubation of plates for 16–24 hours, the zones of

growth inhibition around each of the antibiotic disks are measured. The susceptibility of the strain is determined by the diameter of the growth inhibition zone. For both the broth dilution method and disk diffusion assay, the MIC values and diameters of the zone of growth inhibition, respectively, are compared to the clinical breakpoints published the Clinical and Laboratory

Standards Institute which will determine if the isolate is considered “Resistant”, “Intermediate

Resistant”, or “Sensitive” [20]. In general, resistant isolates will have either a higher MIC or small zone of inhibition than their sensitive counterparts. It should be noted here that these assays as well as the definition of “resistance” in general describes growing bacteria and how bacteria can still grow under stressed (e.g. antibiotics) conditions.

The Phenomenon of Antibiotic Persistence

Despite long periods of antibiotic treatments, numerous bacterial species such as

Mycobacterium tuberculosis, Escherichia coli, Pseudomonas aeruginosa, Borrelia burgdorferi,

Chlamydia trachomatis, Treponema pallidum, and S. aureus are able to persist inside the host

[21, 22]. Bacterial persistence can take place in different tissues within the host and lead to both asymptomatic, symptomatic infections, and periodic latent and active states of disease. For example, at least 10% patients who are infected with M. tuberculosis can have relapse during their lifetime [23]. Similarly, 27% of the women who are infected with uropathogenic E. coli

(UPEC) will have recurrent infections [24]. Persistent infections are associated with at least half of the infections pertaining to indwelling or prosthetic medical devices as bacteria can form into biofilms, a protective extracellular matrix containing secretions of polysaccharides and extracellular DNA that shields the bacteria from external stresses (e.g. immune response, antibiotics).

Infections persist in the host due to several factors with the obvious ones being ineffective clearance by the host and ineffective clearance by antibiotics. Bacterial pathogens can induce anti-inflammatory response or harbor virulence factors that evade the host immune response. For example, M. tuberculosis can induce IL-10 to suppress IFN response and produce superoxide dismutase that allows the bacteria to avoid degradation in phagosomes and detoxification of reactive oxygen species [25]. Ineffective clearance by antibiotics can be attributed to the high prevalence of infections caused by antibiotic resistant strains as well as poor patient compliance and poor pharmacokinetics of drugs used [26, 27]. However, another phenomenon underlying why persistent infections exist is the role of persister cells.

Persisters and Persistent Infections

Gladys Hobby in 1942 first discovered that penicillin was only able to kill 99%, leaving 1% of the bacterial population despite drug treatment [28]. In 1944, Joseph Bigger coined the term

“persisters” after he had a similar observation in which penicillin could not kill 1% of the

S. aureus cultures [29]. Additionally, he observed that the small surviving population of bacteria was susceptible to penicillin again upon regrowth of the bacteria. His hypothesis that these

“persisters” harbor a phenotypic switch between dormancy and growing states during drug- treated and drug-removed conditions, respectively, was supported 60 years later through a microfluidics live imaging experiment that showed that a single bacteria persister cell tolerant to a cidal antibiotic was not growing [30]. Hence, growth arrest and a decrease in metabolic activity are often associated with persister cells.

Increasing evidence from clinical studies now show that bacterial persisters have a role in chronic infections and post-treatment relapse. From clinical isolates of patients with cystic fibrosis, high-persistence mutants were found in samples obtained later in the infection than

samples from an earlier infection [31]. Similarly, high persistence mutants of E. coli were isolated from urinary tract infections and persisters of Candida albicans were also found in patients with prolonged therapy of topical chlorhexidine [32]. Given the difficulty in treating biofilm infections, it is no surprise that a link exists between the persistence of persister cells and biofilm formation of both bacterial and fungal pathogens [32-35].

Biofilms

Chronic infections such as cystic fibrosis are associated with P. aeruginosa biofilms and patients with UTIs can have UPEC biofilm formation on the bladder epithelium [36, 37].

S. aureus also readily form biofilms on both biotic and abiotic surfaces such as the bone on osteomyelitis patients, heart valves from endocarditis patients, and implants from patients with prosthetic joint infections, and cardiac device infections; all these disease listed are notoriously recalcitrant to antibiotic treatment [5].

While it is widely recognized that biofilms hinder penetration of antibiotics to reach the bacterial cells, there have been studies that suggest that antibiotics can penetrate biofilm with relatively great efficiency [38, 39]. Hence, another hypothesis to why biofilms are difficult to treat is the presence of persisters within the biofilms [34, 35]. For example, persister cells of C. albicans that are tolerant to amphotericin B and chlorhexidine were detected in biofilms and not in exponentially growing or stationary-phase planktonic populations [33]. Persister cells are heavily enriched in stationary phase of growth due to nutrient deprivation which can decrease cellular metabolism. Within a biofilm, similar conditions exist; the bacteria cells are in high cell densities and nutrient and oxygen deprivation within the biofilm induces for a reduction in growth rates and a drop in intracellular ATP production and persister cell formation [40, 41] .

And, as a matter of fact, the three-dimensional architecture of a biofilm is analogous to the heterogeneity of a bacterial population [42] .

The Yin-Yang Model of Persistence

To describe the dynamics and heterogeneity of cells with different and interconverting metabolic states of the bacterial population, a Yin-Yang model has been proposed [21]. In red,

(i.e. Yang), are the growing cells and in black, (i.e. Yin), are the non-growing persister cells.

Within the Yang, there are remains of Yin that causes latent diseases and within the Yin, there is the Yang that can cause relapses in infection once the non-growing cell reverts back to a growing state. This Yin Yang model can also be attributed to the heterogonous population in a biofilm.

Using fluorescent tags for specific metabolic markers, studies have shown that cells at the surface of the biofilm are highly active and rapidly growing and the activity gradually decreases towards the center of the biofilm [42]. Another group showed that oxygen levels are depleted by as much as 30-fold near the center of the biofilm [41]. As such, the interplay of bacteria surviving in various environmental conditions, whether in planktonic or biofilm forms, is a rather complex continuum with growing and dormant states that can be described and simplified using the Yin-Yang model.

Aside from persister cells, other persistent forms include the slow-growing small-colony- variants (SCV) of S. aureus which persist in tissues, L-forms, and the viable but non-culturable

(VBNC) bacterial forms found in environmental samples such as Vibrio, Shigella , Listeria, and

Francisella [43-45]. These types of persistent forms contribute to persistence whether it is in the host or environment and can be grown or "resuscitated” upon ideal growth conditions.

Mechanisms of Persisters

While persistence is not a phenomenon due to a genetic change but is rather marked by

changes in epigenetic factors causing the bacterial cell to induce expression of persister related genes, most molecular understanding of persistence is done through a mutagenesis approach [21].

Similarly, other groups found that genes important for sporulation were also identified by mutagenesis approach even though the sporulation process is known to be an epigenetic change, also due to environmental stress cues. Performing a mutagenesis screen, Moyed et al. identified a mutant in E. coli that formed a greater proportion of persister cells. The mutation mapped to a toxin-antitoxin module, hipAB, where HipA is the toxin that phosphorylates the glutamyl-tRNA synthetase and as a result inhibits the loading of tRNAGlu and arrests protein synthesis [46]. As toxin-antitoxin modules encode both a toxin and antidote that neutralizes the activity of the toxin, the possibility for a population to contain both growing and non-growing cells can be explained due to a toxin-antitoxin’s bistablity. Studying the role of other TA modules revealed modules such as relBE, mazEF, tisAB, mqsR, cspP, past, hhA, and hokA all to be involved in persister cell formation [21, 22].

In summary, a review of the literature (Ref 19-20) regarding mechanisms of persistence, focused primarily on E. coli, has suggested that most genes involved in persister formation fall under several pathways such as toxin-antitoxin modules (hipBA, relBE, mazEF, tisAB), energy production (ubiF, sucB, glpD), cellular metabolism (phoU, obgE), stringent response/DNA repair (relA, dskA, lexA, recA, dps, xerC, xerD), antioxidant defense (catalase, superoxide dismutase), and trans-translation processes (ssrA, smpB, rpsA) [21, 22].

However, recent studies reveal that in E. coli and S. aureus, persistence is not necessarily due to TA-modules and perhaps, metabolic regulation is the main driver of persister formation and TA-modules are downstream effectors that are used to control and maintain decrease in metabolic status of the cell [47]. While there is a plethora of genes potentially involved in

persistence and debates for whether TA-modules are the main drivers of persistence, an agreed upon characteristic is that ATP depletion and decrease in cellular metabolism are important for persistence [47].

Mechanisms of persister formation are not well understood because persisters represent a small subpopulation of the cells, persister formation can be heterogeneous, transient and dependent on many factors such as the age of bacterial culture, the starting inoculum size, the type (or class) of antibiotics used, antibiotic concentrations, length of antibiotic exposure, and aeration [21, 48]. Environmental cues such as nutrient limitation, extreme pH, and DNA damage or the complex environments within macrophages can induce for persister cell formations as well.

Hence, the persisters under one condition may not be the same persisters as induced by other stressors. While many specific genes have been identified, what and how bacterial cells sense and coordinate the activation (in a timely and sequential manner) of these pathways remains elusive. Additionally, it appears that mechanisms used to form L-forms, biofilms, and even cancer stem cells may arise from similar pathways [21, 44]. While the genes do not show significant homology, the pathways have similar functions. Hence, core proteins, regulators, and players of the persister pathways are imperative to understanding persistence and designing drug targets for killing persistent forms.

Methods of isolating persisters

The current assays to isolate persisters are mainly based on exposing bacterial cultures to bactericidal antibiotics (cell wall inhibitors, aminoglycosides, or quinolones) for a period of time and enumerating the amount surviving bacteria by colony-forming units. Exposure times of short intervals such as 2-6 hrs or longer times of up to 20 days have been done to induce persister formation [48-50]. Some studies will apply the stress to stationary phase cultures directly

whereas some resuspend or dilute stationary phase cultures but the latter method will result in fewer persister cells as bacteria are being reverted back to growing states after being refreshed in rich media which restores their susceptibility to drugs [51]. For persister counts, washes to remove drugs and plating on agar plates with no drugs are crucial in recovering the surviving bacteria as leaving drugs in the media will not allow the cells to revert back in growing forms.

Since persisters are cells with lower metabolic status, groups have developed unstable

GFP-variants or use fluorescent dyes as a proxy for metabolic activity to separate dormant and growing populations [52]. With technological advancements and increased usage of -omics, promoters and proteins activated in stationary phase have been identified [47]. Strains with fluorescently tagged persistence reporters are made and then flow cytometry and fluorescence- activated cell sorting (FACS) techniques are employed to enrich for persister cells. Single-cell techniques such as microfluidic devices also allow researchers to follow the trajectory of one cell during stress exposure [30].

Cell Death

In mammalian systems, programmed cell death processes such as apoptosis, necroptosis, pyroptosis, and ferroptosis eliminate cells during developmental processes or to rid of defective cells following stress-induced cell damage and pathological states and infection [53, 54]. PCD usually describes a series of genetically regulated processes that are analogous to “cell suicide”, which can be important for survival of multicellular organisms. Elimination of cells is essential during embryonic development such as finger and organ development as well as maintaining the homeostasis of many tissues, notably the immune system which can lead to autoimmune diseases or cancer [55, 56]. Furthermore, damaged or defective cells through exposure to infectious

agents, chemicals, or environmental factor such as heat, mutagens, or toxins need to be prevent their potentially damaging effects to the organisms as a whole.

Hallmarks of Cell Death in Eukaryotes

During the process of cell death in eukaryotes, a series of biochemical and morphological events occur to accompany the dying process. One of the better characterized pathways of eukaryotic PCD is apoptosis, a form of cell death that is currently defined as caspase-3 dependent [53, 57]. In mitochondria-mediated apoptosis, depolarization of the mitochondrial membrane related to the opening of an outer mitochondrial membrane pore to releases cytochrome c into the cytoplasm to initiate the apoptosis pathway [58]. Upon induction of apoptosis, downstream molecules and pathways lead to formation of reactive oxygen species

(ROS) [59]. Hallmarks of eukaryotic apoptosis include chromatin condensation, DNA fragmentation, permeability of the cytoplasmic membrane, membrane blebbing and exposure of phosphatidylserine phospholipid on the outer membrane [53, 60-62]

To regulate mammalian cell death, there are three major groups of proteins that fall under the B cell lymphoma 2 (BCL-2) family. The pro-apoptotic proteins include BAX, BAK, and

BOK; the anti-apoptotic proteins BCL-2, BCL-XL, BCL-W, MCL1); the essential initiators of apoptosis that antagonize the anti-apototic proteins and/or activate the pro-apoptotic proteins are the “BH3-only proteins”. Anti-apoptotic proteins inhibit the function of Bax proteins. However,

BH3-only proteins can regulate apoptosis by blocking the effect of Bcl-2 proteins or by directly stimulating the pro-apoptotic Bax proteins [63]. Once Bax oligomerizes in the mitochondrial outer member, mitochondrial outer membrane permeabilization occurs followed by the release of cytochrome c to activate the apoptotic caspase cascade [64].

It is important to note here that many other death pathways exist. Necroptosis is death that is mediated by kinases such as RIPK1 and RIPK3 causing cell lyses through the activation of pore-forming function of mixed lineage kinase domain-like (MLKL) [65, 66]. Pyroptosis is caspase-1 dependent. The cells die when Gasdermin proteins are cleaved and activated in the cytoplasm to form pores in the cell membrane [67]. The release of IL-1beta leads to cell death.

Lastly, ferroptosis is the less well defined of them all. It is believed that the cell membrane is damaged through a process that is iron-dependent [54]. Hence a complex regulatory control of proteins working in tandem is needed to signal the cell to activate death pathways.

Cell Death in Unicellular Eukaryotes

Cell death pathways were defined characterized in multicellular organisms. The yeast genome of Saccharomyces cerevisiae does not contain any homologues of apoptotic regulators such as BAX or BCL2 family proteins. Because the current definition of mammalian apoptosis is defined as caspase-3 dependent cell death, the term of apoptosis cannot be correctly described in yeast as yeast do not harbor true caspases. Although fungi encode metacaspases, they are not readily comparable to mammalian caspases and do not undergo the same regulation [68].

However, fungal species have been reported to exhibit features of apoptotic mammalian cells

(e.g. chromatin condensation, phosphatidylserine exposure) but unlike mammalian cells where these can be attributed to the actions of caspase-3, there is currently no compelling evidence that these events in S. cerevisiae are due to the evolutionarily selected mechanisms [69-71].

Several studies have suggested the importance of the transcription factor Haa1p and the

MAPK pathway consisting of Hog1p to play a role in regulating PCD of S. cerevisiae during acetic acid stress [72-74]. To perform a more unbiased, thorough search of genes that may

regulate cell death in yeast, genome-wide screens using a yeast knock-out haploid mutant collection were executed. The screens showed that metabolism, specifically lipid catabolism, was a main regulator of death in yeast. The other pathways of significant importance such as proteins involved in mitochondrial function, vesicular trafficking, amino acid transport and biosynthesis, oxidative stress response, cellular metabolism, and protein translation and post-translational modifications [75-77].

Cell Death in Bacteria

If the purpose behind cell death pathways are considered to be “cell suicide” or with altruistic properties, then it may not be immediately clear as to why unicellular bacterial organisms will commit suicide and for what benefits will they gain. However, it is important to recognize that bacteria live and die in complex communications that have properties of a multicellular population. For example, communication among bacterial cells via release of pheromones or peptides, known as quorum sensing, allow the bacteria to adapt to changing environments, especially during times of stress [78]. Furthermore, the release of virulence factors upon cell lysis can increase the fitness of the bacterial population and adaptability of bacteria as they pick up extrachromosomal elements among themselves. Hence, upon exposure to a damaging event or agent, the self-sacrifice of a proportion of the bacterial cells to donate their nutrients instead of depleting limited resources can help enhance the survival of the few that can remain. These complex communications can include biofilms or fruiting bodies of bacteria with different morphological forms [79].

Biofilms are bacteria that are encapsulated within a matrix of exopolysaccharides, proteins, and released DNA from bacterial cells. The added layer provides a physical protection

of the intracellular cells within the biofilms from access to antibiotics and recognition of the immune response; the structural development as demonstrated by the protective barrier in a biofilm is made possible by the release of cellular components from altruistic self-sacrifice of a proportion of the bacterial cells [80].

Similar to multicellular organisms where developmental processes require temporal regulation of cell death pathways, the Gram-positive bacterium Streptomyces forms long, multi- nuclei hyphae that eventually will die and differentiate into chains of spores. The dying hyphae can be observed to have disorganization of the nucleoid structure, DNA digestion, and retraction of the cell membrane from the cell wall, all of which resembles mammalian cell death [81].

Similarly, the sporulation process of Bacillus subtilis is also regulated by an autolytic cell death program. The mother cell undergoing autolysis under stressed conditions to release the spores may provide the kin cells extra nutrients and energy source to sporulate [82].

Inducers of Bacterial Cell Death

To study cell death in bacteria, most studies utilize bactericidal antibiotics, drugs that induce cell death, as opposed to bacteriostatic antibiotics, drugs that inhibit cell growth. Most of the bactericidal antibiotics fall under the categories of DNA, RNA, cell wall, or protein synthesis inhibitors.

The direct effects of the bactericidal antibiotics to their respective targets are well characterized. The quinolones class of antibiotics inhibits DNA synthesis by binding to topoisomerases and trapping the enzyme at the DNA cleavage stage and preventing the strands from rejoining. Hence, the DNA replication machinery becomes arrested at the replication fork which will inhibit DNA synthesis [83]. The rifamycin class of antibiotics inhibits RNA synthesis.

Rifamycins inhibit DNA-dependent RNA transcription by binding to the β-subunit of the RNA polymerase [84]. This high affinity binding of the drug can sterically inhibit RNA strand initiation and negatively impact cellular transcription. β-lactams and glycopeptides are antibiotics that interfere with cell wall synthesis. The β-lactam antibiotics block the crosslinking of the peptidoglycan units by inhibiting the formation of the peptide bond that is mediated by penicillin-binding proteins (PBP). The β-lactam is actually an analogue of the terminal D-alanyl-

D-alanine dipetitide of the peptidoglycan units and acts as a substrate of PBP, disabling the active site of the enzyme crucial for cell wall crosslinking. Similarly, cell wall synthesis inhibitors also include glycopeptides, such as vancomycin. In contrast, glycopeptides bind the peptidolygcan units at the D-alanyl-D-alanine site and block the activity of transglycosylase and

PBP. Glycopeptides can be considered as steric inhibitors of the peptidoglycan maturation [85,

86]. Lastly, the aminoglycoside class of antibiotics inhibits protein synthesis. These drugs bind the 16S rRNA of the 30S ribosome and induce a conformational change in the complex formed between an mRNA codon and the charged aminoacyl tRNA. Consequently, protein mistranslation due to tRNA mismatching and incorporation of inappropriate amino acids ultimately contributes to cell death [87].

Mechanisms of Bacterial Cell Death after Cidal Drug Treatment

Despite understanding and knowing the drug-target interactions of bactericidal drugs, the specific sequence of events after the drug binds to its respective target are not necessarily clear.

Using high-throughput genetic screens or bioinformatics analyses of gene expression profiling studies, a potential common pathway for bacterial cell death can be suggested [88]. All bactericidal antibiotics, despite the differences in their main drug target, can induce lethal hydroxyl radicals in Gram-negative and Gram-positive bacteria. Central metabolism consisting

of increased TCA cycle and cellular respiration can destabilize superoxide-mediated iron-sulfur clusters and induce the Fenton reaction [89]. Additionally, cell envelope and two-component signaling pathways activated by redox stress are affected by the distinct class of bactericidal drugs as well. Bacteria were less susceptible to bactericidal antibiotics when iron chelators were added. Mutants that have impaired formation of iron-sulfur clusters were also less sensitive to bactericidal antibiotics [89]. In summary, a common oxidative damage cell death pathway appears to be significant in the killing activity of all the bactericidal drugs.

While the specific mechanisms of how bacteria cells die upon cidal stress is unknown, there appear to be some parallels between mammalian PCD and bacterial cell death. Upon treatment of ampicillin, norfloxacin, and gentamicin, three drugs belonging to functionally distinct bactericidal drug categories, E. coli exhibited physiological and biochemical hallmarks of cell death. Specifically, treatment of the three drugs resulted in ROS generation, DNA fragmentation, chromosomal condensation, membrane depolarization, and phosphtidlyserine exposure. Treatment with bactericidal drugs has been shown to induce expression of a protein with specificity for caspases [90]. However, the reporter used in these studies can be highly promiscuous and can report the activity of many cysteine proteases across species.

One of the well-characterized mechanistic explanations for bacterial cell death is through activation of TA modules. It was originally observed that extrachromosomal elements such as plasmids that harbored TA modules were capable of preventing bacteria from being killed by a stable toxin (e.g. protease) that is produced inside the cell. It was then revealed that TA pairs contain a high expression of an antitoxin, the antidote for the toxin gene that is transcribed at low levels [91]. The regulatory nature of TA modules’ in bacterial cell death prompted much

investigation into the mechanisms of TA - the best characterized being MazEF- in causing cell death.

In the MazEF module, the toxin is MazF and the antitoxin is MazE. The MazEF pathway is activated upon stress signals such as nutrient starvation, heat, DNA damage, oxidative stress, and antibiotics [91, 92]. After degradation of MazE (the antitoxin), MazF (the toxin) as a site- specific endoribonuclease cleaves mRNA at ACA sequences and as a result, inhibit translation of protein products within the bacteria. It has been proposed that MazF can selectively cleave transcripts that encode for anti-death proteins through unclear mechanisms [93]. Since the identification of MazEF’s role in cell death, other types of TA modules have been shown to induce a cell death phenotype. The entericidin locus of E. coli and Citrobacter freundii contains

EcnA and EcnB which are two envelope lipoproteins that promote bacteriolysis under stationary phase and high osmolarity conditions [94]. In B. subtilis, the spollS locus contains the toxin

SpollSA and putative antitoxin SpollSB, which is degraded upon activation of the sporulation process and allow for SpollSA to induce for lysis after septation [95].

To search for more molecular and genetic mechanisms of bacterial cell death, researchers turned to findings from mammalian cell death. As described previously, members of the BCL-2 protein family are crucial drivers and inhibitors of the cell death pathway in the eukaryotic process. The CidA and LrgA proteins, which were identified in S. aureus, are holin proteins that have analogous functions to the pro-apoptotic and anti-apoptotic functions of BCL-2 proteins, respectively [96, 97].

It has also been shown through several studies that bacterial cell death can be induced by localization of the death effector domain of Fas-associated death domain protein (FADD), a

mediator of apoptosis [98, 99]. Additionally, expression of human proteins BAX and BAK have been shown to oligomerize in the cell membrane to induce cell death in bacteria [100]. Co- expression of anti-apoptotic Bcl-XL showed inhibitory effects of bacterial cell death suggesting that mechanisms of cell death among eukaryotes and prokaryotes may be conserved and CidA and LrgA may well be regulators of bacterial cell death utilizing similar mechanisms as in eukaryotes.

Both CidA and LrgA oligomerize in the bacterial cell membrane. It has been proposed that LrgA mediate inhibition of death effector protein CidA. Upon SOS responses or activation of TA systems, LrgA is released and CidA, through a poorly defined mechanism, alters the posttranslational regulation of peptidoglycan hydrolase and nuclease activity. A decrease in bacterial cytoplasmic membrane depolarization ultimately leads to cell death by autolysis.

Exposure of cidal drugs such as penicillin, vancomycin, and rifampin to bacterial strains with a mutation in CidA or overexpression of LrgA showed increase tolerance and reduced rate of cell death suggesting that there is a specific cell death response that is dependent on CidA expression

[96, 97].

Methods of Studying Cell Death in Unicellular Organisms

To detect mammalian cell death, the assays used include visualization of cell death morphological and phenotypical hallmarks through staining of the phosphatidylserine of the plasma membrane (e.g. Annexin V staining), nuclear DNA fragmentation (e.g. TUNEL labeling), activation of caspase-like proteins (e.g. FITC-tagged fluorescent probes), and cytochrome c release from the mitochondria [101]. Despite being used in yeast and bacteria, the accuracy of these assays in detecting cell death in such unicellular organisms has been decisively rejected

[102]. Both yeast and bacteria do not contain true caspases and the physiological role of cytochrome c release in yeast remains unknown.

In bacteria, most cell death assays include exposure to bactericidal drugs or extreme levels of environmental stresses followed by determination of colony forming units (CFU) count.

Similar to bacteria researchers, yeast researchers define yeast cell survival as colony growth.

However, growth on solid medium only determines whether the cell is growing and is not a readout for cell viability. It does not fully depict whether a cell is truly dead or in the process of dying since colony formation may be not possible. (In bacteria, there are at least 85 species that are viable but non-culturable and among these 85 species, 67 species are pathogenic [103].) The cell death assays developed in yeast by Teng et al have capabilities of performing high- throughput screens and uses both CFU formation and viability staining to quantify cell viability

[104, 105].

One assay developed by Teng et al. utilized a thermocycler with narrow temperature variance that is programmed to slowly ramp up the temperature [105]. This assay was designed to account for a death process that may be an active process (and not a direct assault) which allows further insight into the detailed mechanisms by which regulation of cell death can occur.

Upon growth of yeast strains chosen at the stage of growth according to researcher’s interest, a standardized amount of cells was normalized by OD600 and ramped from room temperature to

30oC immediately, held at 30 oC for 1 min, and then ramped from 30 oC to either 55oC or 62oC depending on the growth phase of the starting inocula [105]. Teng et al. developed another assay using acetic acid stress as a death stimulus. Upon growth of cultures to the growth stage desired, a final concentration of 199 mM or 242 mM of acetic acid was added to the culture and incubated at 30 oC for 4 hours on a rotator. Due to the lack of mammalian apoptotic proteins in

yeast, Teng et al. optimized their conditions by using a yeast clone lacking Dmn1/Drp1, a dynamin-like GTPase with pro-death functions, to ensure that cell death program (and not simply cell assault) is being activated [104].

Treatment of Persistent S. aureus Infections

Different types of persistent infections caused by different bacteria may require their own respective treatment strategies. Complete cure is a combination of the appropriate drug treatment and host immune response. While antibiotic resistance is a problem, doctors usually prescribe antibiotics that have activity toward the strain isolated from the patient. Hallmarks of such persistent infections usually include treatment with drug combinations, prolonged antibiotic treatment, as seen in the treatment of tuberculosis [106].

Despite such therapeutic regimens, the rate of recurrence, reactivation of latent disease, and failure to achieve complete clearance is high among those infected with persistent infections.

A cohort study revealed that among patients with S. aureus bacteremia, a relapse rate of 8% was found in those who received less than two weeks of therapy [107]. Patients with osteomyelitis caused by MRSA receiving less than 8 weeks of treatment were 4.8 times more likely to relapse than those receiving longer therapy [108]. In extreme cases for those with prosthetic implants such as in joint implants, debridement procedures are practiced in addition to antibiotic therapy.

Yet, even with invasive therapies, the cure rate ranges from 52-80% of the patients [109-111].

Current clinical treatments for persistent and chronic MRSA infections are summarized in Table 1.2 [112] . The recommended time to treat a typical infection for S. aureus is usually around 1 week of antibiotic therapy. However, chronic infections may require any duration ranging from at least 1 week (for recurrent soft tissue infections) and up to 3 months (for osteomyelitis). During the course of the antibiotic therapy, the patient may need to continuously

check for bacterial load and in cases of prosthetic implants, invasive debridement or surgical drainage procedures are recommended.

Table 1.2. Recommended treatment for persistent infections caused by S. aureus

Condition Therapy Bacteremia Vancomycin once daily for 4 weeks Endocarditis, Native Valve Vancomycin or daptomycin once daily for 6 weeks Endocarditis, Vancomycin plus rifampin every 8 h for at least 6 Prosthetic Valve weeks plus gentamicin every 8 h for 2 weeks Rifampin-based combination therapy with TMP- SMX, doxycycline-minocycline, clindamycin, or Osteomyelitis a fluoroquinolone for 1-3 months See "osteomyelitis" therapy plus rifampin daily for 2 weeks followed by rifampin plus a Device related fluoroquinolone, TMP-SMX, a tetracycline or osteoarticular infections clindamycin for at least 3 months Skin and Soft Tissue Vancomycin, or linezolid twice daily, or Infections daptomycin once daily for 1 week Recurrent soft tissue Oral antibiotics (e.g. doxycycline) + rifampin for infections at least 1 week

Approaches in Eradicating Persisters

For development of more effective treatments for persistent infections, many different approaches have been proposed. Under the “one drug, one target” approach that has been commonplace for drug discovery, the more direct way is to discover drugs that can directly kill persisters or target a persistence pathway. For example, the acyldepsipeptide ADEP4 can activate the ClpP protease and degrade many proteins resulting in death of S. aureus persister cells [113].

Another strategy is to enhance the activity of current antibiotics by altering the metabolic state of the persister to a growing or higher metabolic state so they respond to treatment. For example, adding the sugar mannitol can enhance killing of E. coli persisters by aminoglycosides in vitro and in vivo in mouse models [114]. This potentiation of the antibiotic is due to increase in the proton motive force needed for uptake of the antibiotic. Another version of utilizing

“woken up” cells is by the process of drug pulsing [115]. After a period of drug treatment, the drugs are removed so the persisters will wake up and then another dose of drugs is given. The cycle is repeated for several times with the hopes that more persisters are killed after each cycle of pulsing. However, this approach works only for cidal antibiotics and not for static antibiotics

[116].

Lastly, due to the heterogeneous population within a bacterial infection, another reliable approach to ensure all the forms are being killed within the population is using a drug combination that can kill both the growing and dormant forms. A good example of this approach is the treatment designed to treat tuberculosis. Tuberculosis is treated with a combination of drugs, isoniazid, rifampin, pyrazinamide and ethambutol. After pyrazinamide as a persister drug

[117] was added to the combination, the duration of TB therapy was shortened from 9-12 to 6 months. Research has shown that pyrazinamide is an unconventional drug as it acts upon dormant persisters while the other drugs act upon growing stages of M. tuberculosis [118, 119].

Hence, the power of pyrazinamide in shortening TB therapy illustrates the important principle that a strategic combination of drugs that can target both growing and non-growing persisters in the form of Yin-Yang model will lead to more effective and short treatment for chronic infections [21].

Conclusions

Persistent S. aureus infections are extremely burdensome to treat as demonstrated by prolonged treatment periods and utilization of different drug combinations. One of the underlying reasons why persistent infections occur is due to the presence of persister cells.

Despite the discovery of persister cells from Staphylococcal cultures in 1940s, the mechanisms of persister cell formation for S. aureus are largely unknown as most research is done in E. coli.

To develop better therapies to treat persistent infections caused by S. aureus, identification of druggable targets to reduce persister cell formation and increase cell death is crucial. Upon identification of anti-persister drugs, more effective drug combinations, that consists of agents targeting both growing and non-growing populations, can be made.

CHAPTER 2

GENETIC SCREEN REVEALS THE ROLE OF PURINE METABOLISM IN

STAPHYLOCOCCUS AUREUS PERSISTENCE TO RIFAMPICIN

This Chapter has been published previously and is copyrighted by MDPI in: Yee R., Cui P., Shi W., Feng J., Zhang Y. 2015. Genetic Screen Reveals the Role of Purine Metabolism in Staphylococcus aureus Persistence to Rifampicin. Antibiotics. 4(4), 627- 642; https://doi.org/10.3390/antibiotics4040627

Introduction

Staphylococcus aureus (S. aureus) infections cause an intense burden on healthcare throughout the world, as the methicillin-resistant S. aureus (MRSA) strain accounts for a majority of the infections present in hospital environments [120]. Infections with MRSA are difficult to treat due to their resistance to multiple antibiotics and highly invasive nature [121].

In particular, USA300 is a MRSA strain that was first isolated in infections among football players in the state of Pennsylvania in the United States in 2000 and has since been discovered in Europe, South America and Australia [122, 123]. This invasive strain has been shown to cause infections in community-associated infections, especially in traditionally low risk groups such as children in daycare, inmates in prisons, and military officials [124].

Many S. aureus infections, including those caused by USA300, can develop into persistent and recurrent infections such as endocarditis and biofilm infections due to the presence of bacterial persisters.

Persisters are quiescent organisms that survive exposure to bactericidal drugs and stresses but are still susceptible to drugs and stresses upon exiting dormancy. Persisters are genetically identical to the rest of the population of cells but exhibit a phenotype that allows them to be drug-tolerant [21]. Most studies on persistence mechanisms are done in Escherichia coli (E. coli). Thus far, pathways such as toxin-antitoxin, SOS response and

DNA repair, signal transduction, membrane stress, energy production, phosphate metabolism, and protein degradation are important for persister formation [21].

A majority of the literature that implicates mechanisms of persister formation in S. aureus stems from understanding the development of small colony variants (SCVs) that are isolated in persistent infections from patients. The unstable nature of SCVs that revert back to

the normal phenotype provides a mechanism for relapsing infections [45]. Earlier works suggest that electron transport and thymidylate biosynthesis were associated with SCV formation. SCVs have defects in genes involved in the respiratory chain such as hemB, which is involved in hemin biosynthesis, and menB, which is involved in menaquinone production

[125]. Additionally, S. aureus can form cell-wall deficient bacteria called L-form bacteria, named after the Lister Institute, which is where these morphological variants were discovered, that can evade antibiotic activity and the host immune response [126]. Using a transposon mutant library of S. aureus to identify mutants defective in unstable L-form formation, Han et al. identified that glpF involved in glycerol uptake was critical for L-form formation and persistence to antibiotics in S. aureus [127]. Although SCV and L-form bacteria are implicated in causing relapsing and persistent infections, the mechanisms of S. aureus persister formation are largely unknown.

Since the discovery of the persister phenomenon in Staphylococcal cultures in 1940s, there is renewed interest in understanding S. aureus persistence and persister biology. Here, we performed a systematic, high-throughput mutant screen against mutants of all non- essential genes in the clinically relevant USA300 S. aureus strain to identify genes that play a role in bacterial persistence in S. aureus. Mutants involved in metabolite production or regulation were found to be required for bacterial persistence to stresses and antibiotics.

Methods

Culture Media, Antibiotics, and Chemicals

Ampicillin, chloramphenicol, rifampicin, gentamicin, and erythromycin were obtained from Sigma-Aldrich Co. (St. Louis, MO, USA). Stock solutions were prepared in the laboratory, filter-sterilized and used at indicated concentrations. Bacterial strains and plasmids used in this

study include the library of transposon mutants of USA300, a clinical MRSA isolate, offered by the Network on Antimicrobial Resistance in Staphylococcus aureus (NARSA). S. aureus strains were cultivated in tryptic soy broth (TSB) and tryptic soy agar (TSA) with the appropriate antibiotics and growth conditions, as mentioned, and E. coli strains were cultivated in Luria-

Bertani (LB) broth or agar at 37 °C.

Library Screens to Identify Mutants with Defect in Persistence

The NARSA library consisting of 1,952 transposon mutants of S. aureus USA300 was grown in TSB containing 50 μg/mL erythromycin at 37 °C in 384-well plates overnight without shaking. The library was grown to stationary phase in tryptic soy broth (TSB) with erythromycin, the antibiotic selective marker of the mutants from the NARSA library [128]. Rifampicin at 2

μg/mL was added to overnight cultures in the wells. The plates were further incubated for 24 h when the library was replica transferred to TSA plates to score for mutants that failed to grow after drug exposure. The antibiotic exposure was carried out over of a period of at least 6 days.

Susceptibility of Mutants to Various Antibiotics and Stresses

The susceptibilities of stationary-phase mutants and the parent strain USA300 cultures to antibiotic rifampicin (2 μg/mL, >10× MIC) and gentamicin (60 μg/mL, >10× MIC) were evaluated in drug exposure experiments. The antibiotic exposure was carried out over the course of 8 days at 37 °C without shaking. The susceptibilities of the mutants and parent strains to low pH stress was tested by diluting overnight culture 1:100 in citric acid buffer, pH 4, at 37 °C. To test susceptibilities to heat, undiluted overnight cultures were placed into a bathtub of water at

58 °C. At different time points, aliquots of bacterial cultures exposed to the antibiotics were taken out, washed in saline, serial-diluted, and plated for viable bacteria (cell forming unit, CFU) on TSA plates.

Complementation of S. aureus Mutants

The wildtype genes of interest, purB, and purM from S. aureus USA300, were amplified by PCR using primers listed. The PCR primers contained restriction sites BglII and EcoRI. The

PCR parameters were: 94 °C for 15 min, followed by 35 cycles of 94 °C for 30 s, 55 °C for 30 s, and 72 °C for 2 min, followed by a final extension at 72 °C for 10 min. The PCR products were digested with BglII and EcoRI and were cloned into plasmid shuttle vector pRAB11, which harbors a tet operator that is induced by anhydrotetracycline (ATc) [129], and were cut with the same enzymes. The ligated products were chemically transformed into DH5alpha cells and spread on LB plates containing 100 μg/mL ampicillin and grown overnight at 37 °C. Colonies were selected for the correct construct, isolated for introduction into S. aureus RN4220 by electroporation (voltage = 2.5 kV, resistance = 100 Ω, capacity = 25 μF) using a MicroPulser

Electroporation Apparatus (Bio-Rad, Hercules, CA, USA), spread onto TSA plates containing 10

μg/mL chloramphenicol, and grown overnight at 37 °C. Afterward, pRAB11 plasmid DNA was isolated from RN4220, and along with the pRAB11 vector alone control was then introduced into the respective S. aureusUSA300 mutant strain by electroporation. Positive clones were identified by restriction digestion, PCR, and DNA sequencing.

For complementation of the purB and purM mutants, purB mutant-pRAB11 vector control, purB mutant pRAB11-purB complemented strain, purM mutant-pRAB11 vector control and the purM mutant pRAB11-purM complemented strain were grown in TSB broth with 5

μg/mL chloramphenicol overnight to stationary phase. The strains were then refreshed into 1:100

TSB only and grown to log phase at OD600 of 0.6. The cells were subsequently induced with 25 ng/mL of anhydrotetracycline. Cells were then washed twice with PBS and then re-suspended into MOPS buffer to perform drug exposure and persister assays as described [48].

Results

Identification of New Persister Genes

To better understand the mechanisms of persister formation, we performed a genetic screen using the Network on Antimicrobial Resistance in Staphylococcus aureus (NARSA) mutant library, which contained mutations in all the 1,952 non-essential genes in the genome of the USA300 strain. We exposed all mutants in the library to rifampicin, which is a bactericidal agent for S. aureus that has been used to treat serious MRSA [130]. Persister isolation and analysis is shown schematically in Figure 1. The stationary phase culture was exposed to rifampicin (2 μg/mL, >10× MIC). Using a 384-well pin replicator, bacteria were stamped onto tryptic soy agar (TSA) plates and mutants with defective persistence, as indicated by the lack of growth, were recorded (Fig. 2.1B).

The screen identified 124 mutants that failed to grow on TSA plates after exposure with rifampicin (Fig. 2.2). Of the mutants that showed defective persistence after rifampicin exposure, 29% of them had mutated genes that play a role in the metabolism in pathways such as carbohydrate metabolism, amino acid, and purine biosynthesis. Among the rest of the mutants, 14% of the candidates represented pathways that play a role in genetic processes such as DNA replication and repair, transcription and translation, 11% of the mutants were involved in environmental signaling processes (e.g., transporters and sensor kinases), 9% were enzymes (e.g., proteases, hydrolases, etc.), and a minority of the gene hits played a role in bacterial pathogenesis such as toxin production (5%), originated from phage derived proteins (4%), or were involved in cell envelope development and drug efflux pathways (both

2%) (Fig. 2.2A,B).

Mutants Involved in Regulating Purine Biosynthesis are More Susceptible to a Variety of

Stresses Including Antibiotics, Heat and Low pH

Metabolic genes account for 29% of the gene hits that our screen identified. Our data imply that metabolic pathways and genes play a crucial role in persistence. Out of the metabolic genes, a prominent group of mutants, which includes purB, purF, purH, purM, and

SAUSA300_0147, all play a role in purine biosynthesis (Fig. 2.2C,D). To further validate the role of purine biosynthesis and its possible involvement in persistence and tolerance to antibiotics and stress, we chose two genes, purB and purM, adenylosuccinate lyase and phosphoribosylaminoimidazole synthetase, respectively, for further confirmation.

The purB and purM mutants not only showed defective persistence to rifampicin (Fig. 2.2) but were also identified as playing a role in persistence to gentamicin using a similarly high throughput method. We further evaluated purB and purM, since these two genes could potentially be core regulators in persistence. We first performed a growth curve study to exclude the possibility that these mutants have growth defects compared to the parental strain.

After we showed that the mutants had no growth defect in log phase and stationary phase under non-stressed conditions (Fig. 2.3A,B), we performed a persister assay, where bacterial cells were first exposed to stresses. Afterward, upon each time point, the stress from the bacteria was removed by several washes and subsequently serial diluted and plated for colony forming unit count (CFU) (Fig. 2.1). Both of the selected mutants purB and purM showed increased susceptibility and defective persistence to rifampicin (2 μg/mL, >10× MIC) and gentamicin (60 μg/mL, >10× MIC). Initially, the mutants were killed as much as the parental strain during the first two days but showed increased susceptibility to rifampicin such that, by day 8, the two mutants purB, and purM had 1 × 107 CFU/mL left, whereas USA300

maintained 1 × 109 CFU/mL (Fig. 2.3C). Under gentamicin (60 μg/mL, >10× MIC) exposure, the remaining CFU/mL of the purB and purM mutants were 1 × 105 CFU/mL, whereas the parental strain USA300 had 1 × 108 CFU/mL upon 8 days post exposure (Fig. 2.3D). We also determined the minimal inhibitory concentration (MIC) for both rifampicin and gentamicin against the mutants and USA300. Our data show that the MIC was the same for the mutants and USA300, suggesting that neither the mutants nor the parental strain USA300 are resistant to either drug. The MIC for rifampicin and gentamicin is 0.03 μg/mL and 6 μg/mL, respectively. Our study suggests that the mutants compared to USA300 are not any more susceptible to the drugs.

To determine the effect of heat on the survival of the mutants, we subjected the mutants and the parental strain to a heat stress at 58 °C. The mutants were much more sensitive to heat treatment, as demonstrated by the poorer growth seen in the purB and purM mutants, compared to USA300 (Fig. 2.3E). Compared to all of the other drug and stress exposures, the starting bacterial inoculum size for low pH stress was lowered in order to prevent neutralization of the media by large inocula. Nonetheless, all of the mutants were also more sensitive than USA300 to buffered low pH solutions of 4.0. At a low pH of 4.0, no viable bacteria were recovered from the mutants after 10 h of exposure, whereas the parent strain had about 1 × 103 CFU/mL surviving bacteria (Fig. 2.3F). Throughout the course of both the heat and low pH exposure experiments, the poorer survival in all of the mutants was evident.

Complementation Studies to Confirm the Role of Purine Pathways in Persistence

To confirm that purB and purM are responsible for defective persistence, we complemented the purB and purM mutants with the wildtype purB and purM gene,

respectively, using the S. aureus-E. coli shuttle vector pRAB11. The shuttle vector pRAB11 harbors a tet operator that is induced by anhydrotetracycline (ATc) [129]. Our findings suggest that both purB and purM are important for rifampicin persistence (Fig. 2.4). Under prolonged rifampicin exposure, the complemented purBmutant restored partial persistence to rifampicin on day 5 (Fig. 2.4A) with an average of 1 × 104 CFU/mL compared to the parental strain that harbored 1 × 107 CFU/mL. The purB mutant transformed with the pRAB11 vector control showed a defect in persistence, and no difference was observed among any of the strains under non-stressed conditions (Fig. 2.4A). Similarly, the complemented purM mutant restored rifampicin persistence at a level similar to the parental strain. On day 5, the complemented purM mutant had an average of 1 × 106 CFU/mL compared to the parental strain that had 1 × 107 CFU/mL (Fig. 2.4B). The purM mutant transformed with the pRAB11 vector control had no colonies by day 5, and no phenotypic difference appeared under non- stressed conditions (Fig. 2.4B). The complemented purB mutant and purM mutant both restored persistence in heat exposure (58 °C) (Fig. 2. 4C,D). After forty minutes of heat exposure, the purB-complemented mutant and purM-complemented mutant had roughly 7.8 ×

104 CFU/mL and 1.64 × 104CFU/mL, respectively, which is comparable to USA300, which had 7.5 × 104 CFU/mL. Additionally, the complemented purB mutant and purM mutant restored persistence when exposed to a low pH of 4 (Fig. 2. 4E,F). After 9 h of exposure to a pH of 4.0, the purB complemented and purM complemented mutants had roughly 0.4 ×

104 CFU/mL and 2.5 × 104 CFU/mL, respectively, comparable to USA300, which had 2.1 ×

104CFU/mL. In contrast, the mutants harboring the empty vector control had a significantly lower number of cells under heat and acid stress conditions compared to the respective complemented strains.

Discussion

Despite the discovery of S. aureus persisters in 1944 and the prevalence of persistent infections caused by S. aureus, the molecular mechanisms underlying S. aureus persister survival has remained largely unknown [29]. The convenience of a comprehensive transposon-mutant library has allowed us to perform the first whole genome-wide mutant analysis and to provide insight into the molecular basis of S. aureus persisters. This study represents the first systemic analysis of persister genes and pathways in a circulating clinical isolate of S. aureus.

We identified 124 different mutants that showed defects in rifampicin persistence. A majority of the mutants belongs to metabolite production and regulation pathways such as purine biosynthesis. Our findings revealed that five purine biosynthesis genes

(purB, purF, purH, purM, and SAUSA300_0147) are involved in S. aureus persistence. In particular, we chose purB and purM, which encode adenylosuccinate lyase and phosphoribosylaminoimidazole synthetase, respectively, for confirmation, and, indeed, complementation studies show that genes involved in purine biosynthesis are important for antibiotic and stress tolerance in S. aureus. Although the molecular mechanisms by which purine pathway mediate persistence remain to be determined, our findings are consistent with previous research suggesting that defective purine synthesis has decreased biofilm formation and attenuated virulence in persistent infections such as endocarditis [131]. In addition, purine biosynthesis mutants (purL and purM) of Burkholderia fail to colonize in the host symbiotic organ and exhibit decreased biofilm formation [132]. The role of pur genes in Staphylococcal persistence is novel, and future animal studies are needed to confirm the virulence and persistence of the S. aureus pur mutants in vivo.

Several mechanisms may be underlying the defect in persistence seen in purine biosynthesis mutants. For example, a defect in the purine biosynthesis pathway may lead to decreased downstream energy production, amino acid biosynthesis and urea cycle activation, which may be responsible for the defect in persistence observed in this study to rifampicin and other antibiotics and stresses (Figure 2.5). Purine biosynthesis has also been shown to be associated with survival in stressed conditions such as vancomycin and daptomycin in other strains of S. aureus [44, 133]. The final step in purine nucleotide synthesis leads to AMP formation, increased AMP, and thus ATP energy levels, which were observed in antibiotic resistant strains [133]. It is hypothesized that increased purine biosynthesis would allow for more energy used in generating polymers, one of the most energy demanding process in bacteria. In S. aureus and other Gram-positive bacteria, the peptidoglycan layer is the most abundantly large polymer. While mechanisms regarding purine biosynthesis in persistence remain to be further elucidated, our hypothesis is consistent with the participation of other genes in persistence from our data. Mutants such as accC, budA, Cap5J, gltA, gntK, leuD, murA, mqo, each of which play a role in carbohydrate metabolism, SAUSA300_0918, which plays a role in lipid metabolism, fmtC, SAUSA300_0729, and SAUSA300_1677 in cell wall and membrane synthesis all showed defects in rifampicin persistence, which suggests a possible role of purine metabolism and of cell envelope formation in persister formation.

Other general mechanisms involved in persistence may help explain the role of purines in persister biology. Studies have shown that pathways involved in DNA repair mechanisms,

SOS response, and energy production affect persister survival [44]. Purines are important substrates for both DNA synthesis and hence will alter DNA repair processes that make it difficult for the bacteria to repair their genetic material upon stress damage. Additionally,

purines are starting compounds for GTP synthesis and thus play an important role for energy production in regulating cell growth [134]. Bacterial cells can be sent into a dormant-like stage through modulation of energy and alarmone (p)ppGpp production [135]. However, imbalance GTP synthesis and metabolism is associated with decreased (p)ppGpp synthesis.

Since purines play a crucial role in producing (p)ppGpp substrates, our pur mutants may experience the lack of (p)ppGpp modulation into a persistence state. This phenomenon has been shown in Pseudomonas, where (p)ppGpp levels affect the persistence in starvation, biofilm formation, and oxidative stress [136]. Similar mechanisms may explain why different pur mutants were unable to persist in our screen.

While our studies focused primarily on genes involved in purine biosynthesis, it is important to note that, based on our screen, metabolic processes showed the largest role in rifampicin tolerance and should be further explored in future studies. For example, arginine biosynthesis has recently been proposed as contributing to successful S. aureus infection

[137]. Most research focused on genes encoded by the arginine catabolic mobile element

(ACME), which does not include argJ, an acetyltransferase [138, 139] that was identified as important to persistence based on our screen. Nonetheless, studies suggested that arginine synthesis might be pivotal in allowing S. aureus to colonize the skin and survive in abscesses

[137]. The bacterial cells use arginine to increase the pH by ammonia production of the extracellular milieu on the skin, which may aid the survival of S. aureus [137, 139]. As another example, gltS, which encodes a sodium/glutamate symporter, has been shown to be upregulated during biofilm formation [140]. Glutamate is required for the development of arginine and produce byproducts such as ornithine. Ornithine is a pivotal compound in cell metabolism in making prolines, polyamines, antibiotics, proteins, and peptidoglycan. The

products driven by ornithine synthesis enhance bacterial growth and pathogenesis [141]. As for the gltS transporter, acquisition of amino acid glutamate allows for the production of

NH3 in the periplasm of the cell to allow for the bacteria to adapt to the acidic environment, similar to the skin [142]. The role of glutamate in regulating bacterial stress includes both the upregulation of GcrR [143], which controls the acid tolerance resistance in Streptococcus mutans, and the downregulation of MarR, which regulates virulence factors, in response to antibiotics and oxidative stresses [144].

Despite the significant findings of this study, there are some limitations. First, due to

S. aureus’s ability to cause wound infections, endocarditis and systemic disease, different animal models (e.g., skin model, intraperitoneal injection, and murine model for urinary tract) should be considered to validate our findings [145-147]. These will be addressed in future studies. Secondly, a mutant screen only explores persister formation at the DNA and RNA level. Further proteomic, metabolomic and even epigenetic analyses (as our screen indicated genes with acetyl-, and methyl-transferase activities to be important) would offer more comprehensive insight into the effects of gene mutations on persistence. Thirdly, while this screen is comprehensive in the sense that all the non-essential genes of the clinical isolate of

MRSA are included, their effects on the persistence of these mutants are determined by the conditions of our assays. Persistence can also be affected by variables such as different aged cultures, concentrations of drugs and bacteria, inoculum size, and the length of the drug exposure [21]. Our screen reveals potential candidates specific to our conditions and may underscore certain genes that can still play a role despite not being revealed from our screen.

We performed our complementation studies using an E. coli-S. aureus shuttle vector pRAB11, which contains an anhydrotetracycline (ATc) inducible promoter [129]. The reason to use a

vector with an inducible promoter is to rule out possible unwanted regulation by the bacterial host and to examine the effect of the inducible gene on persister formation more clearly and effectively. We acknowledge that using an inducer Atc introduces a variable that could complicate data analysis. Future studies will include the use of the constructs with native promoters. We included a negative control by introducing the empty vector into a mutant, and our data suggests that there is a statistically significant difference (using student t-test) in persistence between the empty vector control and the complemented mutants. Lastly, secondary mutations may have occurred in some mutants that could affect the phenotypic outcomes. However, our results on the mutant phenotypes are reproducible, confirmed by complementation studies, and the same subculture from the same stock was also used each time to ensure reproducibility among all replicates.

In conclusion, we report the molecular basis of persistence in S. aureus. Our study encompasses the consideration of all the non-essential genes in S. aureus, and, to our knowledge, this is the first report on identification of S. aureus persistence genes from a whole genome perspective. Our high throughput protocol identified many mutants with defects in persister formation and survival, and the selection of several mutants validated our findings. Persister bacteria pose enormous public health problems due to relapse and can cause the development of genetic drug resistance from prolonged treatment with antibiotics.

Our studies will not only help understand new mechanisms underlying persister formation but also provide novel therapeutic and vaccine targets for developing more effective treatment and prevention for S. aureus infection.

)LJXUH:RUNIORZIRULVRODWLRQRIPXWDQWVZLWKGHIHFWLQSHUVLVWHQFH $ 7KH6DXUHXV

86$WUDQVSRVRQOLEUDU\IURP1$56$ZDVJURZQLQWU\SWLFVR\EURWKIRUKWRUHDFK

VWDWLRQDU\SKDVH5LIDPSLFLQZDVDGGHGWRWKHVWDWLRQDU\SKDVHFXOWXUHDWȝJP/ !;0,&

%HIRUHULIDPSLFLQH[SRVXUHEDFWHULDZHUHVWDPSHGRQWRWU\SWLFVR\DJDUSODWHV 76$ XVLQJD

SLQUHSOLFDWRUDVWKHLQSXWFRQWURO2YHUWKHFRXUVHRIDGD\DQWLELRWLFH[SRVXUHEDFWHULD

ZHUHZDVKHGDQGSODWHGRQWR76$SODWHVGDLO\DQGDOOPXWDQWVWKDWVKRZHGGHIHFWLYHJURZWKZDV

recorded and tallied until Day 4; (B) Representative agar plate images with mutant library after the 4-day rifampicin exposure where some mutants show defect in growth.

)LJXUH $ ,GHQWLILFDWLRQDQGFKDUDFWHUL]DWLRQRIWKHJHQHVLQYROYHGLQULIDPSLFLQ

SHUVLVWHQFH % 3DWKZD\VLQYROYHGLQULIDPSLFLQSHUVLVWHQFH & )XUWKHUEUHDNGRZQRIWKH

PHWDEROLFSDWKZD\VLQYROYHGLQSHUVLVWHQFHUHYHDOHGDVSHFLILFJURXSRIPXWDQWVEHORQJLQJWR

SXULQHELRV\QWKHVLV ' 7KHORFDWLRQRIWKHWUDQVSRVRQLQVHUWLRQLQWKHJHQHVLQYROYHGLQSXULQH

ELRV\QWKHVLVLVGHSLFWHG$OOSDWKZD\DQDO\VHVZHUHSHUIRUPHGXVLQJWKH.\RWR(QF\FORSHGLDRI

*HQHVDQG*HQRPHV .(** GDWDEDVHDQGWKHVFUHHQUHVXOWVZHUHUHSHDWHGDQGZHUH

UHSURGXFLEOH

()

)LJXUH $% 3DUHQWDOVWUDLQ86$DQGSXUPXWDQWVVKRZHGQRGHIHFWVLQORJSKDVHRU

VWDWLRQDU\SKDVHXQGHUQRQVWUHVVHGFRQGLWLRQVSXUPXWDQWVKDGGHIHFWLYHSHUVLVWHQFHLQ

H[SRVXUHWR & ULIDPSLFLQ ȝJP/!;0,& ' JHQWDPLFLQ ȝJP/!;0,& (

KHDW & DQG ) DORZS+RI

()

)LJXUH5HVWRUDWLRQRISHUVLVWHQFHSKHQRW\SHLQFRPSOHPHQWHGPXWDQWV6XUYLYDORI

VWDWLRQDU\SKDVHFXOWXUHVRI6DXUHXVSXU%DQGSXU0PXWDQWVWUDQVIRUPHGZLWKYHFWRUS5$%

RUS5$%ZLOGW\SHSXU%DQGSXU0DQGWKHSDUHQWVWUDLQ86$XSRQ $% ULIDPSLFLQ

exposure (2 μg/mL, >10X MIC) at 5 days post-exposure; (C, D) heat exposure (58 °C for 40 min) and (E, F) low pH of 4.0 for 9 h. The S. aureus purB and purM mutants complemented by their corresponding wildtype genes partially restored the persistence phenotype of the mutants. No change in persistence was observed in stress-free conditions over the course of 5 days. (Student t-test, * = p-value < 0.05, ** = p-value < 0.005).

Table 2.1. Primers used in this study. The underlined sequences AGATCT and GAATTC represent the BglII and EcoRI restriction sites incorporated for cloning the wildtype gene into shuttle vector pRAB11 for complementation.

Primer Name Sequence Source or Reference

purBF 5′-GCAAGATCTATGATTGAACGCTATTCTAG-3′ This study

purBR 5′-ACGGAATTCTTATGCTAATCCAGCGCGTTCG-3′ This study

purMF 5′-GCTAGATCTATGTCTAAAGCATATGAACAATC-3′ This study

purMR 5′-ACGGAATTCTTATACCCCCAACAATTCAAT-3′ This study

CHAPTER 3

IDENTIFICATION OF A NOVEL GENE ARGJ INVOLVED IN ARGININE

BIOSYNTHESIS CRITICAL FOR PERSISTER FORMATION IN

STAPHYLOCOCCUS AUREUS

This Chapter has been published previously on Cold Spring Harbor Laboratory's bioRxiv The Preprint Server for Biology: Yee R., Cui P., Xu T, Shi W., Feng J., Zhang W., Zhang Y. 2017. Identification of a Novel Gene argJ involved in Arginine Biosynthesis Critical for Persister Formation in Staphylococcus aureus. doi: https://doi.org/10.1101/114827

Introduction Persisters are metabolically quiescent cells that are tolerant to antibiotics or stresses but can revert back to a growing state upon antibiotic stress removal and remain susceptible to the same antibiotic [29]. Persister cells are implicated in persistent infections and can cause relapse in various bacterial infections. While S. aureus often causes acute infections, it can also cause chronic recurrent infections such as peritonitis, endocarditis, osteomyelitis, wound and soft tissue infections, and infections from indwelling medical devices [148]. Most studies on persister cell mechanisms have been conducted using Escherichia coli (E. coli) as a model organism. In E. coli, the toxin-antitoxin (TA) systems such as HipBA and MazF cause persister formation by inhibiting protein synthesis through phosphorylation of Glu-tRNA synthase and cleaving of mRNA, respectively [149]. Unlike E. coli, it has been shown that S. aureus persister formation does not involve TA systems but is dependent on ATP production [47]. Additionally, we have also previously shown that pathways involved in protein synthesis, efflux/transporter and metabolism and energy production are heavily involved in persister cell formation in S. aureus

[150]. We have shown that upon prolonged exposure to rifampicin, over one hundred genes that play a role in rifampicin persistence were identified, where approximately one-third of the genes play a role in the metabolism of amino acids, lipids, vitamins, carbohydrates and purine biosynthesis [150]. To find core genes and pathways that play a role in persister cell formation with different drugs, we performed an unbiased high-throughput screen using gentamicin against a mutant library of S. aureus clinical isolate USA 300 [151]. We identified the argJ gene as a core gene that plays an important role in persister formation under various antibiotics and stresses. We report for the first time the importance of ArgJ for S. aureus persister cell formation and also virulence and survival in C. elegans and mice.

Methods

Culture media, chemicals, and antibiotics

S. aureus strains were cultivated in tryptic soy broth (TSB) and tryptic soy agar (TSA) and E. coli strains were cultivated in Luria-Bertani (LB) broth or agar at 37oC with the appropriate antibiotics. Citric acid monohydrate, the antibiotics ampicillin, chloramphenicol, rifampicin, gentamicin, erythromycin and amino acids L-Arginine, L-Histidine, L-Lysine were obtained from Sigma-Aldrich Co. Stock solutions were prepared and sterilized through filtration or autoclaved, if necessary, and used at indicated concentrations. Library screen to identify mutants with defective persistence

Persister assays to measure susceptibility to various antibiotics and stresses

Overnight stationary phase cultures were exposed to selected drugs or stresses and colony forming units per milliliter (CFU/ml) were measured through serial dilutions and plating onto

TSA plates. The antibiotic exposure was carried out over the course of 6 days at 37oC as previously described [150]. To measure the susceptibilities in low pH stress, the overnight culture was diluted 1:100 and incubated in buffered acid solution with pH = 4 at 37 oC. To measure the susceptibility to heat, undiluted overnight cultures were placed in a 58 oC water bath.

At different time points, 100 µl of bacterial suspension was removed and washed in 1X PBS and enumerated for CFU/ml. For amino acid supplementation, overnight bacterial cultures were refreshed 1:100 into TSB with a supplementation of the indicated amino acids to the growth media and grown for 16 hours at 37 oC before use. This step was repeated before persister assays were performed.

Complementation of ArgJ mutant

The wildtype argJ gene from S. aureus USA300 was amplified by PCR. The primers contained restriction sites KpnI and EcoRI. The forward primer used had the sequence 5'

GCAGGTACCATGAAACATCAAGAAACGAC 3' and the reverse primer used had the sequence 5' GCCGAATTCTTATGTTCGATATGATGCGTT 3'. The PCR parameters used were as follows: 94°C for 15 min, followed by 35 cycles of 94°C for 30 s, 55°C for 30 s, and 72°C for

2 min, and a final extension at 72°C for 10 min. The PCR products were digested with KpnI and

EcoRI and cloned into S. aureus-E. coli shuttle vector pRAB11 [129]. Ligation mixtures were transformed into chemically competent E. coli DH5α cells (Invitrogen) and spread onto LB agar plates containing ampicillin (100 µg/ml) and grown overnight at 37oC. Upon confirmation of the transformants by DNA sequencing, plasmid DNA was isolated using Lysostaphin (Sigma290

Aldrich) lysis (2mg/ml) followed by purification with QIAprep Spin Miniprep Kit. The plasmid was introduced into S. aureus RN4220 by electroporation (voltage = 2.5 kV, resistance = 100Ω, capacity = 25 µF) using MicroPulser Electroporation Apparatus (Bio-Rad) followed by plating onto TSA plates containing chloramphenicol (10 µg/ml) and incubation overnight at 37 oC. To induce for ArgJ expression, bacterial cultures were grown in TSB containing chloramphenicol overnight then refreshed 1:100 into TSB only. When the cells reached OD600 of 0.5, the cells were induced with anhydrotetracycline (25 ng/ml) overnight with shaking at 220 rpm in 37 oC.

Overnight cells were washed twice with 1X PBS and resuspended in MOPS buffer to perform persister assays as described above.

RNA preparation and real-time PCR (qRT-PCR)

Samples were prepared for RNA extraction based on the instructions described in the

"Enzymatic Lysis and Proteinase K Digestion of Bacteria" protocol of the RNAprotect Bacteria

Reagent Handbook with the addition of incubation with lysostaphin before purification using the

RNeasy mini kit (Qiagen). cDNA was synthesized from 1 µg of RNA with random primers using

QuantiTech Reverse Transcription Kit (Qiagen). Quantitative RT-PCR (qRT-PCR) was performed in a 20 µl reaction mixture using SYBR Green PCR Master Mix (Life Technologies) and 0.2 µM (each) of gene-specific primers (Table S2). Amplification and detection of specific products were performed using StepOnePlus Real-Time PCR Systems (Applied Biosystems).

The PCR parameters used were as follows: 95°C for 10 min, followed by 40 cycles of 95°C for

15 s and 60°C for 1 min. Relative gene expression levels were calculated using the comparative threshold cycle (CT) method (2−ΔΔCT method) with 16s rRNA as the internal control gene for normalization of gene expression to basal levels.

Nematode-killing assay

S. aureus nematode-killing assay was performed as described [152]. Briefly, S. aureus strains were grown overnight at 30 oC in TSB containing the appropriate antibiotics as needed.

One spot of overnight S. aureus culture (70 µl) was dropped onto nematode growth agar containing 5- Fluoro-2'-deoxyuridine (100 µM). The prepared plates were incubated at 37 oC overnight and then allowed to equilibrate to room temperature (20 oC) for 60 minutes before being seeded with C. elegans N2 Bristol worms (Caenorhabditis Genetics Center). The worms were synchronized to the same growth stage by treatment with alkaline hypochlorite solution as described [153]. Worms of the adult stage were recovered in 15 ml tubes with M9 buffer. The worms were washed twice to remove the residual bacteria in their diet by centrifugation at 1500 rpm for 2 minutes at room temperature in a table top centrifuge. Bleaching solution with 5% hypochlorite was then added and incubated with the worms for 9 minutes to lyse the adult stages but keeping the eggs intact. The lysing reaction was stopped when M9 buffer was added. Bleach

was removed by centrifugation at 1500 rpm for 1 minute followed by three more washes with

M9 buffer. To induce hatching of eggs, M9 buffer was added to the pellet and incubated at 20 oC with gentle agitation and proper aeration. After 24 hours, worms were pelleted with a 2-minute spin at 1500 rpm at room temperature and seeded onto OP50 seeded plates. L4 stage worms were obtained after 48 hours at 20 oC. In each assay, 10-20 L4-stage nematodes were added to each plate and each assay was carried out at least twice. The plates were incubated at 20 oC and scored for live and dead worms every 24 hours. A worm was considered dead when it failed to respond to touch.

Mouse intraperitoneal challenge

Overnight cultures of S. aureus were subcultured into fresh TSB (1:100) and grown for

2.5 hours with shaking (220 rpm) at 37 oC. The cells were washed with 1X PBS. Adult (7-8 weeks old) female Swiss-Webster mice (Charles River Laboratories) were infected via intraperitoneal injection with an inoculum size of 7 x 107 CFU. Mice were housed in cages under standard BSL2 housing conditions. Mice infected after 3 days were euthanized and spleens and kidneys were homogenized for CFU enumeration.

Results argJ is a novel persistence gene

To identify genes and pathways involved in persister formation, we performed a genomewide screen using the saturated Nebraska transposon mutant library (NTML) [151] to isolate mutants with defective persistence which is defined as a decrease in the number of persister cells relative to the parental strain USA300. We exposed stationary phase cultures of the mutant library to gentamicin (60 µg/ml, 10X MIC) (Fig. 3.1A). Over the course of six days,

the screen identified the ArgJ mutant that failed to grow on tryptic soy agar (TSA) plates after gentamicin exposure, showing at least a three-log fold decrease which is a significance decrease compared to other mutant hits. While it is known that the argJ gene is required for the arginine biosynthesis cycle [138, 154] and encodes an acetyltransferase that synthesizes N- acetylglutamate and ornithine (Fig. 3.1B), the importance of ArgJ in antibiotic and stress tolerance has not been explored. To establish the importance of ArgJ in stress tolerance, we first excluded the possibility that altered growth dynamics is a confounding factor. Our growth curve study suggested that the ArgJ mutant and USA300 had similar growth patterns (Fig 3.2A) and similar colony forming unit per milliliter (CFU/ml) in normal growth medium even up to 8 days

(Fig 3.2B). Next, to confirm that a mutation in ArgJ causes a defect in persister formation in stressed conditions, we performed a persister assay by exposing stationary phase cultures of the

ArgJ mutant and USA300 control strain to different antibiotics and stresses. At different time points, the cells were washed and then enumerated for CFU. As early as day 2 post-gentamicin exposure, the ArgJ mutant harbored 1x 107 CFU/ml as opposed to USA300 with 1x109 CFU/ml.

By day 6, the ArgJ mutant had 1 x106 CFU/ml compared to USA300 with 1 x 109 CFU/ml, a significant difference of about 3-logs (Fig. 3.2C). Similarly, by day 6 of rifampicin exposure, there was also a significant three-log difference in CFU between the ArgJ mutant and USA300

(Fig. 3.2D). To confirm the specific role of ArgJ in persistence, we also measured the persister levels in mutants with mutations in other proteins of the Arg pathway that were not identified as a hit in my screens (e.g. ArgB and ArgF) and TrpA, protein involved in tryptophan biosynthesis, as a control. Our results indicated that after gentamicin exposure of 6 days, there were no significant differences in CFU/ml among the ArgB, ArgF, TrpA mutants and the control strain

USA300 (Fig. 3.2E). We, therefore, concluded that a mutation in ArgJ is specific in causing a defect in persister formation in S. aureus.

In our separate recent study, we showed that mutations in metabolic pathways in S. aureus have defective persistence to different antibiotics, low pH, and heat stress [150]. To explore if ArgJ mediates persister cell formation in other stresses besides antibiotics, we subjected the ArgJ mutant to heat stress at 58 oC. The difference in heat tolerance between the

ArgJ mutant and USA300 was statistically significant. After 80 minutes, the ArgJ mutant had a mean of 1.1 x 103 CFU/ml while USA300 had 4.6 x 104 CFU/ml (Fig. 3.2F). Additionally, the

ArgJ mutant was also significantly less tolerant to low pH (pH = 4) compared to USA300.

Unlike the other stress exposures, the starting bacterial inoculum concentration for acid pH exposure was standardized to 1 x 105 CFU/ml in order to prevent neutralization of the acid pH due to neutralization of acid pH by a high bacterial inoculum [48]. Nonetheless, after 24 hours of exposure in a low pH environment, the ArgJ mutant had only about 4.6 CFU/ml left while the

USA300 had about 4.4 x 103 CFU/ml, indicating that the ArgJ mutant is more susceptible to low pH (Fig. 3.2G).

To confirm that the mutation in argJ is responsible for the defective persistence, we complemented the ArgJ mutant with the wildtype argJ gene from the USA300. We used an

S. aureus- E. coli shuttle vector pRAB11 to insert the wildtype argJ gene back into the mutant

[129]. After 6 days of gentamicin exposure, the ArgJ mutant with an empty vector had 6.4 x 103

CFU/ml whereas the ArgJ complemented strain and USA300 had 7.1 x 104 CFU/ml and 2.0 x

107 CFU/ml, respectively (Fig. 3.4A). Similarly, after 6 days of rifampicin exposure, the ArgJ mutant with an empty vector had 3.1 x 103 CFU/ml whereas the ArgJ complemented strain had

5.0 x 106 CFU/ml, only one-log fold less than USA300 with 1.7 x 107 CFU/ml. These findings indicate that a genetic mutation in ArgJ confers a defect in persister formation.

Arginine biosynthesis via the Arg pathway is important for persistence

To determine if the arginine pathway is important for persistence, we supplemented exogenous L-arginine into TSB growth medium when the bacterial cultures reached mid- expotential phase. The ArgJ mutant grown without any L-arginine supplementation had 1.2 x 104

CFU/ml under gentamicin stress. However, when L-arginine (30 mM) was supplemented into the growth medium of the ArgJ mutant, the amount of cells on day 6 post-gentamicin exposure was

6.0 x 106 CFU/ml, which is similar to the parent strain USA300 with 2.8 x 106 CFU/ml (Fig.

3.3C), indicating that L-arginine supplementation complemented the defect in persistence in the

ArgJ mutant. Because arginine is a positively-charged amino acid, we tested if persistence restoration is specific to arginine and not achieved by other positively-charged amino acids. Thus, we supplemented the growth medium with two positively charged amino acids L-histidine and

L-lysine. Our results suggest that histidine and lysine did not restore the persistence of the ArgJ mutant indicating that L-arginine is specifically important sfor persistence (Fig. 3.3C).

Activity of arginine pathway genes in relation to persistence

S. aureus has the ability to synthesize arginine using secondary carbon sources such as glutamate (via the Arg pathway) or proline (via PutA and ProC) (Fig. 3.1B) [155]. It has been suggested that arginine production under normal growth conditions is mainly due to the proline precursor pathway [155] . However, the activity of the Arg pathway under stress conditions such as stationary phase and antibiotic exposure is unknown. To evaluate if the Arg pathway is induced under stress conditions, we performed qRT-PCR to compare the levels of gene

expression of genes from the Arg pathway (argCG) versus genes involved in arginine synthesis from a proline precursor (ProC and putA). Since the ArgJ mutant showed defective persistence

(Fig. 3.2) compared to USA300, we compared the gene expression fold-change of USA300 and the ArgJ mutant. Genes argC and argG were at least 2-fold more over-expressed in USA300 than the ArgJ mutant (Fig. 3.4A) in stationary phase (when persister cells enrich due to limiting nutrients) compared to log phase (when growing cells are heavily populated). Under gentamicin treatment, we also observed the same genes, argC and argG, having at least 2-fold higher expression in USA300 than in the ArgJ mutant (Fig. 3.4B). Collectively, our data suggest that arginine production through the ArgJ pathway is more expressed in stationary and drug-treated cells than log phase cells and untreated cells, respectively.

ArgJ plays a role in virulence in vivo

There is a general lack of research establishing the relationship between the mechanisms of persister cell formation and virulence in different bacterial pathogens including S. aureus.

Given that our results suggest the importance of ArgJ in persistence, we decided to explore the role of ArgJ in virulence. To test for virulence, we used a nematode C. elegans model, an accepted model for bacterial pathogenesis research [152]. To test the hypothesis that the ArgJ mutant has attenuated virulence, we examined the survival of C. elegans after S. aureus infection.

Our results showed that C. elegans killing was highly attenuated after infection with the ArgJ mutant as opposed to USA300. The first death caused by USA300 was observed at day 4 as opposed to day 6 for the ArgJ mutant. By day 8, there was a 76% survival in C. elegans exposed to the ArgJ mutant as opposed to the 56% survival in worms exposed to USA300 (Fig. 3.5A). No change in survival was seen in worms exposed to nonpathogenic E. coli strain.

To further confirm the role of ArgJ in virulence, we then utilized an S. aureus peritonitis mouse model [146]. Briefly, Swiss-Webster mice were infected by intraperitoneal injection with

7 x 107 CFU of the ArgJ mutant, the ArgJ complemented strain and USA300. After 3 day post infection, the CFU counts in the spleens and kidneys were enumerated. Our results indicate that mice infected with the ArgJ mutant harbored an average of 5 CFU/g of spleen while mice infected with USA300 and the ArgJ complemented strain had 3.5 x 104 CFU/g of spleen and 1.6 x 103 CFU/g of the spleen, respectively (Fig. 3.5B). Similarly, all mice infected with the ArgJ mutant resulted in no bacteria in the kidney while mice infected with USA300 and the ArgJ complemented strain had 1.0 x 104 CFU/g and 3.6 x 103 CFU/g of the kidney, respectively (Fig.

3. 5C), indicating that an ArgJ mutation caused statistically significant attenuation of virulence in mice.

Discussion

While there is renewed interest in persister biology recently, the molecular mechanisms of persistence are largely derived from the model organism E. coli and the mechanisms crucial to

S. aureus persister formation remain poorly understood. After we performed a comprehensive genetic screen to identify persister genes and pathways in the clinically relevant strain USA300, we identified ArgJ as important for persistence to multiple drugs and also stresses. To our knowledge, this is the first study that provided insights into the physiological impact of ArgJ and its related Arg pathway in stress tolerance and persistence in bacterial systems. Supplementation of arginine but not other amino acids into the growth media conferred increased persistence of the ArgJ mutant to gentamicin. Our data suggest that arginine increases tolerance and formation of persister cells and that ArgJ regulates persistence in S. aureus is further supported by our qRT-PCR results. It is important to note that under normal growth conditions, our results are

consistent with previous findings that suggest arginine synthesis from the proline precursor pathway (using PutA and ProC) is the canonical pathway (Fig. 3.2B) [155].However, the ArgJ- mediated Arg pathway appears to play a role in stationary phase and during stress conditions

(Fig. 3.5).

While the molecular mechanisms by which the arginine biosynthesis pathway mediated persistence remains to be determined, there are several possibilities. One proposed mechanism of

ArgJ-mediated persistence is through the direct generation of arginine. The bacterial cell can then catabolize the arginine through the arginine deiminase pathway to synthesize ammonia which mitigates against hydroxyl radicals produced during antibiotic action that promote cell death [156]. Persister cells are known to have increased capacity to deal with reactive oxygen species. Conlon et al. have shown that in stationary phase S. aureus cells, the araC gene which is involved in arginine deamination is over-expressed in stationary phase cultures [47].

Additionally, the downstream products of arginine production such as ornithine and polyamines are shown to increase the cell’s fitness and survival [157]. Within the arginine metabolic pathway, the metabolites glutamate and ornithine can be decarboxylated into glutamate-γ- aminobutyrate, and putrescene which have all been show to help the bacterial cell survive under acetic conditions due to the removal of hydrogen ions inside the cell [158, 159]. The downstream products of polyamines also modulate the translation and expression of key proteins in biofilms, an exopolysaccharide structure that contains persister cells [160].

Because persisters are non-growing cells with decreased energy state, we speculate that an ArgJ mutation causes defective persistence because the mutant bacteria need to resort to a more energy unfavorable pathway to produce arginine. In S. aureus, ArgJ is preferred over ArgE, which produces the same end products of ornithine and acetate, due to favorable energy kinetics

[154]. Thus, under growth-limiting conditions where the cells are more energetically inactive,

ArgJ may be preferentially expressed to facilitate persister survival. Hence, the altered cellular energetic state could impede the cells to reach dormancy and allow the ArgJ mutant to be killed more easily by antibiotics and stresses.

ArgJ is a bifunctional enzyme involved in de novo as well as recycling pathway for arginine biosynthesis (Fig. 3.1B). The role of ArgJ in the de novo pathway of arginine synthesis is to catalyze the first step of the linear arginine production, synthesizing N-acetylglutamate from glutamate and acetyl-CoA as the acetyl donor. In the recycling pathway, ArgJ helps generate ornithine by trans-acetylation of the acetyl group from N(2)-acetylornithine to glutamate. Our finding that mutations in de novo pathway genes (argB, argF) did not cause a defect in persistence (Fig. 3.3) indicates that the de novo arginine biosynthesis pathway is not important for persistence in S. aureus but rather the recycling function of ArgJ may be important for persistence. Further biochemical and genetic studies such as site-directed mutagenesis on the binding and active sites of the ArgJ protein to separate the bifunctional activity of the protein are needed to confirm the importance of the recycling pathway in persistence. In addition to ArgJ, 6 of the genes that we have identified playing a role in gentamicin persistence were also transferases (miaA, trmB, SAUSA300_ 0689, 1111, 1669, 2232) . Acetyl-transferases can help drive bistable gene expression and changes in DNA and protein modifications in tolerating stress and adapting to environmental changes [161]. Studies have shown that acetylated proteins in prokaryotes play a role in stress response through chemotaxis and cell cycle control [162].

Protein homology analyses suggest that the binding site of acetyltransferase SAUSA300_2232, also identified as a gene important for persister formation, has a similar amino acid sequence to

ArgJ. More studies will be needed to explore the role of these transferases in epigenetic control

of persistence in S. aureus. Our complementation studies showed that the argJ gene is important for maintaining persistence in S. aureus. However, the complemented ArgJ mutant achieved partial restoration of persistence. In fact, in virtually all S. aureus complementation studies, none could achieve full complementation. S. aureus possesses different restriction modification systems that may destroy exogenous plasmids [163], has endonucleases targeting specific sequences [164] and can methylate exogenous DNA [165] causing inactivation of exogenous

DNA. These could serve as possible explanations for the partial complementation of ArgJ mutant in this study.

While this study revealed novel insights into the mechanisms of S. aureus persister formation, the results from the screen is dependent on the conditions of our assays. Persisters can be affected by variables pertaining to the drugs administered such as the drug concentrations, drug exposure time, inoculum size and the age of the culture [48] . However, the reported persister genes in this study were identified twice from two independent screens with both rifampicin and gentamicin and can thus be considered reproducible core genes involved in persister formation in S. aureus.

In conclusion, we identified a comprehensive list of genes and pathways that play a role in establishing persistence in S. aureus. For the first time, we identified a novel mechanism of persistence in S. aureus mediated by ArgJ in maintaining persistence to different antibiotics and stresses and also virulence in-vivo. For S. aureus, tackling persistence may be a solution to reducing the rate of drug resistance. Our findings not only improve our understanding of mechanisms of persistence but also provide insights into novel therapeutic targets for developing new and more effective drugs that eradicate persistent S. aureus infections.

)LJXUH6XPPDU\RIWKHJHQHWLFVFUHHQ $ :RUNIORZRIWKHJHQHWLFVFUHHQ %

Identification of 75 genes involved in gentamicin persistence (see Table S1). All pathway analyses were performed using Kyoto Encyclopedia of Genes and genomes (KEGG) database.

The results from the screen were reproducible and performed twice. (C) The arginine biosynthesis pathway involved in persister formation in S. aureus.

Figure 3.2. Comparison of persistence phenotypes of the ArgJ mutant and parental strain

USA300. (A & B) Parental strain USA300 and the ArgJ mutant had no difference in growth dynamics. The ArgJ mutant showed defective persistence to both (C) gentamicin (60

µg/ml, >10X MIC) and (D) rifampicin (2 µg/ml, > 10XMIC). (E) Strains with mutations in ArgB,

ArgF, and TrpA did not show a defect in persistence. More killing was seen in the ArgJ mutant by (F) heat stress of 58oC and (G) low pH of 4.0. Data are representative of three independent experiments. * = p< 0.05 by two-way ANOVA (where indicated) or Student’s t-test.

Figure 3.3. Persistence to antibiotics can be restored in the ArgJ mutant by complementation.

(A) The ArgJ mutant was transformed with the wildtype argJ gene using shuttle vector pRAB11. Upon 6 day exposure of (A) gentamicin and (B) rifampicin, the ArgJ mutant complemented with argJ gene showed partial restoration of persistence compared to the empty

vector. (C) The supplementation of L-arginine (30 mM) can restore gentamicin persistence. The effect of amino acid supplementation to rescue ArgJ's persistence phenotype is specific to the amino acid arginine. Amino acids with similar chemical properties to arginine (e.g. histidine and lysine) did not restore persistence. Data are representative of three independent experiments.

Figure 3.4. Arginine biosynthesis through the Arg pathway plays a role in persistence. In stationary phase (B) or gentamicin-treated (C) cultures, there is at least 2-fold more expression of

Arg pathway genes argC and argG in USA300 compared to the ArgJ mutant. Differences among the expression of the arg genes and both the expression of proC and putA are also statistically significant. Data are representative of three independent experiments. ** = p < 0.005, * = p <

0.05 by Student’s t-test.

B C

Figure 3.5. The ArgJ mutant has attenuated virulence in vivo. (A)The ArgJ mutant showed attenuated killing of C. elegans compared to parent strain USA300. Three days post-infection, mice (n=4) infected with the ArgJ mutant had a lower bacterial load in both the (B) spleens and

(C) kidneys. Genetic complementation with the wildtype argJ gene restores virulence in the ArgJ mutant. (* = p < 0.05 by Student’s t-test).

Table 3. 1. Oligonucleotide primers used for qRT-PCR

Gene Forward Primer (5' to 3') Reverse Primer ( 5' to 3' ) argB AATCGAGCCACACTTTGTTAATG CTGCAATGAGCGTGTGTTTAG argC TTCAGAATGGCAATCGTTTGATA GGGAAACAGCCAGGATTAGAAA argG AAGCATTAGAAACGATTACGTTAACG CAAATTGCTTCTCAATGATTGGTT argJ TGGTGGTATGCACATCGGTTT AGACGATGAGTAAATCCATCCAAAG proC TGCCAAAATCCAGTTGCTAGAA GCCAGTAACAGAGTGTCCAACTTG putA CCTTATGGCGATGATTGGTTTG CCAGCAGGTTTCACAAATTCTT

16s CCGCATGGTTCAAAAGTGAAA GCAGCGCGGATCCATCTAT

CHAPTER 4

IDENTIFICATION OF GENES REGULATING CELL DEATH IN

STAPHYLOCOCCUS AUREUS

Introduction

With the increasing prevalence of infections with antibiotic resistant bacteria and the medical burden of treating persistent bacterial infections, novel therapeutic approaches are in dire need. One of the major players of bacterial infections is S. aureus which is colonized in over one-third of the human population and can cause over 80% of the skin and soft tissue infections.

Upon spreading into the bloodstream, S. aureus can also cause infections such as bacteremia, pneumonia, meningitis and persistent infections such as osteomyelitis, endocarditis, and those related to biofilm formation such as on prosthetic implants [5]. Due to emerging resistance and high risk of nosocomial infections and community-acquired infections, S. aureus infections are a major public health concern. The risk of hospital mortality in patients with a S. aureus infection can be increased up to five-fold compared to patients uninfected with S. aureus [4].

To develop better treatments for clearing S. aureus and other bacterial infections, understanding how bacterial cells die is crucial. By definition, bacteria treated with bactericidal antibiotics such as the β-lactams, quinolones, and aminoglycosides are killed. Much is known about how the drug makes contact with its target in the bacterial cell. β-lactams bind to PBP disrupting proper cell wall synthesis; quinolones bind to topoisomerase blocking DNA replication and aminoglycosides bind to ribosomal proteins resulting in mistranslated proteins

[15, 83, 166]. Despite having different drug-target interactions, the downstream effects of drug lethality are similar. Bacteria treated with cidal antibiotics have pathways such as SOS response,

TCA cycle, ROS formation, and Fe-S cluster formation activated suggesting that regardless of the drug target, the bacteria has a complex yet similar process leading to cell death [88, 89].

Mechanisms pertaining to bacterial cell death were mainly characterized in toxin- antitoxin systems. One of the better characterized TA modules in the context of bacterial cell death is the MazEF module found in E. coli, and other species such as Salmonella, Neisseria,

Streptococcus and Mycobacterium. Upon exposure to stresses such as nutrient depletion, DNA damage, temperature, antibiotics, and oxidative radicals, the MazF anti-toxin is degraded and hence, the MazE toxin can degrade cellular mRNA causing cellular shutdown [91, 93]. In particularly to S. aureus, the CidA and LrgA proteins, which are holin-like proteins, were proposed to play a role in death and lysis of S. aureus [96, 97]. CidA is a holin that has a positive effect on cell death due to oligomerization into the cell membrane causing membrane disruption whereas LrgA is an antiholin that inhibits the activity of CidA [100]. However, the specific process to how CidA and LrgA regulate cell death is poorly defined. Meanwhile, central metabolic pathways consisting acetate metabolism through activation of CidC, pyruvate oxidase, or inactivation of the phosphotransacetylase-acetate kinase (Pta-AckA) pathway were also suggested to cause cell death in S. aureus [167, 168].

Even though, cell death is such an important process of a living organism, little is known about the mechanisms. High-throughput screens have been developed to study the cell death mechanism of unicellular organisms such as S. cerevisiae upon stress signals from temperature and acetic acid, both of which also induce death in bacteria. Here, we performed a high- throughput genetic screen using a transposon-mutant library of USA300 to identify genes involved in the mechanisms of cell death in S. aureus [151]. Under multiple death stimuli, we identified 27 genes whose mutations caused the bacteria to be more stress-tolerant, death- resistant and 10 genes whose mutations caused the bacteria to be more death-sensitive.

Methods

Culture Media, Antibiotics, and Chemicals

Meropenem, moxifloxacin, rifampicin, gentamicin, and erythromycin were obtained from Sigma-Aldrich Co. (St. Louis, MO, USA). Stock solutions were prepared in the laboratory, filter-sterilized and used at indicated concentrations. Bacterial strains used in this study include USA300 and the Nebraska-Transposon Mutant Library (NTML) [151].

S. aureus strains were cultured in tryptic soy broth (TSB) and tryptic soy agar (TSA) with the appropriate antibiotics and growth conditions. Transposon-insertion mutants grew in erythromycin (50 µg/ml), the antibiotic selective marker.

Genetic Screen to identify cell death mutants

For the heat-ramp, we performed the assay as described [105]. Briefly, we normalized the concentration of the bacteria to OD600= 0.5 using PBS as the diluent. Then, we placed the samples in the thermocycler with a protocol following: 30oC for 1 minute, ramp from 30 oC to

62 oC with a step-interval of 0.5 oC per 30 seconds. For acid stress, acetic acid (6 mM) were added into stationary phase cultures and incubated overnight. To enumerate for cell counts, the mutant library was replica transferred to TSA plates to score for mutants that failed to grow after stress. For viability staining, SYBR Green I/PI staining was performed as described [169, 170]. SYBR Green I (10,000× stock, Invitrogen) was mixed with PI (20 mM,

Sigma) in distilled H2O with a ratio of 1:3 (SYBR Green I to PI) in 100 ul distilled H2O and stained for 30 minutes in room temperature. Prior to the heat-ramp, the SYBR Green I/PI dye was added to the bacteria as the heat-ramp impaired the uptake of the dye. For the acetic acid stress, the dyes were added at the end of the exposure [105]. The green and red fluorescence

intensity was detected using a Synergy H1 microplate reader by BioTek Instruments (VT, USA) at excitation wavelength of 485 nm and 538 nm and 612 nm for green and red emission, respectively. The live/dead ratio was calculated by dividing the green/red fluorescence.

Persister Assays

Selected drugs were added to overnight stationary phase cultures for the 6 days. At the selected time points, bacterial cultures were washed with 1X PBS to rid of stress, serially diluted, and plated onto TSA with no drugs for cell enumeration [150].

Nematode-killing Assay

C. elegans N2 Bristol worms (Caenorhabditis Genetics Center) were synchronized to the same growth stage by treatment with alkaline hypochlorite solution as described [153]. Worms of adult stage were washed and suspended in bleaching solution with 5% hypochlorite for 9 minutes to lyse all the adult stages but keeping the eggs intact. Bleach was removed by centrifugation at 1,500 rpm for 1 minute and washed three times with M9 buffer. The pellet was incubated in M9 buffer at 20 oC with gentle agitation and proper aeration. L4 stage adult worms were obtained after 48 hours at 20 oC. For each assay, 20-30 worms were added to liquid M9 buffer supplemented with 5- Fluoro-2'-deoxyuridine (10 µM) to inhibit progeny formation. S. aureus (106CFU) that were grown overnight at 30 oC in TSB containing the appropriate antibiotics as needed were added into the buffer containing the worms. The sample were scored for live and dead worms every 24 hours.

Results

Identification of Genes Important for Cell Death Resistance

To better understand the mechanisms of cell death, we performed a genetic screen using the Nebraska Transposon Mutant Library (NTML) which contained mutations in all the non-

essential genes of USA300, the most common circulating community-acquired MRSA strain in the United States [151]. To design our assay, we utilized the heat-ramp assay that have been used to evaluate cell death programs in yeast [105] . To determine viability, we employed both the traditional agar replica plating for visualization of viable growth on solid media but also stained cells with SYBR Green I/PI, a viability stain that can detect both live and dead cells [169, 170].

Using a cidA mutant [97] which has been shown to be death-resistant control and the parental strain of USA300 as a death sensitive control, we optimized the condition of our ramp to show the biggest difference between both the death-resistant and death-sensitive phenotype based on agar plating and the live/dead ratio from viability staining with SYBR Green I/PI.

For identification of death-resistant mutants, we searched for clones that are still viable on agar plating (as opposed to USA300 that no longer show colony growth from replica plating) and a live/dead ratio that is higher than our death-resistant mutant, cidA. After the heat ramp exposure, we identified 74 that were death-resistant (Fig. 4.1A). Because we cannot pinpoint a specific gene to be the ultimate regulator of cell death, we generated a list of potential core regulators of cell death. In order to identify core genes and pathways, we exposed the whole transposon-mutant library to acetic acid stress (Fig. 4.1A).

Acetic acid has been shown to induce cell death in S. aureus and has been used for this purpose in high-throughput screens in yeast as well [70, 97, 104, 171]. Upon treatment with acetic acid (6 mM) overnight, 28 out of the74 heat-ramp resistant mutants were also acetic acid resistant. A majority of the genes identified were transporters (n=9), involved in transcription

(n=4), metabolism (n=3), encode peptidases (n=2) and phosphatases & kinases (n=2). For genes encoding transferases, and proteins involved in stress response, nucleic acid synthesis and protein synthesis, one candidate was found in each category (Fig. 4.1C, Table 4.1).

Death-Resistant Strains are more virulent in vivo

Upon infection inside a host, in addition to drug stress, the bacteria are exposed to various types of stresses such as oxidative stress especially in the phagosomes of immune cells.

We then tested if our death-resistant mutants were more virulent in causing an infection inside the host. After infection of C. elegans with the top 4 death-resistant mutants (folD,

USA300HOU_0997, sspA, USA300HOU_0232) and parental strain USA300, we observed that all four mutants significantly decreased the survival of the C. elegans and killed the worms faster than USA300 (Fig. 4.2). By 2 days post-infection, the survival of worms infected with our death- resistant mutants had a survival rate of 36% or lower while worms infected with USA300 had a survival of 50%. The most virulent strain was Tn::USA300HOU_0232, a mutation in an iron transporter, as it caused the greatest mortality in the worms, resulting with only 22% survival of the worms by day 2 (Fig. 4.2). Our data suggest that bacteria that are more death-resistant could potentially cause more serious infections.

Identification of Genes Important for Cell Death Sensitivity

To more fully understand the regulatory networks of cell death pathways, it is crucial to examine the genes whose mutations cause the cells to be more sensitive to stress as well. Using the data from our two screens (heat-ramp and acetic acid stress), we adjusted the parameters of data analysis to distinguish mutants that were cell death sensitive. Unlike the cut-offs mentioned previously, cell death sensitive mutants were identified as having no viable growth on agar media and a live/dead ratio that was lower than USA300.

Our screen revealed that 92 mutants were hypersensitive to the heat-ramp, of which 10 were also sensitive to acetic acid stress (Fig. 4.1B). Transporters were the more abundant (n=4)

followed by genes involved in metabolism (n=2), and lastly with genes involved in transcription

(n=1) and cell wall synthesis (n=1). Two of the candidates were hypothetical proteins.

Interestingly, two (USA300HOU_2029, gltT ) of the top four death-sensitive mutants

( USA300HOU_0329, USA300HOU_2029, USA300HOU_2386, gltT ) harbored mutations pertaining to glutamate levels inside the cell (Fig. 4.1D, Table 4.2).

Death-Sensitive Strains are Less Persistent in vitro

One of the reasons why S. aureus can cause persistent and recalcitrant infections is due to its ability to form persister cells. Persisters are dormant cells that are formed during stressed conditions and upon stress removal, the bacteria can revert back to a growing state and consequently, cause a relapse in infection [21]. Bacterial persistence can also be viewed as cells with a strong anti-death program. Given now that we have identified genes whose mutation render the bacteria death-sensitive, we then wanted to know if these mutations also lead to defective in persistence, forming lower amounts of persister cells. Such mutations could then potentially be drug targets for clearing persistent infections.

Persisters are enriched by treating stationary phase bacteria with high concentrations of bactericidal antibiotics (usually at least 10X MIC). Cells are then washed to rid of stress and plated on solid medium with no drug for CFU enumeration [150]. We exposed the top 4 death- sensitive mutants to bactericidal antibiotics with different mechanism of actions: gentamicin, meropenem, rifampin, and moxifloxacin. Upon 6-days post exposure of antibiotics, all 4 mutants showed a defect in persistence when exposed to all different classes of antibiotics (Fig. 4.3). The amount of persisters is dependent on the type of stress which can be seen here since the absolute amount of persisters changes among the drugs tested [21]. However, the overall amount of

persister cells formed by the death-sensitive mutants was significantly lower than USA300. The defect in persistence was the most prominent for gentamicin stress. Under gentamicin stress, both mutants of ZnuA and USA300HOU_2029 were completely killed by 4 days-post exposure while USA300 still have over 107 CFU/ml (Fig. 4.3). In non-stressed conditions, no growth defects and decrease in viability of cells were observed.

Discussion

To our knowledge, this is the first comprehensive study to identify genes and pathways that play a role in anti-death and pro-death programs in S. aureus. These data suggest that not only are transporters important in regulating cell death pathways in S. aureus, but in particular, glutamate metabolism and glutamate transport are important for transformation of a bacteria cell into a both a more death-resistant and a persistent phenotype under stressed conditions. Our findings show that mutations in genes involved in intracellular glutamate level such as

USA300HOU_2029 and gltT can decrease cell viability and persistence under antibiotic stresses, and environmental stresses such as temperature changes and low pH.

Currently, relevant cell death pathways identified as important for S. aureus cell death during cidal stress are those related to ROS generation and radicals released from the Fenton

Reaction Pathway. One of the top death-resistant mutants harbored a mutation in

USA300HOU_0232 which is an iron transporter. Considering how ROS causes cell death [89], a mutated iron transporter may cause the cells to be death-resistant by limiting the amount of iron that can catalyzes the Fenton reaction and thus limit the amount of ROS inside the cell [172].

While other studies have searched for cell death genes in S. aureus, the specific pathway of glutamate metabolism has never been identified as core cell death proteins. Although further

research is needed to explore how glutamate metabolism plays a protective role in cell death, hypotheses based on what is currently known about cell death can offer insights to how intracellular glutamate levels could fit in the program of S. aureus cell death. Sadykov et al. identified that inactivation of the phosphotransacetylase-acetate kinase (Pta-AckA) pathway which normally generates acetate from acetyl-CoA leads to cell death in S. aureus under glucose and oxygen excess. In bacteria, glutamate fermentation can occur via 3-methylaspartate which produces pyruvate followed by acetyl-CoA. Considering that acetate can be produced from acetyl-CoA, our findings may help explain the events that occur upstream of Pta-AckA activation [173]. Additionally, the lethality induced by cidal antibiotics has been shown to be due to ROS generation and radical generation from the Fenton reaction suggesting that death mechanisms result in oxidative responses within the cell [89]. In Francisella, glutamate transporter GadC has been shown to neutralize reactive oxygen species [174]. A study performed to evaluate the bactericidal effect of CO-releasing molecules (CO-RMs) showed that CO-RMs stimulated the production of intercellular ROS in the bacteria which was abolished when glutamate was supplemented to the culture [175].

These data further suggest that cell-death resistance and cell persistence could be facilitated by similar molecular programs within the cell. While the specific mechanisms regarding S. aureus persistence is largely unknown, it appears that pathways that affect ATP synthesis, cellular metabolism, DNA repair, and stress response pathways all play a role [150].

Similar to the findings of this study, the overall functional groups that appear to play a role in cell death are similar to cell persistence. Interestingly, two of the four top hits in my screen for cell death participated in glutamate synthesis. Our results from a screen performed to identify genes involved in persistence to rifampicin (See Chapter 2) showed that genes gltS, gltD, gltA,

all of which are involved in glutamate synthesis, were important. Intriguingly, we found the protein ArgJ to be a potential core regulator for S. aureus persistence in various stresses

(different classes of drugs, heat, and low pH) (See Chapter 3). Glutamate is the substrate for

ArgJ and since mutants with impaired glutamate biosynthesis and transport showed both death- sensitive and defective persistence phenotypes, it can be speculated that glutamate can be the main driver of anti-death programs where arginine synthesis is a potential downstream effector pathway.

It should be noted here that this is not the first time mechanisms of bacterial cell death are shown to overlap with mechanisms of persistence.. For example, our lab has shown that tmRNA involved in trans-translation which tags truncated proteins for degradation plays an important role in persistence in bacteria such as E. coli and M. tuberculosis under exposure to various antibiotics and stresses such as acid, heat, and peroxide [118, 176]. With regards to cell death, a study has shown that under severe stress, the ribosomal elongation factor 4 (EF4) can inhibit the tmRNA/ClpP degradative pathway and generate ROS. Additionally, EF4 acts in the same pathway as MazF, a toxin with proposed cell death roles in bacteria [177]. The connection among EF4, MazF, ROS, and tmRNA demonstrates how proteins that have a role in cell death may also play a role in cell persistence.

A similar approach of a “heat-ramp” stress was used in one study on B. subtilis using a water bath [178]. Bacteria were placed in a water bath at 65 oC and at different temperatures as the bacteria warmed up (40, 50, 57, 58, 59, 60 oC), RNA was extracted to evaluate the transcriptomic profile of gene expression during heat stress. Kort et al. revealed that heat shock proteins, sporulation, competence, and carbon metabolism were important [178]. It is worth noting that while they identified sporulation factors as top genes involved in cell death,

sporulation is a process that is limited to Bacillus and Clostridia species. However, it does not undermine the importance of stress response pathways, as sporulation is activated under nutrient limiting conditions. While we did not identify similar genes, carbon metabolism was identified in both our screens. Heat shock proteins were heavily enriched in their study but not ours which can be attributed to the candidates in our library. Our transposon mutant library only contains mutants of non-essential genes in the S. aureus genome and only two candidates out of 1,952 mutants in the library were heat shock proteins [128].

This work adapted assays that are used in the yeast community to identify genes involved in yeast cell death. In the high-throughput screen consisting of yeast mutants and even in yeast strains of different backgrounds, categories of genes that were redundant among the studies include carbohydrate metabolism, transcription factors, and ion transport [74, 75, 179].

Interestingly, amino acid transport was the most significantly enriched term for genes involved in positive regulation of acetic acid-induced death; two of the identified transporters were involved in metabolism of glutamate (GDH1 and GDH2) [75].

Despite the significant findings of this study, there are some limitations. First, while we generated a list of the mutants that are death-sensitive and death-resistant to both heat ramp and acetic acid stress, the screens performed here are only a snapshot of what is happening in the cell. The phenotypes seen here were determined by the condition of our assays and can be affected by the level of stress (e.g. concentration of the acetic acid, ramping time and temperature for the heat ramp) [105]. Although we included both a wild type

USA300 strain and a CidA mutant, a known death-resistant mutant, to optimize our conditions, it is important to recognize that our screen may not be comprehensive in finding all the mediators of cell death. While our goal was more conservative and was to search for core

regulators, the significant role of other genes and pathways involved in cell death that were not classified as core regulators should not be undermined. Our screen only explored cell death at the DNA level and further protein studies would provide more comprehensive insight into the effects of gene transposon insertions in regulating cell death. This mutant library only contained all the non-essential genes in USA300. The fact that cell death is an important program for any living organism should not be overlooked. Cell death regulators may be essential genes which are not included in the library. As with any mutant library, secondary mutations may have occurred in some mutants that could affect the phenotypic outcomes.

However, whole genome sequencing was performed on all identified candidates and no other mutations besides the one listed were found. Additionally, the screens were performed twice and were followed by individual confirmation. The same subculture from the same stock was also used each time to ensure reproducibility among all replicates.

To our knowledge, this is the first report on identification of S. aureus cell death genes from a whole genome perspective. Our extensive screen also offers insights to common core mechanisms that are relevant to not only cell death but bacterial persistence, a phenomenon that’s at the core of recalcitrant infections and biofilm formations. Our studies provide insights to possible new druggable targets, biomarkers for recalcitrant infections for diagnostic purposes and novel vaccine targets for prevention of bacterial infections. The similarity in functional groups found between our study and other yeast studies suggest that our work also sheds light into cell death pathways of eukaryotic systems such as pathogenic fungi and cancer stem cells.

)LJXUH,GHQWLILFDWLRQRIJHQHVLQYROYHGLQFDXVLQJEDFWHULDOFHOOGHDWKUHVLVWDQFHDQGFHOO

GHDWKVHQVLWLYLW\ $ 6XPPDU\RIWKHQXPEHURIFDQGLGDWHVLGHQWLILHGIURPWKH1HEUDVND

7UDQVSRVRQ0XWDQW/LEUDU\ 170/ DVEHLQJPRUH $ GHDWKUHVLVWDQWDQG % GHDWKVHQVLWLYHLQ

heat-ramp and acetic acid stress. (B) The breakdown of the respective categories of genes whose mutations cause (C) death-resistance and (D) death-sensitivity in both stresses.

)LJXUH6WUDLQVWKDWDUHPRUHGHDWKUHVLVWDQWFDXVHGPRUHYLUXOHQWLQIHFWLRQVLQ&HOHJDQV

&HOHJDQV Q LQIHFWHG &)8 ZLWKPXWDQWVVKRZLQJUHVLVWDQFHWRFHOOGHDWKVKRZHG

LQFUHDVHGPRUWDOLW\WKDQ&HOHJDQVLQIHFWHGZLWKSDUHQWDOVWUDLQ86$

)LJXUH6WUDLQVWKDWDUHPRUHGHDWKVHQVLWLYHDOVRVKRZGHFUHDVHGSHUVLVWHQFHWREDFWHULFLGDO

DQWLELRWLFV0XWDQWVWKDWZHUHPRUHFHOOGHDWKVHQVLWLYHVKRZHGGHIHFWLYHSHUVLVWHQFHWR

JHQWDPLFLQ JPO PHURSHQHP JPO PR[LIOR[DFLQ JPO DQGULIDPSLQ

JPO XSRQSURORQJHGGUXJH[SRVXUHXSWRGD\V

Table 4.1. Top 4 genes whose mutations result in resistance to cell death in both heat-ramp and acetic acid stress

Accession Number Gene name Function KEGG Ontology (if applicable) USA300HOU_1008 folD Methylenetetra- Metabolism of cofactors and hydrofolate vitamins; One carbon pool by dehydrogenase; folate methenyltetrahydrofolate cyclohydrolase USA300HOU_0997 Bifunctional N- Metabolism acetylmuramoyl-L- alanine amidase/mannosyl- glycoprotein endo- beta-N- acetylglucosaminidase USA300HOU_0996 sspA Glutamyl Cellular community endopeptidase prokaryotes; Quorum sensing; Serine Peptidases of chymotrypsin family USA300HOU_0232 Iron ABC transporter Environmental Information membrane binding Processing; Membrane protein transport; ABC transporters; Mineral and organic ion transporters; Iron transporter

Table 4.2. Top 4 genes whose mutations result in sensitivity to cell death in both heat-ramp and acetic acid stress

Accession Number Gene name Function KEGG Ontology (if applicable) USA300HOU_0329 ABC transporter- ATP Protein families: binding protein signaling and cellular processes; Transporters USA300HOU_2029 Amidohydrolase Metabolism: Amino acid metabolism; Alanine, aspartate and glutamate metabolism USA300HOU_2386 znuA Zinc transport system Environmental substrate-binding protein Information Processing; Membrane transport; Metallic cation, iron- siderophore and vitamin B12 transporters; Zinc transporter USA300HOU_2366 gltT, gltP Proton glutamate Protein families: symport protein signaling and cellular processes; Electrochemical potential-driven transporters

CHAPTER 5

DRUG COMBINATIONS TARGETING GROWING AND PERSISTER CELLS

ERADICATE CHRONIC STAPHYLOCOCCUS AUREUS INFECTION

This Chapter is currently in preparation for publication as: Yee, R., Yuan Y., Tarff A., Brayton C., Gour N., Feng J., Shi W., Zhang Y. 2018. Drug Combinations Targeting Growing and Persister Cells Eradicate Chronic Staphylococcus aureus Infection. In preparation.

Introduction

Methicillin-resistant Staphylococcus aureus (MRSA) strains are highly prevalent in healthcare and community-acquired staphylococcal infections [180]. The mortality rate associated with MRSA infection is as high as 40% [181, 182]. As an opportunistic pathogen, S. aureus is the most common cause of skin infections and can also cause chronic diseases such as endocarditis, osteomyelitis, and prosthetic joint infections [183-185]. In particular, indwelling devices are conducive to biofilm formation, complicating treatment and leading to prolonged infections. Persistent and chronic infections are a burden to public health as they increase the length of hospital stay, relapse, cost of treatment, and risk of death by at least three-fold [4].

Bacteria in biofilms are more resistant to antibiotics compared to planktonic cells [186].

Studies have shown that antibiotics do indeed penetrate the biofilm but they do not always kill the bacteria, suggesting that tolerance to treatment is not due to impaired antibiotic penetration

[38, 39], but mostly due to dormant non-growing or slowing growing persister bacteria. Bacteria inside the biofilm grow slowly, are representative of stationary phase bacteria, and can comprise persister cells due to the high cell density, nutrient and oxygen limiting environment inside the biofilm matrix [78].

First described in 1942, Hobby et al. found that 1% of S. aureus cells were not killed by penicillin and these were called persister cells [28]. The persister cells were not resistant to penicillin and hence, did not undergo genetic changes; these cells were phenotypic variants that became tolerant to antibiotics [29]. Similarly, a clinical observation was also made as penicillin failed to clear chronic infections due to the presence of persister cells found in patients [29].

While the mechanisms of S. aureus persistence were largely unknown for a long time, recent studies have shown that pathways involved in quorum sensing, pigmentation production, and

metabolic processes such as oxidative phosphorylation, glycolysis, amino acid and energy metabolism [47, 150, 187-189] are responsible. Understanding the pathways of persistence would facilitate development of novel drugs and therapeutic approaches to more effectively eradicate persistent bacterial infections.

An effective drug combination approach to eradicate persistent infections may be implicated by the drug combination principle for treatment of tuberculosis. Tuberculosis treatment was shortened from 9-12 months to 6 months after inclusion of the persister drug pyrazinamide (PZA) in the drug combination [190]. PZA’s activity in killing persisters, unlike the other drugs used to treat tuberculosis, is crucial in developing a shorter treatment [118, 190-

192]. A drug like PZA validates an important principle of use of a persister drug in combination with other drugs targeting both persisters and growing cells in formulating an effective therapy for chronic persistent infections [193]. More recently, a similar approach has been used to identify effective drug combinations to eradicate biofilm-like microcolony structures consisting of heterogeneous cells of Borrelia burgdorferi [194].

Using this approach, in a recent study aimed to identify drugs targeting non-growing persisters, we used stationary phase S. aureus as a drug screen model and identified drugs with high activity against S. aureus persisters [195]. However, their activities in killing biofilms and utility in drug combinations have not been evaluated for activity against biofilms in vitro or in related infections caused by S. aureus in vivo. Therefore, we developed drug combinations that can more effectively eradicate S. aureus biofilms by formulating drug combinations that have high activities against growing bacteria and non-growing persisters in biofilm model in vitro. In addition, we established a chronic skin infection mouse model for S. aureus using “biofilm seeding” and evaluated drug combinations in clearing the infection in this persistent skin

infection model. Here, we show that combining meropenem, daptomycin targeting growing bacteria, and clinafloxacin targeting persister bacteria led to complete eradication of S. aureus biofilm not only in vitro but more importantly also in vivo in a murine model of persistent skin infection.

Methods

Culture Media, Antibiotics, and Chemicals

Staphylococcus aureus strain Newman, USA300, CA-409, CA-127, GA-656, were obtained from American Type Tissue Collections (Manassas, VA, USA). S. aureus strains were cultivated in tryptic soy broth (TSB) and tryptic soy agar (TSA) from Becton Dickinson

(Franklin Lakes, NJ, USA) at 37°C. Vancomycin, gentamicin, rifampicin, levofloxacin, ciprofloxacin, moxifloxacin, and oritavancin were obtained from Sigma-Aldrich Co. (St. Louis,

MO, USA). Daptomycin, meropenem, tosufloxacin, and clinafloxacin were obtained from AK

Scientific, Inc. (Union City, CA, USA). Stock solutions were prepared in the laboratory, filter- sterilized and used at indicated concentrations.

Microtiter plate biofilm assays.

S. aureus strains grown overnight in TSB were diluted 1:100 in TSB. Then, 100-μl aliquots of each diluted isolate were placed into a 96-well flat-bottom microtiter plate and statically incubated for 24 h at 37°C [196]. Planktonic cells were removed and discarded from microtiter plates. Drugs at the indicated concentrations were then added onto the biofilms at a total volume of 100 μl in TSB. To determine the cell and biofilm density, the supernatant was removed from the well and the biofilms were washed twice with PBS (1X). To enumerate

bacterial cell counts, the biofilms in the wells were resuspended in TSB and scraped with a pipette tip before serial dilution and plating. To assess cell viability using the ratio of green:red fluorescence to determine the ratio of live:dead cells, respectively, the biofilms were stained with

SYBR Green I/Propidium Iodide dyes as described [169, 170]. Briefly, SYBR Green I (10,000× stock) was mixed with PI (20 mM) in distilled H2O at a ratio of 1:3, respectively. The SYBR

Green I/PI staining mix was added to each sample at a ratio of 1:10 (10 µl of dye for 100 µl of sample). Upon incubation at room temperature in the dark for 20 min, the green and red fluorescence intensity was detected using a Synergy H1 microplate reader by BioTek

Instruments (Winooski, VT, USA) at excitation wavelength of 485 nm and 538 nm and 612 nm for green and red emission, respectively. To visualize biofilm biomass, biofilms were stained with crystal violet (0.1%) for 15 minutes in room temperature. Excess dyes were washed with water and the biofilms were left to air dry. Images were recorded using Keyence BZ-X710

Microscope and were processed using BZ-X Analyzer provided by Keyence (Osaka, Japan)

Mouse skin infection model

Female Swiss-Webster mice of 6 weeks of age were obtained from Charles River. They were housed 3 to 5 per cage under BSL-2 housing conditions. All animal procedures were approved the by the Johns Hopkins University Animal Care and Use Committee. S. aureus strain USA300 were used in the mouse experiments. Mice were anesthetized and then shaved to remove a patch of skin of approximately 3 cm by 2 cm. Bacteria of indicated inoculum size and age were subcutaneously injected into the mice. For log phase inoculum, bacteria grown overnight were diluted 1:100 in TSB and grown for 2 hrs in 37°C at 220 RPM. For stationary phase inoculum, overnight cultures of bacteria grown at 37°C were used. For an inoculum of cells recovered from biofilms, biofilms were first grown in microtiter plates as described

previously, and then resuspended and scraped up with a pipette tip. Quantification of all inoculum was performed by serial dilution and plating. After 1 week post-infection, treatment was started (Table 5.1). Skin lesion sizes were measured at indicated time points using a caliper.

Mice were euthanized after 1 week post-treatment and skin tissues were removed, homogenized, and serial diluted for bacterial plating and counting. Throughout the study, mice were monitored daily for signs of distress (e.g. reduced movement, hunched posture, piloerection, weight loss).

Histology

Skin tissues were dissected, laid flat, and fixed for 24 hrs with neutral buffered formalin.

Tissues were embedded in paraffin, cut into 5-um sections, and mounted on glass slides. Tissue sections were stained with hematoxylin and eosin for histopathological scoring. Tissue sections were evaluated for crust formation, ulcer formation, hyperplasia, inflammation, gross size, and bacterial count and were assigned a score on a 0–3 scale (0 = none, 1 = mild, 2 = moderate, and

3 = severe). The cumulative pathology score represented the sum of each individual pathology parameter. Scoring was performed by an observer in consultation with a boarded veterinary pathologist. Representative images were taken using a Keyence BZ-X710 Microscope.

Statistical Analyses

Statistical analyses were performed using two-tailed Student’s t-test and two-way

ANOVAs where appropriate. Mean differences were considered statistically significant if p was

<0.05. All experiments were performed in triplicates. Analyses were performed using GraphPad

Prism and Microsoft Office Excel.

Results

Commonly used treatments for MRSA have poor activity against biofilms in vitro

While vancomycin is extremely effective to kill MRSA bacteria in vitro, monotherapy with vancomycin may not be the most effective to clear chronic infections with S. aureus. For conditions like osteomyelitis and prosthetic joint infections, treatment with vancomycin as a monotherapy or in combination for at least 6 weeks are recommended. Combinations such as doxycycline + rifampin for up to 10 days, vancomycin + gentamicin + rifampin for at least 6 weeks are recommended to treat chronic infections such as recurrent tissue infections and endocarditis on prosthetic valves, respectively [112]. We first evaluated the activity of the above drugs in killing biofilms in vitro using traditional bacterial cell counts (Fig. 5.1A), viability assessment by SYBR Green I/Propidium Iodide staining that has been developed to screen for drugs targeting borrelia persister bacteria [169] (Fig. 5.1B), and staining of absolute biofilm (Fig.

5.1C-G). We found that such clinically used combinations are not completely effective against biofilms. After 4-days post treatment, biofilm bacteria were not completely eradicated as shown by significant numbers of bacteria remaining (Fig. 5.1).

Identification of drug combinations with strong anti-biofilm activity

To address the clinical unmet need of better treatments against chronic infections, we hypothesize that a drug combination that includes drugs that act on growing bacteria such as cell wall (e.g. vancomycin, meropenem) or cell membrane inhibitors (e.g. daptomycin) plus a drug that acts on persister bacteria will be a more potent drug combination in eradicating biofilm bacteria. Previous studies from our lab identified tosufloxacin and clinafloxacin as having strong anti-persister activity against S. aureus [195]. To identify a potent combination, we tested

various drug combinations that include drugs against both growing bacteria and non-growing persisters in an in vitro biofilm model. The proposed drug combination consists of cell wall inhibitors which have great activity against growing bacteria, a cell membrane inhibitor which can have both activity against growing and persister bacteria, and lastly, an anti-persister drug that targets the persister population. Biofilms of S. aureus strain USA300, a common circulating strain of community acquired-MRSA (CA-MRSA), were grown in 96-well microtiter plates to allow biofilm formation on the bottom of the wells [196]. While we previously showed that tosufloxacin had robust activity against persister cells, the drug combination of vancomycin/meropenem + daptomycin + tosufloxacin achieved only partial eradication, with 105

CFU/ml in biofilms after treatment (Fig. 5.2A-B). In contrast, combination of vancomycin/meropenem + daptomycin + clinafloxacin showed absolute eradication of biofilms by 4-days post-treatment as indicated by 0 CFU and a live/dead ratio below the limit of detection

(Fig. 5.2A-B). Although we used same molar concentrations of each drug (50 uM each of drug) in our drug screen for comparison of relative drug activity, to evaluate the activity of the combination in a more clinically relevant manner, we treated the biofilms with the drugs at their

Cmax concentrations (Table 5.1). Our findings with Cmax drug concentrations were confirmatory as the combination of vancomycin/meropenem + daptomycin + clinafloxacin still achieved complete eradication while our no treatment control and the clinically used combination of doxycycline plus rifampin could not. The clearance of biofilms was confirmed by both viable CFU counts and viability staining (Fig. 5.2C-D).

We then tested the eradicating potential of the drug combination of meropenem + daptomycin + clinafloxacin against different MRSA S. aureus strains. These included other CA-

MRSA clinical isolates CA-409, CA-127 and hospital-acquired MRSA strain GA-656. Complete

eradication (0 CFU/ml) and undetectable levels of live cells (under the limit of detection) were found for all of the MRSA strains tested after 4 days of treating biofilms with our combination meropenem + daptomycin + clinafloxacin (Fig. 5.2E-F).

Clinafloxacin is a member of the fluoroquinolone class of antibiotics and hence, the drug inhibits DNA replication by binding to DNA gyrase. As our results suggest, clinafloxacin is a powerful anti-persister drug. We then wanted to rank the anti-biofilm activity of different fluoroquinolones to determine whether the robust anti-biofilm activity of clinafloxacin used in combination is unique to the drug itself or can be replaced by other members of fluoroquinolone antibiotics. To do so, we used the S. aureus Newman strain due to its susceptibility to many fluoroquinolones as we wanted to eliminate any confounding factors due to inherent drug resistance. While other fluoroquinolones such as ciprofloxacin, levofloxacin, and moxifloxacin have certain anti-persister or anti-biofilm activity when used in combination with meropenem and daptomycin after 4-days of treatment, the drug combination with clinafloxacin was indeed the most active and was the only combination that achieved complete eradication. By contrast, biofilms treated with combinations consisting of other quinolones still harbored 104-108 CFU/ml.

When used in combination, the activity of the quinolones from strongest to weakest as ranked by both viability assessment and viable cell counts was as follows: clinafloxacin, ciprofloxacin, moxifloxacin, and levofloxacin (Table 5.2). Hence, clinafloxacin has unique potent activity against persisters compared to its fluoroquinolone counterparts.

Thus far, our data suggest that inclusion of a drug with great anti-persister activity can be beneficial in killing biofilms. To identify other potential anti-persister candidates, we turned to the new generation of lipoglycopeptides such as oritavancin and dalbavancin. These drugs have multiple mechanisms of action: inhibition of transglycosylation, transpeptidation, and cell

membrane disruption, a property of persister drugs [190].We first tested the activity of oritavancin and dalbavancin in killing persisters against its parent vancomycin counterpart, and the results revealed oritavancin was the best in killing persisters among the three drugs compared

(Fig. 5.3A). After 6-day post drug exposure, oritavancin killed 106 CFU/ml of persisters as compared with dalbavancin or vancomycin which killed only about 102 CFU/ml.

Since oritavancin showed strong anti-persister activity, we next evaluated oritavancin's activity in drug combinations. After replacing clinafloxacin with oritavancin, we observed that the combination of meropenem + daptomycin + oritavancin exhibited partial activity against biofilms, a decrease of 105 CFU/ml, which was much better than the activity achieved by treatment with single drugs or two-drug combinations, but still inferior to the clinafloxacin combination (Fig. 5.4B).

Due to oritavancin's strong activity against growing phase S. aureus (MIC of 0.03 mg/L)

[197] and its dual mechanism of action that mimics cell wall + cell membrane inhibitors in our drug combination, we tested oritavancin in place of meropenem and daptomycin. Surprisingly, the combination of oritavancin + clinafloxacin was also able to achieve complete eradication of biofilms suggesting that oritavancin can replace the component in our drug combination that targets actively growing bacteria (Fig. 5.3C). It is also important to note that single drug of oritavancin cannot kill biofilms (no change after 4 day treatment), which further validates the importance of drug combinations in biofilm bacteria.

To compare the activity of the three combinations tested thus far with clinafloxacin, we performed a time-course experiment which revealed that oritavancin + clinafloxacin can kill all biofilms by 2-day treatment whereas 4 days were needed for meropenem/vancomycin + daptomycin + clinafloxacin (Fig. 5.3D). Overall, our data suggest that inclusion of an anti-

100

persister drug in a drug combination to treat biofilms is paramount and these combinations possess better activity than current regimens based on our in vitro studies.

Establishing a persistent skin infection model in mice

Patients with skin and soft tissue infections due to MRSA have recurrence rates up to 45%

[198-200]. Hence, to develop better therapies to treat persistent skin infections, a preclinical, animal model that can mimic persistent skin and tissue infections is paramount. To establish a severe, persistent skin infection model in mice, we evaluated the clinical outcomes of Swiss-

Webster mice subcutaneously infected with different forms of S. aureus (log phase, stationary phase, and biofilm bacteria) (Fig. 5.4). Mice infected with stationary phase bacteria developed skin lesions that required at least one week longer to heal compared to mice infected with log phase bacteria (Fig. 5.4B). Additionally, despite inoculation with equivalent amount of bacteria

(108 CFU), mice infected with bacteria from cultures with more persister cells such as stationary phase or biofilms harbored elevated bacterial loads (Fig. 5.4C) recovered from the skin injection sites and had increased histopathology (Fig. 5.4D) than mice infected with log phase bacteria.

Mice infected with high dose (108 CFU) of bacteria from biofilms showed crust formation, hyperplasia, immune cell infiltration and focal lesion/abscess formation in tissue histology whereas mice infected with log phase showed low levels of immune cell infiltration and inflammation or lesions (Fig. 5.4E-H). Overall, these in-vivo findings suggest that infection of mice with persistent forms of bacteria (i.e. stationary phase bacteria and biofilm bacteria) leads to a more prolonged persistent infection with more severe lesions.

The drug combination meropenem + daptomycin +clinafloxacin eradicated chronic skin infections in mice

Given the robust activity of our drug combinations in eradicating biofilms in vitro (see

101

above), we were interested to know if our combination can also eradicate persistent infections in- vivo. To evaluate the efficacy of our drug combinations in treating the persistent skin infection, we chose to infect mice with bacteria from biofilms of S. aureus strain USA300 as we observed clinical presentations most representative to chronic infections in a host. We allowed the infection to develop for 7 days, followed by treatment of 7 days (Fig. 5.5A). Administration of the combination of doxycycline + rifampin (a control group treated with a clinically used therapy) or drug combination vancomycin + daptomycin + clinafloxacin decreased the bacteria load

(about 1-log of bacteria) but did not clear the infection (Fig. 5.5B). Other therapies which were claimed to eradicate chronic S. aureus infections such as ADEP4+rifampin [113] or frustose+gentamicin [114] did not show sterilizing activity in our model and instead, had increased lesion size and inflammation (Fig. 5.5C). Remarkably, our combination of meropenem

+ daptomycin + clinafloxacin cleared the infection completely, decreased the sizes of lesions, and reduced histopathology (Fig. 5.5D-G). Because our in-vivo experiments were done using

USA300 (MRSA), we wanted to infect mice with Newman strain (MSSA) to ensure that our drug combination is effective toward other bacterial strains but also, further show that clinafloxacin's activity, compared to other quinolones, is superior despite the sensitivity background of the bacterial strain. Despite moxifloxacin and clinafloxacin having the same MIC in the Newman strain, the combination of meropenem + daptomycin + moxifloxacin was not effective in clearing the infection whereas the combination of meropenem + daptomycin + clinafloxacin cleared the infection completely (Fig. 5.5H).

Skin infections caused by S. aureus are cleared by neutrophil recruitment driven by IL-17 production [201]. To evaluate any potential immunopathology consequences of our treatment, we measured the levels of IL-17 and proinflammatory cytokine IL-1β at the infection site and

102

performed histology. Skin tissues of mice treated with our drug combination of meropenem + daptomycin + clinafloxacin produced the lowest amount of IL-17 (Fig. 5.6A) and IL-1 (Fig. 5.6B) and showed the least amount of gross inflammation (Fig. 5.6C) compared to other control treatments. These data support our hypothesis that a drug combination targeting both growing

(e.g. meropenem, daptomycin) and persister cells (e.g. clinafloxacin) is essential in clearing chronic infections in-vivo such as a skin infection caused by S. aureus.

Discussion

Previous in vitro studies have shown how resilient biofilms are to antibiotic treatments.

Given the nature of persister cells that are enshrouded in the biofilm, many groups have attempted to identify novel treatment regimens and synthetic compounds to kill S. aureus persisters [202, 203]. Some approaches include resuscitating or altering the metabolic status of persisters [106, 204] or enhancing the activity of current bactericidal agents with sugars [114].

Although these new therapeutic regimens showed promising results in vitro and in some cases, in-vivo, not all treatments achieved sterilizing activity and/or were tested in a mouse model of a chronic infection up to 2 weeks [113, 202, 205]. In this study, to identify more effective regimens to treat chronic S. aureus infections, we first identified several drug combinations that are more active in killing biofilms in vitro than currently recommended regimens (e.g. vancomycin alone, doxycycline + rifampin, and vancomycin + gentamicin + rifampin) used clinically. Then, we confirmed the potent activity of the combination meropenem + daptomycin

+ clinafloxacin in our newly established chronic, skin infection mouse model.

103

Despite the resurgence to search for better anti-persistence therapies, most drug screens are conducted in vitro and there is a lack of research validating the efficacy of treatments in vivo

[206]. Here, we showed that meropenem + daptomycin + clinafloxacin achieved sterilizing activity in both in vitro and in a chronic, skin infection model in mice. The strong activity of meropenem or daptomycin against S. aureus growing bacteria is indisputable as most S. aureus strains have a relatively low MIC to these drugs [20, 207]. The inclusion of drugs with such strong activity against active bacteria allows for rapid killing of growing bacteria in the population. To kill non-growing biofilm persisters, we identify the unique anti-persister activity of clinafloxacin and the significant role this drug has in drug combinations against biofilm infections. Aside from the treatment used to treat tuberculosis which consists of the anti-persister drug pyrazinamide, no other approved treatment or combination suggested by clinicians take into account the persister population until the study here [21, 118, 206].

Additionally, it is important to note that the chronic infection status of our mice is a key component to our disease model. The phenotype observed in the severity of infection in the host caused by persistent forms cannot be ignored and our model could potentially better mimic chronic infections in humans. Here, we show that more persistent forms of bacteria (e.g. stationary phase and biofilm bacteria) cause chronic infections in murine models and actively growing bacteria (e.g. log phase bacteria) cause lesions that are easily cleared and healed. In typical mouse models for S. aureus, a mouse is infected with a low dose of log phase S. aureus and antibiotic therapy begins after a few hours post-infection [208]. A caveat of those study designs is that antibiotics can be very effective as the infection has not been developed inside the host.

104

Similarly, Conlon et al. also used a high dose to infect their mice and caused a deep- seated infection, they only allowed the infection to develop for 24 hours before treatment and the mice were made neutropenic [113], a condition that may not translate to a majority of patients who suffer from chronic S. aureus infections as such patients have infections for weeks to months [5]. Such differences in our models may explain why ADEP4 + rifampin did not have sterilizing activity in our model. The combination of an aminoglycoside + sugar was shown to be effective in a E. coli urinary tract infection model [114] but unfortunately this approach was not effective in our biofilm infection model. Allison et al. showed that gentamicin + fructose can reduce 1.5 fold of S. aureus biofilms in vitro after 4 hours of treatment [114] but was not tested in animals. In our study here, we showed that mice treated for seven days with gentamicin + fructose still harbored 105 CFU/ml in skin tissues and showed an increase in lesion size. In both these examples, the discrepancy can be due to the disease model, as ours is a chronic skin infection model, and the bacterial strain. In our mouse model, we used a rather high inoculum

(108 CFU) and we injected bacteria derived from biofilms. While such conditions can be deemed as artificial, previous animal studies were able to recover a bacterial load of 108 CFU from biofilms formed on heart valves in rabbits with endocarditis, another chronic infection caused by S. aureus [167].

While meropenem and daptomycin are our agents directed at killing growing bacteria, it is important to note that these drugs also have some activity against persisters. Meropenem used in combination with polymyxin B has been shown to eradicate persisters in Acinetobacter baumannii strains [209]. Similarly, daptomycin has been shown to be active against S. aureus biofilms found on implants [210] and in combination with doxycycline and cefoperazone have been shown to kill biofilm-like microcolonies of B. burgdorferi [194, 211]. Its mechanism of

105

action in disrupting membrane structure and rapid depolarization of the membrane may impact the viability of persisters and thus, play an important role in the combination [212].

Although not commonly used, clinafloxacin administration drastically improved the condition of a cystic fibrosis patient who was not responding to different antibiotics and had a chronic Burkholderia cenocepacia infection [213]. A human trial with patients having native or prosthetic valve endocarditis also showed that clinafloxacin was an effective treatment [214]. As a quinolone, clinafloxacin inhibits bacterial DNA gyrase and topoisomerase IV but not all quinolones have anti-persister activity (Table 5.1). Comparing the chemical structure of clinafloxacin to the other quinolones that have weak anti-persister activity (ciprofloxacin, tosufloxacin, moxifloxacin, and levofloxacin), a chloride group attached to the benzene ring appears to be unique to only clinafloxacin (Fig. 5.7). Further studies to explore the mechanism of clinafloxacin's unique ability to kill persisters requires further investigation.

Previously, we identified both clinafloxacin and tosufloxacin as having robust activity against S. aureus persisters [195]. However, in our drug combination studies, clinafloxacin used in combination displayed greater activity against biofilms than tosufloxacin. This could potentially be explained by the genetic background and inherent antibiotic resistance of the strain tested. Previously, tosufloxacin was identified to have great persister activity in the background of the Newman strain [195], a methicillin-sensitive S. aureus strain, whereas our biofilm experiments conducted here used USA300, a MRSA strain. The ability of each of these strains to form biofilms may also be a contributing factor [215]. Combinations of any anti-persister drug

(e.g. clinafloxacin, tosufloxacin, or oritavancin) with drugs that have robust activity in killing growing bacteria can eradicate more bacteria in biofilms than currently-approved regimens,

106

confirming the importance of targeting the heterogeneous of bacterial populations in developing more effective treatments.

In our study, clinafloxacin was shown to have robust anti-persister activity but the specific mechanism that clinafloxacin uses to kill persisters remains to be determined. While clinafloxacin is a quinolone drug, it was shown here that other quinolones have poor anti- persister activity suggesting that inhibition of DNA synthesis is not the reason why clinafloxacin is a great persister drug. Based on other studies regarding persister drugs, clinafloxacin may potentially alter the cell membrane permeability which could affect cell membrane potential and energy kinetics [191, 202]. Additionally, comparing the chemical structure of clinafloxacin to other quinolones revealed a chloride group that is not found in other quinolones (Fig. 5.7). The reactivity of the chloride group at position eight of the chemical structure could be targeting the cell membrane. Chemical modifications at the eight position of fluoroquinolone structure have been shown to be important for the ability of quinolones to kill anaerobic bacteria [216, 217]. It could be speculated that the chloride group on clinafloxacin causes the drug to selectively target cells with different energetic metabolic pathways such as persister cells.

The combination of meropenem + daptomycin + clinafloxacin showed sterilizing activity in mice after one week based on the concentrations of drugs and dosing regimens commonly found in literature (Table 5.2). While no overt toxicity (e.g. lack of appetite, weight loss, or behavior change) were observed, additional work is needed to better establish the safety profiles of the drug combinations intended for use in people. Higher, yet safe, concentrations of drugs should be tested to see whether or not a shorter treatment period can achieve similar eradication, which would be beneficial for patients undergoing therapy. Further PK/PD studies of our drug combination are needed. Additionally, drugs in this study were administered intraperitoneally

107

which allows the drugs to bypass gut processing and hence, impacting the pharmacokinetics and pharmacodynamics of drug processing and elimination. Thus, the route of administration used here may influence the clinical activity of these drugs in human usage.

Moreover, meropenem, daptomycin, and clinafloxacin are intravenous drugs and not convenient to administer. Future studies to develop oral regimens as effective as the identified combinations are needed for more convenient administration. Our in vitro data suggested that oritavancin used in combination with clinafloxacin had robust activity against biofilms, killing all the bacteria in the biofilms (107 CFU) after a short treatment of 2 days. The administration of oritavancin is a single 1200-mg dose given in a slow, 3 hour infusion, which may also be of interest for patients due to the ease of administration and dosing schedule. Hence, preclinical studies in mice to test oritavancin's activity in chronic infections need to be performed carefully.

Currently-used regimens are lengthy and the inability to clear the bacteria in a timely fashion may also increase the chances of developing antibiotic resistance. A drug combination that has both activities against growing and persister cells have promising potential in being a more effective and faster therapy for treating chronic infections. This treatment algorithm takes into account the heterogeneous population of bacterial cells that exists upon encountering stress.

With this principle in mind, this study reports novel drug combinations that are effective in killing S. aureus biofilms and treating chronic infections. We established a chronic skin infection mouse model that more appropriately mimics human chronic disease. Then, we developed a triple drug combination of meropenem (or vancomycin) + daptomycin + clinafloxacin or two drug combination of oritavancin + clinafloxacin that can achieve sterilizing activity in vitro. We also show that administration of meropenem + daptomycin + clinafloxacin allowed the mice with chronic skin infections to completely clear the bacterial load, heal lesions completely, and show

108

reduced pathology and inflammation. Our approach of combining drugs targeting both growing and non-growing bacteria with persister drugs to completely eradicate biofilm infections may have implications for developing better treatments against other persistent infections by other bacterial pathogens and fungi.

109

)LJXUH&OLQLFDOO\UHFRPPHQGHGWUHDWPHQWVIRUFKURQLF6DXUHXVLQIHFWLRQVSDUWLDOO\

HUDGLFDWHELRILOPVLQYLWUR7UHDWPHQWV DOOX0 RIYDQFRP\FLQDORQHGR[\F\FOLQHULIDPSLQ

DQGWKHFRPELQDWLRQRIYDQFRP\FLQJHQWDPLFLQULIDPSLQIRUGD\VZHUHHYDOXDWHGIRUELRILOP

NLOOLQJE\ $ &)8HQXPHUDWLRQDQG % YLDELOLW\VWDLQLQJXVLQJ6<%5*UHHQ,3, &)

,PDJHVRIELRILOPELRPDVVDQG * TXDQWLILFDWLRQRIDEVROXWHELRILOPELRPDVVDIWHUUHVSHFWLYH

DQWLELRWLFWUHDWPHQWV 9DQFRP\FLQ9$1'R[\F\FOLQH'2;5LIDPSLQ5,)*HQWDPLFLQ

*(16WXGHQW¶VWWHVW Sௗ ௗ Sௗ ௗ Sௗ ௗ

)LJXUH,GHQWLILFDWLRQRIGUXJFRPELQDWLRQVWKDWNLOO056$ELRILOPV9DULRXVGUXJ

FRPELQDWLRQVFRQVLVWLQJRIGUXJV DWX0 KLJKO\DFWLYHDJDLQVWJURZLQJSKDVHVDQGSHUVLVWHU

FHOOVZHUHWHVWHGDQGHYDOXDWHGIRUWKHLUDQWLELRILOPDFWLYLW\E\ $ 6<%5*UHHQ,3,YLDELOLW\

VWDLQLQJDQG % &)8HQXPHUDWLRQ &' 'UXJFRPELQDWLRQVZLWKVWHULOL]LQJDFWLYLW\DJDLQVW

86$ELRILOPVZHUHWHVWHGDWFOLQLFDOO\DFKLHYDEOHFRQFHQWUDWLRQV &PD[ () 9DOLGDWLRQRI

PHURSHQHPGDSWRP\FLQFOLQDIOR[DFLQLQNLOOLQJELRILOPVRIYDULRXVFOLQLFDOLVRODWHVRI

056$ 9DQFRP\FLQ9$10HURSHQHP0(5'DSWRP\FLQ'$37RVXIOR[DFLQ7268

&OLQDIOR[DFLQ&/,1$'R[\F\FOLQH'2;5LIDPSLQ5,)6WXGHQW¶VWWHVW Sௗ ௗ

Sௗ ௗ Sௗ ௗ

)LJXUH(YDOXDWLRQRIRULWDYDQFLQLQNLOOLQJELRILOPVDVDVLQJOHGUXJRULQFRPELQDWLRQ $

&RPSDULQJQRYHOOLSRJO\FRSHSWLGHVRULWDYDQFLQDQGGDOEDYDQFLQWRYDQFRP\FLQLQWKHLUDFWLYLW\

WRNLOOSHUVLVWHUV DWX0 (YDOXDWLQJRULWDYDQFLQDVWKHDJHQW % NLOOLQJSHUVLVWHUVLQ

FRPELQDWLRQZLWKPHURSHQHPGDSWRP\FLQDQGWKHDJHQW & NLOOLQJJURZLQJSKDVHEDFWHULDLQ

FRPELQDWLRQZLWKFOLQDIOR[DFLQ DW&PD[FRQFHQWUDWLRQV ' 7LPHUHVSRQVHNLOOFXUYHRI

ELRILOPVFRPSDULQJWKHWRSGUXJFRPELQDWLRQVFDQGLGDWHV DW&PD[FRQFHQWUDWLRQV

0HURSHQHP0(5'DSWRP\FLQ'$32ULWDYDQFLQ25,7$&OLQDIOR[DFLQ&/,1$

'R[\F\FOLQH'2;5LIDPSLQ5,)7ZRZD\$129$DQG6WXGHQW¶VWWHVW Sௗ ௗ

Sௗ ௗ Sௗ ௗ

)LJXUH(VWDEOLVKLQJDFKURQLF6DXUHXVVNLQLQIHFWLRQPRXVHPRGHO $ *URZWKFXUYHRI

86$WRHVWDEOLVKWKHWLPHSRLQWVRIGLIIHUHQWSKDVHVRIJURZWK % 7KHVL]HVRIVNLQOHVLRQV

IURPPLFHLQIHFWHGZLWKORJDQGVWDWLRQDU\SKDVHEDFWHULDZHUHPRQLWRUHGRYHUWLPH$IWHU

GD\VSRVWLQIHFWLRQ & EDFWHULDOORDGZDVHQXPHUDWHGDQG ' KLVWRORJ\ZHUHSHUIRUPHGLQVNLQ

WLVVXHVRILQIHFWHGPLFH+LVWRSDWKRORJ\RIPLFHZLWK ( QRLQIHFWLRQLQIHFWLRQZLWK ) ORJ

SKDVH * VWDWLRQDU\SKDVHDQG + EDFWHULDIURPELRILOPZDVSHUIRUPHG,PDJHVZHUHWDNHQDW

[PDJQLILFDWLRQ %OXHDUURZVLQGLFDWHFUXVWIRUPDWLRQUHGEUDFNHWVLQGLFDWHK\SHUSODVLDDQG

FHOOXODULQILOWUDWLRQDQGEODFNDUURZLQGLFDWHVIRFDOOHVLRQGHYHORSPHQW7ZRZD\$129$DQG

6WXGHQW¶VWWHVW Sௗ ௗ Sௗ ௗ Sௗ ௗ

)LJXUH9DOLGDWLRQRIGUXJFRPELQDWLRQVLQFKURQLFVNLQLQIHFWLRQPRGHO $ 6WXG\GHVLJQ

RILQYLYRWUHDWPHQWVWXGLHV % %DFWHULDORDG & FKDQJHVLQOHVLRQVL]HVDQGKLVWRORJ\RIVNLQ

WLVVXHRIPLFHWUHDWHGGD\VZLWKGUXJFRPELQDWLRQVDQGUHVSHFWLYHFRQWUROVZHUHPHDVXUHG

+LVWRSDWKRORJ\RI ' XQLQIHFWHGPLFHPLFHUHFHLYLQJGD\VRI ( QRWUHDWPHQW )

GR[\F\FOLQHULIDPSLQDQG * PHURSHQHPGDSWRP\FLQFOLQDIOR[DFLQZDVSHUIRUPHG +

%DFWHULDORDGLQVNLQWLVVXHVRIPLFHLQIHFWHGZLWK066$1HZPDQVWUDLQDQGWUHDWHGZLWKGD\V

ZLWKGUXJFRPELQDWLRQVDQGFRQWUROWUHDWPHQWVZHUHHQXPHUDWHG,PDJHVZHUHWDNHQDW[

PDJQLILFDWLRQ 0HURSHQHP0(5'DSWRP\FLQ'$37RVXIOR[DFLQ7268&OLQDIOR[DFLQ

&/,1$'R[\F\FOLQH'2;5LIDPSLQ5,)*HQWDPLFLQ*(1%OXHDUURZVLQGLFDWHFUXVW

IRUPDWLRQUHGEUDFNHWVLQGLFDWHK\SHUSODVLDDQGFHOOXODULQILOWUDWLRQ6WXGHQW¶VWWHVW Sௗ ௗ

Sௗ ௗ Sௗ ௗ

)LJXUH0HURSHQHP'DSWRP\FLQ&OLQDIOR[DFLQUHGXFHGLPPXQHUHVSRQVHRILQIHFWHG

VNLQWLVVXHV $ ,/ % ,/DQG & JURVVSDWKRORJ\RIVNLQWLVVXHRIPLFHWUHDWHGGD\V

ZLWKGUXJFRPELQDWLRQVDQGUHVSHFWLYHFRQWUROVZHUHHYDOXDWHG 0HURSHQHP0(5

'DSWRP\FLQ'$37RVXIOR[DFLQ7268&OLQDIOR[DFLQ&/,1$'R[\F\FOLQH'2;5LIDPSLQ

5,)*HQWDPLFLQ*(1

)LJXUH&KHPLFDOVWUXFWXUHRIWKHTXLQRORQHVWHVWHG7KHFKORULGHJURXS FLUFOHG LVXQLTXHWR FOLQDIOR[DFLQ

Table 5.1. Drug Dosage, Scheduling, and Administration

Drug Dosage Route Times Cmax tested Treated (clinically achievable conditions) Vancomycin 110 mg/kg Intraperitoneally Twice/daily 20 µg/ml Daptomycin 50 mg/kg Intraperitoneally Once/daily 80 µg/ml Meropenem 50 mg/kg Intraperitoneally Once/daily 20 µg/ml Clinafloxacin 50 mg/kg Intraperitoneally Once/daily 2 µg/ml Doxycycline 100 mg/kg Oral Twice/daily 5 µg/ml Rifampin 10 mg/kg Oral Twice/daily 5 µg/ml Moxifloxacin 100 mg/kg Oral Once/daily 4 µg/ml ADEP4 25 mg/kg Intraperitoneally Twice/Daily and 35 mg/kg Rifampin 30 mg/kg Intraperitoneally Once/daily (for ADEP4 combination) Gentamicin 20 mg/kg Intraperitoneally Once/daily Fructose 1.5 g/kg Intraperitoneally Once/daily Oritavancin ------5 µg/ml Ciprofloxacin ------10 µg/ml Levofloxacin ------10 µg/ml

117

Table 5.2. Ranking of Fluoroquinolones Based on Their Activity in Killing Biofilms

Treatment (Cmax) Log Live/Dead Ratio Ranking of (CFU/ml) combination (1= best, 4= worst) None 9.23 ± 0.4 11.95 ± 0.1 Meropenem 7.34 ± 0.7 8.11 ± 0.04 Daptomycin 9.34 ± 0.3 7.92 ± 0.2 Meropenem+Daptomycin 6.77 ± 0.6 4.52 ± 0.03 Mer+Dap+Clina 0 1.03 ± 0.07 Ŧ 1 Mer+Dap+Cipro 3.87 ± 0.3 1.34 ± 0.08 Ŧ 2 Mer+Dap+Moxi 4.99 ± 0.6 1.24 ± 0.06 Ŧ 3 Mer+Dap+Levo 8.91 ± 0.5 7.19 ± 0.1 Ŧ 4

Ŧ = Limit of detection has been reached for the SYBR Green I/PI viability assay

Meropenem, MER; Daptomycin, DAP; Clinafloxacin, CLINA; Ciprofloxacin, CIPRO;

Moxifloxacin, MOXI; Levofloxacin, LEVO

118

CHAPTER 6

CONCLUSIONS AND FUTURE DIRECTIONS

119

There has been increased attention and interest in bacterial persisters and antibiotic tolerance due to the burden of chronic bacterial infections, especially those that are caused by drug resistant bacteria like S. aureus. Despite the discovery of "persisters" in staphylococci cultures in the 1940s, mechanisms of S. aureus persistence remain incompletely understood. In our genetic screens, we identified 124 genes involved in rifampicin persistence, 75 genes involved in aminoglycosides persistence, and 8 genes that overlapped between the two data sets

(alr, argJ, fmtC, gltS, miaA, murA, purB, purM, SAUSA300_1969) (Fig. 6.1, Table 6.1). Our findings reveal that pathways such as amino acid metabolism, energy metabolism, carbon metabolism, DNA repair, and regulation of transcription factors are important to persistence, which supports current findings in persister biology in other organisms (Fig. 6.2). The mechanisms of persistence are multifactorial, complex but likely to be partially redundant among different bacterial species [21]. Important genes involved in persistence commonly include those involved in amino acid metabolism and those encoding for transporters and transcription factors.

These genes were also identified by other research groups to be important in cell death for an eukaryotic organism like S. cerevisiae [75, 179] (Fig. 6.2).

It is also important to note that the role of TA modules, such as HipAB and MazEF were attributed to bacterial persistence and cell death, respectively [46, 91]. However, it is now being debated that TA modules are not the main drivers of persistence and cell death, which is also supported by our work. Mutations in TA modules in S. aureus did not result in decrease in persistence and while we identified 2 candidates with mutations in TA modules

(SAUSA300_2026, SAUSA300_2352) that resulted in defect in gentamicin persistence, no TA module was identified as a core regulator in both persistence and cell death screens. This suggests that TA modules may not be important for persistence. Additionally, the misconception

120

that it is necessary to delete multiple TA modules to see an effect on persistence was highlighted in an review wrote by Kim and Wood [218]. Wood et al. quoted the findings from Conlon et al. that claim "Although deletion of individual TA has no phenotype, a knockout of ten TAs produced a decrease in persisters in both a growing culture and in the stationary phase.” Wood emphasizes that TA modules have other roles in the cell and due to the varying targets of the different TA modules, it is reasonable that not all, and not only, TA modules can affect persistence. Our findings here emphasize the fact that TA modules are not the core regulators of persistence in S. aureus.

Additionally, the role of MazF in causing programmed cell death has been debated as further research suggested flaws in the pioneering study that showed MazF's role in cell death.

For example, the E. coli strain used actually overexpressed MazF via a lambda PL promoter [219,

220]. Similarly, cell death via MazF does not occur in Myxococcus xanthus but only in the strain with a defect in the PilQl secretin [221]. Additionally, we also did not uncover any TA modules as important for cell death in our study further suggesting that TA modules may not be important for cell death.

Mapping the 8 core regulators in a metabolic pathway schematic (Fig. 6.1, Table 6.1) suggests that glutamate concentration within the cell can be a potential starting substrate for the metabolic cascade that could link the other core regulators such as GltS, PurB, PurM, ArgJ,

MurA, and Alr. Intracellular glutamate through GltS transporter feeds into the arginine biosynthesis pathway through ArgJ which also produces products such as ornithine and L- alanine; L-alanine is generated through MurA and can be used to form peptidoglycan by Alr.

Meanwhile, the ornithine produced makes citrulline, followed by arginosuccinate and fumarate which is involved in purine biosynthesis and thiamine metabolism via PurB and PurM, forming

121

more glutamate to recycle the process. Of note, this is our suggested model with respect to our screen findings and further work is required to validate the interconnected role of all the core regulators. Some potential approaches include performing metabolomics to determine the levels of the specific aforementioned metabolites under stressed conditions, generating strains with several core persister genes overexpressed or knocked down to evaluate the restoration or defects in persistence, respectively, and/or supplementing metabolites, in addition to arginine, (Chapter

3) to restore persistence will provide more compelling evidence.

Intriguingly, mutations in genes involved in glutamate biosynthesis also caused S. aureus to die quicker in the presence of our death stimulus, a controlled heat-ramp and acetic acid stress.

The data from the cell persistence and cell death screen only suggests that there could be a significant role of glutamate but the findings published here are preliminary yet promising.

Speculations of how glutamate could mediate cell persistence or cell death were described previously (See Chapter 4 - Discussion Section). Further characterization needs to be done. If glutamate is important for an anti-death program that also mirrors persistence pathways, then testing the outcomes of the death stimulus (e.g. heat-ramp or acetic acid stress) after performing the aforementioned experiments in the previous paragraph would also offer more insights into the mechanistic aspects of cell death in bacteria.

Generation of core regulators of persistence or mediators of cell death-resistance can be extremely valuable for diagnoses of chronic, bacterial infections. Many bacteria are difficult to culture, and some species classified as viable but nonculturable [103]. The bacterial burden in liquid specimen samples, such as synovial, cerebral spine fluid, or bronchoalveolar lavage fluid may be low and the absolute volume recovered from pediatric patient populations is limited. For example, 40% of clinical samples of synovial joint fluid recovered from prosthetic joint

122

infections are culture negative [222, 223]. Thus, clinical labs are in dire need for better diagnostic tools to help determine if an infection is occurring. The genes identified here in this thesis could be used as biomarkers or antigens on immunochromatogenic assays to detect for bacterial pathogens in people with symptoms resembling a bacterial infection but cannot be confirmed using traditional culture methods. Perhaps, the antigenic targets ad biomarkers currently used may not be appropriate for bacteria causing persistent infections and our data may provide a panel of more appropriate biomarkers.

While the combinatorial treatment for tuberculosis encompasses the Yin-Yang model of persistence, thus far, no other treatment therapy for persistent bacterial infections utilize this philosophy in formulating a treatment regimen [21]. We show here that adding an anti-persister drug, such as clinafloxacin, in a drug combination can more effectively clear a chronic skin infection than currently-used antibiotics. Not only is the duration of treatment decreased but absolute killing and clearance of bacterial load was observed. The robust activity of clinafloxacin was irreplaceable by other quinolones. While one can argue that clinafloxacin is routinely not used in the United States due to toxicity, even though recently there was a case report describing a patient from Maine who used this drug [213], the mere fact that its superior activity we identified may be a promising start for medicinal chemists to use clinafloxacin's structure as a starting point to design more non-toxic derivatives. Synthesizing compounds randomly and then performing drug screens for anti-persister activity may not be as ideal since clinafloxacin is already included in a FDA approved library and have been used in patient care previously [213].

Much time can be saved if drugs are repurposed for their usages. Additionally, with the increasing interest in persisters, more studies have been done to identify compounds with anti-

123

persister activity in S. aureus such as 5-Iodoindole, ADEP4, boromycin, halogenated phenazines, quinolone-derivative HT61, and lipidated lysine [206].

We believe that single therapy may not be the ultimate solution to clear chronic infections and that combination therapies hold promise and should be tested in vitro and in vivo. It is also important to note that monotherapy with just clinafloxacin did not clear the chronic infection and neither did two-drug treatments with clinafloxacin. The precise combination of meropenem + daptomycin + clinafloxacin further validates the importance of ensuring that an effective combination needs activity against both growing and persistent bacteria With the rise in technological development and big data, methods that can model pharmacokinetic and pharmacodynamic parameters in the host and in vitro efficacy may be new approaches to help systematically search for more combinations as options for persistent infections.

Aside from designing drug combinations with anti-persister activity, our genetic screens to understand cell death in bacteria shed light on another potential component to be added to a drug combination. Given that we know the killing activity of antibiotics extend beyond the drug- target interactions, understanding the effector proteins and downstream events of bacterial cell death can help provide novel drug treatment approaches for bacterial infections. For example, bacterial cell death programs could be pharmacologically manipulated to disrupt bacterial cell death in a similar fashion to how apoptotic pathways are exploited to treat cancer cells. Recently, it was shown that extracellular death peptides produced and released by both E. coli and P. aeruginosa can induce the toxic endoribonucleolytic activities of MazEF in M. tuberculosis suggesting promising therapeutic outcomes upon manipulation of the cell death mediators in bacteria [224]. Similarly, since ROS formation serves an important role in the lethality of cidal antibiotics, a drug combination with metabolic perturbations may enhance killing of currently

124

used antibiotics. For example, causing defects in peroxide-detoxifying enzymes have been shown to increase antimicrobial lethality [225]. On the other hand, it is of concern that more than half of the population in the US consumes nutritional supplements such as vitamins which are antioxidants [226] and the potential antagonistic effects of an antioxidant diet in a patient taking antibiotics will require further investigation.

As persisters can be recovered upon stress removal, Sharma et al. proposed a method of drug pulse dosing to kill persisters which they found to be effective against B. burgdorferi [115].

Conceptually, drug pulsing may be a sound approach but Feng et al. showed that ceftriaxone pulse dosing fails to clear biofilm-like microcolonies, a more persistent form, of B. burgdorferi

[115]. Instead, Feng et al. showed that a triple drug combination of daptomycin (a persister drug)+doxycycline+ceftriaxone was more effective in killing these biofilm-like aggregates than pulse dosing. My study here did not evaluate pulse dosing in S. aureus and my findings presented here neither support nor refute the activity of pulse dosing on S. aureus persistent infections. It is possible that pulse dosing regimens using clinically approved drugs for S. aureus infection (See "Introduction" Table 2) or even drugs from our robust combination of meropenem

+ daptomycin + clinafloxacin could yield important knowledge regarding treatment approach and duration of relevance in clinical settings. Whether pulse dosing in S. aureus is an effective method to clear persistent infections and whether this approach leads to antibiotic resistance needs to be further explored.

The novelty in the work done here is two folds: the chronic skin infection model in mice and also the drug combination that can clear the chronic skin infection. Currently, mouse models modeling "chronic" infections let the infection develop for 1-2 days post-inoculation and treatment regimens are administered for less than one week [113, 114]. If we were to evaluate

125

our treatment against persistent infections, we believe establishment of a chronic infection is important. Waiting only 1-2 days after inoculation may not be appropriate as bacteria may not have enough time to cause a persistent infection [21]. Use of an immunomodulatory agent to dampen the immune response in order to get a mouse S. aureus infection is also not ideal as the immune response can induce for persister cell formation [113]. Here, we showed that infecting mice with more persistent forms (biofilm or stationary phase bacteria) caused a more chronic skin lesion and even one week treatment of clinically recommended drugs such as doxycycline + rifampin was not effective in clearing the infection. On the other hand, mice infected with log phase, actively growing, bacteria developed lesions that were healed much quicker than mice infected with stationary phase bacteria, or persister-enriched biofilmbacteria.

While we were able to show that bacteria isolated from biofilms established a more chronic infection in mice compared to stationary phase or log phase bacteria, the reason that allowed the more persistent bacterial forms to do so still need to be explored. To associate the absolute amount of persisters in these persistent forms as the culprit of chronic infections, we need to observe the clinical outcomes of mice in which they are infected with different amounts of isolated persisters (e.g. through cidal drug treatment of 4-6 hours). The anticipated result is that mice infected with more persisters develop more severe chronic infections, possibly in a persister dose-dependent manner. From a mechanistic perspective, further RNA-seq or proteomics study comparing the transcripts and proteins in biofilm forms compared to stationary or log phase forms can offer more insights to potential virulence factors found specifically in persistent forms that allows for chronic infections inside the host.

Meanwhile, host-pathogen interactions cannot be overlooked. Here, we used Swiss-

Webster mice that were previously employed by other researchers to evaluate treatment for a

126

deep-seated tissue infection of S. aureus and mice that are routinely used to test for drug compounds [113]. However, literature search reveals that studies of S. aureus skin infections can also be performed using Balb/C and C57Black/6 mice, both of which have opposing immunological backgrounds [227, 228]. In humans, patients with hyper-IgE syndrome and those who are TH17/IL17-deficient are at higher risk for recurrent skin infections [229]. Additionally,

HIV patients with low CD4 counts frequently have recurrent SSTI [230]. Hence, the inflammatory states of the host during infection [227] would offer crucial information in development of a better treatment for chronic infection where immunomodulatory agents could be beneficial and also, development of a better mouse model to test for therapies for chronic infections. It is also important to note here that other reasons as to why persistent infections occur may include repeated environmental exposures through contaminated inanimate objects or communication with infected individuals and also, the microbiome composition of the host which can affect bacterial colonization as well as the immune response to pathogenic organisms.

In our work, we show that there is a potential relationship between the bacteria’s susceptibility to die and their degree of virulence in several model systems. First, mice that were infected with the ArgJ mutant that was defective in persistence had a lower virulence and lower bacterial load than mice infected with parental strain USA300. Second, mice infected with more persistent forms (e.g. stationary phase bacteria) developed skin lesions that took longer to heal, were much larger, and have more severe pathology than the lesions developed in mice infected with actively growing, log phase bacteria. Third, C. elegans infected with more death-resistant mutants showed increased mortality. As we did not see any changes in the growth rate of these strains in vitro, we believe that the growth capacity of the bacteria may not be the main reason.

Then, it needs to be explored if the increased bacterial load is due to the bacteria evading the host

127

immune system (through canonical or non-conical ways) or if molecular changes in the bacteria's death program also induces epigenetic changes in the virulence gene expression or even fitness of the bacteria. All in all, these observations call for more experiments to understand how persistence and anti-death phenotypes of a bacterial cell affects clinical outcomes.

In conclusion, the scope of this thesis is to offer some insights to broad scientific questions regarding mechanisms of persistence in S. aureus and persistent S. aureus infections.

How does S. aureus persist and not die? What are the clinical outcomes of hosts infected with persistent forms of bacteria? What are some better treatment strategies in eliminating chronic S. aureus infections? The findings of our research advance our knowledge of the genetic mechanisms used by bacteria to persist in unfavorable environments and how we can eliminate persistent bacterial infections. The public health burden that persister cells have on clinical care warrant new approaches in drug screens, drug development, and even pathogen detection methods. This thesis provides the groundwork in revolutionizing clinical care of recalcitrant persistent bacterial infections.

128

)LJXUH$PRGHORISRWHQWLDO6DXUHXVSHUVLVWHQFHDQGFHOOGHDWKSDWKZD\VEDVHGRQRXU

JHQHWLFVFUHHQUHVXOWV

)LJXUH2YHUODSSLQJSDWKZD\VLQ6DXUHXVSHUVLVWHQFHDQGFHOOGHDWKIRXQGLQRXUJHQHWLF

VFUHHQV

Table 6.1. Overlapping genes found in our persistence screens.

Gene Function alr alanine racemase argJ arginine biosynthesis bifunctional protein ArgJ fmtC oxacillin resistance-related FmtC protein gltS sodium/glutamate symporter miaA tRNA delta(2)-isopentenylpyrophosphate transferase murA UDP-N-acetylglucosamine 1-carboxyvinyltransferase purB adenylosuccinate lyase purM phosphoribosylaminoimidazole synthetase

phi77 ORF011-like protein, SAUSA300_1969 phage transcriptional repressor

131

APPENDIX

SUPPLEMENTARY TABLES

132

Table 1. Genes involved in S. aureus persistence to rifampicin.

Gene name Function Accessory Number (if applicable) Cell Envelope oxacillin resistance-related fmtC FmtC protein SAUSA300_1255 integral membrane protein SAUSA300_0729 cell wall surface anchor family protein SAUSA300_1677 Pathogenesis antibiotic epidermin epiD biosynthesis protein SAUSA300_1764 lukE leukotoxin LukE SAUSA300_1769 rarD RarD protein SAUSA300_2628 sek enterotoxin K SAUSA300_0800 putative exotoxin 4 SAUSA300_1060 superantigen-like protein SAUSA300_1061 Phage Derived Proteins phiSLT ORF110-like protein SAUSA300_1399 putative prophage protease SAUSA300_1402 terminase, large subunit SAUSA300_1404 putative DNA-binding protein SAUSA300_1430 phage transcriptional repressor SAUSA300_1969 Enzymes (Proteases, hydrolases, etc) htrA serine protease SAUSA300_0923 tRNA delta(2)-isopentenyl miaA pyrophosphate transferase SAUSA300_1195 anaerobic ribonucleotide reductase, nrdG small subunit SAUSA300_2550 splB serine protease SplB SAUSA300_1757 splC serine protease SplC SAUSA300_1756 sspA V8 protease SAUSA300_0951 putative endoribonuclease L-PSP SAUSA300_0474 RluA family pseudouridine synthase SAUSA300_0909

133

staphopain A SAUSA300_1890 truncated amidase SAUSA300_1921 putative ATP-dependent Clp proteinase SAUSA300_2486 Drug Efflux Pumps MATE efflux family protein SAUSA300_0335 DedA family protein SAUSA300_2450 Transporters ferrichrome transport fhuA ATP-binding protein SAUSA300_0633 sodium/glutamate gltS symporter SAUSA300_2291 molybdenum ABC transporter, ATP-binding modC protein SAUSA300_2228 oligopeptide ABC transporter, permease oppB protein SAUSA300_0895 phosphonate ABC transporter, permease phnE protein SAUSA300_0142 NLPA lipoprotein SAUSA300_0437 putative Na+/H+ antiporter, MnhE component SAUSA300_0614 iron compound ABC transporter, permease protein SAUSA300_0719 oligopeptide ABC transporter, ATP-binding protein SAUSA300_0894 ABC transporter permease SAUSA300_1515 CamS sex pheromone cAM373 SAUSA300_1884 iron compound ABC transporter iron compound-binding protein SAUSA300_2235 putative drug transporter SAUSA300_2389 oligopeptide ABC transporter ATP-binding protein SAUSA300_2407 Replication, Transcription, Translation cassette chromosome ccrA recombinase A SAUSA300_0038

134

DNA-directed RNA rpoE polymerase, delta subunit SAUSA300_2082 RNA polymerase sigma rpoF factor SigB SAUSA300_2022 spsA signal peptidase IA SAUSA300_0867 tnp IS1181, transposase SAUSA300_2115 tRNA pseudouridine truB synthase B SAUSA300_1164 GntR family transcriptional regulator SAUSA300_0258 MarR family transcriptional regulator SAUSA300_0334 IS200 family transposase SAUSA300_1309 endonuclease IV SAUSA300_1517 putative transposase SAUSA300_1732 putative endodeoxyribonuclease RusA SAUSA300_1955 single-strand binding protein SAUSA300_1958 MerR family transcriptional regulator SAUSA300_2160 transcription regulatory protein SAUSA300_2326 transcriptional regulator, MerR family SAUSA300_2445 putative chromosome partioning protein, ParB family SAUSA300_2643

Metabolism Purine Biosynthesis purB adenylosuccinate lyase SAUSA300_1889 amidophosphoribosyltrans purF ferase SAUSA300_0972 bifunctional purine purH biosynthesis protein SAUSA300_0975 phosphoribosylaminoimid purM azole synthetase SAUSA300_0973 5' nucleotidase family protein SAUSA300_0147

Carbohydrate acetyl-CoA carboxylase, accC biotin carboxylase SAUSA300_1563 budA alpha-acetolactate SAUSA300_2536

135

decarboxylase capsular polysaccharide Cap5J biosynthesis protein SAUSA300_0161 gltA citrate synthase II SAUSA300_1641 gntK gluconate kinase SAUSA300_2443 isopropylmalate isomerase leuD small subunit SAUSA300_2013 UDP-N-acetylglucosamine murA 1-carboxyvinyltransferase SAUSA300_2078 malate:quinone mqo oxidoreductase SAUSA300_2312 PTS system, fructose- specific enzyme II, BC component SAUSA300_0239 L-ribulokinase SAUSA300_0537 2-oxoglutarate ferredoxin oxidoreductase subunit beta SAUSA300_1183 catalase SAUSA300_1232 geranyltranstransferase SAUSA300_1470 phosphoglucomutase/phos phomannomutase family protein SAUSA300_2433 glyoxalase family protein SAUSA300_2461 D-lactate dehydrogenase SAUSA300_2496 phosphotransferase system, fructose-specific IIABC component SAUSA300_2576 putative N- acetyltransferase SAUSA300_2631

Amino Acid alr alanine racemase SAUSA300_2027 alpha-acetolactate alsS synthase SAUSA300_2166 arginine biosynthesis argJ bifunctional protein ArgJ SAUSA300_0185 glycine betaine aldehyde betB dehydrogenase SAUSA300_2546 D-alanine dat aminotransferase SAUSA300_1696 glutamate synthase subunit gltD beta SAUSA300_0446 branched-chain amino acid ilvE aminotransferase SAUSA300_0539

136

Orn/Lys/Arg decarboxylase SAUSA300_0458 acetyltransferase SAUSA300_0451

Lipid diacylglycerol glucosyltransferase SAUSA300_0918

Vitamin and Co factors 6,7-dimethyl-8- ribH ribityllumazine synthase SAUSA300_1712 TENA/THI-4 family protein SAUSA300_2050 uroporphyrin-III C-methyl transferase SAUSA300_2344

137

Table 2. Genes involved in S. aureus persistence to gentamicin

Gene name Function Accessory Number (if applicable) Phage Proteins phiSLT ORF80-like protein SAUSA300_1419 phi77 ORF044-like protein SAUSA300_1926 phiPVL ORF39-like protein SAUSA300_1962 phi77 ORF014-like protein SAUSA300_1966 phi77 ORF011-like protein SAUSA300_1969 putative phage infection SAUSA300_2578 protein Toxin-

Antitoxin PemK family protein SAUSA300_2026 addiction module antitoxin SAUSA300_2352

Digestive Enzymes splE serine protease SplE SAUSA300_1754 amidohydrolase family protein SAUSA300_2517 hydrolase family protein SAUSA300_2518

DNA replication DNA polymerase I polA SAUSA300_1636 superfamily xerC tyrosine recombinase xerC SAUSA300_1145 ComE operon protein 1 SAUSA300_1549 Holliday junction resolvase- SAUSA300_1573 like protein Membrane lantibiotic epidermin epiC SAUSA300_1765 biosynthesis glucosamine-1-phosphate fmtB SAUSA300_2109 synthesis phosphatidylglycerol fmtC SAUSA300_1255 lysyltransferase tandem lipoprotein SAUSA300_0416 YibE/F-like protein SAUSA300_0443 Transferases tRNA delta(2)- miA isopentenylpyrophosphate SAUSA300_1195 Transferase

138

anti-sigma-B factor, serine- rsbW SAUSA300_2023 protein kinase tRNA (guanine-N(7)-)- trmB SAUSA300_1694 methyltransferase glycosyl transferase SAUSA300_0689 ribosomal RNA large subunit SAUSA300_1111 methyltransferase N aminotransferase, class V SAUSA300_1669 Acetyltransferase SAUSA300_2232 Transcription codY transcriptional repressor CodY SAUSA300_1148 rsbU sigma-B regulation protein SAUSA300_2025 scrR sucrose operon repressor SAUSA300_ 1995 transcriptional regulator SAUSA300_ 0195 iron-dependent repressor SAUSA300_0621 transcriptional regulator, Fur SAUSA300_1842 family lactose phosphotransferase system SAUSA300_2156 repressor putative transcriptional SAUSA300_ 2259 regulator

Cell growth fnbB fibronectin binding protein B SAUSA300_2440

Signaling glpF glycerol uptake facilitator SAUSA300_1191 sensor histidine kinase SAUSA300_0646 Pathogenesis penicillin-binding protein 2'- mecA SAUSA300_0032 pathogen sak Staphylokinase- meta SAUSA300_1922 superantigen-like protein SAUSA300_0395 truncated beta-hemolysin SAUSA300_1918 Transporter amino acid transport carrier brnQ SAUSA300_0306 protein gltS sodium/glutamate symporter SAUSA300_2291 lspA lipoprotein signal peptidase SAUSA300_1089 ABC transporter permease SAUSA300_ 0797 protein putative transporter - SAUSA300_ 2139 membrane amino acid ABC transporter SAUSA300_ 2359 139

ABC transporter SAUSA300_ 2633 Translation rpsA 30S ribosomal protein S1 SAUSA300_1365

Metabolism Amino Acid alr alanine racemase SAUSA300_2027 arginine biosynthesis argJ SAUSA300_0185 bifunctional protein ArgJ argR arginine repressor SAUSA300_1469

Carbohydrate arcC carbamate kinase - metabolic SAUSA300_2567 acetolactate synthase 1 ilvN SAUSA300_2008 regulatory subunit UDP-N-acetylglucosamine, murA SAUSA300_2055 1-carboxyvinyltransferase 3-hydroxyacyl-CoA SAUSA300_ 0226 dehydrogenase zinc-binding dehydrogenase SAUSA300_ 0244 branched-chain alpha-keto acid, SAUSA300_0995 dehydrogenase subunit E2

Energy F0F1 ATP synthase subunit atpH SAUSA300_2061 delta pyruvate dehydrogenase E1 pdhB component, SAUSA300_0994 beta subunit 2-oxoglutarate dehydrogenase, sucA SAUSA300_1306 E1 component 2-oxoglutarate ferredoxin SAUSA300_1183 oxidoreductase

Lipid glpK glycerol kinase SAUSA300_ 1192 choloylglycine hydrolase SAUSA300_ 0269 family protein

Nucleotide purB adenylosuccinate lyase SAUSA300_1889 phosphoribosylaminoimidazole purM SAUSA300_0973 synthetase pyrF orotidine 5'-phosphate SAUSA300_1097

140

decarboxylase inosine-uridine preferring SAUSA300_ 2234 nucleoside hydrolase

141

Table 3. Genes whose mutation resulted in cell death resistance to both heat-ramp stress and acetic acid stress in S. aureus

Gene name Function Accessory Number (if applicable) Transporters ABC transporter ATP- vraF binding protein USA300HOU_0682 potassium-transporting kdpB ATPase subunit B USA300HOU_2071 ABC transporter ATP- binding protein USA300HOU_0152 oligopeptide ABC transporter ATP-binding protein USA300HOU_2455 ABC transporter ATP- binding protein USA300HOU_0154 iron ABC transporter membrane binding protein USA300HOU_0232 PTS system ascorbate- specific transporter subunit IIC USA300HOU_0350 ABC transporter ATP- binding protein USA300HOU_2386 Transcription Regulators transcription-repair coupling mfd factor USA300HOU_0497 transcriptional regulator USA300HOU_0044 transcription regulator USA300HOU_2004 GntR family transcriptional regulator USA300HOU_1209 Peptidases sspA glutamyl endopeptidase USA300HOU_0996 Peptidase USA300HOU_2590 Metabolism methylenetetrahydrofolate dehydrogenase; methenyltetrahydrofolate folD cyclohydrolase USA300HOU_1008 bifunctional N- acetylmuramoyl-L-alanine amidase, mannosyl-glycoprotein endo-beta-N- acetylglucosaminidase USA300HOU_0997 phytoene dehydrogenase USA300HOU_2559

142

Phosphatases & Kinases nisin susceptibility- associated sensor histidine nsaS kinase USA300HOU_2623 HAD family phosphatase USA300HOU_0930 Transferases Glycosyltransferase USA300HOU_0142 Stress Response hslO Hsp33-like chaperonin USA300HOU_0506 Nucleic Acid Synthesis type I site-specific deoxyribonuclease hsdR1 restriction subunit USA300HOU_0033 Protein Synthesis rRNA %28cytosine-5-%29- methyltransferase USA300HOU_1154

Hypothetical Proteins USA300HOU_2684 USA300HOU_0868 USA300HOU_0369 USA300HOU_0420

143

Table 4. Genes whose mutations resulted in cell death sensitivity after heat-ramp and acetic acid stress in S. aureus

Gene name Function Accessory Number (if applicable) Transporters Proton glutamate symport gltT protein USA300HOU_2366 monovalent cation antiporter mnhG1 subunit G USA300HOU_0649 ABC transporter ATP- binding protein USA300HOU_0329 ABC transporter ATP- binding protein USA300HOU_2386 Metabolism isochorismatase USA300HOU_0200 amidohydrolase USA300HOU_2029 Transcription transcriptional regulator USA300HOU_0336 Cell wall capsular polysaccharide capA biosynthesis protein USA300HOU_2664

Hypothetical Proteins USA300HOU_1780 USA300HOU_2496

144

REFERENCES

145

1. McCaig, L.F., et al., Staphylococcus aureus-associated skin and soft tissue infections in ambulatory care. Emerg Infect Dis, 2006. 12(11): p. 1715-23. 2. Moran, G.J., et al., Methicillin-resistant S. aureus infections among patients in the emergency department. N Engl J Med, 2006. 355(7): p. 666-74. 3. Klevens, R.M., et al., Invasive methicillin-resistant Staphylococcus aureus infections in the United States. Jama, 2007. 298(15): p. 1763-71. 4. Noskin, G.A., et al., The burden of Staphylococcus aureus infections on hospitals in the United States: an analysis of the 2000 and 2001 Nationwide Inpatient Sample Database. Arch Intern Med, 2005. 165(15): p. 1756-61. 5. Tong, S.Y., et al., Staphylococcus aureus infections: epidemiology, pathophysiology, clinical manifestations, and management. Clin Microbiol Rev, 2015. 28(3): p. 603-61. 6. CDC, Emerging Infections Program Network Report, Methicillin-Resistant Staphylococcus aureus, 2015. 2018. 7. Krishna, S. and L.S. Miller, Innate and adaptive immune responses against Staphylococcus aureus skin infections. Semin Immunopathol, 2012. 34(2): p. 261-80. 8. CDC, Antibiotic Resistance Threats in the United States, 2013. 2013. 9. Hampton, T., Report reveals scope of US antibiotic resistance threat. Jama, 2013. 310(16): p. 1661-3. 10. Blair, J.M., et al., Molecular mechanisms of antibiotic resistance. Nat Rev Microbiol, 2015. 13(1): p. 42-51. 11. Katayama, Y., T. Ito, and K. Hiramatsu, A new class of genetic element, staphylococcus cassette chromosome mec, encodes methicillin resistance in Staphylococcus aureus. Antimicrob Agents Chemother, 2000. 44(6): p. 1549-55. 12. Gardete, S. and A. Tomasz, Mechanisms of vancomycin resistance in Staphylococcus aureus. J Clin Invest, 2014. 124(7): p. 2836-40. 13. Hafer, C., et al., Contribution of selected gene mutations to resistance in clinical isolates of vancomycin-intermediate Staphylococcus aureus. Antimicrob Agents Chemother, 2012. 56(11): p. 5845-51. 14. Kato, Y., et al., Genetic changes associated with glycopeptide resistance in Staphylococcus aureus: predominance of amino acid substitutions in YvqF/VraSR. J Antimicrob Chemother, 2010. 65(1): p. 37-45. 15. Arthur, M. and P. Courvalin, Genetics and mechanisms of glycopeptide resistance in enterococci. Antimicrob Agents Chemother, 1993. 37(8): p. 1563-71. 16. Floyd, J.L., et al., LmrS is a multidrug efflux pump of the major facilitator superfamily from Staphylococcus aureus. Antimicrob Agents Chemother, 2010. 54(12): p. 5406-12. 17. Long, K.S., et al., The Cfr rRNA methyltransferase confers resistance to Phenicols, Lincosamides, Oxazolidinones, Pleuromutilins, and Streptogramin A antibiotics. Antimicrob Agents Chemother, 2006. 50(7): p. 2500-5. 18. Jorgensen, J.H. and M.J. Ferraro, Antimicrobial susceptibility testing: a review of general principles and contemporary practices. Clin Infect Dis, 2009. 49(11): p. 1749-55. 19. Bauer, A.W., et al., Antibiotic susceptibility testing by a standardized single disk method. Am J Clin Pathol, 1966. 45(4): p. 493-6. 20. CLSI, Performance Standards for Antimicrobial Performance Standards for Antimicrobial Susceptibility Testing-26th Edition: CLSI supplement M100S. NCCLS, Wayne, PA, USA. 2016.

146

21. Zhang, Y., Persisters, Persistent Infections and the Yin-Yang Model. Emerging Microbes and Infection, 2014. 3: p. 10. 22. Fisher, R.A., B. Gollan, and S. Helaine, Persistent bacterial infections and persister cells. Nat Rev Microbiol, 2017. 15(8): p. 453-464. 23. Gomez, J.E. and J.D. McKinney, M. tuberculosis persistence, latency, and drug tolerance. Tuberculosis (Edinb), 2004. 84(1-2): p. 29-44. 24. Foxman, B., Recurring urinary tract infection: incidence and risk factors. Am J Public Health, 1990. 80(3): p. 331-3. 25. Redpath, S., P. Ghazal, and N.R. Gascoigne, Hijacking and exploitation of IL-10 by intracellular pathogens. Trends Microbiol, 2001. 9(2): p. 86-92. 26. Kardas, P., Patient compliance with antibiotic treatment for respiratory tract infections. J Antimicrob Chemother, 2002. 49(6): p. 897-903. 27. Goncalves-Pereira, J. and P. Povoa, Antibiotics in critically ill patients: a systematic review of the pharmacokinetics of beta-lactams. Crit Care, 2011. 15(5): p. R206. 28. Hobby, G.L., K. Meyer, and E. Chaffee, Observations on the mechanism of action of penicillin. Proc Soc Exp Biol NY, 1942. 50: p. 281-285. 29. Bigger, J.W., Treatment of staphylococcal infections with penicillin by intermittent sterilisation. Lancet, 1944. 244: p. 4. 30. Balaban, N.Q., et al., Bacterial persistence as a phenotypic switch. Science, 2004. 305(5690): p. 1622-5. 31. Mulcahy, L.R., et al., Emergence of Pseudomonas aeruginosa strains producing high levels of persister cells in patients with cystic fibrosis. J Bacteriol, 2010. 192(23): p. 6191-9. 32. Lafleur, M.D., Q. Qi, and K. Lewis, Patients with long-term oral carriage harbor high- persister mutants of Candida albicans. Antimicrob Agents Chemother, 2010. 54(1): p. 39-44. 33. LaFleur, M.D., C.A. Kumamoto, and K. Lewis, Candida albicans biofilms produce antifungal-tolerant persister cells. Antimicrob Agents Chemother, 2006. 50(11): p. 3839- 46. 34. Waters, E.M., et al., Convergence of Staphylococcus aureus Persister and Biofilm Research: Can Biofilms Be Defined as Communities of Adherent Persister Cells? PLoS Pathog, 2016. 12(12): p. e1006012. 35. Conlon, B.P., S.E. Rowe, and K. Lewis, Persister cells in biofilm associated infections. Adv Exp Med Biol, 2015. 831: p. 1-9. 36. Davies, J.C., Pseudomonas aeruginosa in cystic fibrosis: pathogenesis and persistence. Paediatr Respir Rev, 2002. 3(2): p. 128-34. 37. Hunstad, D.A. and S.S. Justice, Intracellular lifestyles and immune evasion strategies of uropathogenic Escherichia coli. Annu Rev Microbiol, 2010. 64: p. 203-21. 38. Stewart, P.S., W.M. Davison, and J.N. Steenbergen, Daptomycin rapidly penetrates a Staphylococcus epidermidis biofilm. Antimicrob Agents Chemother, 2009. 53(8): p. 3505-7. 39. Kirker, K.R., S.T. Fisher, and G.A. James, Potency and penetration of telavancin in staphylococcal biofilms. Int J Antimicrob Agents, 2015. 46(4): p. 451-5. 40. Sternberg, C., et al., Distribution of bacterial growth activity in flow-chamber biofilms. Appl Environ Microbiol, 1999. 65(9): p. 4108-17.

147

41. de Beer, D., et al., Effects of biofilm structures on oxygen distribution and mass transport. Biotechnol Bioeng, 1994. 43(11): p. 1131-8. 42. Davies, D., Understanding biofilm resistance to antibacterial agents. Nat Rev Drug Discov, 2003. 2(2): p. 114-22. 43. Ramamurthy, T., et al., Current Perspectives on Viable but Non-Culturable (VBNC) Pathogenic Bacteria. Front Public Health, 2014. 2: p. 103. 44. Glover, W.A., Y. Yang, and Y. Zhang, Insights into the molecular basis of L-form formation and survival in Escherichia coli. PLoS One, 2009. 4(10): p. e7316. 45. Proctor, R.A., et al., Staphylococcus aureus Small Colony Variants (SCVs): a road map for the metabolic pathways involved in persistent infections. Front Cell Infect Microbiol, 2014. 4: p. 99. 46. Moyed, H.S. and K.P. Bertrand, hipA, a newly recognized gene of Escherichia coli K-12 that affects frequency of persistence after inhibition of murein synthesis. J Bacteriol, 1983. 155(2): p. 768-75. 47. Conlon, B.P., et al., Persister formation in Staphylococcus aureus is associated with ATP depletion. 2016. 1: p. 16051. 48. Li, Y. and Y. Zhang, PhoU is a persistence switch involved in persister formation and tolerance to multiple antibiotics and stresses in Escherichia coli. Antimicrob Agents Chemother, 2007. 51(6): p. 2092-9. 49. Hansen, S., K. Lewis, and M. Vulic, Role of global regulators and nucleotide metabolism in antibiotic tolerance in Escherichia coli. Antimicrob Agents Chemother, 2008. 52(8): p. 2718-26. 50. Hong, S.H., et al., Bacterial persistence increases as environmental fitness decreases. Microb Biotechnol, 2012. 5(4): p. 509-22. 51. Luidalepp, H., et al., Age of inoculum strongly influences persister frequency and can mask effects of mutations implicated in altered persistence. J Bacteriol, 2011. 193(14): p. 3598-605. 52. Manina, G., N. Dhar, and J.D. McKinney, Stress and host immunity amplify Mycobacterium tuberculosis phenotypic heterogeneity and induce nongrowing metabolically active forms. Cell Host Microbe, 2015. 17(1): p. 32-46. 53. Kerr, J.F., A.H. Wyllie, and A.R. Currie, Apoptosis: a basic biological phenomenon with wide-ranging implications in tissue kinetics. Br J Cancer, 1972. 26(4): p. 239-57. 54. Fearnhead, H.O., P. Vandenabeele, and T. Vanden Berghe, How do we fit ferroptosis in the family of regulated cell death? Cell Death Differ, 2017. 24(12): p. 1991-1998. 55. Meier, P., A. Finch, and G. Evan, Apoptosis in development. Nature, 2000. 407(6805): p. 796-801. 56. Li, F., et al., Apoptotic cells activate the "phoenix rising" pathway to promote wound healing and tissue regeneration. Sci Signal, 2010. 3(110): p. ra13. 57. Galluzzi, L., et al., Molecular mechanisms of cell death: recommendations of the Nomenclature Committee on Cell Death 2018. Cell Death Differ, 2018. 25(3): p. 486-541. 58. Green, D.R. and J.C. Reed, Mitochondria and apoptosis. Science, 1998. 281(5381): p. 1309-12. 59. Buttke, T.M. and P.A. Sandstrom, Oxidative stress as a mediator of apoptosis. Immunol Today, 1994. 15(1): p. 7-10. 60. Sahara, S., et al., Acinus is a caspase-3-activated protein required for apoptotic chromatin condensation. Nature, 1999. 401(6749): p. 168-73.

148

61. Coleman, M.L., et al., Membrane blebbing during apoptosis results from caspase- mediated activation of ROCK I. Nat Cell Biol, 2001. 3(4): p. 339-45. 62. Segawa, K., et al., Caspase-mediated cleavage of phospholipid flippase for apoptotic phosphatidylserine exposure. Science, 2014. 344(6188): p. 1164-8. 63. Tait, S.W. and D.R. Green, Mitochondria and cell death: outer membrane permeabilization and beyond. Nat Rev Mol Cell Biol, 2010. 11(9): p. 621-32. 64. Walensky, L.D. and E. Gavathiotis, BAX unleashed: the biochemical transformation of an inactive cytosolic monomer into a toxic mitochondrial pore. Trends Biochem Sci, 2011. 36(12): p. 642-52. 65. Degterev, A., et al., Chemical inhibitor of nonapoptotic cell death with therapeutic potential for ischemic brain injury. Nat Chem Biol, 2005. 1(2): p. 112-9. 66. Degterev, A., et al., Identification of RIP1 kinase as a specific cellular target of necrostatins. Nat Chem Biol, 2008. 4(5): p. 313-21. 67. Shi, J., W. Gao, and F. Shao, Pyroptosis: Gasdermin-Mediated Programmed Necrotic Cell Death. Trends Biochem Sci, 2017. 42(4): p. 245-254. 68. Minina, E.A., et al., Metacaspases versus caspases in development and cell fate regulation. Cell Death Differ, 2017. 24(8): p. 1314-1325. 69. Ludovico, P., et al., Cytochrome c release and mitochondria involvement in programmed cell death induced by acetic acid in Saccharomyces cerevisiae. Mol Biol Cell, 2002. 13(8): p. 2598-606. 70. Ludovico, P., et al., Saccharomyces cerevisiae commits to a programmed cell death process in response to acetic acid. Microbiology, 2001. 147(Pt 9): p. 2409-15. 71. Hardwick, J.M., Do Fungi Undergo Apoptosis-Like Programmed Cell Death? MBio, 2018. 9(4). 72. Mollapour, M. and P.W. Piper, Hog1p mitogen-activated protein kinase determines acetic acid resistance in Saccharomyces cerevisiae. FEMS Yeast Res, 2006. 6(8): p. 1274-80. 73. Fernandes, A.R., et al., Saccharomyces cerevisiae adaptation to weak acids involves the transcription factor Haa1p and Haa1p-regulated genes. Biochem Biophys Res Commun, 2005. 337(1): p. 95-103. 74. Mira, N.P., et al., Genome-wide identification of Saccharomyces cerevisiae genes required for tolerance to acetic acid. Microb Cell Fact, 2010. 9: p. 79. 75. Sousa, M., et al., Genome-wide identification of genes involved in the positive and negative regulation of acetic acid-induced programmed cell death in Saccharomyces cerevisiae. BMC Genomics, 2013. 14: p. 838. 76. Kim, H., A. Kim, and K.W. Cunningham, Vacuolar H+-ATPase (V-ATPase) promotes vacuolar membrane permeabilization and nonapoptotic death in stressed yeast. J Biol Chem, 2012. 287(23): p. 19029-39. 77. Martin, D.C., et al., New regulators of a high affinity Ca2+ influx system revealed through a genome-wide screen in yeast. J Biol Chem, 2011. 286(12): p. 10744-54. 78. Kavanaugh, J.S. and A.R. Horswill, Impact of Environmental Cues on Staphylococcal Quorum Sensing and Biofilm Development. J Biol Chem, 2016. 291(24): p. 12556-64. 79. Rice, K.C. and K.W. Bayles, Death's toolbox: examining the molecular components of bacterial programmed cell death. Mol Microbiol, 2003. 50(3): p. 729-38. 80. Asally, M., et al., Localized cell death focuses mechanical forces during 3D patterning in a biofilm. Proc Natl Acad Sci U S A, 2012. 109(46): p. 18891-6.

149

81. Miguelez, E.M., C. Hardisson, and M.B. Manzanal, Hyphal death during colony development in Streptomyces antibioticus: morphological evidence for the existence of a process of cell deletion in a multicellular prokaryote. J Cell Biol, 1999. 145(3): p. 515-25. 82. Doyle, R.J. and A.L. Koch, The functions of autolysins in the growth and division of Bacillus subtilis. Crit Rev Microbiol, 1987. 15(2): p. 169-222. 83. Drlica, K., et al., Quinolone-mediated bacterial death. Antimicrob Agents Chemother, 2008. 52(2): p. 385-92. 84. Campbell, E.A., et al., Structural mechanism for rifampicin inhibition of bacterial rna polymerase. Cell, 2001. 104(6): p. 901-12. 85. Ge, M., et al., Vancomycin derivatives that inhibit peptidoglycan biosynthesis without binding D-Ala-D-Ala. Science, 1999. 284(5413): p. 507-11. 86. Kahne, D., et al., Glycopeptide and lipoglycopeptide antibiotics. Chem Rev, 2005. 105(2): p. 425-48. 87. Davies, J., L. Gorini, and B.D. Davis, Misreading of RNA codewords induced by aminoglycoside antibiotics. Mol Pharmacol, 1965. 1(1): p. 93-106. 88. Kohanski, M.A., D.J. Dwyer, and J.J. Collins, How antibiotics kill bacteria: from targets to networks. Nat Rev Microbiol, 2010. 8(6): p. 423-35. 89. Kohanski, M.A., et al., A common mechanism of cellular death induced by bactericidal antibiotics. Cell, 2007. 130(5): p. 797-810. 90. Dwyer, D.J., et al., Antibiotic-induced bacterial cell death exhibits physiological and biochemical hallmarks of apoptosis. Mol Cell, 2012. 46(5): p. 561-72. 91. Engelberg-Kulka, H. and G. Glaser, Addiction modules and programmed cell death and antideath in bacterial cultures. Annu Rev Microbiol, 1999. 53: p. 43-70. 92. Sat, B., et al., Programmed cell death in Escherichia coli: some antibiotics can trigger mazEF lethality. J Bacteriol, 2001. 183(6): p. 2041-5. 93. Amitai, S., et al., Escherichia coli MazF leads to the simultaneous selective synthesis of both "death proteins" and "survival proteins". PLoS Genet, 2009. 5(3): p. e1000390. 94. Bishop, R.E., et al., The entericidin locus of Escherichia coli and its implications for programmed bacterial cell death. J Mol Biol, 1998. 280(4): p. 583-96. 95. Adler, E., I. Barak, and P. Stragier, Bacillus subtilis locus encoding a killer protein and its antidote. J Bacteriol, 2001. 183(12): p. 3574-81. 96. Groicher, K.H., et al., The Staphylococcus aureus lrgAB operon modulates murein hydrolase activity and penicillin tolerance. J Bacteriol, 2000. 182(7): p. 1794-801. 97. Rice, K.C., et al., The Staphylococcus aureus cidAB operon: evaluation of its role in regulation of murein hydrolase activity and penicillin tolerance. J Bacteriol, 2003. 185(8): p. 2635-43. 98. Thorenoor, N., et al., Localization of the death effector domain of Fas-associated death domain protein into the membrane of Escherichia coli induces reactive oxygen species- involved cell death. Biochemistry, 2010. 49(7): p. 1435-47. 99. Lee, S.W., et al., Death effector domain of a mammalian apoptosis mediator, FADD, induces bacterial cell death. Mol Microbiol, 2000. 35(6): p. 1540-9. 100. Ranjit, D.K., J.L. Endres, and K.W. Bayles, Staphylococcus aureus CidA and LrgA proteins exhibit holin-like properties. J Bacteriol, 2011. 193(10): p. 2468-76. 101. Hardwick, J.M. and W.C. Cheng, Mitochondrial programmed cell death pathways in yeast. Dev Cell, 2004. 7(5): p. 630-2.

150

102. Vachova, L. and Z. Palkova, Physiological regulation of yeast cell death in multicellular colonies is triggered by ammonia. J Cell Biol, 2005. 169(5): p. 711-7. 103. Zhao, X., et al., Current Perspectives on Viable but Non-culturable State in Foodborne Pathogens. Front Microbiol, 2017. 8: p. 580. 104. Teng, X. and J.M. Hardwick, Reliable method for detection of programmed cell death in yeast. Methods Mol Biol, 2009. 559: p. 335-42. 105. Teng, X. and J.M. Hardwick, Quantification of genetically controlled cell death in budding yeast. Methods Mol Biol, 2013. 1004: p. 161-70. 106. Zhang, Y., W.W. Yew, and M.R. Barer, Targeting persisters for tuberculosis control. Antimicrob Agents Chemother, 2012. 56(5): p. 2223-30. 107. Chong, Y.P., et al., Treatment duration for uncomplicated Staphylococcus aureus bacteremia to prevent relapse: analysis of a prospective observational cohort study. Antimicrob Agents Chemother, 2013. 57(3): p. 1150-6. 108. Park, K.H., et al., Clinical characteristics and therapeutic outcomes of hematogenous vertebral osteomyelitis caused by methicillin-resistant Staphylococcus aureus. J Infect, 2013. 67(6): p. 556-64. 109. Senneville, E., et al., Outcome and predictors of treatment failure in total hip/knee prosthetic joint infections due to Staphylococcus aureus. Clin Infect Dis, 2011. 53(4): p. 334-40. 110. Romano, C.L., et al., Value of debridement and irrigation for the treatment of peri- prosthetic infections. A systematic review. Hip Int, 2012. 22 Suppl 8: p. S19-24. 111. Peel, T.N., et al., Outcome of debridement and retention in prosthetic joint infections by methicillin-resistant staphylococci, with special reference to rifampin and fusidic acid combination therapy. Antimicrob Agents Chemother, 2013. 57(1): p. 350-5. 112. Liu, C., et al., Clinical practice guidelines by the infectious diseases society of america for the treatment of methicillin-resistant Staphylococcus aureus infections in adults and children. Clin Infect Dis, 2011. 52(3): p. e18-55. 113. Conlon, B.P., et al., Activated ClpP kills persisters and eradicates a chronic biofilm infection. Nature, 2013. 503(7476): p. 365-70. 114. Allison, K.R., M.P. Brynildsen, and J.J. Collins, Metabolite-enabled eradication of bacterial persisters by aminoglycosides. Nature, 2011. 473(7346): p. 216-20. 115. Sharma, B., et al., Borrelia burgdorferi, the Causative Agent of Lyme Disease, Forms Drug-Tolerant Persister Cells. Antimicrob Agents Chemother, 2015. 59(8): p. 4616-24. 116. Feng, J., et al., Ceftriaxone Pulse Dosing Fails to Eradicate Biofilm-Like Microcolony B. burgdorferi Persisters Which Are Sterilized by Daptomycin/ Doxycycline/Cefuroxime without Pulse Dosing. Front Microbiol, 2016. 7: p. 1744. 117. Zhang, Y., et al., Mechanisms of Pyrazinamide Action and Resistance. Microbiol Spectr, 2013. 2(4): p. 1-12. 118. Shi, W., et al., Pyrazinamide inhibits trans-translation in Mycobacterium tuberculosis. Science, 2011. 333(6049): p. 1630-2. 119. Zhang, Y., et al., Mechanisms of Pyrazinamide Action and Resistance. Microbiol Spectr, 2014. 2(4): p. Mgm2-0023-2013. 120. Vandenesch, F., et al., Community-acquired methicillin-resistant Staphylococcus aureus carrying Panton-Valentine leukocidin genes: worldwide emergence. Emerg Infect Dis, 2003. 9(8): p. 978-84.

151

121. Chambers, H.F. and F.R. Deleo, Waves of resistance: Staphylococcus aureus in the antibiotic era. Nat Rev Microbiol, 2009. 7(9): p. 629-41. 122. McDougal, L.K., et al., Pulsed-field gel electrophoresis typing of oxacillin-resistant Staphylococcus aureus isolates from the United States: establishing a national database. J Clin Microbiol, 2003. 41(11): p. 5113-20. 123. Barrett, T.W. and G.J. Moran, Update on emerging infections:news from the Centers for Disease Control and Prevention. Methicillin-resistant Staphylococcus aureus infections among competitive sports participants--Colorado, Indiana, Pennsylvania, and Los Angeles County, 2000-2003. Ann Emerg Med, 2004. 43(1): p. 43-5; discussion 45-7. 124. Tenover, F.C. and R.V. Goering, Methicillin-resistant Staphylococcus aureus strain USA300: origin and epidemiology. J Antimicrob Chemother, 2009. 64(3): p. 441-6. 125. von Eiff, C., et al., Phenotype microarray profiling of Staphylococcus aureus menD and hemB mutants with the small-colony-variant phenotype. J Bacteriol, 2006. 188(2): p. 687- 93. 126. Clasener, H., Pathogenicity of the L-phase of bacteria. Annu Rev Microbiol, 1972. 26: p. 55-84. 127. Han, J., et al., Glycerol uptake is important for L-form formation and persistence in Staphylococcus aureus. PLoS One, 2014. 9(9): p. e108325. 128. Bose, J.L., P.D. Fey, and K.W. Bayles, Genetic tools to enhance the study of gene function and regulation in Staphylococcus aureus. Appl Environ Microbiol, 2013. 79(7): p. 2218-24. 129. Helle, L., et al., Vectors for improved Tet repressor-dependent gradual gene induction or silencing in Staphylococcus aureus. Microbiology, 2011. 157(Pt 12): p. 3314-23. 130. Korzeniowski, O. and M.A. Sande, Combination antimicrobial therapy for Staphylococcus aureus endocarditis in patients addicted to parenteral drugs and in nonaddicts: A prospective study. Ann Intern Med, 1982. 97(4): p. 496-503. 131. Ge, X., et al., Identification of Streptococcus sanguinis genes required for biofilm formation and examination of their role in endocarditis virulence. Infect Immun, 2008. 76(6): p. 2551-9. 132. Kim, J.K., et al., Purine biosynthesis-deficient Burkholderia mutants are incapable of symbiotic accommodation in the stinkbug. Isme j, 2014. 8(3): p. 552-63. 133. Mongodin, E., et al., Microarray transcription analysis of clinical Staphylococcus aureus isolates resistant to vancomycin. J Bacteriol, 2003. 185(15): p. 4638-43. 134. Kriel, A., et al., Direct regulation of GTP homeostasis by (p)ppGpp: a critical component of viability and stress resistance. Mol Cell, 2012. 48(2): p. 231-41. 135. Potrykus, K., et al., ppGpp is the major source of growth rate control in E. coli. Environ Microbiol, 2011. 13(3): p. 563-75. 136. Khakimova, M., et al., The stringent response controls catalases in Pseudomonas aeruginosa and is required for hydrogen peroxide and antibiotic tolerance. J Bacteriol, 2013. 195(9): p. 2011-20. 137. Thurlow, L.R., et al., Functional modularity of the arginine catabolic mobile element contributes to the success of USA300 methicillin-resistant Staphylococcus aureus. Cell Host Microbe, 2013. 13(1): p. 100-7. 138. Diep, B.A., et al., Complete genome sequence of USA300, an epidemic clone of community-acquired meticillin-resistant Staphylococcus aureus. Lancet, 2006. 367(9512): p. 731-9.

152

139. Diep, B.A., et al., The arginine catabolic mobile element and staphylococcal chromosomal cassette mec linkage: convergence of virulence and resistance in the USA300 clone of methicillin-resistant Staphylococcus aureus. J Infect Dis, 2008. 197(11): p. 1523-30. 140. Scherr, T.D., et al., Global transcriptome analysis of Staphylococcus aureus biofilms in response to innate immune cells. Infect Immun, 2013. 81(12): p. 4363-76. 141. Xu, S. and A.D. Chisholm, Methods for skin wounding and assays for wound responses in C. elegans. J Vis Exp, 2014(94). 142. Leduc, D., et al., Coupled amino acid deamidase-transport systems essential for Helicobacter pylori colonization. Infect Immun, 2010. 78(6): p. 2782-92. 143. Dunning, D.W., et al., SloR modulation of the Streptococcus mutans acid tolerance response involves the GcrR response regulator as an essential intermediary. Microbiology, 2008. 154(Pt 4): p. 1132-43. 144. Fiorentino, G., et al., MarR-like transcriptional regulator involved in detoxification of aromatic compounds in Sulfolobus solfataricus. J Bacteriol, 2007. 189(20): p. 7351-60. 145. Kennedy, A.D., et al., Targeting of alpha-hemolysin by active or passive immunization decreases severity of USA300 skin infection in a mouse model. J Infect Dis, 2010. 202(7): p. 1050-8. 146. Rauch, S., et al., Abscess formation and alpha-hemolysin induced toxicity in a mouse model of Staphylococcus aureus peritoneal infection. Infect Immun, 2012. 80(10): p. 3721-32. 147. Hung, C.S., K.W. Dodson, and S.J. Hultgren, A murine model of urinary tract infection. Nat Protoc, 2009. 4(8): p. 1230-43. 148. Archer, N.K., et al., Staphylococcus aureus biofilms: properties, regulation, and roles in human disease. Virulence, 2011. 2(5): p. 445-59. 149. Smith, P.A. and F.E. Romesberg, Combating bacteria and drug resistance by inhibiting mechanisms of persistence and adaptation. Nat Chem Biol, 2007. 3(9): p. 549-56. 150. Yee, R., et al., Genetic Screen Reveals the Role of Purine Metabolism in Staphylococcus aureus Persistence to Rifampicin. Antibiotics, 2015. 4(4): p. 627. 151. Fey, P.D., et al., A genetic resource for rapid and comprehensive phenotype screening of nonessential Staphylococcus aureus genes. MBio, 2013. 4(1): p. e00537-12. 152. Darby, C., Interactions with microbial pathogens. WormBook, 2005: p. 1-15. 153. Porta-de-la-Riva, M., et al., Basic Caenorhabditis elegans methods: synchronization and observation. J Vis Exp, 2012(64): p. e4019. 154. Xu, Y., B. Labedan, and N. Glansdorff, Surprising arginine biosynthesis: a reappraisal of the enzymology and evolution of the pathway in microorganisms. Microbiol Mol Biol Rev, 2007. 71(1): p. 36-47. 155. Nuxoll, A.S., et al., CcpA regulates arginine biosynthesis in Staphylococcus aureus through repression of proline catabolism. PLoS Pathog, 2012. 8(11): p. e1003033. 156. Lindgren, J.K., et al., Arginine deiminase in Staphylococcus epidermidis functions to augment biofilm maturation through pH homeostasis. J Bacteriol, 2014. 196(12): p. 2277-89. 157. Wortham, B.W., C.N. Patel, and M.A. Oliveira, Polyamines in bacteria: pleiotropic effects yet specific mechanisms. Adv Exp Med Biol, 2007. 603: p. 106-15.

153

158. Chattopadhyay, M.K. and H. Tabor, Polyamines are critical for the induction of the glutamate decarboxylase-dependent acid resistance system in Escherichia coli. J Biol Chem, 2013. 288(47): p. 33559-70. 159. Ferreira, A.B., et al., Involvement of the ornithine decarboxylase gene in acid stress response in probiotic Lactobacillus delbrueckii UFV H2b20. Benef Microbes, 2015. 6(5): p. 719-25. 160. Wortham, B.W., et al., Polyamines are required for the expression of key Hms proteins important for Yersinia pestis biofilm formation. Environ Microbiol, 2010. 12(7): p. 2034- 47. 161. Hu, L.I., B.P. Lima, and A.J. Wolfe, Bacterial protein acetylation: the dawning of a new age. Mol Microbiol, 2010. 77(1): p. 15-21. 162. Okanishi, H., et al., Acetylome with structural mapping reveals the significance of lysine acetylation in Thermus thermophilus. J Proteome Res, 2013. 12(9): p. 3952-68. 163. Monk, I.R. and T.J. Foster, Genetic manipulation of Staphylococci-breaking through the barrier. Front Cell Infect Microbiol, 2012. 2: p. 49. 164. Stobberingh, E.E., R. Schiphof, and J.S. Sussenbach, Occurrence of a class II restriction endonuclease in Staphylococcus aureus. J Bacteriol, 1977. 131(2): p. 645-9. 165. Xu, S.Y., et al., A type IV modification-dependent restriction enzyme SauUSI from Staphylococcus aureus subsp. aureus USA300. Nucleic Acids Res, 2011. 39(13): p. 5597- 610. 166. Davis, B.D., Mechanism of bactericidal action of aminoglycosides. Microbiol Rev, 1987. 51(3): p. 341-50. 167. Thomas, V.C., et al., A central role for carbon-overflow pathways in the modulation of bacterial cell death. PLoS Pathog, 2014. 10(6): p. e1004205. 168. Sadykov, M.R., et al., Inactivation of the Pta-AckA pathway causes cell death in Staphylococcus aureus. J Bacteriol, 2013. 195(13): p. 3035-44. 169. Feng, J., et al., An optimized SYBR Green I/PI assay for rapid viability assessment and antibiotic susceptibility testing for Borrelia burgdorferi. PLoS One, 2014. 9(11): p. e111809. 170. Feng, J., et al., A Rapid Growth-Independent Antibiotic Resistance Detection Test by SYBR Green/Propidium Iodide Viability Assay. Front Med (Lausanne), 2018. 5: p. 127. 171. Windham, I.H., et al., SrrAB Modulates Staphylococcus aureus Cell Death through Regulation of cidABC Transcription. J Bacteriol, 2016. 198(7): p. 1114-22. 172. Birben, E., et al., Oxidative stress and antioxidant defense. World Allergy Organ J, 2012. 5(1): p. 9-19. 173. Herrmann, G., et al., Energy conservation via electron-transferring flavoprotein in anaerobic bacteria. J Bacteriol, 2008. 190(3): p. 784-91. 174. Ramond, E., et al., Glutamate utilization couples oxidative stress defense and the tricarboxylic acid cycle in Francisella phagosomal escape. PLoS Pathog, 2014. 10(1): p. e1003893. 175. Tavares, A.F., et al., Reactive oxygen species mediate bactericidal killing elicited by carbon monoxide-releasing molecules. J Biol Chem, 2011. 286(30): p. 26708-17. 176. Li, J., et al., Trans-translation mediates tolerance to multiple antibiotics and stresses in Escherichia coli. J Antimicrob Chemother, 2013. 68(11): p. 2477-81. 177. Li, L., et al., Ribosomal elongation factor 4 promotes cell death associated with lethal stress. MBio, 2014. 5(6): p. e01708.

154

178. Kort, R., et al., Transcriptional activity around bacterial cell death reveals molecular biomarkers for cell viability. BMC Genomics, 2008. 9: p. 590. 179. Kawahata, M., et al., Yeast genes involved in response to lactic acid and acetic acid: acidic conditions caused by the organic acids in Saccharomyces cerevisiae cultures induce expression of intracellular metal metabolism genes regulated by Aft1p. FEMS Yeast Res, 2006. 6(6): p. 924-36. 180. DeLeo, F.R., et al., Community-associated meticillin-resistant Staphylococcus aureus. Lancet, 2010. 375(9725): p. 1557-68. 181. Selton-Suty, C., et al., Preeminence of Staphylococcus aureus in infective endocarditis: a 1-year population-based survey. Clin Infect Dis, 2012. 54(9): p. 1230-9. 182. Bae, I.G., et al., Heterogeneous vancomycin-intermediate susceptibility phenotype in bloodstream methicillin-resistant Staphylococcus aureus isolates from an international cohort of patients with infective endocarditis: prevalence, genotype, and clinical significance. J Infect Dis, 2009. 200(9): p. 1355-66. 183. Chan, L.C., et al., Innate Immune Memory Contributes to Host Defense against Recurrent Skin and Skin Structure Infections Caused by Methicillin-Resistant Staphylococcus aureus. Infect Immun, 2017. 85(2). 184. David, M.Z. and R.S. Daum, Community-associated methicillin-resistant Staphylococcus aureus: epidemiology and clinical consequences of an emerging epidemic. Clin Microbiol Rev, 2010. 23(3): p. 616-87. 185. Dryden, M.S., Complicated skin and soft tissue infection. J Antimicrob Chemother, 2010. 65 Suppl 3: p. iii35-44. 186. de Oliveira, A., et al., Antimicrobial Resistance Profile of Planktonic and Biofilm Cells of Staphylococcus aureus and Coagulase-Negative Staphylococci. Int J Mol Sci, 2016. 17(9). 187. Wang, W., et al., Transposon Mutagenesis Identifies Novel Genes Associated with Staphylococcus aureus Persister Formation. Front Microbiol, 2015. 6: p. 1437. 188. Xu, T., et al., The Agr Quorum Sensing System Represses Persister Formation through Regulation of Phenol Soluble Modulins in Staphylococcus aureus. Front Microbiol, 2017. 8: p. 2189. 189. Sahukhal, G.S., S. Pandey, and M.O. Elasri, msaABCR operon is involved in persister cell formation in Staphylococcus aureus. BMC Microbiol, 2017. 17(1): p. 218. 190. Zhang, Y. and D. Mitchison, The curious characteristics of pyrazinamide: a review. Int J Tuberc Lung Dis, 2003. 7(1): p. 6-21. 191. Zhang, Y., et al., Mode of action of pyrazinamide: disruption of Mycobacterium tuberculosis membrane transport and energetics by pyrazinoic acid. J Antimicrob Chemother, 2003. 52(5): p. 790-5. 192. Zhang, S., et al., Mutations in panD encoding aspartate decarboxylase are associated with pyrazinamide resistance in Mycobacterium tuberculosis. Emerg Microbes Infect, 2013. 2(6): p. e34. 193. Coates, A.R. and Y. Hu, Targeting non-multiplying organisms as a way to develop novel antimicrobials. Trends Pharmacol Sci, 2008. 29(3): p. 143-50. 194. Feng, J., et al., Eradication of Biofilm-Like Microcolony Structures of Borrelia burgdorferi by Daunomycin and Daptomycin but not Mitomycin C in Combination with Doxycycline and Cefuroxime. Front Microbiol, 2016. 7: p. 62.

155

195. Niu, H., Peng Cui, Rebecca Yee, Wanliang Shi, Shuo Zhang, Jie Feng, David Sullivan, Wenhong Zhang, Bingdong Zhu, Ying Zhang, A Clinical Drug Library Screen Identifies Tosufloxacin as Being Highly Active against Staphylococcus aureus Persisters. Antibiotics, 2015. 4(3): p. 329-336. 196. O'Toole, G.A., Microtiter dish biofilm formation assay. J Vis Exp, 2011(47). 197. Pfaller, M.A., et al., Oritavancin in vitro activity against gram-positive organisms from European and United States medical centers: results from the SENTRY Antimicrobial Surveillance Program for 2010-2014. Diagn Microbiol Infect Dis, 2018. 91(2): p. 199- 204. 198. Anderson, E.J., et al., A series of skin and soft tissue infections due to methicillin- resistant Staphylococcus aureus in HIV-infected patients. J Acquir Immune Defic Syndr, 2006. 41(1): p. 125-7. 199. Shastry, L., J. Rahimian, and S. Lascher, Community-associated methicillin-resistant Staphylococcus aureus skin and soft tissue infections in men who have sex with men in New York City. Arch Intern Med, 2007. 167(8): p. 854-7. 200. Skiest, D., et al., Community-onset methicillin-resistant Staphylococcus aureus in an urban HIV clinic. HIV Med, 2006. 7(6): p. 361-8. 201. Cho, J.S., et al., IL-17 is essential for host defense against cutaneous Staphylococcus aureus infection in mice. J Clin Invest, 2010. 120(5): p. 1762-73. 202. Kim, W., et al., A new class of synthetic retinoid antibiotics effective against bacterial persisters. Nature, 2018. 556(7699): p. 103-107. 203. Mohamed, M.F., A. Abdelkhalek, and M.N. Seleem, Evaluation of short synthetic antimicrobial peptides for treatment of drug-resistant and intracellular Staphylococcus aureus. Sci Rep, 2016. 6: p. 29707. 204. Joers, A., N. Kaldalu, and T. Tenson, The frequency of persisters in Escherichia coli reflects the kinetics of awakening from dormancy. J Bacteriol, 2010. 192(13): p. 3379-84. 205. Mandell, J.B., et al., Elimination of Antibiotic Resistant Surgical Implant Biofilms Using an Engineered Cationic Amphipathic Peptide WLBU2. Sci Rep, 2017. 7(1): p. 18098. 206. Defraine, V., M. Fauvart, and J. Michiels, Fighting bacterial persistence: Current and emerging anti-persister strategies and therapeutics. Drug Resist Updat, 2018. 38: p. 12- 26. 207. Humphries, R.M., S. Pollett, and G. Sakoulas, A current perspective on daptomycin for the clinical microbiologist. Clin Microbiol Rev, 2013. 26(4): p. 759-80. 208. Kim, H.K., D. Missiakas, and O. Schneewind, Mouse models for infectious diseases caused by Staphylococcus aureus. J Immunol Methods, 2014. 410: p. 88-99. 209. Gallo, S.W., C.A.S. Ferreira, and S.D. de Oliveira, Combination of polymyxin B and meropenem eradicates persister cells from Acinetobacter baumannii strains in exponential growth. J Med Microbiol, 2017. 66(8): p. 1257-1260. 210. Carpenter, C.F. and H.F. Chambers, Daptomycin: another novel agent for treating infections due to drug-resistant gram-positive pathogens. Clin Infect Dis, 2004. 38(7): p. 994-1000. 211. Feng, J., P.G. Auwaerter, and Y. Zhang, Drug combinations against Borrelia burgdorferi persisters in vitro: eradication achieved by using daptomycin, cefoperazone and doxycycline. PLoS One, 2015. 10(3): p. e0117207.

156

212. Pogliano, J., N. Pogliano, and J.A. Silverman, Daptomycin-mediated reorganization of membrane architecture causes mislocalization of essential cell division proteins. J Bacteriol, 2012. 194(17): p. 4494-504. 213. Balwan, A., et al., Clinafloxacin for Treatment of Burkholderia cenocepacia Infection in a Cystic Fibrosis Patient. Antimicrob Agents Chemother, 2016. 60(1): p. 1-5. 214. Levine, D.P., et al., Clinafloxacin for the treatment of bacterial endocarditis. Clin Infect Dis, 2004. 38(5): p. 620-31. 215. Cue, D., et al., SaeRS-dependent inhibition of biofilm formation in Staphylococcus aureus Newman. PLoS One, 2015. 10(4): p. e0123027. 216. Hooper, D.C., Structure of grepafloxacin relative to activity and safety profile. Clin Microbiol Infect, 1998. 4 Suppl 1: p. S15-s20. 217. Domagala, J.M., Structure-activity and structure-side-effect relationships for the quinolone antibacterials. J Antimicrob Chemother, 1994. 33(4): p. 685-706. 218. Kim, J.S. and T.K. Wood, Persistent Persister Misperceptions. Front Microbiol, 2016. 7: p. 2134. 219. Song, S. and T.K. Wood, Post-segregational Killing and Phage Inhibition Are Not Mediated by Cell Death Through Toxin/Antitoxin Systems. Front Microbiol, 2018. 9: p. 814. 220. Aizenman, E., H. Engelberg-Kulka, and G. Glaser, An Escherichia coli chromosomal "addiction module" regulated by guanosine [corrected] 3',5'-bispyrophosphate: a model for programmed bacterial cell death. Proc Natl Acad Sci U S A, 1996. 93(12): p. 6059- 63. 221. Lee, B., et al., Myxococcus xanthus developmental cell fate production: heterogeneous accumulation of developmental regulatory proteins and reexamination of the role of MazF in developmental lysis. J Bacteriol, 2012. 194(12): p. 3058-68. 222. Corvec, S., et al., Epidemiology and new developments in the diagnosis of prosthetic joint infection. Int J Artif Organs, 2012. 35(10): p. 923-34. 223. Moran, E., I. Byren, and B.L. Atkins, The diagnosis and management of prosthetic joint infections. J Antimicrob Chemother, 2010. 65 Suppl 3: p. iii45-54. 224. Nigam, A., S. Kumar, and H. Engelberg-Kulka, Quorum Sensing Extracellular Death Peptides Enhance the Endoribonucleolytic Activities of Mycobacterium tuberculosis MazF Toxins. MBio, 2018. 9(3). 225. Wang, X. and X. Zhao, Contribution of oxidative damage to antimicrobial lethality. Antimicrob Agents Chemother, 2009. 53(4): p. 1395-402. 226. Radimer, K., et al., Dietary supplement use by US adults: data from the National Health and Nutrition Examination Survey, 1999-2000. Am J Epidemiol, 2004. 160(4): p. 339-49. 227. Montgomery, C.P., M.Z. David, and R.S. Daum, Host factors that contribute to recurrent staphylococcal skin infection. Curr Opin Infect Dis, 2015. 28(3): p. 253-8. 228. Montgomery, C.P., et al., Protective immunity against recurrent Staphylococcus aureus skin infection requires antibody and interleukin-17A. Infect Immun, 2014. 82(5): p. 2125- 34. 229. Milner, J.D., et al., Impaired T(H)17 cell differentiation in subjects with autosomal dominant hyper-IgE syndrome. Nature, 2008. 452(7188): p. 773-6. 230. Vyas, K.J., et al., Trends and factors associated with initial and recurrent methicillin- resistant Staphylococcus aureus (MRSA) skin and soft-tissue infections among HIV- infected persons: an 18-year study. J Int Assoc Provid AIDS Care, 2014. 13(3): p. 206-13.

157

REBECCA YEE Johns Hopkins Bloomberg School of Public Health Molecular Microbiology and Immunology 615 North Wolfe Street, Baltimore, MD 21205 [email protected]

EDUCATION Johns Hopkins Bloomberg School of Public Health, Baltimore, MD Fall 2018 Ph.D, Molecular Microbiology and Immunology

University of Pennsylvania, Philadelphia, PA May 2013 B.A, Biology

RESEARCH EXPERIENCES

Johns Hopkins Bloomberg School of Public Health Ph.D Candidate Department of Molecular Microbiology and Immunology 2013-2018 Research Advisor: Ying Zhang, MD, PhD University of Pennsylvania School of Medicine Research Assistant Department of Microbiology 2006- 2013 Research Advisor: Jun Zhu, PhD The Johns Hopkins University School of Medicine Research Assistant Department of Biological Chemistry 2012 Research Advisor : Michael Wolfgang, PhD Carnegie Mellon University Research Assistant Pennsylvania Governor School of Sciences Participant 2008 Department of Biological Sciences Research Advisor: Allison Marciszyn, PhD

LEADERSHIP EXPERIENCES

Medical and Educational Perspectives, Inc Board Member Baltimore, MD 2015-2019

MERIT Health Leadership Academy Mentor Baltimore, MD 2015-2018

Graduate Coating Committee, Johns Hopkins Medical Campus Member Baltimore, MD 2015-2017

Johns Hopkins University Student Research Forum Coordinator Baltimore, MD 2014-2017

158

Phi Sigma Biological Sciences Honor Society UPenn Chapter Co-founder & President Philadelphia, PA 2012-2013

Gates Millennium Scholars Program Philadelphia City Representative Washington, DC 2010 - 2013

Pennsylvania Junior Academy of Sciences Lead Judge Hosted by Pennsylvania State University, College Park, PA 2010-2012

PATENT Application filed Zhang, Y, Feng, J., Yee, R., Zhang, S., Zhang, W. 2016. Novel Methodology for Identifying Anti-Persister Activity and Antimicrobial Susceptibility For Borrelia burgdorferi and Other Bacteria. United States Patent Application US20170058314 A1 Filed Mar 2, 2017.

PUBLICATIONS Niu, H., Yee, R., Cui, P., Zhang, S., Tian, L., Shi, W., Sullivan, D., Zhu, B., Zhang, W., Zhang, Y. Identification and Ranking of Clinical Compounds with Activity against Log-phase Growing Uropathogenic Escherichia coli. Curr Drug Discov Technol. 2018 Feng, J., Yee, R.,* Zhang, S., Tian, L., Shi, W., Zhang, W., Zhang, Y. A Rapid Growth- Independent Antibiotic Resistance Detection Test by SYBR Green/Propidium Iodide Viability Assay. Front. Med. 2018 *co-first author

Niu, H., Yee, R., Cui, P., Tian, L., Zhang, S., Shi, W., Sullivan, D., Zhu, B., Zhang, W., Zhang, Y. Identification of Agents Active against Methicillin-Resistant Staphylococcus aureus USA300 from a Clinical Compound Library. Pathogens. 2017

Yee, R., Cui, P., Xu, T., Shi, W., Feng, J., Zhang, W., Zhang, Y. Identification of a Novel Gene argJ involved in Arginine Biosynthesis Critical for Persister Formation in Staphylococcus aureus. bioRxiv 114827; doi:https://doi.org/10.1101/114827

Pastrana-Mena, R., Matthias, D., Delves, Rajaram, K., M., King, J.G, Yee, R., Trucchi, B., Verotta, L., Dinglasan, R. A malaria transmission-blocking (+)-usnic acid derivative prevents Plasmodium zygote-to-ookinete maturation in the mosquito midgut. ACS Chem Biol. 2016

Yee, R., Cui, P., Shi, W., Feng, J., Zhang, Y. Genetic Screen Reveals the Role of Purine Metabolism in Staphylococcus aureus Persistence to Rifampicin. Antibiotics 2015

Niu, H., Cui, P., Yee, R., Shi, W., Zhang, S., Feng, J., Sullivan, D., Zhang, W., Zhu, B., Zhang, Y. A Clinical Drug Library Screen Identifies Tosufloxacin as Being Highly Active against Staphylococcus aureus Persisters. Antibiotics 2015

159

Yee,R., Patel, N., Zhang, L., Brooks, E., Sherman, E., Jayakumar, K., Nawrocki, E., K. Donley. A Comparative Survey of 16S Ribosomal RNA Diversity in Soil Bacteria under Various Environmental Conditions. Journal of the Pennsylvania Governor’s School for the Sciences 2009

ABSTRACTS / CONFERENCES

Yee. R., Cui, P., Shi, W., Feng, J., Zhang, Y. The Role of Purine Metabolism in Staphylococcus aureus Persistence to Rifampicin. Poster Presentation, American Society of Microbiology Maryland Branch Meeting, Baltimore, MD. June 2018 Tarff, A., Yee, R., Pfrommer, B., Simner, P., Zhang, Y., Behrens, A. Amoebicidal activity of oxygenized riboflavin photoactivation: BPerox-UVA as a novel combination therapy for keratitis. Poster Presentation, Association for Research in Vision and Ophthalmology Meeting, Honolulu, HI. May 2018 Behrens, A., Tarff, A., Yee, R., Zhang, Y. Treatment for fungal corneal melts by photochemical therapy using BPerox as a new light sensitive molecule in vitro. Poster Presentation, Association for Research in Vision and Ophthalmology Meeting, Honolulu, HI. May 2018 Reyes, L. Z., Tarff, A., Yee, R., Zhang, Y., Behrens, A. Photochemical antimicrobial therapy for keratitis with an improved riboflavin conjugate (BPerox) for highly resistant bacteria in vitro. Poster Presentation, Association for Research in Vision and Ophthalmology Meeting, Honolulu, HI. May 2018 Tarff, A., Yee, R., Zhang, Y., Behrens, A. Bactericidal peroxidation of riboflavin (BPerox) photoactivated by UVA for stationary-phase MRSA persister cells causing keratitis in vitro. Paper talk, American Society of Cataract and Refractive Surgery (ASCRS), Washington DC. April 2018 (Awarded the best paper of session) Tarff, A., Yee, R., Zhang, Y., Behrens, A. High-irradiance accelerated CXL for Keratitis with chromophore photoactivation by low levels of hydrogen peroxide (BPerox). Paper talk, American Society of Cataract and Refractive Surgery (ASCRS), Washington DC. April 2018 Tarff, A., Yee, R., Zhang, Y., Behrens, A. Conventional CXL for keratitis with chromophore photoactivation magnified by an active for of oxygen from low levels of hydrogen peroxide (BPerox). Paper talk, American Society of Cataract and Refractive Surgery (ASCRS), Washington DC. April 2018 Tarff, A., Yee, R., Zhang, Y., Behrens, A. High-intensity accelerated corneal cross-linking as a therapy for keratitis: chromophore-photoactivation with low levels of hydrogen peroxide (BPerox). Poster Presentation, 30TH Biennial Cornea Conference: Harvard Medical School, Department of Ophthalmology, Boston, MA. Oct 2017 Tarff, A., Yee, R., Pfrommer, B., Simner, P., Casadevall, A., Zhang, Y., Behrens, A. Photochemical therapies for corneal infections… the sequel. Poster Presentation, Pan American Association of Ophthalmology, Lima, Peru. Aug 2017

160

Yee. R., Cui, P., Shi, W., Feng, J., Zhang, Y. The Role of Purine Metabolism in Staphylococcus aureus Persistence to Rifampicin. Poster Presentation and Rapid Fire Talk, American Society of Microbiology Microbe, New Orleans, LA. June 2017 Behrens, A., Yee, R., Tarff, A., Cano, M., Gupta, P., Di Meglio, L., May, W., Casadevall, A., Zhang, Y. Bactericidal activity of photoactivated riboflavin, ultraviolet light A, and hydrogen peroxide against stationary phase methicillin-resistant Staphylococcus aureus (MRSA) for potential use in bacterial keratitis. Poster Presentation, Association for Research in Vision and Ophthalmology Meeting, Baltimore, MD. May 2017 Reyes, L.Z., Tarff, A., Yee, R., del Valle Cano, M., Di Meglio, L., Gupta, P., May, W., Casadevall, A., Zhang, Y., Behrens, A. Efficacy of high fluence ultraviolet light A (UVA) photoactivation of riboflavin in vitro as a therapy for keratitis. Poster Presentation, Association for Research in Vision and Ophthalmology Meeting, Baltimore, MD. May 2017 Di MEglio, L., Tarff, A., Yee, R., Gupta, P., Annadanam, A., del Valle Cano, M., May, W., Behrens, A. Potential role of topical bovine colostrum in remodeling corneal epithelial cells after an acute ocular alkali burn in mice. Poster Presentation, Association for Research in Vision and Ophthalmology Meeting, Baltimore, MD. May 2017 Tarff, A., Yee, R., Di Meglio, La., Gupta, P., del Valle Cano, M., May, W., Casadevall, A., Zhang, Y., Behrens, A. Bactericidal ultraviolet light A (UVA) photoactivation of riboflavin potentiated by low levels of hydrogen peroxide in vitro: a potential novel therapy for methicillin sensitive and multidrug resistant-methicillin resistant Staphylococcus aureus keratitis. Poster Presentation, Association for Research in Vision and Ophthalmology Meeting, Baltimore, MD. May 2017 Tarff, A., Yee, R., del Valle Cano, M., Zhang, Y., Behrens, A. Antibacterial Activity of Riboflavin Combined with a Low Power Density of Ultraviolet Light A in Vitro: A Therapy for Methicillin Sensitive and Resistant Staphylococcus Aureus Corneal Keratitis. Paper Talk, American Society of Cataract and Refractive Surgery, Los Angeles, CA. May 2017 (First place awarded: residents and fellows paper session) Tarff, A., Yee, R., del Valle Cano, M., Zhang, Y., Behrens, A. Paper Talk, Antibacterial Activity of Riboflavin Combined with High Fluence Ultraviolet Light A in Vitro: A Therapy for Methicillin Sensitive and Resistant Staphylococcus Aureus Corneal Keratitis. American Society of Cataract and Refractive Surgery, Los Angeles, CA. May 2017 Yee, R., Yang, M., Zhu, J. Virulence Genes Regulation by Bile Acids in Vibrio cholerae. Poster Presentation, Annual Biomedical Research Conference for Minority Students (ABRCMS) by ASM, San Jose, CA. November 2012 (Recipient of Poster Presentation award) Yang, M., Liu, Z., Hughes, C., Yee, R., Fenical, W., Zhu, J. Host Signal Induced Intermolecular Disulfide Bond Formation Promotes V. cholerae Virulence. Poster Presentation, American Society of Microbiology 112th General Meeting, San Francisco, CA. June 2012 HONORS AND AWARDS

Johns Hopkins Bloomberg School of Public Health Teaching Award 2017

161

The Eleanor A. Bliss Honorary Fellowship 2017 The Eleanor A. Bliss Honorary Fellowship 2016 Frederik B. Bang Fellowship 2016 The Ruth G. Wittler Scholarship 2015 The Martin Frobisher Fellowship 2014 ASM Poster Presentation Award at ABRCMS 2012 Philadelphia Mayor’s Scholarship ($ 112,000) 2009-13 Carol Einiger Trustee Scholarship ($40,000) 2009-13 AstraZeneca Pharmaceuticals Certificate of Achievement 2009 United States Department of Agriculture Certificate of Achievement 2009 TEACHING EXPERIENCES Pathogenesis of Bacterial Infections, Teaching Assistant & Instructor 2014-2019  Lecturer on Clinical Plate Rounds: Case studies on Staphylococcal, Mycobacterium and fungal species  Lecturer on Mechanisms of Antibiotics Resistance  Lecturer on Helicobacter pylori

Medical Entrepreneurship, Course Organizer 2017-2019 ASSOCIATION MEMBERSHIPS

American Association for the Advancement of Science 2017- present Association for Research in Vision and Ophthalmology 2016- present American Society of Microbiology 2012-present Phi Sigma Biology Sciences Honors Society 2009-present

162